Amazonian Languages. Volume 2 Language Isolates II: Kanoé to Yurakaré: An International Handbook 9783110432732, 9783110440744

Citation preview

Amazonian Languages: Language Isolates An International Handbook HSK 44.2

Handbooks of Linguistics and Communication Science Handbücher zur Sprach- und Kommunikationswissenschaft

Founded by Gerold Ungeheuer Co-edited by Hugo Steger 1985–2001 Founded and edited by Herbert Ernst Wiegand 1982–2018 Edited by Jeroen Darquennes and Patience Epps

Volume 44.2

Amazonian Languages Language Isolates II: Kanoé to Yurakaré An International Handbook Edited by Patience Epps and Lev Michael

ISBN 978-3-11-044074-4 e-ISBN (PDF) 978-3-11-043273-2 e-ISBN (EPUB) 978-3-11-043284-8 ISSN 1861-5090 Library of Congress Control Number: 2022935912 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. © 2023 Walter de Gruyter GmbH, Berlin/Boston Cover design: Martin Zech, Bremen Typsetting: Meta Systems Publishing & Printservices GmbH, Wustermark Printing: CPI books GmbH, Leck www.degruyter.com

Contents Volume II Abbreviations

ix

Laércio Nora Bacelar and Hein van der Voort 659 14 Kanoé Hein van der Voort 719 15 Kwaza Raoul Zamponi and Chris Rogers 767 16 Máku Katharina Haude 807 17 Movima Lev Michael, Stephanie Farmer, Gregory Finley, Karina Sullón Acosta, Christine Beier, Alexandra Chanchari Icahuate †, Donalia Icahuate Baneo, and Melchor Sinti Saita 851 18 Muniche Bernat Bardagil 893 19 Mỹky Zachary O’Hagan 939 20 Omurano Raiane Salles 21 Pirahã (Apáitisí)

957

Zachary O’Hagan 995 22 Taushiro Katherine Bolaños 1029 23 Tinigua Raquel Guirardello-Damian 24 Trumai 1077

vi

Contents

Knut J. Olawsky 1107 25 Urarina Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón 1143 26 Wã́nsöjöt (Puinave) Alexia Z. Fawcett 27 Wao Terero

1191

Andrés Romero-Figueroa and Konrad Rybka 1243 28 Warao Esteban Emilio Mosonyi and Raoul Zamponi 1283 29 Yaruro (Pumé) Rik van Gijn 30 Yurakaré (Yurújare) Index

1323

Contents

Volume I Preface

v

Abbreviations

xi

Patience Epps and Lev Michael Introduction: The languages of Amazonia

xvii

Hein van der Voort and Joshua Birchall 1 Aikanã 1 Rafael Fischer and Kees Hengeveld 2 A'ingae (Cofán) 65 Jon Landaburu 3 Andoke 125 Christine Beier and Lev Michael 4 Aʔɨwa 173 Raoul Zamponi 5 Betoi-Jirara

223

Mily Crevels and Pieter Muysken † 6 Cayubaba 263 Sandy Ritchie and Jeanette Sakel 7 Chimane-Mosetén 301 Astrid Alexander-Bakkerus and Kelsey Caitlyn Neely 8 Cholón 371 Kristina Balykova, Gustavo Godoy, and Adair Pimentel Palácio † 9 Guató 409 An Van linden 10 Harakmbut Mily Crevels 11 Itonama

441

483

vii

viii

Contents

Jorge Emilio Rosés Labrada 547 12 Jodï Simon E. Overall 13 Kandozi-Chapra Index

615

Abbreviations ♀ ♀ref 1 2 3 2d 3d a abl abs absnt acc acc.irr acc.real accord acq.abil.aux act act.advr act.nmlz add add.foc ade adj adjr adl admon adv AdvCl advr advrs af aff aff.emph affr ag ag.nmlz agr alien all aln.poss am an ana ana.loc and

female genderlect female reference first person second person third person two-dimensional three-dimensional most agentive argument of canonical transitive verb ablative absolutive absential accusative irrealis accusative realis accusative accord clause marker acquired ability auxiliary active active adverbializer action nominalizer additive additive focus adessive adjective adjectivalizer adlative admonitory adverb(ial) adverbial clause adverbializer adversative affected affect affective emphasis affirmative epistemic modality agent agentive nominalizer agreement alienator allative alienable possession associated motion animate anaphoric reference to entity or event anaphoric reference to location or time andative

https://doi.org/10.1515/9783110432732-201

anph ant antg antip apfv appl appr aprx art asp ass assc assev atten attr aug aux avrt

anaphoric anteriority antagonistic antipassive anti-perfective applicative apprehensive approximative article aspect assertive associative asseverative attenuative attributive augmentative auxiliary avertive

ben

benefactive

cat.neg caus caus.indr caus.soc cd clf clf.rptr cmnd cmpr cnsq cnt.int cntf cntl cntr.foc cntr.top cntrx coll com comm comp CompCl compl conc cond conj cont coord

categorical negation causative indirect causative sociative causative classifying directional classifier repeater classifier commandative comparative consequential constituent question counterfactual continual contrastive focus contrastive topic counterexpectational collective comitative commitment complementizer complement clause completive concessive conditional conjunction continuative coordinator

x

Abbreviations

cop coprt coref cos crtn cvb

copula coparticipant applicative coreferential cosubordinate certainty converb

dat dec decl def deic del del.inch dem dem.dist dem.prox deon dep depr desid det detrns dim dir dirr disc disc.con dist distr down drct drtv drv ds dstr.quant du dub dur dv

dative deceased declarative definite deictic delimiter delayed inchoative demonstrative distal demonstrative proximal demonstrative deontic dependent deprecative desiderative determiner detransitivizer diminutive directional double Irrealis discourse discursive connector distal distributive vertical down and downriver locative direct directive derivational different subject distributive quantifier dual dubitative durative dummy vowel

e ec egr elat em.caus emph ep.c ep.v

extension to core semantically empty classifier egressive elative emotional causative emphatic epenthetic consonant epenthetic vowel

epist epist.necc er erg evnt.nmlz evo excl exclm exco exh exist

epistemic modal epistemic necessity semantically empty or dummy root ergative event nominalization evocative exclusive exclamative exocentric compound interfix exhaustive existential

f firsth foc frq frus func fut

feminine firsthand evidential focus frequentative frustrative function future

gen gen.clf ger go.aimlessly goal hab hab.nmlz hod.pst hon hort hrs hum hum.pl hyp

genitive genitive classifier gerund go aimlessly goal habitual habitual nominalizer hodiernal past honorific hortative hearsay evidential human human plural hypothetical

ident ideo ignr ill ilt imp imprs inact inan inc inch incl incorp.n ind

identificative ideophone ignorative illative in-law-talk imperative impersonal inactive inanimate incompletive inchoative inclusive incorporated noun indicative

Abbreviations

indet indf iness inf infl infr ingr init ins ins.com instant int intn intr intr.abil.aux intrj ints intzr inv io ipfv irr irs iter

indeterminate indefinite inessive infinitive inflection(al) inferential ingressive initiatory instrumental instrumental-comitative instantaneous aspect interrogative intentional intransitive intrinsic ability auxiliary interjection intensifier intransitivizer inverse indirect object imperfective irrealis indirect reported speech iterative

juss

jussive

lb lim ln loc loc.nmlz

lexical base limitative linking nasal locative locative nominalizer

m mal man med med.pst mel mid mir mn/pth mod mot mult

masculine malefactive malefactivemanner medial medial past main event line middle voice mirative manner/path modality motion multiple participant pluractional

n nar ncl

neuter narrative noun class marker

nec neg neg.aff neg.cop neg.dep neg.emph neg.exist neg.fnl neg.hab neg.pred neut.spat new.sit new.top nf nfin nfirsth nh nintn nm nml nml.pst nmlz nom npr nr.fut nsbj nspc nstd ntr ntr.asp num nvis nvol

necessity negation negative affect negative copula dependent clause negation emphatic negative negative existential final negation habitual negation negative predicate neutral spatial marker new situation new topic non-feminine non-finite non-firsthand evidential non-human non-intentional non-masculine nominal nominal past nominalizer, nominalization nominative non-present demonstrative near future non-subject non-specific non-standing neutral verb marking neutral aspect numeral non-visible non-volitional

o

patient-like argument of canonical transitive verb object oblique optative ordinal number overland

obj obl opt ord overland p

p.appl part pass pass.crtn

patient; most patientive argument of a canonical transitive verb patient applicative particle passive passive + epistemic certainty portmanteau

xi

xii pass.prob pat pauc paus per peri.caus perm pers pert pfv phat pity pl plc.nm plr.int pluract polit posp poss possd possr pot prec precul pred prep pres pres.dist prf prior priv prj.time prms pro proc prog proh prol prop.n prosp prosp.aux prox prox.temp prs prsc prxm pst pst.nmlz

Abbreviations

passive + epistemic probability portmanteau patientive paucal pausative perlative periphrastic causative permanent person marker pertains to perfective phatic pity plural place name polar interrogative pluractional polite postposition possessive possessed possessor potential precondition preculminative predicative preposition presentational distant presentational perfect prioritive privative projected time permissive pronoun procrastinative progressive prohibitive prolative proper noun prospective prospective auxiliary proximal/proximate (spatial) proximate temporal present prosecutive proximate (grammatical relation) past past nominalizer

psych.impf ptcp purp

psychological verb imperfective participle purposive

q qnt qual quot

question particle/marker quantifier quality quotative

r rcp rcrd rdc.ag real reas rec.pst reconc recp refl reg rel rel.imp RelCl rem.fut rem.pst rem.tns rep res res.pos rest rev rn rpt rsg

recipient recipient postposition/case recordative reduced agent trace realis reason recent past reconciliative reciprocal reflexive regressive relative/relativizer relayed imperative relative clause remote future remote past remote tense repetitive resultative resulting position restrictive reversed relational nominal reportive resignation

s

single argument of canonical intransitive verb speech act participant subject subjunctive sensory deixis separative sequential singular single-actional similative simultaneous Spanish loan spatial

sap sbj sbjv sens sep seq sg sgact sim simul sp spat

Abbreviations

spcl spec src ss stat stat.advr stat.compl stat.nmlz sub suff supe swr symp

speculative specific source same subject stative stative adverbializer stative completive stative nominalization subordinator suffix superessive switch reference mood sympathetic

t tag tel temp temp.nmlz term th

theme semantico-syntactic role interrogative tag telic temporal temporal nominalizer terminative thematic suffix

thm top tr trnc trns trnsl

theme topic transitive truncated element transitivizer translative

undir univ.quant unposs up

undirected motion universal quantifier unpossessed upward

v.neg v.pl vbz vbz.inch vcls vind voc voi vol

negative verb verbal plural verbalizer inchoative verbalizer verb class marker vindictive vocative voice volitional

xiii

Laércio Nora Bacelar and Hein van der Voort

14 Kanoé 1 2 3 4 5 6 7 8 9 10 11

Classification, demography, and sociolinguistic context Phonology Morphology and morphosyntax The noun phrase The verb phrase Syntactic structure of simple sentences Syntactic structure of complex sentences Discourse and information structure Conclusion Acknowledgments References

1 Classification, demography, and sociolinguistic context Kanoé or Kanoê (kano1245) is a Brazilian indigenous language isolate. In historical documents and traditional linguistic literature it is referred to as Canoee and Kanoä (Becker-Donner, 1955) and/or Canoê (Zack, 1943) and Canoé (Pottier 1983: 422, 441).1 It also appears under the names Kapishana or Kapišana (Nimuendajú 1955: 179– 197), Kapixaná, and Capixana (Campbell 1997: 173, 327), names that are not recognized by the remaining population. Although the total number of ethnic Kanoé is approximately 325, the Kanoé language currently has only three speakers (Bacelar 2010): two monolingual adult speakers and a teenage bilingual speaker of Kanoé and Akuntsu (akun1241), the latter being a Tuparí language (Tupi stock). The Kanoé inhabit part of their traditional lands in the municipalities of Corumbiara and Chupinguaia, in the southern region of the State of Rondônia, Brazil, more precisely, in the Rio Omeré Indigenous Land (T. I. Rio Omeré, recognized on 18 April 2006). This small group of Kanoé at the Omeré River was officially contacted in 1995. First contact with the Kanoé people in general occurred in the early 20th century. In the subsequent colonial process, the Kanoé were semi-enslaved and spread throughout the state of Rondônia, losing their language and many aspects of the original culture. The majority of the Kanoé ethnic group now live in the Rio Guaporé

1 The name Kanoé is pronounced as [kanoˈɛ]. Its origin and etymology are unknown. According to traditional mythology, the Kanoé received their name when they were formed out of clay. Laércio Nora Bacelar, Independent researcher Hein van der Voort, Museu Paraense Emílio Goeldi https://doi.org/10.1515/9783110432732-001

660

Laércio Nora Bacelar and Hein van der Voort

14 Kanoé

661

and Sagarana Indigenous Lands, municipality of Guajará-Mirim, Rondônia, on the border between Brazil and Bolivia, while a smaller part of the ethnic group lives in the Rio Branco Indigenous Land, municipality of Alta Floresta d’Oeste, Rondônia. Among the Kanoé of the Rio Guaporé and Sagarana Indigenous Lands, there were until recently also two potential speakers, very elderly, who no longer had the opportunity to use Kanoé and retained it only in their memory. Neither has had the opportunity to use the language since they were children and, for this reason, it was very difficult for them to maintain a conversation or produce a narrative in Kanoé. Oral testimonies (Bacelar 1992, 2004) – supplemented by early documents and maps (e.g., Nimuendajú 1955; Zach 1943; Becker-Donner 1955: 297; Caspar 1975: 8, Dequech 1942) – report that the Kanoé people have always lived in the region between the headwaters of the rivers Tanaru (a tributary of the Pimenta Bueno River) and Omeré (tributary of the Corumbiara River), in southern Rondônia. The T. I. Rio Omeré, where the last three speakers currently live (shared with the Akuntsu), represents only part of the Kanoé people’s original territory. Kanoé territory is part of a wider region that spans the border between Rondônia and the Bolivian Amazon in which approximately 40 languages ​still survive, ten of which are considered isolates (see Crevels & van der Voort 2008). In this region, there are several multilingual cultural complexes; Kanoé culture can be considered part of the “Marico cultural complex” (cf. Maldi 1991), together with, for example, the cultures of the Aikanã (aika1237), Sakurabiat (saki1248), and Kwaza (kwaz1243). Although the languages of these neighboring ​groups share with Kanoé – and with each other – several similar phonetic, lexical-semantic, and (sparse) morphological properties (e.g., nominal classifiers, person marking traits), these similarities are not sufficiently consistent to support the hypothesis of a genealogical relationship among them (van der Voort 2005). The first Kanoé documentation of Kanoé comes from the first half of the 20th century. Word lists were collected by Curt Nimuendajú in 1928 (published in 1955) and by Stanislau Zack in a 1943 manuscript. Etta Becker-Donner (1955) published a short ethnographic description and a word list, Willem Bontkes collected a word list in 1967, and Denny Moore collected a word list in 1988. An audio CD with stories was published by Galucio et al. (2008). A comprehensive language description and documentation initiatives began in the 1990s with Laércio Bacelar’s projects, resulting in several works (Bacelar 1992; Bacelar 2005; Bacelar & Silva 2003), a descriptive grammar (2004), and a substantial collection of audiovisual cultural and linguistic documentation (2009–2013).

2 Phonology This section presents Kanoé’s phonological system, which is composed of 26 phonemes; its syllable structure; and its prosodic phenomena.

662

Laércio Nora Bacelar and Hein van der Voort

2.1 Vowel phonemes The vowel subsystem is composed of seven oral and seven nasal vowels, as shown in Table 14.1. Note that /ɨ/, which is represented orthographically as , is realized phonetically as IPA [ɨ] or [ə]. Its nasal counterpart /ĩ ̵/, realized as [ĩ ̵] or [ə̃] and represented as , is rare. The phonemes /e/, /ẽ/, /o/ and /õ/ tend to be realized as [ɛ], [ɛ̃], [ɔ] and [ɔ̃], respectively. Tab. 14.1: Kanoé vowel phonemes.

close close-mid open

front unrounded

central unrounded

back rounded

oral

nasal

oral

nasal

oral

nasal

i e æ

ĩ ẽ æ̃

ɨ

ĩ ̵

u o

ũ õ

a

ã

2.2 Consonant phonemes The Kanoé consonant subsystem has 12 phonemes, as shown in Table 14.2. The phoneme /w̟ / varies phonetically between [w] and [ß] and contrasts with /w/: [w̟ æˈtʃi] ‘heavy rain’ - [wæˈtʃi] ‘grass’. Tab. 14.2: Kanoé consonant phonemes.

plosive nasal affricate fricative flap approximant

bilabial

alveolar

palatal

p m

t n ts

ɲ

velar k

w̟

h ɾ

w

j

The glottal stop [ʔ] has no phonemic status. Glottalization occurs occasionally on the phonetic level between adjacent identical vowel phonemes. Thus, /kwaatsie/ ‘frog’ becomes [kwaʔatʃi’e].

2.3 Syllabic patterns Kanoé syllable patterns include: /V/, /VC/, /CV/, /CVC/, /CCV/. As /V/ and /CV/ occur most frequently and do not present phonotactic restrictions, (C)V can be taken

14 Kanoé

663

as the canonical syllable, with other syllable patterns resulting from insertion of a glide, either /w/ or /j/. The syllable configurations of Kanoé can be summarized in the general formula: (C1) (C2) V (C3), in which: a) in (C1) any consonants can occur, except /w/ and /j/; b) in (C2) only /w/ or /j/ can occur; c) in V, any vowel may occur; d) in (C3) only /w/ or /j/ appear and, in restricted cases, /m/ or /n/, due to morphophonological processes (see Bacelar: 2004, 90–102).

2.4 Prosodic phenomena Kanoé is a non-tonal language, since pitch does not alter the semantic content of morphemes. Changes in intonation have an expressive function in relation to verbal mood affect. In daily conversation intonation is generally level, but it can gradually rise, due to expressive needs, such as, for example, emphasis on certain words, intensification of verbal expressions or to suggest greater or lesser spatial-temporal distance from referents. The vowel quantity or duration has no phonological function or grammatical or lexical-semantic implications. Vowel duration is expressive. Stress is fixed and predictable, systematically falling on the last syllable of nominal or verbal words, thus having a merely demarcating function (Weiss 1980: 68). It delimits boundaries between words and within verbal constructions, as in (1) through (3). (1) (2) (3)

[kwiˈkaj] [jaˈtɔ mũȷ ]̃ [ɲa ikuˈta iw̟ aeˈɾe]

/kwikaj/ /jato muj/ /ɲa ikuta iw̟ a eɾe/

‘sun’ ‘our mother’ ‘My head hurts.’

In cases of derivational suffixation, as in (4) versus (5) and (6); or as part of a syntactic operation, (7) versus (8), main word stress falls on the suffix. (4) (5) (6) (7) (8)

[mũtɨˈɾɨ] [mũtɨɾɨˈkɔ] [mũtɨɾɨkɔˈmũ] [piˈkɔ] [tʃupikɔˈtɔ nɛˈɾɛ]

/mutɨɾɨ/ /mutɨɾɨ-ko/ /mutɨɾɨ-ko-mu/ /piko/ /tsupikoto n-e-ɾe/

‘açaí’ ‘açaí fruit’ ‘açaí juice’ ‘nail’ ‘He is scratching.’

In addition to the existence of nasal vowels in the phonological system, nasalization of vowels at the phonetic level can occur when adjacent to a nasal consonant, although this does not always happen. In other cases, nasal spreading is attested, that is, the occurrence of a nasal vowel in some words can result in the nasalization of some or all the other vowels and semivowels in the same word.

664

Laércio Nora Bacelar and Hein van der Voort

(9) [mamãȷ ̃tɔˈkɔ] ~ [mãmãȷ ̃tɔˈkɔ] (10) [nĩõpuˈo] ~ [nĩõpũˈõ]

/mamajtoko/ /niopuo/

‘hawk’ ‘red-and-black ant sp.’

3 Morphology and morphosyntax 3.1 Word classes Kanoé word classes include nouns (free and bound), pronouns (personal, possessive, demonstrative, and indefinite), numerals and quantifiers, adverbs, and stative verbal roots (with descriptive or qualifying content) or dynamic verbal roots (referring to actions). In addition, Kanoé also has some grammatical particles. Kanoé does not have a discrete class of adjectives. Instead, descriptive and quality concepts are expressed by verb roots. There are no definite or indefinite articles or adpositions. Verbal roots head verbal structures, which are morphosyntactically complex, with obligatory person and number agreement and mood marking and with optional incorporation2 of nominal classifiers that exhibit anaphoric functions. Furthermore, nouns can be derived from verbs through nominalization, and verbal structures can be derived from nouns through verbalization. The only open classes in Kanoé are nouns and verbs.

3.2 Morphological typology Kanoé is a polysynthetic, agglutinating, and predominantly suffixing language. Verbs are often rather complex, using productive transitivizing and nominalizing derivations and incorporation of nominal classifiers and bound nominal roots with very specific semantic content. There are no derivational prefixes, but a specific class of verbal roots requires prefixation of inflectional person and number markers (see Section 4.1.1). Otherwise, inflectional morphology tends to be suffixed.

3.2.1 Word formation processes The most productive processes of word formation in Kanoé are (a) compounding; (b) derivation; and (c) reduplication.

2 In this article we use the term “incorporation” informally to refer to classifiers and nouns that form part of verbs. In Kanoé, the insertion of these semi-lexical elements can be considered a derivational operation that does not modify the verb’s argument structure.

14 Kanoé

665

3.2.1.1 Composition of lexical roots Compound words can contain multiple juxtaposed lexical roots, either nominal as in (11) and (12), or nouns combined with stative verbs as in (13) through (15). (11) oky-kũti snake-mortar ‘anaconda’ (Eunectes sp.) (12) ary-tepy bird-ground ‘grey-winged trumpeter’ (Psophia crepitans) (13) ej-turo-e big-space-nmlz ‘distance’ (14) uru-jati new-flower ‘flower bud’ (15) uru-kani new-child ‘baby, newborn, little child’ Bound nominal roots form a special category; they require the noun-forming root i (glossed er) in order to act as free nouns. They can also occur as second element in combination with verbal or nominal roots (Section 4.3.3). Example (16) shows a compound involving a noun as first element. (16) ytse + i-katsi > ytse-katsi tree er-root tree-root ‘tree’ ‘root’ ‘tree root’

3.2.1.2 Derivational affixes Derivational processes involve several types of affixes, including a complex system of nominal and verbal classifier suffixes that produce derived nouns and verbs. Note the examples of nominal classifiers in (17) through (20). (17) pætæj ‘buriti palm tree’ (Mauritia flexuosa)

666

Laércio Nora Bacelar and Hein van der Voort

(18) pætæj-ko buriti-clf:fruit ‘buriti coconut’ (19) pætæj-ko-mu buriti-clf:fruit-clf:liquid ‘juice or fermented beverage made of buriti nuts’ (20) aj kometaw uru-taw o-õ-taw õ-e-re 1sg.pro bean new-clf:bean eat-1-clf:bean 1-decl-aux ‘I eat green beans.’ (Lit. ‘I am eating green beans.’) Other derivational processes include sex marking (not grammatical gender, see Section 4.3.1) as in (21), optional collective or plurality marking (which is not a grammatical plural, see Section 4.2.1) as in (22), and nominalization (see Section 4.4.2), expressed by the suffix -e, as in (23). (21) e uru-nake woman new-f ‘girl, unmarried woman’

(22) uruã-te boy-coll ‘boys’

(23) jato u-ro-e topi e-re poss.1pl eat-ag-nmlz rotten decl-aux ‘Our food is rotten.’ 3.2.1.3 Reduplication Reduplication is productive in many verbal roots, denoting iterativity, that is, iterative aspect, shown in (24) and (25). (24) aj oroe mana~mana õ-e-re 1sg.pro clay press~red 1-decl-aux ‘I am kneading clay.’ (25) aj tsero mañu~mañu õ-e-re mænæræ-nĩ 1sg.pro chicha squeeze~red 1-decl-aux sieve-obl ‘I am straining chicha3 with the sieve.’ The reduplication in the examples above is iconic, of the repetition of the action denoted by the verb. Non-reduplicated use of, for example, mana means ‘press, tighten, squeeze’, as in (26). 3 Traditional fermented beverage based on manioc, yam, maize etc.

14 Kanoé

667

(26) ña i-kwamu mana-to e-re poss.1sg er-arm press-tr decl-aux ‘I squeezed my arm.’ It is important to distinguish reduplication as a productive process of word formation from repetition of syllables, as in (27) and (28). Although in these cases the word also iconically symbolizes an action, a sound, or another repetitive aspect, there is no non-repeated form of the root. (27) kwaatsie pæ̃ pæ̃ -e-ro n-e-re frog jump-nmlz-ag 3-decl-aux ‘The frog is hopping.’ (28) Joãw kykyo papa-ro-to n-e-re João nail beat-ag-tr 3-decl-aux ‘João is hammering a nail.’ Syllable repetition in the names of certain animals is clearly onomatopoeic, mimicking or symbolizing the sound they typically emit, as in (29) and (30). (29) kiki (30) kẽkẽ

‘parakeet’ (Psittacidae sp.) ‘white-eared puffbird, jucuru’ (Nystalus chacuru)

However, not all syllable repetition is onomatopoeic, as some may symbolize other characteristic features of a referent, such as flashing light, multiplicity of thorns, typical behavior, distribution, etc., as in (31) through (34). (31) (32) (33) (34)

varyvary pũpũ kunukunu perepere

‘star’ ‘pente de macaco tree’ (Apeiba tibourbou sp.) ‘ring-tailed coati’ (Nasua nasua) ‘alate (winged) termite’

3.2.2 Inflectional morphology Kanoe does not use nominal inflection aside from the possessive morpheme -o (see Section 4.5.2) and the oblique case marker -ni (Section 4.5.3). Verbs, on the other hand, are obligatorily inflected for person (Sections 4.1.1, 5.1, and 5.2) and mood (Section 6.2). Inflectional morphology occurs word finally, with the exception of person marking on a specific set of verbs (Section 5.1.1). Furthermore, verbal inflection is often distributed over the main verb root and the auxiliary complex (Section 5.2). This may result in double marking of subject person, which has a progressive temporal-aspectual function (Section 5.3.1). The same person markers express object

668

Laércio Nora Bacelar and Hein van der Voort

person, but only on the main verb and always to the right of the subject person marker (Section 5.5.3). Third person is usually unmarked, but in the progressive aspect-tense it is expressed by the morpheme n-, which occurs mainly prefixed on the auxiliary complex.

3.3 Syntactic order Kanoé is predominantly head-final. The order of noun phrase constituents is discussed in Section 4.5. The order of verb phrase constituents is discussed in Sections 4.5 and 6.1. In Section 7.4, iconic aspects of subordinate clause ordering are mentioned.

4 The noun phrase 4.1 Pronouns and demonstrative distinctions 4.1.1 Personal pronouns and verbal agreement Kanoé has a symmetrical paradigm of free personal pronouns in which the plural forms are derived by suffixation of collective -te (also applicable to certain common nouns) to the singular forms, as in Table 14.3. The use of a free pronoun is in principle redundant if the argument is also marked inflectionally on the verb, as we will see later. In the phrasal environment, free personal pronouns generally exercise the syntactic functions of subject (S) or object (O), as observed in the contrast between the pairs of examples in (35) through (40). (35) aj mi i-memu-ro õ-e-re 1sg.pro 2sg.pro 1-like-ag 1-decl-aux ‘I like you.’ Tab. 14.3: Free personal pronouns. person

singular

1 2 3

aj mi oj oj nake4

plural ‘I’ ‘you’ ‘s/he’ ‘she’

ajte mite ojte ojte nake

‘we’ ‘you’ ‘they’ ‘they (f)’

4 The female forms of 3sg or 3pl, marked by nake feminine are less common and are used only when the speaker wants to emphasize the sex of a human referent.

14 Kanoé

669

(36) ajte oj i-memu-ro õ-e-re 1pl.pro 3sg.pro 1-like-ag 1-decl-aux ‘We like him.’ (37) mi aj pi-memu-ro mi-tsi 2sg.pro 1sg.pro 2-like-ag 2-int ‘Do you like me?’ (38) mite oj pi-memu-ro mi-tsi 2pl.pro 3sg.pro 2-like-ag 2-int ‘Do you like him?’ (39) oj aj memu-ro k-e-re 3sg.pro 1sg.pro like-ag neg-decl-aux ‘He doesn’t like me.’ (40) ojte mite memu-ro e-re 3pl.pro 2pl.pro like-ag decl-aux ‘They like you.’ Bound verbal first- and second-person morphemes agree with their respective free pronouns, as in (35) through (38). The third person, either singular or plural, is usually unmarked, or null. Prefixation of subject person markers, at the beginning of the verbal structure, is restricted to verbs that denote mental processes (understanding, knowing, thinking etc.) or affect (liking); they may be intransitive, as in (41), or transitive, as in (42) and (43).5 In addition, the auxiliary complex always takes prefixed person markers, as many examples show (see also Section 5.2). (41) aj i-pateñu õ-e-re 1sg.pro 1-know 1-decl-aux ‘I know.’ (42) aj mi i-pateñu-pe-to õ-e-re 1sg.pro 2sg.pro 1-know-2-tr 1-decl-aux ‘I know you.’ (43) mi ũko itevæ pi-patenũ-to k-e-re 2sg.pro dem.dist person 2-know-tr neg-decl-aux ‘You don’t know that man.’ 5 This property of Kanoé verb inflection is remarkably similar to that of the class of verbs that allow prefixes in the neighboring language Aikanã (isolate).

670

Laércio Nora Bacelar and Hein van der Voort

Tab. 14.4: Personal pronouns and inflections.6 person

number

pronoun

inflection prefix

suffix

prefix

1

sg pl

aj ajte

i-

-õ

õ-

2

sg pl

mi mite

pi-

-p, -pe, -mi

p-, pe-, mi-

3

sg pl

oj ojte

ø-

-ø, -n

ø-, n-, ni-

distribution scope

verb root main verb root

(neg)-decl-aux auxiliary

However, with other verbs, both stative and active, personal markers are always suffixed, following the verbal root, as in (20), (44), (45) and many other examples in this article. Table 14.4 displays all pronouns and person markers. For both intransitive and transitive verbs, if the verbal form is clearly marked by person inflection, as with first and second person, overt subject and direct object pronouns, as in (41) and (42), respectively, can be omitted, leaving the respective syntactic position in the sentence structure empty, as (44) and (45) show. (44) vara-õ-ro õ-e-re speak-1-ag 1-decl-aux ‘I am speaking.’ (45) vara-õ-ro-pe-to õ-e-re speak-1-ag-2-tr 1-decl-aux ‘I am speaking with you.’

4.1.2 Possessive personal pronouns Expressing possession involves the suffix -o (see Section 4.5.2). This suffix also occurs in possessive pronouns, and it is especially transparent in the case of the singular, possessive third person ojo (based on the third-person pronoun oj). Table 14.5 shows the complete set of possessive pronouns.

6 Verbs that denote natural phenomena (rain, wind, thunder, lightning etc.) are impersonal and have the unmarked third-person form.

14 Kanoé

671

Tab. 14.5: Possessive pronouns. person

singular

1 2 3

ña pja ojo

plural

‘my, mine’ ‘your, yours’ ‘his, her, hers’

jato pjato ojoto

‘our, ours’ ‘your, yours’ ‘their, theirs’

As seen in Table 14.5, the system of possessive personal pronouns is similar to that of free personal pronouns, and likewise, plural forms are derived from singular forms. In the plural forms, collective -te and possessive -o have apparently undergone the process of assimilation illustrated by (46). (46) jato muj < ja-te-o muj poss.1pl mother poss.1sg-coll-poss mother ‘our mother’

4.1.3 Indefinite pronouns Kanoé has four indefinite pronoun roots: nuvi ‘someone, nobody’, tsyke ‘other(s)’, naj ‘something, some’, and tsoke ‘nothing’. When referring to human beings, indefinite pronouns may receive optional gender marking with masculine -kỹj and feminine -nake, which refer to the sex of the referent. One of the indefinite pronouns, tsoke ‘nothing’, is negative and refers exclusively to the absence of non-human beings. For negative indefinite reference to humans, positive indefinite nuvi is used in a negative grammatical context. Table 14.6 displays all categories. Indefinite pronouns can have the syntactic functions of subject or object, as in (47) through (51). (47) nuvi mõ-kỹj e-re indf sleep-eye decl-aux ‘No one slept.’/‘Someone slept.’

Tab. 14.6: Indefinite pronouns. meaning/use

human beings

positive

nuvi (-m/-f)

non-human beings and things ‘who, someone, no-one’

negative alternative

tsyke (-m/-f)

‘other’

naj

‘some, something’

tsoke

‘nothing’

tsyke

‘something else’

672

Laércio Nora Bacelar and Hein van der Voort

Tab. 14.7: Interrogative pronouns. human beings nuvi-tsi

‘Who?’

things naj-tsi

‘What?’

(48) nuvi ja-kjũ k-e-re indf want-npr neg-decl-aux ‘No one likes him.’/‘Someone does not like him.’ (49) nuvi mõ-kỹj k-e-re indf sleep-eye neg-decl-aux ‘Someone did not sleep.’ (50) tsyke-nake oj ja-kjũ k-e-re indf-f 3sg.pro want-npr neg-decl-aux ‘The other one does not like him.’ (51) oj tsyke-nake ja-nake e-re 3sg.pro indf-f want-f decl-aux ‘He likes the other one (woman).’ Two of the indefinite pronouns can have interrogative functions when used with interrogative -tsi (Section 6.2.2), and they can occur in focus position at the beginning of the sentence, as in (52), (53), (182), and (196). Table 14.7 shows their meanings. In an interrogative sentence, in which a possessor is questioned, the indefinite pronoun nuvi receives the possessive marker -o, and then the interrogative marker -tsi, resulting in nuviotsi, a form translatable as ‘Whose (is it)?’. Compare (52) and (53). (52) nuvi-tsi jũ mapi po-ro-e7 indf-int dem.prox arrow make-ag-nmlz ‘Who made this arrow?’ (53) nuvi-o-tsi jũ mapi-ka-re indf-poss-int dem.prox arrow-clf:hard-aux ‘Whose is this arrow?’

7 Alternatively, the interrogative marker can also occur at the end of the sentence: nuvi jũ mapi po-ro-e-tsi.

14 Kanoé

673

4.1.4 Demonstrative pronouns There are only two demonstrative determiners with deictic functions: jũ ‘this’, for living beings and objects close to the speaker, and ũko ‘that’, for living beings and objects remote from the speaker, as summarized in Table 14.8.

Tab. 14.8: Demonstrative pronouns. proximate

distal

jũ

ũko

Both pronouns are invariant in terms of gender and number and occur to the left of the head they modify, as shown by (54) and (55). (54) jũ uruã ereã k-e-re dem.prox boy big neg-decl-aux ‘This boy is small.’ (Lit. ‘This boy is not tall.’) (55) ũko ævo ereã e-re dem.dist man big decl-aux ‘That man is big.’

4.1.5 Non-present demonstrative There is a bound demonstrative morpheme that does not express any spatial meaning or person/number/gender agreement, -kjũ. It refers to a third person who is absent. Etymologically, it appears to be a fossilization of negative k- with the proximate demonstrative -jũ. This form is used as an anaphoric pronoun, and it also occurs as a syntactic and redundant element of subordination, functioning as a kind of copula for clauses that are embedded in more complex syntactic structures (Section 7.3.3). In its demonstrative capacity, as in (56) and (57), -kjũ creates a type of relative clause that modifies one of the predicate’s arguments. In other occurrences of -kjũ, as in in example (61), it refers to the head of the modified argument. (56) oj po-ro-e-kjũ munaw e 3sg.pro make-ag-nmlz-npr hammock woman ‘Who made the hammock is a woman.’ (Lit. ‘That one who made the hammock (is) a woman.’)

674

Laércio Nora Bacelar and Hein van der Voort

(57) aj tsere-õ-ro-kjũ kjũ-re itevæ pæ-kuta kjũ-re 1sg.pro see-1-ag-npr npr-aux person white-head npr-aux ‘I met a bald person.’ (Lit. ‘That person that I met, that (is) one who (is) bald.’)

4.2 The expression of nominal number, numerals, and quantifiers 4.2.1 Nominal number In Kanoé, collective -te, which conveys the idea of plurality or collectivity, is not used very frequently. It is used for explicit reference to sets of beings or items and, although productive, as in (62), it is not an obligatory grammatical morpheme. (58) uruã-te boy-coll ‘many boys, a group of boys’ (59) kanĩ-te > kante / kãte child-coll children ‘many children, a bunch of children’ Example (59) shows that -te may be lexicalizing in frequently used expressions. As seen in Tables 14.3 and 14.4, -te is also used as a pronominal plural marker.

4.2.2 Numerals and quantifiers Kanoé has two simplex numerals, pja [pja] ‘one’ and mõw [mõw̃ ] ‘two’, from which the numbers from three to nine can be constructed via addition, as shown by Table 14.9. For the numbers 10, 15, and 20, the logical basis is the fingers of one or both hands (itso ‘hand’) and one or two feet (itsotsi ‘foot’). In practice, however, there are hardly any situations in which numerals referring to quantities or measures

Tab. 14.9: Numerals. 1 2 3 4 5

pja mõw mõw pja mõw mõw mõw mõw pja

 6  7  8  9 10

mõw mõw mõw mõw mõw mõw pja mõw mõw mõw mõw mõw mõw mõw mõw pja mõw itso / itso mõw

14 Kanoé

675

greater than ten are, in fact, used, given the non-acquisitive character of Kanoé culture. The numbers six to ten were exclusively attested in elicitation. Preceding a noun, the high frequency term for indefinite quantity is arakere, which in fact represents a fixed litotic expression: ara-k-e-re (few/little-neg-declaux) = ‘many/much’ (Lit. ‘there are/is not few/little.’). Numerals and the arakere ‘many/much’ quantifier (or intensifier) occur as determiners in a noun phrase, as in (60) through (65). (60) aj de õ-e-re pja-ro-k-e uromu 1sg.pro kill 1-decl-aux one-ag-neg-nmlz cayman ‘I killed several caymans.’ (Lit. ‘I killed not (just) one cayman.’) (61) aj kani mow-kani e-re 1sg.pro child two-child decl-aux ‘I have two children.’ (62) aj mow-kỹj kani õ-tsi mo-e-re 1sg.pro two-m child 1-have appl-decl-aux ‘I have two sons.’ (63) oj mow-nake kani tsi mo-e-re 3sg.pro two-f child have appl-decl-aux ‘He has two daughters.’ (64) arakere kani tevaj e-re much child play decl-aux ‘Many children are playing.’ (65) oj arakere kani tsi mo-e-re 3sg.pro much child have appl-decl-aux ‘He has many children.’ The term arakere is not only used as a quantifier but also as an adverb of intensity, when modifying the verbal phrase, either in initial or in final position, as in (66) and (67). (66) aj iriri-ro e-re arakere 1sg.pro run-ag decl-aux much ‘I ran much.’ (67) oj ore-ry e-re arakere 3sg.pro tire-refl decl-aux much ‘He is very tired.’

676

Laércio Nora Bacelar and Hein van der Voort

4.3 Nominal classification 4.3.1 Gender and sex Kanoé has no gender inflection. For human referents, biological sex is semantically inherent to nouns, as in ævo ‘man, male, husband’ versus e ‘woman, female, wife’, papa ~ pa ‘father’ versus mũj ‘mother’, tsusu ‘grandfather’ versus keke ‘grandmother’, etc. In gender-neutral nouns, such as kanĩ ‘child’, the same terms are usually preposed in the noun phrase to express sex: ævo kanĩ ‘boy’ versus e kanĩ ‘girl’, as in (68) and (69). However, in the case of pronouns, like in (70) and (71), and animal referents, the same terms are postposed: kurakura ævo ‘rooster’ versus kurakura e ‘hen’, as in (72) and (73). (68) ævo kani tũvo n-e-re man child fall.ill 3-decl-aux ‘The boy is sick.’ (69) e kani tũvo n-e-re woman child fall.ill 3-decl-aux ‘The girl is sick.’ (70) ojte ævo ã-kỹj more-kỹj e-re 3pl.pro man tall-m good-m decl-aux ‘They are tall and handsome.’ (71) ojte nake ã-nake more-nake e-re 3pl.pro woman tall-f good-f decl-aux ‘They are tall and beautiful.’ (72) kurakura ævo vaj~vaj-ro e-re chicken man sing~red-ag decl-aux ‘The rooster is crowing.’ (73) kurakura e vaj~vaj-ro e-re chicken woman sing~red-ag decl-aux ‘The hen is cackling.’ There are, however, three nominal classifiers that mark gender: masculine -kỹj / -kwæ̃ versus feminine -nake. The masculine form is always bound, while feminine -nake can occur as a free form, synonymous with e ‘woman’: nake ‘girl, woman, single woman’. However, in the case of the two masculine forms, -kỹj also occurs in verbal structures, having a wider distribution than -kwæ̃ , which occurs only as a

14 Kanoé

677

masculine suffix in kinship terms, (74) and (75). These classifiers may also be employed to form nouns denoting mythical entities, as in (76) and (77). (74) tu-kwæ̃ older.sibling-m ‘older brother’ (75) tu-nake older.sibling-f ‘older sister’ (76) aki-moã-kỹj stone-spirit-m ‘male stone spirit’ (77) aki-moã-nake stone-spirit-f ‘female stone spirit’ Like other nominal classifiers (Section 4.3.2), these morphemes can play a role in verbal subject agreement, and they occur frequently. The gender may be inherent in nouns with human referents, as in (78) and (79), or even explicitly marked in the case of pronouns, as seen in (70) and (71). (78) jato papa tẽpũ-kỹj e-re poss.1pl father old-m decl-aux ‘Our father is old.’ (79) ña muj tẽpũ-nake e-re poss.1sg mother old-f decl-aux ‘My mother is old.’

4.3.2 Nominal classifiers As detailed in Bacelar (2004: 119–135), and similar to other Amazonian languages (Grinevald & Seifart 2004; Seifart & Payne 2007; van der Voort 2015, 2018), Kanoé also has a complex system of nominal classifiers. These are bound elements that refer to salient or physical properties of nominal referents. As demonstrated by (81) and (83), a word may bear several classifiers.

678

Laércio Nora Bacelar and Hein van der Voort

(80) atiti-kaña maize-clf:flour ‘maize flour’ (81) atiti-tekwa-kaña maize-clf:seed-clf:flour ‘rice flour’ (82) mapi-ka arrow-clf:hard ‘bow’ (83) mapi-ka-ty arrow-clf:hard-clf:thread ‘bowstring’ Classifiers can establish agreement relationships between nouns and verbs, shown in (84). (84) aj epy-kwã o-õ-kwã õ-e-re 1sg.pro nut-clf:small eat-1-clf:small 1-decl-aux ‘I eat peanuts.’ Kanoé has two types of morphemes that can function as nominal classifiers: true nominal classifiers and bound nouns. Table 14.10 shows a number of true nominal classifiers of Kanoé. Nominal classifiers form a small group. They are productive morphemes with a wide distribution, as they occur in a variety of nominal and verbal constructions. Some of them also occur as free nouns. The classifier -taw ‘bean’, as in (20), is probably derived from the final syllable of the corresponding noun with the same meaning, while the etymology of the others is not so clear. In addition to classifiers, there are some recurring fixed elements in certain words, such as -ky, which is found in words referring to conical and elongated objects and insects, for example, taruaky ‘stick insect’, which is possibly etymologically related to kyj ‘sting’. Another element often found is -nu / -ñu in names referring to sharp objects and protuberances, as in numunu ‘club, cudgel’ and ikañu ‘nose, bird beak’. Such elements occur regularly in the Kanoé lexicon, but they are not productive and do not occur in typical environments for nominal classifiers. Due to the semantics of the words in which such elements are found, it can be assumed that they represent ancient classifiers which today are fossilized in certain roots and are not synchronically analyzable.

14 Kanoé

679

Tab. 14.10: Nominal classifiers. classifier

observations

-ka

classified items

example

wood, hard as wood

mapi-ka arrow-clf:wood ‘bow’

-kaj

< kwikaj ‘sun’

sun; perhaps also as part of a fixed expression in verbs for ‘roast’

itsaj-kaj night-clf:sun ‘nightfall’ pejkaj ‘roast’

-kaña

< kaña ‘flour’

flour, porridge, paçoca (crushed mixture)

atiti-kaña maize-flour ‘maize flour’

-ky

fruits; fixed in several fruit names

po-væ-ky eat.fruit-du-clf:fruit ‘we eat fruit’ tɔky ‘papaya’

-ko

coconuts, hard shell fruits, etc.; fixed in many insect names

karana-ko patauá-clf:fruit ‘patauá fruits’ nãko ‘beetle’

-kwa

small objects; fixed in some nouns

etsivi-kwa-e urinate-clf:small-nom ‘bladder’ pyryrykwa ‘piaba fish’

-mu

liquid, beverage

æj-mu leaf-clf:liquid ‘tea’

-taw

< kometaw ‘bean’

bean, green bean pod

uru-taw new-clf:bean ‘green bean’

-tekwa / -tikwa

< tekwa ‘seed’

seed, grain, kernel, etc.

atiti-tekwa maize-seed ‘rice’ mapi-tikwa arrow-clf:seed ‘cartridge of firearm’

-tæmu

beiju pancake (i.e. pancake made of dry flour)

atiti-tæmu maize-clf:beiju ‘beiju of maize’

-tinu

soup, paste, mud, etc.

ivo-tinu yam-clf:paste ‘yam soup’

-ty

thread, line, cord, vine

oti-ty cotton-clf:thread ‘cotton thread’

680

Laércio Nora Bacelar and Hein van der Voort

4.3.3 Bound nominal roots The other group of morphemes with classificatory capacities is much larger than the set of classifiers described in Section 4.3.2. It is a group of approximately 80 bound nominal roots, most of which refer to body parts, such as -kuta ‘head’, -kỹj ‘eye’, -taw ‘tongue’, -ako ‘back’, -ry ‘liver’, and -tso ‘finger’; some refer to parts of plants, such as -ati ‘flower’, -ki ‘branch’, -katsi ‘root’, and -tæ ‘concave, bark’. To act as independent nouns, they must be attached to the morpheme i-, which we consider as a semantically empty root. Since this element appears as part of nouns that refer mainly to parts of the human or animal body, it may be analyzed historically as an inherent possession prefix (see Section 4.5.2). Bound nominal roots also act as productive classifiers and occur in the same environments as nominal classifiers. Example (85), involving the verb ‘wound’, shows classifier agreement similar to (84). (85) opera voro-n-e-o i-kuta rwa-kuta e-re jaguar black-3-nmlz-poss er-head wound-head decl-aux ‘The head of the black jaguar is wounded.’ A number of bound noun roots appear to be composed of several classificatory morphemes, such as -kañutinu ‘mucus’ < -kañu ‘nose, beak’ + -tinu ‘paste’; -kỹjmu ‘tear’ < -kỹj ‘eye’ + -mu ‘liquid’. Several cases are not completely transparent, and the origin of only one part is known, such as -tæ ‘concave, round’ in -tæmu ‘beiju, paddle’ and -kyetæ ‘bark, shell’; and -ati ‘flower’ + -tæ ‘concave’ in -atitsetæ ‘petal’; and -tekwa ‘belly’ in -oroetekwa ‘kidney’.

4.3.4 The intransitive and agentive verbal classifiers -o and -ro Kanoé has two suffixes that we consider verbal classifiers, since they occur only within the verbal complex and are attested only with certain verbs. Their distribution is therefore different from the (wide) distribution of nominal classifiers. The intransitive classifier -o occurs only with a very limited number of verbs, mainly process verbs such as tũ-o-e-re ‘die’, where it is segmentable, and tũvo-n-e-re ‘fall ill’, where it is fossilized. Agentive -ro occurs only in verbal structures that, from a semantic point of view, denote voluntary actions attributable to an agent. It also less frequently occurs on verbs that denote physiological processes, movements or displacements in space, and on those denoting some natural phenomena. In other words, a good part of verbal forms that are actually active, that is, interpretable as actions that are attributable to an agent or to a force – as in vuj ‘blow (wind)’ shown in (221) – require this suffix. It appears immediately following subject inflection, which may be zero, as in (86), and before other morphemes such as object inflection and the transitivizer,

14 Kanoé

681

as in (28), (45), (114), (128), and (212), if any are present.8 In some verb roots, such as po-ro ‘make’ it is probably lexicalized. (86) opera i-utã o~o-ro n-e-re jaguar er-bone eat~red-ag 3-decl-aux ‘The dog is gnawing on a bone.’

4.4 Other issues related to the noun: Diminutives, intensity, and nominalization 4.4.1 Diminutive and augmentative Kanoé has a very high-frequency nominal diminutive suffix -tsĩkwa which is composed of the sequence of -tsĩ ‘little, thin, narrow’ and -kwa clf:small. As a bound form, this suffix is ​highly productive in the formation of nouns that refer to species and subspecies of animals and plants: tekoko ‘pacu fish’ (Myloplus rhomboidalis) > tekokotsĩkwa ‘pacumirim fish’ (Mylossoma duriventris); opera ‘jaguar’ > operatsĩkwa ‘margay cat’ (Leopardus wiedii); õkwa ‘bird’ > õkwatsĩkwa ‘little bird’, etc. In addition, diminutive -tsĩkwa also applies to human beings and material objects in general, expressing a size smaller than “normal”: e ‘woman’ > etsĩkwa ‘little woman, little girl’; kani ‘child’ > kanitsĩkwa ‘little child, baby, infant’; tekwa ‘seed’ > tekwatsĩkwa ‘tiny seed’, etc. Furthermore, the diminutive meaning of the suffix -tsĩkwa itself can be intensified to varying degrees, corresponding to the duration of the nasal vowel ĩ: [ĩ] ~ [ĩː] ~ [ĩːːː] ~ [ĩːːːː]. Thus, the longer the duration, the smaller the referential object. Likewise, in other contexts, -tsĩ ‘little, thin’ can mean ‘little/thin; very little/thin; very very little/thin, etc’, according to the duration of the nasal vowel. There is no augmentative morpheme, but the verbal root ej ‘large, very’, is usually part of such expressions, in combination with nouns; see (87) through (89). (87) ava ej-kañu macaw big-beak ‘red-and-green macaw’ (Ara chloropterus) (88) pærætæ ej-rũ pan big-bowl ‘big pan’

8 Alternatively, -ro might be considered a generic classifier that occurs instead of a specific nominal classifier. It is never found in combination with a nominal classifier. However, its distribution is narrower than that of the nominal classifiers, which occur on both nouns and verbs.

682

Laércio Nora Bacelar and Hein van der Voort

(89) ej-turo e-re big-space decl-aux ‘It is a (relatively) large space/distance/time.’ Given the productivity of litotes in Kanoé (Section 6.3.2), the direct opposite of the diminutive is expressed through negation, compare (90) and (91). (90) ej e-re big decl-aux ‘It is big.’ (91) ej k-e-re big neg-decl-aux ‘It is small.’ (Lit. ‘It is not big.’)

4.4.2 Nominalization As mentioned in Section 3.2.1, one derivational process in Kanoé is nominalization. Nominalization of verb phrases is very productive and is expressed by the nominalizer -e. The nominalized verb can perform any grammatical function, like other nouns, such as that of subject in (92c) and that of object in (93c). (92) a. u-ro n-e-re eat-ag 3-decl-aux ‘He eats / is eating.’

b. u-ro-e eat-ag-nmlz ‘food’

c. jato u-ro-e muj n-e-re poss.1pl eat-ag-nmlz end 3-decl-aux ‘Our food is finished.’ (93) a. oj-twa e-re flu-throat decl-aux ‘He has got a flu.’

b. oj-twa-e flu-throat-nmlz ‘influenza, cold’

c. oj oj-twa-e tsi mo-e-re 3sg.pro flu-throat-nmlz have appl-decl-aux ‘He has got the flu.’

4.5 Structural characteristics of the noun phrase As mentioned in Section 3.3, Kanoé constituent order tends to be head-final. The language does not have morphological marking of subject and object functions (see

14 Kanoé

683

Sections 4.1.1 and 5.4). Thus, the syntactic function of a noun phrase, subject or object, is determined by the preferred syntactic order, AOV, which occurs with very high frequency, or AVO, which is less frequent. As the language has no adpositions to mark syntactic functions, the syntactic order, the semantic logic, and the speech context are fundamental for avoiding ambiguities of interpretation (Section 6.1). In functions of adverbial nature (temporal, locative, instrumental, comitative, and similar adverbial complements), a noun phrase receives the oblique suffix -nĩ. The canonical order of constituents within the noun phrase is: (Modifier) Noun (Modifier), as in (94) and (95). (94) nã tsyke-kỹj-o kani poss.1sg indf-m-poss child ‘the children of my friend’ (95) jũ kani e-tsĩkwa dem.prox child woman-dim ‘this (little) girl’ Possessives, demonstratives, and quantifiers can occupy the determiner position, to the left of the head noun. On the right, in dependent position, there are modifying nominal roots or descriptive or qualifying adjectival verbal roots. The position for gender specification, male or female, varies (either before or after the head noun), while the position of gender-neutral nouns referring to human beings, such as kanĩ ‘child’, vary only when they are marked by the free forms ævo ‘man, male’ or e ‘woman, female’.

4.5.1 Nominal modification The lack of a category of true adjectives does not prevent attributive modification of nouns. In Kanoé, nouns can be modified by other nouns, as in (96). Furthermore, a number of stative verbs with adjectival (descriptive or qualifying) semantic content can function as modifiers when nominalized; see (97). In both situations, the modifier follows the modified head noun. (96) peutæ tsuæ-tinu beiju manioc-clf:paste ‘beiju of soft manioc mass’ (97) mõw opera voro-n-e two jaguar black-3-nmlz ‘two black jaguars’

684

Laércio Nora Bacelar and Hein van der Voort

4.5.2 Possessive attributive modification As we saw in Section 4.1.2, attributive possession within the nominal phrase in Kanoé can be expressed by a possessive pronoun in possessor position, preceding the possessum. In addition, attributive possession can be realized by attaching the possessive suffix -o (which is lexicalized in most possessive pronouns) to the possessor noun, as in (98) through (101). (98)

Toãkũ-o mũj koko-ro n-e-re Toãkun-poss mother sweep-ag 3-decl-aux ‘Toãkun’s mother is sweeping the floor.’

(99)

Pura-o mũj tũ-o e-re mĩke Purá-poss mother die-intr decl-aux already ‘Purá’s mother has already deceased.’

(100) João-o ævo kani tũvo n-e-re João-poss man child fall.ill 3-decl-aux ‘João’s son is sick.’ (101) ña i-tsotsi-o piko iva e-re poss.1sg er-foot-poss nail hurt decl-aux ‘My toenail hurts.’ For a discussion of the interaction between possession and classification and the relationship between possession, beneficiary, and comitative functions in Kanoé, see van der Voort (2009: 365–367). In addition to a transparent system of possession marking, there is also a remnant of grammatical inalienability marking. The majority of the names for parts and organs of the human body and of animals, along with some plant parts, are bound nouns (Section 4.3.3). In order for these bound nouns to occur as free nouns, they must be preceded by the element i-. Here this morpheme is analyzed as and glossed as er ‘empty root’, a strictly grammatical and semantically empty morpheme, which allows bound nouns to appear as free nouns. (102) jato i-kuta poss.1pl er-head ‘our head’ (103) æky-o i-rãw banana-poss er-meat ‘banana pulp’

14 Kanoé

685

(104) ña i-kỹj iva-õ-kỹj e-re poss.1sg er-eye hurt-1-eye decl-aux ‘My eye hurts.’ (105) pja i-kuta iva-pe-kuta mi-tsi poss.2sg er-head hurt-2-head 2-int ‘Does your head hurt?’ Bound nouns attached to the empty root can occur in any function or construction, not only in possessive phrases, as demonstrated by (86) and (209).

4.5.3 Oblique argument marking As previously mentioned, the syntactic functions of subject and object in Kanoé are not indicated morphologically on the noun. In the absence of adpositions, non-core arguments, such as those with locative, instrumental, temporal and, in some cases, comitative or benefactive functions, receive the generic oblique suffix, -ni [nĩ], as shown in (106) through (110). (106) ajte kwini po õ-e-re kuni-mu-ni ajũkoe-ni 1pl.pro fish capture 1-decl-aux water-clf:liquid-obl lake-obl ‘We fished on the river and on the lake.’ (107) aj kuni-ni vyry-õ-tu-mu õ-e-re 1sg.pro water-obl move-dir:out-clf:liquid 1-decl-aux ‘I am getting out of the water.’ (108) aj ipæ õ-e-re iry mapi-ni 1sg.pro kill 1-decl-aux monkey arrow-obl ‘I kill monkeys using arrows.’ (109) itsaj-e-ni ajte pæ-væ-ja õ-e-re night-nmlz-obl 1pl.pro dance-du-dir:down 1-decl-aux ‘At night we will dance.’ (110) aj ña koro-ni tyvæ-ro õ-e-re 1sg.pro poss.1sg brother-obl hunt-ag 1-decl-aux ‘I hunt with my brother.’ The combination of possessive -o with the oblique -ni generates the benefactive sense, as in (111), and comitative sense, as in (112).

686

Laércio Nora Bacelar and Hein van der Voort

(111) pa-pe-t-ãw æky aj-o-ni throw-2-tr-imp banana 1sg-poss-obl ‘Throw the bananas to me!’ (112) aj uræ ipæ õ-e-re mi-o-ni 1sg.pro pigs kill 1-decl-aux 2sg-poss-obl ‘I hunt pig with you.’ Note that possessive pronouns are not used in this construction but instead the standard personal pronouns are used in combination with possessive marking.9

5 The verb phrase The verb consists minimally of a lexical root and verb inflection. Often the root is followed by person inflection, subsequently by derivational suffixes, and finally by other inflectional morphemes. In addition, certain verb classes require prefixing of person inflection (as seen in Section 4.1.1).

5.1 Verb classes and morpheme order in verbal structures There are five verbal subclasses, which depend on valence and semantic traits related to morpheme order.

5.1.1 Verbs of cognition Explicit first- and second-person marking on verbs that relate to mental, cognitive, or emotional processes prototypically occurs prefixed to the main verbal root, rather than suffixed (see also Section 4.1.1). This is illustrated by the verbal roots pateñu ‘know’, as in (41)–(43), and memu ‘like’, in (113). (113) mĩ oj pi-memu-ro-to k-e-re 2sg.pro 3sg.pro 2-like-ag-tr neg-decl-aux ‘You don’t like him.’

9 A similar interaction between possessive and benefactive is attested in Aikanã (see van der Voort & Birchall, this volume).

14 Kanoé

687

5.1.2 Non-incorporating verbs For non-incorporating, intransitive verbs such as i ‘cry’ and atsi ‘fear’, or transitive itæ ‘drink’, po ‘capture’, and tserero ‘see, look’, which are verbs that never incorporate nominal classifiers, the subject person inflection occurs after the main verb root and before optional morphemes, such as the verbal classifier -ro and the transitivizer -to, shown in (114). (114) pejake aj mi vara-õ-ro-pe-to no-e-re tomorrow 1sg.pro 2sg.pro speak-1-ag-2-tr fut-decl-aux ‘Tomorrow I want to speak with you.’

5.1.3 Incorporating verbs For incorporating, intransitive verbs such as mokỹj ‘sleep’ or transitive verbs, such as o ‘eat’ and kotso ‘wash’, which tend to incorporate bound nouns or nominal classifiers, person marking occurs between the main verbal root and incorporated morphemes, as in (115). (115) aj atiti-kaña o-õ-kaña õ-e-re 1sg.pro maize-flour eat-1-flour 1-decl-aux ‘I eat paçoca10 of maize.’

5.1.4 Reflexive verbs With reflexive verbs that refer to physiological processes, such as breathing, getting tired, and resting, person marking occurs in between the main verbal root and reflexive -ry (which may be etymologically related to the bound root -ry ‘liver’), as in (116) and, notably, in (211). (116) aj ore-õ-ry e-re ara-k-e-re 1sg.pro tire-1-refl decl-aux little-neg-decl-aux ‘I am very tired.’ In combination with the dual morpheme -væ, reflexive -ry expresses reciprocity, as in (158).

10 An indigenous crushed dry food mixture that gave rise to popular Brazilian nut-based candies and salty mixtures.

688

Laércio Nora Bacelar and Hein van der Voort

5.1.5 Process verbs With intransitive verbs describing processes, overt first- or second-person markers are obligatorily part of the auxiliary complex, as in (117) and (118). Process verbs include those marked by the intransitive morpheme -o (see, e.g., (141) and (251)); onomatopoetic verbs that refer to biological processes or actions, such as spitting, burping, vomiting, urinating, sneezing, or sobbing; and verbs that refer to natural phenomena, such as wind, thunder, and lightning. (117) ajte imujkaj õ-e-re 1pl.pro spit 1-decl-aux ‘We spit.’ (118) aj ew~ew õ-e-re 1sg.pro burp~red 1-decl-aux ‘I am burping.’

5.2 The auxiliary complex Declarative sentences are almost always include the auxiliary morpheme -re at the end of the verbal complex. Subject person markers and certain other morphemes are often prefixed to the auxiliary. This can be seen in the majority of declarative examples in this chapter. In addition, auxiliary -re is also used as a verbalizer in relative clause-like constructions such as (57), (187), and (242)–(244). It also appears in somewhat rare copula-like constructions, creating non-verbal predicates, as in (53), (119), and (120) (see also Bacelar 2004: 227). (119) aj-re Kanoe 1sg-aux Kanoé ‘I am Kanoé.’ (120) ojte ævo-re 3sg.pro man-aux ‘They are men.’ The auxiliary morpheme -re does not occur in non-declarative sentences, such as interrogative and imperative ones (Section 6.2). Thus, -re seems to function as a declarative auxiliary on which verbal inflection and negation appear separately from the verb root. As discussed more extensively by Bacelar (2004: 169), auxiliary -re, with associated person and mood, can be analyzed as a separate auxiliary verb which supports the lexical verb to form the head of a verb phrase.

14 Kanoé

689

Other mood morphemes may sometimes also occur separately from the main verb and bear person marker prefixes, as in interrogatives, seen in (105) and (146), and imperatives, as in (201) and (202).

5.3 Tense, aspect, modality and evidentiality In Kanoé, tense and aspect are verbal grammatical categories although there are also lexical expressions of temporal reference. Modality and evidentiality are not expressed grammatically. Instead, information source is only expressed lexically.

5.3.1 Tense and aspect There is a morphological contrast between future and non-future tense. There is no grammatical verbal marking of non-future tense, while the future tense is marked on the auxiliary cluster by nu- or no-, which and can be further specified by a temporal adverb, as in (122). (121) aj vara-õ-ro e-re 1sg.pro speak-1-ag decl-aux ‘I spoke.’ (122) pejake aj vara-õ-ro-pe-to nu-e-re tomorrow 1sg.pro speak-1-ag-2-tr fut-decl-aux ‘Tomorrow I will speak with you.’ Note that the future also has connotations of inchoative (imminent) aspect and volitional or desiderative modality. The eventual meaning of the future tense morpheme also depends on the phrasal and speech context, as shown by (123) through (127). (123) kanĩ mõ-kỹj no-e-re child sleep-eye fut-decl-aux ‘The child is going to / wants to sleep.’ (124) aj mõ-õ-kỹj no-k-e-re minĩ 1sg.pro sleep-1-eye fut-neg-decl-aux now ‘I am not going to / I don’t want to sleep now.’ (125) aj ivo pejkaj u no-e-re 1sg.pro yam bake eat fut-decl-aux ‘I want to eat baked yam!’

690

Laércio Nora Bacelar and Hein van der Voort

(126) aj nũvæ õ-no-e-re 1sg.pro defecate 1-fut-decl-aux ‘I have to defecate!’ (127) aj etsivi õ-no-e-re 1sg.pro urinate 1-fut-decl-aux ‘I have to urinate!’ With regard to present and past tenses, the situation is more complex. In principle, the simplest form of the verb is interpreted as reflecting non-future tense, as in (121). On the other hand, the notion of present tense in Kanoé is inseparable from the notion of progressive aspect, as in (128). The distinction between non-future tenses in Kanoé is given by the occurrence or not of person marking within the auxiliary complex. Thus, present progressive is marked by the presence of person inflection on both the main verb and the auxiliary complex, while past tense is expressed by the presence of inflection on the main verb only. (128) aj vara-õ-ro õ-e-re 1sg.pro speak-1-ag 1-decl-aux ‘I speak.’ (Lit. ‘I am speaking.’) As a consequence, the double marking of person in (128) is not redundant but expresses a tense-aspect distinction. Furthermore, in the third person, progressive tense-aspect involves the person marker n- on the auxiliary complex, while third person continues zero marked on the main verb, as in, for example, (28), (92), (143), (216), (217), etc. With stative and property verbs, however, the distinction between non-future tenses is not visible in the third person, which is always zero marked, as in (135). Another way of expressing progressive aspect involves the use of the nominalizing morpheme -e, as seen in (129) and (198). (129) akiki-e-mi-ro-to-ku pja papa scream-nmlz-2-ag-tr-neg poss.2sg father ‘Do not scream at your father!’ Temporal reference can also be expressed by adverbs or adverbial phrases. They have syntactic mobility: they can occur at the beginning of the sentence, as in (130); immediately before the verb, as in (131); or after the verb, as in (132). (130) kamitsi wæ tsi e-re yesterday rain have decl-aux ‘Yesterday it rained.’

14 Kanoé

691

(131) jũ opera tũvo ara-k-e-re neunu pejake tũ-o dem.prox jaguar fall.ill little-neg-decl-aux maybe tomorrow die-intr e-re decl-aux ‘This dog is very sick; maybe it will be dead tomorrow.’ (132) kani pa e-re mini child be.born decl-aux today ‘The baby was born today.’

5.3.2 Modality and evidentiality Modals and evidentials are also expressed by adverbs, adverbial clauses, verbs, or a combination of these. In (131), we have already observed the use of the dubitative adverb neunu ‘maybe’ to express epistemic possibility. Notions of adversity and frustration of intention are often expressed through the verb ja ‘want’ in combination with the adverb tsoke ‘but’ in a complex sentence, as in (220) and (251). The notion of purpose is expressed by an adverbial clause, as in (240) and (241), treated in Section 7.3. There is no clear grammatical expression of modality. Evidentiality is also not expressed grammatically. Instead, lexical verbs such as patẽñu ‘know’, as in (41), and vara ‘say, speak, talk’, as in (245), are used.

5.4 Verbal agreement Some characteristics of the interaction between person marking and pronouns were presented in Section 4.1.1. Kanoé verbal structures are morphosyntactically complex in the sense that there is no clear boundary between morphological and syntactic structures. In most cases, this complexity stems from a double system of agreement at the subject level and the level of the direct object. At the subject level, nouns and free personal pronouns agree by cross-reference with prefixed or suffixed person markers in predicate structures, as in (133). Transitive verbs can also show object agreement, as in (134). (133) mi oj pi-pateñu-to k-e-re 2sg.pro 3sg.pro 2-know-tr neg-decl-aux ‘You don’t know him.’ (134) oj mi pateñu-pe-to k-e-re 3sg.pro 2sg.pro know-2-tr neg-decl-aux ‘He doesn’t know you.’

692

Laércio Nora Bacelar and Hein van der Voort

Note that third-person inflection in certain verbal classes is zero (as mentioned in Sections 3.2.2 and 5.3.1), and the overt expression of arguments is realized only by noun phrases. In addition, classifiers and bound nouns (Section 4.3.3) representing the subject can be incorporated into certain verbal structures without changing the verb’s valence or the grammatical function of its arguments; see (135). (135) ña i-tsotsi pæ-tsotsi e-re poss.1sg er-foot white-foot decl-aux ‘My foot is clean. ‘ (Lit. ‘My foot is white on the foot.’) The incorporated element refers to one of the arguments, but it should not be considered an inflectional agreement marker. In (136), the embedded bound noun kuta ‘head’ refers to the object, while there is verbal inflectional agreement with the second-person subject. (136) pja i-kuta ej-pe-kuta e-re poss.2sg er-head big-2-head decl-aux ‘Your head is big.’ (Lit. ‘You have your head big.’) At the level of the direct object, free nouns and personal pronouns can cross-reference bound person markers on the main verb, shown in (134). In the case of a zeromarked third person – in non-incorporating transitive verbal structures – nouns (with or without a nominal classifier) and free personal pronouns can act as arguments, as in (137). (137) e kani ava tsere-to e-re woman child macaw see-tr decl-aux ‘The girl saw the macaw.’ However, classifiers that appear on nouns can also co-occur with identical classifiers incorporated into verbs, as in (138). (138) aj epy-kwã o-õ-kwã õ-e-re 1sg.pro nut-clf:small eat-1-clf:small 1-decl-aux ‘I eat peanuts.’

5.5 Valence In Kanoé, the following classes of verbs are attested with respect to valence.

14 Kanoé

693

5.5.1 Valence 0 Verbs referring to natural and temporal phenomena, such as ‘thunder’, ‘lightning’ and ‘nightfall’, among others, are always intransitive in Kanoé and have a valence of zero. As a consequence, the sentence lacks a subject position, as demonstrated in (139) and (140). (139) piraraw e-re lightning decl-aux ‘Lightning flashed.’ (140) itsaj e-re night decl-aux ‘Night fell.’

5.5.2 Valence 1 Verbs with a valence of one only require a nominal or pronominal argument in subject function. Intransitive verbs such as ‘die’, ‘get sick’, and ‘spoil’, which are marked by intransitive -o (as in (141) and fossilized in (142)), or unmarked verbs such as ‘cry’, ‘run’ etc. (as in (143)) and reflexive verbs, such as ‘tire’, ‘rest’, ‘sadden’ (as in (144)) are of this type. (141) ævo tũ-o e-re man die-intr decl-aux ‘The man died.’ (Lit. ‘The man is dead.’) (142) oj tũvo e-re 3sg.pro fall.ill decl-aux ‘He is sick.’ (143) kani iriri-ro n-e-re child run-ag 3-decl-aux ‘The boy is running.’ (144) mi ore-pe-ry k-e-re 2sg.pro tire-2-refl neg-decl-aux ‘You’re not tired.’

5.5.3 Valence 2 Transitive verbs minimally require one argument in subject function and another in object function. There are at least two types of predicates headed by transitive ver-

694

Laércio Nora Bacelar and Hein van der Voort

bal roots: underived and derived transitive constructions. Certain verbs appear to be inherently transitive, as they never present transitivizing morphemes, as (145) and (146) show. (145) aj epỹ-kwã mũ-õ-e-re 1sg.pro nut-clf:small plant-1-decl-aux ‘I plant peanuts.’ (146) mi mutyry-ko-mu itæ mi-tsi 2sg.pro ac̹ai-clf:fruit-clf:liquid drink 2-int ‘Do you drink ac̹ai juice?’ Other verbs tend to contain transitivizing -to when they are divalent. Notice the contrasts between the pairs of examples in (147) through (152). (147) aj tsere-vo õ-e-re 1sg.pro see-dir:down 1-decl-aux ‘I am looking down.’ (148) e iry tsere-to e-re woman monkey see-tr decl-aux ‘The woman saw the monkey.’ (149) aj vara-õ-ro õ-e-re 1sg.pro speak-1-ag 1-decl-aux ‘I speak.’ (150) aj mite vara-õ-ro-pe-to õ-e-re 1sg.pro 2pl.pro speak-1-ag-2-tr 1-decl-aux ‘I am speaking with you.’ (151) kani tsere-ro n-e-re child see-ag 3-decl-aux ‘The child woke up.’ (Lit. ‘The child is looking.’) (152) kani ava tsere-to e-re child macaw see-tr decl-aux ‘The child saw the macaw.’ In (154), -to (not to be confused with the directional -to in Section 5.6.1) has an applicative effect. Here it introduces a comitative object, with respect to which the

14 Kanoé

695

event occurs. The sentence can be interpreted literally as: ‘I want to play (with respect to) you’. (153) kani tævæ-ja n-e-re child play-dir:down 3-decl-aux ‘The child is playing.’ (154) aj mi tævæ-ja-õ-to no-e-re 1sg.pro 2sg.pro play-dir:down-1-tr fut-decl-aux ‘I want to play with you.’ Note that incorporated nominal classifiers can also represent one of the verb's arguments. As an example, the verb tyry ‘tie’ requires the transitivizer -to, as in (155), or a nominal classifier, (156). (155) aj ña itsote tyry-õ-to õ-e-re 1sg.pro poss.1sg shoe tie-1-tr 1-decl-aux ‘I am tying my shoe.’ (156) aj ña jy tyry-õ-ty õ-e-re 1sg.pro poss.1sg hair tie-1-clf:thread 1-decl-aux ‘I am tying my hair.’ Several examples in this chapter (e.g., (62), (63), (65), (93c), and (240)) exemplify the applicative mo-. It is of limited productivity; as it occurs exclusively in a fixed combination with the existential verb tsi, meaning ‘have’, or perhaps literally, ‘exist to/with’. When applicative mo- is absent, tsi has an existential interpretation, such as in (130), (236), and (238) through (239). Bacelar (2004: 199–200) deals with applicative mo- in more detail. Another morpheme with valency effects is the dual morpheme -væ, which can be attested in examples (109), (179), (188), (203), (204), and (249). It tends to express comitativity, as in (157), and, in combination with reflexive -ry, reciprocity, as in (158). (157) oj po-væ-turo-pe-to-nu n-e-re 3sg.pro make-du-space-2-tr-fut 3-decl-aux ‘He will be working with you.’ (158) uruã-te i-væ-kuta-ry-ni-ro e-re boy-coll quarrel-du-head-refl-3-ag decl-aux ‘The boys are quarreling.’

696

Laércio Nora Bacelar and Hein van der Voort

5.6 Directional morphemes Kanoé verbs that refer to direction and movement in physical space, intransitive or not, dynamic or not, (e.g., ‘look up’, ‘look down’, etc.), may include one or more directional morphemes. Such directionals follow the verbal root and complement or specify its semantic content. Table 14.11 summarizes the Kanoé directional system. Tab. 14.11: Semantic properties of the directionals.

-to -tu -ja -vo -mu

gloss

inward

outward

downward

upward

dir:in dir:out dir:down dir:down dir:up

+ – – – –

– + – – –

– – + + –

– – – – +

5.6.1 Directional -to versus -tu: inward and outward movements The directional morpheme -to ‘inward, inside of’, occurs suffixed to certain verbal roots that denote movement in physical space. The morpheme is homophonous with the transitivizer -to, but it is not a special function of the latter, for three reasons: 1) transitive -to occurs only with transitive roots, whereas directional -to occurs with roots denoting movement and does not introduce a direct object; 2) the homophonous directional and transitivizer occupy different slots in the verb, as in (171); 3) directional -to contrasts with the directional -tu ‘outward, out of’, as (159) through (162) show. (159) erekanã tyj-ni vyry-to n-e-re elderly house-obl move-dir:in 3-decl-aux ‘The old man is entering the communal house.’ (160) erekanã tyj-ni vyry-tu n-e-re elderly house -obl move-dir:out 3-decl-aux ‘The old man is leaving the communal house.’ (161) kũkoe vawe-ni vyry-to e-re armadillo hole-obl move-dir:in decl-aux ‘The armadillo entered the hole.’ (162) aj kuni-ni uj-õ-tu õ-e-re 1sg.pro water-obl go-1-dir:out 1-decl-aux ‘I’m going to the river.’

14 Kanoé

697

5.6.2 The directional -ja ‘downward motion’ and -mu ‘upward motion’ The directional -ja ‘downward’ is frequent in verbal structures headed by roots that denote some type of action or vertical downward movement. It tends to occur after the verb roots aj ‘sit’, para ‘fall’, pe ‘lay down’, and tsu ‘crouch’, as in (163) and (164). (163) aj aj-õ-ja õ-e-re æræræ-ni 1sg.pro sit-1-dir:down 1-decl-aux bench-obl ‘I am sitting on the bench.’ (164) aj pe-õ-ja õ-e-re 1sg.pro lie-1-dir:down 1-decl-aux ‘I am lying down.’ Furthermore, -ja occurs in structures headed by pæ ‘dance’ (possibly ‘stamp feet’), as in (109) and (165), and teva ‘play’, as we saw in (154). (165) aj pæ-õ-ja õ-e-re 1sg.pro dance-1-dir:down 1-decl-aux ‘I am dancing.’ Less frequent is -mu ‘upward’, not to be confused with the homophonic classifier -mu ‘liquid’. Examples (166) through (169) show that there is a clear opposition between -mu ‘upward’ and -ja ‘downward’. (166) kwikaj ty-to-mu-kaj e-re sun move-dir:in-dir:up-clf:sun decl-aux ‘It is eight o’clock in the morning.’ (Lit. ‘The sun is going up.’) (167) kani ty-ja nu-n-e-re child move-dir:down fut-3-decl-aux ‘The boy will come down.’ (168) iry ty-to-mu n-e-re ytse-ni monkey move-dir:in-dir:up 3-decl-aux tree-obl ‘The monkey went up into the tree.’ (169) iry ty-ja n-e-re monkey move-dir:down 3-decl-aux ‘The monkey is coming down.’

698

Laércio Nora Bacelar and Hein van der Voort

In (170) and (171), three directionals co-occur in the same verb, apparently with the meaning ‘upward’. (170) tsere-to-mu-ja-ãw see-dir:in-dir:up-dir:down-imp ‘Look up!’ (171) opera-tsĩkwa tsere-to-mu-ja-to e-re kiki jaguar-dim see-dir:in-dir:up-dir:down-tr decl-aux parakeet ‘The cat is looking up at the parakeet.’ A less frequent directional morpheme is -vo ‘downward’. It has the same meaning as -ja and it is fossilized in some verbal roots, like tuvo ‘fall ill’. See also (147). (172) itevæ ty-vo n-e-re ytse-ni person move-dir:down 3-decl-aux tree-obl ‘The man is coming down from the tree.’ (173) uru-nake ty-vo n-e-re new-f move-dir:down 3-decl-aux ‘The girl is coming down.’

5.7 Noun incorporation As mentioned at the end of Section 4.3.1, we are not using a strict formal definition of the term ‘incorporation’. In Kanoé, the incorporation of (semi-)lexical elements in the verb is a derivational operation that does not modify the syntactic properties of the verb or the arguments. Partial or total nominal incorporation occurs systematically in the same position as the transitivizer -to (Section 5.5.3) or nominal classifiers (Section 4.3.2), as in (174). (174) mi oroe-tinu topi-e-tinu tapa-pe-tinu-kjũ 2sg.pro mud-clf:paste rot-nmlz-clf:paste step-2-clf:paste-npr ‘Are you stepping in rotten mud?’ Bound nouns (Section 4.3.3) are often incorporated, as in (175), which involves the bound noun -tsotsi ‘foot’, and in (176), which involves -tsonu ‘tail’. (175) aj i õ-e-re tõ y tapa-e-tsotsi e-re 1sg.pro cry 1-decl-aux because thorn step-nmlz-foot decl-aux ‘I am crying because I stepped into a thorn.’

14 Kanoé

699

(176) kani opera-tsĩkwa ojo tsonu tapa-e-tsonu n-e-re child jaguar-dim poss.3sg tail step-nmlz-tail 3-decl-aux ‘The boy is stepping on the cat’s tail.’ Often, incorporated nouns refer to body parts involved in the event described by the verb. In (177) the free noun ytaytsi ‘beard’ is incorporated in its entirety. (177) aj kujve-ytaytsi õ-e-re 1sg.pro scrape-beard 1-decl-aux ‘I’m shaving.’ (Lit. ‘I am scraping the beard.’) In addition to productive incorporation of bound nouns, there are also cases where historically incorporated nouns have become frozen in potentially discontinuous constructions. For example, the verb petso ‘give’ presumably contains the bound noun -tso ‘finger, hand’; the root pe does not exist without it. It is worth mentioning that ‘finger’ cannot be exchanged with any other body part noun or classifier, and that person marking can intervene between the original verb root and the frozen body part term, as in (178). (178) mi aj u-ro-e pe-õ-tsõ mi-e-re 2sg.pro 1sg.pro eat-ag-nmlz give-1-finger 2-decl-aux ‘You give me food.’ Another example is the verb mokỹj ‘sleep’, which consists of an unidentifiable root mo and the bound noun -kỹj ‘eye’. Example (179) shows the insertion of derivational morphology between the verb root and frozen incorporated noun. (179) ajte mo-væ-kỹj õ-e-re 1pl.pro sleep-du-eye 1-decl-aux ‘We sleep together.’ In other cases, the combination of a verbal root with a bound noun is fossilized and does not allow the insertion of any morphemes. The noun jako ‘back’ was probably incorporated as a reduced form -ko in pevako ‘carry on the back’ and tyveko ‘take away on the back’, as in (180), which doesn’t allow person marking. (180) aj tyveko õ-e-re ña uru 1sg.pro take.on.back 1-decl-aux poss.1sg new11 ‘I am taking my little sister on the back.’

11 Reduced form of urunake ‘younger sister’.

700

Laércio Nora Bacelar and Hein van der Voort

6 Syntactic structure of simple sentences 6.1 Direct order and inversions Kanoé is a predominantly head-final AOV language, as seen in most of the previously-cited phrasal examples. In a corpus of 1000 transitive clauses headed by transitive verbs, AOV order occurred in 73.5 % of the cases, as shown in Table 14.12. It is important to note that this corpus included sentences obtained from elderly, acculturated, and bilingual speakers, who for decades had not spoken their native language on a daily basis, Portuguese having become their daily language. This may have been responsible for the relatively high frequency of AVO order in elicited sentences. It should be noted that if only data from the Kanoé of Omeré, who are the last remaining monolingual speakers, are considered, the percentage of AOV sentences is higher, approximately 90 %, in declarative, affirmative, and negative sentences. This can be considered the canonical order, in which the A subject, nominal or pronominal, is topicalized, as in (181) through (183). (181) ña e kope-ko-mu tsoe-to e-re poss.1sg woman coffee-clf:fruit-clf:liquid strain-tr decl-aux a o v ‘My wife made coffee.’ (182) aj kope-ko-mu tsoe-to no-e-re 1sg.pro coffee-clf:fruit-clf:liquid strain-tr fut-decl-aux a o v ‘I wanted to make coffee.’ (183) Opera-o muȷ̃ atiti tutu n-e-re Operá-poss mother maize beat 3-decl-aux a o v ‘Operá’s mother is crushing maize.’ It is perhaps telling that basic AOV order is also attested in the few narrative texts by the aforementioned elderly acculturated speakers, in Bacelar (2004: 300–304). Other syntactic orders are less frequent and conditioned by the imperative (affirmative or negative), exhortative, and volitional moods, or information structure. The question and answer sequence in (184)–(185) shows such alternative ordering. (184) naj-tsi Samuew po-ro-e mini some-int Samuel make-ag-nmlz now (o) a v ‘What is Samuel doing now?’

14 Kanoé

701

Tab. 14: 12: Constituent order in transitive declarative sentences.12 order

AVO

AOV

VAO

VOA

OAV

OVA

total

percentage

24.1

73.5

0.4

0.4

1.3

0.3

100

(185) uræ tsutsi n-e-re Samuew pig cut 3-decl-aux Samuel o v a ‘Samuel is butchering the pig.’ In (184)–(185) the object appears at the beginning of the sentence, because it represents a focused entity. Example (186) shows a sentence with focused object and pronominal subject. (186) mũtyry-ko-mu aj itæ e-re açai-clf:fruit-clf:liquid 1sg.pro drink decl-aux ‘I drank açai juice.’ There is another specific construction used to highlight the direct object. It involves the non-present demonstrative -kjũ (see Section 4.1.5) and places the object in focus position, as in (187). (187) peutæ tsuæ-tinu oj po-ro-e kjũ-re beiju manioc-clf:paste 3sg.pro make-ag-nmlz npr-aux ‘He makes beiju of manioc.’ (Lit. ‘Beiju of manioc paste, which he makes.’) Adverbs, especially temporal adverbs, can occur topicalized in affirmative or negative declarative sentences. Since temporal adverbs scope over the entire sentence, a single adverb can simultaneously affect two coordinated clauses, as in (188). (188) mini ajte pæ-væ-ja vaj~vaj-væ-ro nu-e-re today 1pl.pro dance-du-dir:down sing~red-du-ag fut-decl-aux ‘Today we’re going to dance and sing.’ In interrogative constructions (Section 6.2.2), focusing of the questioned item is common. Indefinite pronouns, either functioning as O or A, always occur in the first position of the sentence, as in (189). This also concerns entire interrogative sentences, as in (52) and (53).

12 The numbers represent both nominal and pronominal arguments.

702

Laércio Nora Bacelar and Hein van der Voort

(189) nũvi-tsi jũ ytsika po-ro-e indf-int dem.prox. stairs make-ag-nmlz ‘Who made these stairs?’ Interrogative adverbs are always focused, which can lead to inversions of constituent order, as in (190). (190) nuũ aj-ja-tsi Masaka-o atsoni where sit-dir:down-int Masaká-poss settlement ‘Where is the Masaká village located?’

6.2 Mood and main sentence types Kanoé verbal inflectional morphology distinguishes several moods which determine the illocutionary status of the matrix sentence. Most of the examples of complete sentences in the present chapter are in declarative mood, which is expressed by the prefix e-. There are two other grammatical moods: interrogative, expressed by -tsi, and imperative, expressed by -ãw. This section describes and exemplifies these moods. There are also some suffixes that express adverbial clause moods, such as conditional -neko, final -hĩ, and the non-present demonstrative -kjũ, which are dealt with in Section 7.3. The Kanoé mood markers also affect the entire sentence on the phonological level and determine its intonation, according to its type. Table 14.13 summarizes the mood marking system in Kanoé matrix clauses. The next subsections deal with the sentence types that were mentioned here. Negation is dealt with in more detail in Section 6.3. The expression of future tense was discussed in Section 5.3.1.

Tab. 14.13: Morphosyntactic markers of different types of matrix sentences. mood

illocutionary force

marker

declarative

affirmative negative and litotic affirmative future / volitive

-e -k + -e -nu \ -no \ -n + -e

interrogative

interrogative

-tsi

imperative

directive negative / prohibitive

-ãw tso… k-ãw

supplicative

mitigated imperative

peõtsõ

exhortative

exhortative

kapeãw

14 Kanoé

703

6.2.1 Declarative sentences Declarative sentences have an ascending-descending intonation and are marked morphologically by the affix e-, as in (191) through (193). (191) ña papa munũ-to e-re poss.1sg father hear-tr decl-aux ‘My father listened (heard).’ (192) oj munũ-to k-e-re 3sg.pro 3-hear-tr neg-decl-aux ‘He didn’t listen (didn’t hear).’ (193) aj mõ-õ-kỹj no-e-re 1sg.pro sleep-1-eye fut-decl-aux ‘I want to sleep!’

6.2.2 Interrogative sentences In interrogative and imperative sentences, intonation always rises. The difference between them is the mood marker. The interrogative mood is expressed by -tsi, which often comes at the end of the sentence, as in (194) through (197). (194) tyjko tsũtsũ-pe-tsi flea suck-2-int ‘Did the flea bite you?’ (195) more-tsi good-int ‘Is it good?’ (196) mĩ pi-munũ-to-tsi naj jũ wa~waj-õ-to-tsi 2sg.pro 2-hear-tr-int some dem.prox speak~red-1-tr-int ‘Did you listen to (hear) what I said?’13 (197) naj vara-ro mi-ro-tsi some speak-ag 2-ag-int ‘What are you saying?

13 It is not very clear why the interrogative occurs twice in (196), which seems unlike (218), where two interrogative clauses are coordinated.

704

Laércio Nora Bacelar and Hein van der Voort

Attached to the indefinite pronoun (e.g., naj ‘something)’, interrogative -tsi can occur at the beginning of the sentence, as in (198). (198) naj-tsi Reginaldo po-ro-e some-int Reginaldo make-ag-nmlz ‘What is Reginaldo doing?’

6.2.3 Imperative sentences In imperative sentences, intonation is always ascending. Morphologically, the interrogative mood is expressed by the suffix -ãw, as in (199) and (200). (199) kutsitsi-e-r-ãw scratch-nmlz-ag-imp ‘Go and scratch yourself!’ (200) tsere-t-ãw ereremũ see-tr-imp pigeon ‘See the pale-vented pigeon!’ The prohibitive mood is expressed by combining imperative -ãw with the negative morpheme -k (Section 6.3), as in (201) and (202). (201) tso n-jũ akiki-e-mi-ro e-k-ãw not 3-dem.prox scream-nmlz-2-ag decl-neg-imp ‘Stop that, don’t scream!’ (202) tso n-jũ iũ-kuta n-e-k-ãw not 3-dem.prox quarrel-head 3-decl-neg-imp ‘Stop that, don’t quarrel!’ The simultaneous use of declarative e- suggests that -ãw behaves more like a deontic modality marker in (201)–(202) than like an imperative mood marker. The word kapeãw ‘Let’s (do it)!, Come on!’ is a fixed expression in which the imperative suffix -ãw is fossilized. It is used for exhortations, as in (203) and (204). (203) kapeãw kere-væ-kuta-e erejkwa-kuta-e let’s smoke-du-head-nmlz rubber-head-nmlz ‘Let’s smoke the rubber!’ (Lit. ‘Let us two smoke the heads of rubber!’)14 14 In the processing of natural rubber the harvested rubber latex was solidified and cured by smoke in the form of large balls.

14 Kanoé

705

(204) kapeãw toky po-væ-ky let’s papaya eat.fruit-du-clf:fruit ‘Let’s have watermelon!’15 For requests one can also use the verb petso ‘give’ in a first-person construction that can be considered a mitigated imperative, as in (205) and (205). (205) pe-õ-tsõ tsutsiraj-tsĩkwa give-1-finger machete-dim ‘Give me the knife!’ (‘Please, pass the knife!’)’ (206) pe-õ-tsõ mapi give-1-finger arrow ‘Give me an arrow!’

6.3 Negation In negative declarative sentences the negative morpheme -k usually occurs affixed to the auxiliary complex ere, as in (207)–(210). In cosubordinated clause chains, as in (211), it can also be attached to the linking particle. (207) aj teva ja õ-k-e-re 1sg.pro play want 1-neg-decl-aux ‘I don’t want to play.’ (208) aj pwã õ-k-e-re 1sg.pro smoke 1-neg-decl-aux ‘I don’t smoke.’ (209) ævjũ i-rãw u-ro k-e-re paca er-meat eat-ag neg-decl-aux ‘The paca doesn’t eat meat.’ (210) oj ja ni-k-e-re mũtyry-ko 3sg.pro want 3-neg-decl-aux ac̹ai-clf:fruit ‘He doesn’t want ac̹ai.’

15 The word for papaya is also used for watermelon, which is not native to the Americas.

706

Laércio Nora Bacelar and Hein van der Voort

(211) oj iry-tsu-ry k-eko tywæ-ro e-hĩ de-ni-k-eko 3sg.pro liver-good-refl neg-cos hunt-ag decl-purp kill-3-neg-cos tynu-eko iry-tsu-ry k-e-re return-cos liver-good-refl neg-decl-aux ‘He is sad, because he went hunting, killed nothing, and returned downcast.’ Negative imperatives, or prohibitives, also involve -k, as seen in (201) and (202). Negation in complex comparative sentences is dealt with in Section 7.2.

6.3.1 Double negative marking in declarative sentences In Kanoé, the negative answer to a polar question can be reduced to tsokere, which means ‘no, nothing, there isn’t, it doesn’t exist, zero’, depending on the question. However, for the occasional purpose of emphasis, the answer can be doubly negative, with tsokere at the beginning and negative -k within the sentence, as in (212). ña (212) tso-k-e-re aj o-õ-ro k-u-n-e-re16 not-neg-decl-aux 1sg take-1-ag neg-fut-3-decl-aux poss.1sg mapi-ka arrow-clf:hard ‘No, I’m not going to take my bow with me.’

6.3.2 Attributive predicates in litotic constructions Except for some rare cases, there are no pairs of descriptive or quality-denoting verbal roots with antonymic semantics, such as oppositions of the type high/low, fat/thin, good/bad, etc. The language makes use of litotes as a descriptive and expressive resource, stating something by negating its opposite (Bacelar & Silva Jr. 2003). Therefore, most predicative attributive oppositions are not expressed by dedicated antonyms. Instead, antonymic meanings are usually expressed by litotic affirmative-negative constructions involving negative -k, as in (213) through (216). (213) pja i-kañu ej-pe-kañu e-re poss.2sg er-nose big-2-nose decl-aux ‘Your nose is big.’

16 Note that the future morpheme nu- loses its onset consonant /n/ when preceded by negative k-.

14 Kanoé

707

(214) ojo i-kañu ej-kañu k-e-re poss.3sg er-nose big-nose neg-decl-aux ‘His nose is small.’ (215) pja i-tokytwa more e-re poss.2sg er-throat good decl-aux ‘Your throat is fine.’ (216) pja i-tokytwa more ni-k-e-re poss.2sg er-throat good 3-neg-decl-aux ‘Your throat is bad.’ There is no clear tendency with regard to which of the antonymic concepts is lexicalized. Examples (228) to (231) in Section 7.2 show the importance of litotes in constructions that express comparative degree.

7 Syntactic structure of complex sentences In complex sentences, grammatical and logical semantic relations between clauses are expressed by coordination, subordination, and co-subordination. In coordination, two or more clauses are either juxtaposed or connected by a copular element, without any syntactic dependence between them. In subordination, one clause is in a dependent relationship with a matrix or main clause; there is an intrinsic relationship of syntactic-semantic subordination. Cosubordination involves coordination at the semantic level and subordination at the syntactic level. Table 14.14 summarizes the forms and functions of the different types of complex sentences in Kanoé.

7.1 Coordination Interclausal coordination in Kanoé occurs in several different constructions and can have multiple functions, such as additive, adversative, explanatory, and others. In juxtaposition coordination, the clauses (or sentences) have an independent syntactic structure and, as there is no coordination marker, they are juxtaposed as in (217) and (218). (217) oj ñomu-ni uj n-e-re tsuæ o-ro n-e-re 3sg.pro field-obl go 3-decl-aux manioc hold-ag 3-decl-aux ‘He went to the swidden field and is bringing manioc.’

cosubordination

V

V-matrix V-e-re V V-matrix V-kjũ V-matrix

nẽko neko

subordination

clause 1 V-matrix V-matrix V-matrix V-matrix

particle

coordination

construction

Tab. 14.14: Types of complex sentenes: forms and functions.

eko

neko

tsoke tõ

part.

V-matrix

V-e V-matrix V-matrix V-e-hĩ (V-)kjũ-re V-e

V-matrix (V-matrix) V-matrix V-matrix

clause 2

maere

part.

additive, temporal

nominal complement concessive conditional final relative causative

additive, comparative, explicative comparative of equality adversative explicative

meaning

708 Laércio Nora Bacelar and Hein van der Voort

14 Kanoé

709

(218) oj vaj~vaj-ni-tsi pæ-ja-ni-tsi 3sg.pro sing~red-3-int dance-dir:down-3-int ‘Is he going to sing and dance?’ Adversative coordination, on the other hand, is expressed by the adversative conjunction tsoke ‘but’, which establishes an oppositional relationship between the coordinated clauses, as in (219). (219) ña muj majkamu-ro e-re tsoke aj i-munu-to poss.1sg mother call-ag decl-aux but 1sg.pro 1-hear-tr k-e-re neg-decl-aux ‘My mother called me, but I did not hear.’ In constructions with a compound main verb, the occurrence of tsoke may result in the deletion of a part of the verb in the second clause, eliminating redundancy, as in (220). (220) oj ja vara-ro-õ-to n-e tsoke ja õ-k-e-re 3sg.pro want speak-ag-1-tr 3-decl but want 1-neg-decl-aux ‘He wants to talk with me but I don’t want to.’ Explicative coordination can be expressed in two ways: in juxtaposed sentences, as in (221) and (222); or in a construction with the particle tõ ‘because’, as in (223) and (224). Note that the explicative clause can not only mention a reason but also a cause. (221) kapeãw iriri-õ-ro vuj~vuj-ro-e ty no-e-re let’s run-1-ag wind~red-ag-nmlz move fut-decl-aux ‘Let’s run, (because) a storm is coming.’ (222) aj mi tævæ-ja-pe-to õ-k-e-re mini aj 1sg.pro 2sg.pro play-dir:down-2-tr 1-neg-decl-aux today 1sg.pro mo-õ-kỹj no-e-re sleep-1-eye fut-decl-aux ‘I am not going to play with you now, (because) I want to sleep.’ (223) opera uryketa e-re tõ tũvo n-e-re jaguar skinny decl-aux because fall.ill 3-decl-aux ‘The dog is skinny because it is sick.’ (224) kani i n-e-re tõ para-ja n-eko rwa-roko e-re child cry 3-decl-aux because fall-dir:down 3-cos break-knee decl-aux ‘The child is crying because it fell and broke its knee.’

710

Laércio Nora Bacelar and Hein van der Voort

Coordination can also express purpose or finality, as in (225). (225) aj ña kani va-õ-ro õ-e-re mo-kỹj e-re 1sg.pro poss.1sg child order-1-ag 1-decl-aux sleep-eye decl-aux ojo munaw-ni poss.3sg hammock-obl ‘I am ordering my son to sleep in his own hammock.’

7.2 The expression of comparison There are no grammaticalized degrees of comparison in Kanoé. Similar to coordination, comparison is expressed through juxtaposed sentences. With regard to equative comparison, the equation involves the adverbial particle maere ‘too’, at the end of the second clause, as in (226). (226) ña tyj ej-turo e-re pja tyj ej-turo e-re poss.1sg house big-space decl-aux poss.2sg house big-space decl-aux maere too ‘My house is as big as yours.’ (Lit. ‘My house is big; your house is big too.’) The particle maere ‘too’ even allows for ellipsis of the gradable predicate, as in (227), in which the second occurrence of tẽpũnake ere ‘is old’ is understood. (227) ña keke tẽpũ-nake e-re pja keke poss.1sg grandmother old-f decl-aux poss.2sg grandmother (tẽpũ-nake) maere17 old-f too ‘My grandmother is as old as your grandmother.’ (Lit. ‘My grandmother is old, your grandmother too.’) Comparatives can also be expressed through juxtaposition. In case of comparative superiority, the first sentence asserts a property attributed to the standard of comparison, while the second sentence, by means of litotes involving the negative morpheme -k, denies the same property to the target of comparison, as in (228) and (229). In cases of comparative inferiority, the construction occurs in the reverse order: the negative sentence comes first and the affirmative comes second, as in (230) and (231).

17 It is possible that maere contains the declarative and auxiliary morphemes -e-re, which could make ellipsis easier. However, ma- does not occur in other combinations and -ere is never affected by permutation or inflection.

14 Kanoé

711

(228) uromu ereã e-re kometakãw ereã k-e-re cayman big decl-aux lizard big neg-decl-aux ‘The cayman is bigger than the lizard.’ (Lit. ‘The cayman is big; the lizard is not big.’) (229) ña i-ta more-ta e-re pja i-ta more-ta poss.1sg er-skin good-skin decl-aux poss.2sg er-skin good-skin ni-k-e-re 3-neg-decl-aux ‘My clothes are better than yours.’ (Lit. ‘My clothes are good; your clothes are not good.’) (230) kykyte ereã k-e-re nãkaw ereã e-re frog big neg-decl-aux toad big decl-aux ‘The frog is smaller than the toad.’ (Lit. ‘The frog is not big; the toad is big.’) (231) ña vae-nake ã-nake k-e-re aj ã-kỹj õ-e-re poss.1sg cousin-f tall-f neg-decl-aux 1sg.pro tall-m 1-decl-aux ‘My cousin is shorter than I.’ (Lit. ‘My cousin is not tall; I am tall.’) In comparative structures, one of the clauses is always the negation of the other, and the syntactic position of the clauses may reflect the meaning of the whole: the negative clause tends to express comparative inferiority, and its position in the sentence determines whether one wants to emphasize superiority (as in (228) and (229)) or inferiority (as in (230) and (231)).

7.3 Subordination In Kanoé, subordination or hypotaxis consists of a syntactic-semantic relationship between at least two sentences, one of which is the matrix clause and one of which is a subordinate clause. There are three major categories of hypotactic clauses: complement clauses, adverbial clauses, and relative clauses.

7.3.1 Complement clauses Kanoé has only one type of complement clause: a subordinate clause that functions as a direct object of the main clause verb. This is a type of nominalized clause that acts as a complement to verbs of cognition (e.g., jẽ ‘know (have knowledge of)’, patẽñu ‘know (be acquainted with, recognize)’), predicates referring to volition (e.g., ja ‘want’), and predicates expressing commands or causation (e.g., va ‘or-

712

Laércio Nora Bacelar and Hein van der Voort

der’). The construction does not involve a conjunction, and the subordinate clause always occurs after the main verb. Complement clauses are nominalized but may be internally complex and contain overt object and subject nouns, as in (232) through (235). (232) aj i-jẽ-to e-re mapi po-ro-e 1sg.pro 1-know-tr decl-aux arrow make-ag-nmlz ‘I know how to make arrows.’ (233) mi pi-jẽ-to k-u twĩ-n-e 2sg.pro 2-know-tr neg-fut swim-3-nmlz ‘You don’t know how to swim?’ (234) aj ña kani va-õ-ro e-re (kwinĩ po-n-e) 1sg.pro poss.1sg child order-1-ag decl-aux fish capture-3-nmlz ‘I told my child (to fish).’ (235) ojo papa va-ro e-re oj mapi-ka po-ro-e poss.3sg father order-ag decl-aux 3sg.pro bow-clf:hard make-ag-nmlz ‘His father told him to make a bow.’

7.3.2 Adverbial clauses In Kanoé, adverbial clauses can simply be juxtaposed or they can be introduced by a conjunction. Syntactically, they can be placed before or after the main verb, which depends on semantic factors. According to the notions they convey, adverbial clauses can be classified into three subcategories: i) concessive; ii) conditional, and iii) final. Concessive clauses are introduced by the conjunction nẽko ‘although, even if, albeit’. From a semantic point of view, this subordinate clause denotes ​an obstacle, but not an impediment, with regard to the realization of the event expressed by the main verb, as in (236) and (237). (236) nẽko væ tsi e-re arakere aj tyvæ-ro õ-e-re although rain have decl-aux much 1sg.pro hunt-ag 1-decl-aux ‘Even though it is raining a lot, I’m going to hunt.’ (237) nẽko na papa ja ni-k-e-re aj twĩ~twĩ although poss.1sg father want 3-neg-decl-aux 1sg.pro swim~red õ-e-re ajũkoe-ni 1-decl-aux lake-obl ‘Even though my father doesn’t want it, I am swimming in the lake.’

14 Kanoé

713

Conditional clauses are usually subordinated to the main clause by the conditional conjunction neko. They describe a precondition for the realization of the event expressed by the main verb, (238) and (239). (238) mini væ tsi-e neko aj tyvæ-ro k-u-n-e-re today rain have-nmlz cond 1sg.pro hunt-ag neg-fut-3-decl-aux ‘If it rains today, I won’t hunt.’ (239) væ tsi ni-k-eko tsoke væ tsi-e neko jato ãtã-o rain have 3-neg-cos but rain have-nmlz cond poss.1pl aunt-poss tyj-ni ajte mo-õ-kỹj õ-e-re house-obl 1pl.pro sleep-1-eye 1-decl-aux ‘It is not raining, but when it rains, let us sleep in our aunt’s house.’ Final clauses convey the purpose of the event expressed by the main verb. They are subordinated to the main clause by the final copulative -hĩ, attached to the final clause verb, as in (240) and (241). (240) ña ãtã veo-õ-to eko aj namu õ-tsi mo-e-re poss.1sg aunt ask-1-tr cos 1sg.pro medicine 1-have appl-decl-aux pe-tso-e-hĩ ña vae-nake give-finger-nmlz-purp poss.1sg cousin-f ‘My aunt asked me whether I have medicine, in order to give to my cousin.’ (241) mapĩkwari itevæ po n-e-re u n-e-hĩ mapinguari person capture 3-decl-aux eat 3-nmlz-purp ‘The mapinguari catches people in order to eat them.’

7.3.3 Relative clauses In relative clauses, both the main verb and the embedded verb bear the non-present (in the sense of absent) demonstrative -kjũ (see Section 4.1.5). Furthermore, -kjũ is repeated in the auxiliary complex, thus establishing agreement relations among the constituents of the complex sentence, as in (242) through (244). (242) æky ajte pepej-õ-ro-kjũ oke uru-ky-kjũ kjũ-re banana 1pl.pro get-1-ag-npr still new-clf:fruit-npr npr-aux ‘The bananas we harvested are still green.’ (243) uko uru-nake aj mi vara-õ-ro-pe-to-kjũ kjũ-re dem.dist new-woman 1sg.pro 2sg.pro speak-1-ag-2-tr-npr npr-aux ‘That is the girl of which I told you.’

714

Laércio Nora Bacelar and Hein van der Voort

(244) aj tsere-õ-kjũ kjũ-re itevæ pæ-kỹȷ̃-kjũ munu-ro-k-e 1sg.pro see-1-npr npr-aux person white-m-npr hear-ag-neg-nmlz kjũ-re npr-aux ‘I knew a man who is bald and deaf.’

7.4 Cosubordination Kanoé has sentence structures that can be called cosubordinate. They are grammatically subordinate, but semantically coordinated, constructions, in which the matrix clause, which occurs at the end of the sentence, is preceded by one or multiple cosubordinate clauses, with the possibility of forming an extensive chain of clauses. Clause chains are linked by the particle eko, as in (245). (245) orymape vara-ro eko ña papa pejake more-kỹj e-re shaman speak-ag cos poss.1sg father tomorrow good-m decl-aux ‘The shaman said that our father will get better tomorrow.’ In general, the clauses tend to follow an iconic sequence, symbolizing the order of the events that they relate, as shown in (246). (246) na kani tsurune-kaña po-ro eko o-õ-kaña tsu-kaña e-re poss.1sg child paçoca-flour make-ag cos eat-1-flour fat-flour decl-aux ‘My daughter made paçoca18 and I ate tasty (paçoca).’ In (246), eko functions as a cosubordinating element. In addition, noun incorporation of kaña ‘flour’, used as a classifier for ‘flour, paste’, contributes to the interconnection of the chained clauses. It creates agreement across the entire construction, and it allows for ellipsis of the original head noun in the subsequent clause. In (247), there is no agreement through classifier incorporation, and the subject of the second clause has been elided altogether. (247) aj kãpore re õ-eko para-ja n-e-re 1sg.pro guan kill 1-cos fall-dir:down 3-decl-aux’ ‘I hit the guan and (it) fell.’ The functions of the co-subordinated clauses are diverse. They can have an additive function, as in (248).

18 Crushed dry food mixture; see note 10.

14 Kanoé

715

(248) kani iriri-ro n-e para-ja n-eko rwa-kwamu e-re child run-ag 3-decl fall-dir:down 3-cos break-arm decl-aux ‘The boy ran, fell, and broke his arm.’ Cosubordinated clauses can also have a temporal function, as in (249) and (250). (249) væ~væ tsi-e pej-o no-eko ivo mu-væ õ-e-re rain~rain have-nmlz halt-intr fut-cos yam plant-du 1-decl-aux ‘When the rain stops, we’re going to plant yams.’ (250) jato koro ty nu-eko itsaj nu-n-e-re poss.1pl brother move fut-cos night fut-3-decl-aux ‘Our brother will arrive at dusk.’ Because syntactic processes are not mutually exclusive, complex sentences may simultaneously include several types of structures. Example (251) includes cosubordinate, adversative, and coordinate clause constructions. (251) aj vavo pejkaj õ-eko ja õ-e-re u-ro-e tsoke vavo 1sg.pro potato bake 1-cos want 1-decl-aux eat-ag-nmlz but potato more ni-k-e-re ẽtẽ-o e-re good 3-neg-decl-aux rot-intr decl-aux ‘I baked a potato and wanted to eat it, but the potato is not good; it’s rotten.’

8 Discourse and information structure Information that is in focus tends to occur at the beginning of the sentence. This can be seen in many examples in this chapter. In Section 6.1 departures from basic AOV word order are shown to be due to focalization. Although predicates contain inflectional person marking, overt nominal and pronominal arguments are usually present, which is clear from many of the examples presented here. Furthermore, in narrative texts and dialogues by rusty heritage language speakers as well as by the recently contacted monolingual speakers (Bacelar 2004: 300–304), verbal person inflection typically co-occurs with overt pronouns. As seen in Section 7.4 and in Bacelar (2004: 300–304), cosubordinated clause chains may involve identical or different subjects, but there is no grammatical switchreference marking system.

716

Laércio Nora Bacelar and Hein van der Voort

9 Conclusion One of the special features of Kanoé is the way in which a relatively simple paradigm of inflectional person marking is used in different functions and positions inside the verb and/or the auxiliary complex. As shown in Sections 5.1 and 5.3.1, these different positions of person markers serve to distinguish verb classes, construction types, or semantic distinctions such as progressive aspect. Another special characteristic of Kanoé is the frequent use of negation in the creation of antonyms, called litotes and treated in Section 6.3.2. Furthermore, there is a special nonpresent demonstrative suffix -kjũ, which refers to a third person that is absent from the speech context (see Section 5.1.4) and which is also used in relative clause constructions, as described in Section 7.3. Finally, the productive system of nominal classifiers and bound nominal roots stands out, and its similarity to comparable systems in unrelated neighboring languages contributes to a picture of the GuaporéMamoré region as a linguístic area. The nature and origin of certain grammatical features of Kanoé require more research. However, the very low number of speakers, the difficult circumstances under which they survive, and the possibility of obsolescence effects cause further investigation to be a challenge. The predicament of the Kanoé people and their fascinating language is due to a century of disease, slavery, deportation, dispossession, and murder, which continue in our times.

10 Acknowledgments We owe thanks to the Kanoé people, especially Munuzinho Kanoé and Teresa Kanoé (in memoriam), Purá and Txinamãty Kanoé. We are highly indebted to the late Dr. Pieter Muysken for material support, guidance, and theoretical contributions to the advancement of the research on Kanoé. Furthermore, we thank Altair Algayer of the Fundação Nacional do Índio (FUNAI) for support and hospitality in the field. We also thank Pattie Epps, Lev Michael and Bernat Bardagil for their important critical observations and suggestions. Any remaining errors are ours alone.

11 References Bacelar, Laércio N. 1992. Fonologia preliminar da língua Kanoê. Brasília: Universidade de Brasília MA thesis. Bacelar, Laércio N. 2004. Gramática da língua Kanoê. Nijmegen: Katholieke Universiteit dissertation. https://repository.ubn.ru.nl/handle/2066/19429 (acessed 30 April 2020). Bacelar, Laércio N. 2005. A first overview of morphology of Kanoê. Signótica: Revista do Mestrado em Letras e Linguística 17(2). 215–231. https://www.revistas.ufg.br/sig/article/view/3728/ 3483 (acessed 30 April 2020).

14 Kanoé

717

Bacelar, Laércio N. 2010, Relatório preliminar do censo sociolinguístico Kanoé. Rio de Janeiro: Museu do Índio, 40 p., manuscript. Bacelar, Laércio N. & Augusto R. Silva Jr. 2003. Tipologia da negação em Kanoê. Signótica: Revista do Mestrado em Letras e Linguística 15(2). 237–247. https://www.revistas.ufg.br/sig/ article/view/3760/3521 (acessed 30 April 2020). Bacelar, Laércio N. 2009–2013. Etnografia e documentação da língua Kanoé. (Projeto de Documentação de Línguas Indígenas (PRODOCLIN).) Rio de Janeiro: Museu do Índio. http:// prodoclin.museudoindio.gov.br/index.php/lingua/41-etnias/kanoe (accessed 25 August 2020). Becker-Donner, Etta. 1955. Notizen über einige Stämme an den rechten Zuflüssen des Rio Guapore. Archiv für Völkerkunde, vol. 10. 275–343. Wien: Wilhelm Braumüller UniversitätsVerlag G.M.B.H. Bontkes, Willem. 1967. Kanoê. Formulário dos vocabulários padrões para estudos comparativos preliminares nas línguas indígenas brasileiras. – Questionário. Rio de Janeiro: Museu Nacional – Divisão de Antropologia – Setor Lingüístico, 11 p., manuscript. Campbell, Lyle. 1997. American Indian languages: The historical linguistics of Native America. New York/Oxford: Oxford University Press. Caspar, Franz. 1975. Die Tuparí: Ein Indianerstamm in Westbrasilien. (Monographien zur Völkerkunde VII.) Berlin/New York: Walter de Gruyter. Crevels, Mily & Hein van der Voort. 2008. The Guaporé-Mamoré region as a linguistic area. In Muysken, Pieter (ed.), From linguistic areas to areal linguistics, 151–179. Amsterdam & Philadelphia: John Benjamins. Dequech, Victor. 1942. Estado de Mato Grosso: Trabalhos da Commissão do Urucumacuan nos rios Apediá e Corumbiára. Manuscript map. Galucio, Ana Vilacy, Antônia Fernanda Nogueira & Elizabeth Santos. 2008. Histórias Kanoê: Narradas por Francisco Kanoê em Kanoê e Português. Audio CD. Belém: Museu Goeldi. Grinevald, Colette & Frank Seifart. 2004. Noun classes in African and Amazonian languages. Linguistic Typology 8(2). 243–285. Maldi, Denise. 1991. O complexo cultural do Marico: Sociedades indígenas dos rios Branco, Colorado e Mequens, afluentes do médio Guaporé. Boletim do Museu Paraense Emílio Goeldi (Antropologia) 7(2). 209–269. Moore, Denny. 1988. Kanoê word list. Belém: Museu Paraense Emílio Goeldi manuscript. Nimuendajú, Curt. 1955. Vocabulários Makuši, Wapičána’, Ipurinã e Kapišana. Journal de la Société des Américanistes, nouvelle série 44. 179–197. Pottier, Bernard. 1983. America latina en sus lenguas indígenas: Coordinación, presentación y documentación. Caracas: UNESCO/Monte Ávila Editores. Seifart, Frank & Doris Payne. 2007. Nominal classification in the North West Amazon: Issues in areal diffusion and typological characterization. In International Journal of American Linguistics 73(4). 381–387. van der Voort, Hein. 2005. Kwaza in a comparative perspective. International Journal of American Linguistics 71(4). 365–412. van der Voort, Hein. 2009. Possessive expressions in the southwestern Amazon. In McGregor, William B. (ed.), The expression of possession, 343–388. Berlin: Mouton de Gruyter. van der Voort, Hein. 2015. Sistemas de classificação nominal no Sudoeste amazônico. MOARA: Revista Eletrônica do Programa de Pós-Graduação em Letras 43(2). 6–22. https:// periodicos.ufpa.br/index.php/moara/article/view/3836/3861 (accessed 30 April 2020). van der Voort, Hein. 2018. Development and diffusion of classifier systems in southwestern Amazonia. In William B. McGregor & Søren Wichmann (eds.), The diachrony of classification systems, 201–240. Amsterdam & Philadelphia: John Benjamins. van der Voort, Hein & Joshua Birchall. 2022. Aikanã. (Volume 1).

718

Laércio Nora Bacelar and Hein van der Voort

Weiss, Helga E. 1980. Fonética articulatória: Guia e exercícios. Brasília: Summer Institute of Linguistics. Zack, Estanislau. 1943. Relatório, vocabulário das tribos Massacá, Salamãin, Coaiá e Canoê. Unpublished manuscript no. 308. Rio de Janeiro: Arquivo do Museu do Índio.

Hein van der Voort

15 Kwaza 1 2 3 4 5 6 7 8 9 10

Classification, demographics, and sociolinguistic background Phonology Word classes and morphological structure The noun phrase The verb phrase Simple clauses Clause-linking, information structure, and discourse Final observations Acknowledgments References

1 Classification, demographics, and sociolinguistic background Kwaza (kwaz1243) is a language isolate, also known in earlier sources as Coaiá or Koaia. It has about 25 speakers out of an ethnic population of about 45. The language is spoken in the southeastern part of Rondônia, a federal Brazilian state that shares its southern and western borders with Bolivia. As demonstrated by Crevels and van der Voort (2008), the larger region which encompasses both sides of the border is likely a linguistic area, the Guaporé-Mamoré region, which holds an enormous genetic linguistic diversity including about 10 linguistic isolates, of which several are on the brink of extinction or have become extinct recently. Although the isolate status of these languages continues to be recognized, recent descriptive research has unveiled various linguistic features that are shared across genetic divisions (van der Voort 2009a, b, 2013, 2016, 2018). Kwaza notably shares lexical and grammatical traits with the neighboring isolate languages Aikanã (aika1237) and Kanoé (kano1245). The issue of a possible long-distance genetic connection was taken up by van der Voort (2005), who concluded that potential evidence for the connection is not sufficient to confirm such a connection. A full description of the language is van der Voort (2004). From written and oral reports, it appears that the Kwaza traditionally lived in the headwaters of the Apediá or Pimenta Bueno river, and were also neighbours to the Salamãy (Tupian, Mondé; sala1273) in the east and to the Sakɨrabiat (Tupian, Tuparian; saki1248) and other Tuparian (Tupian) groups in the west. Different ethnic groups would visit each other, crossing the Parecis plains, where the headwaters

Hein van der Voort, Museu Paraense Emílio Goeldi https://doi.org/10.1515/9783110432732-002

720

Hein van der Voort

15 Kwaza

721

of the Corumbiara, Mekens, Branco and Apediá river systems meet. Neighboring groups would know each other’s music, and multilingualism was not unusual. Direct or indirect prehistoric contacts with Nambikwaran groups further to the east, and with Jabutian (Macro-Jê) and perhaps Chapakuran groups in the remoter west may have been possible. The possibility of such contacts is suggested by occasional shared linguistic details both on the formal and the structural level. There were strong cultural similarities among neighboring groups, due to frequent contact and coexistence. Kwaza material and spiritual culture formed part of a complex that characterized southern Rondônia and that was identified by LéviStrauss (1948) as the Guaporé cultural complex and by Maldi (1991) as the Marico cultural complex. As various early sources describe (e.g., Nordenskiöld 1915, Snethlage 1939, Caspar 1975), the groups of the region shared communal straw thatched houses, subsistence hunting and cultivation of maize, sweet manioc, peanuts, yam, palm tree larvae, fermented beverages based on maize and manioc, the game of head-ball, shamanism and hallucinogenic snuff, specific musical instruments, body paint and specific body adornments (see also van der Voort 2008). Southeastern Rondônia was opened up for rubber exploitation in the early 20th century. A major road was built in the 1960s, leading to almost total destruction of the forests due to lumbering, cattle ranching, and, lately, soybean cultivation. Western material and religious culture have advanced dramatically, and presently evangelical missionaries are actively trying to eradicate specific aspects of indigenous culture. Today there is no longer a single community of Kwaza speakers, but the language survives in three separate families: two in different indigenous reserves among Aikanã linguistic majorities, and one in a local Portuguese-speaking boomtown. Despite this dispersal, the low number of speakers, and the cultural pressure from the surrounding Brazilian population, the language continues to be acquired by the youngest generation in two of the families.

2 Phonology Tables 15.1 and 15.2 give an overview of Kwaza phonemic segments and the alternative symbols that correspond to the orthography used in this chapter. Kwaza has a relatively large set of oral vowels, all of which occur frequently, with the exception of /œ/, which occurs only in a few words. There is a slightly smaller set of nasal vowels. The pronunciation of the unrounded central vowel /ɨ/ varies between [ɨ] and [ə]. The consonant inventory is also relatively large and includes implosive stops /ɓ/ and /ɗ/, as well as a series of apico-alveolar consonants, including a /t/, which contrasts with /t/, and a /s/ (variably [ʃ]), which contrasts with /s/.

722

Hein van der Voort

Tab. 15.1: Kwaza vowel phonemes. oral vowels front close close-mid open-mid open

i e ε

nasal vowels

central

ɨ œ a

back

front

central

u

ĩ ẽ ε͂

ĩ̵

o

back ũ õ

ã

Tab. 15.2: Kwaza consonant phonemes.

plosive implosive affricate fricative nasal trill/tap lateral approximant

labial

laminoalveolar

apicoalveolar

p ɓ

t

t ɗ

ts s m

w

palatal

velar

glottal

k

ɂ

tʃ s n ɾ l

h ɲ

j

Borrowed phonemes, such as the /z/ (from Aikanã) in the ethnonym and the /f/ and /v/ (from Portuguese), have not been included here. Basic syllable structure is (C)V, where C = C, G, or CG, and V = V or VG. Glides (G) in onset position [j] and [w] are interpreted as consonantal phonemes /j/ and /w/. Glides in coda position [i ̭] and [ṷ] are interpreted as allophones of the vowels /i/ and /u/ respectively. This interpretation is justified among other things by the behavior of another semivowel, [ɨ ̭], which is an allophone of the vowel phoneme /ɨ/. Unlike /j/ and /w/, the /ɨ/ does not occur in consonantal onset position or in consonant clusters. Nasality does not systematically spread to neighboring syllables, but if syllables on both sides have a nasal vowel, the vowel of the syllable in between tends to be nasalized. Glides tend to be nasalized under the influence of nearby nasal vowels. Main word stress tends to be placed on the last syllable of the uninflected stem and consequently shifts when the stem is extended by further derivational morphology, which is suffixing. In quotation constructions (Section 5.1.2), inflectional elements can receive stress. Under standard negation (Section 6.4), the derivationallike negative suffix -he attracts stress. There are no word minimality constraints in Kwaza. Kwaza is not a tone language.

15 Kwaza

723

3 Word classes and morphological structure 3.1 Major word classes The open word classes of Kwaza are nouns and verbs. Adverbs form a minor and probably closed word class. In addition, there are several particles and ideophones. Kwaza does not have a grammatically obvious class of adjectives and there are no articles or adpositions. Verbs form the only word class of which the members require additional morphology: obligatory person and mood inflection. Nominalization, verbalization and adverbialization occur frequently due to highly productive word class changing morphology and non-verbal predication.

3.2 Basic morphological profile and formative types Kwaza can be considered as a polysynthetic language, due to the considerable semantic and syntactic load of its rich bound morphology, the recursive properties of its derivation and inflection, the possibility of a large number of bound morphemes in a word, and the possibility of a full sentence consisting of a single word. The language does not meet all the received criteria for polysynthesis (Fortescue 1994, Fortescue et al. 2017), having agglutinating morphology and lacking complex morphophonology. The language has a predominantly suffixing morphology with up to 300 bound morphemes, the majority of which can be considered derivational, and only two real prefixes. Bound person and mood inflections are obligatory for a complete sentence and they always occur together at the end of the verb.

3.3 General observations regarding head- versus dependentmarking tendencies Noun phrases in Kwaza tend to be dependent-marking. In attributive constructions, which consist of juxtaposed nouns, morphological marking is usually found on the modifier. Semantically adjectival and demonstrative verbs bear nominalizing morphology. In attributive possessive constructions, the overt possessor bears nominalizing morphology, although there is an alternative construction in which anaphoric third person possession is marked on the possessum. Verb phrases are usually headmarked and the arguments, often with the exception of animate objects, do not receive case marking. Example (1) shows both attributive modification and verbal argument marking, in this case of the subject. (1)

mangka ki-hĩ̵ ya-da-ki mango ripe-nmlz eat-1sg-decl ‘I ate a ripe mango.’

724

Hein van der Voort

Example (2) illustrates that subject function is marked morphologically on the verb. Note that overt pronominal subjects tend to bear contrastive and identificational focus. (2)

si ãwĩ̵i-da-hĩ̵-ki yerexwa 1sg.pro see-1sg-nmlz-decl jaguar ‘I (and not someone else) saw the jaguar.’

Only non-third-person arguments are marked on the verb. Objects can be marked on the verb by specific inflectional person markers that are different from subject markers. When overt object nouns are present, they may receive case marking, following a pattern of differential object marking. If marked on the verb, objects tend to be animate and definite, and their expression in the form of an overt noun or pronoun usually requires an explicit accusative case marker on the nominal constituent. Example (3) shows object marking both on the verb and the pronoun. (3)

sitsɛ-wã huhui-eteya-xa-ki 1pl.excl-acc beat-1pl.obj-2sg-decl ‘You clubbed us.’

4 The noun phrase 4.1 Pronominal and demonstrative distinctions Kwaza has a category of free personal pronouns. Table 15.3 shows the distinctions between the pronouns. It furthermore shows that their forms are quite different from those of the inflectional person markers on the verb.

Tab. 15.3: Personal pronouns and subject person inflections. singular

plural

general

pro

infl

pro

infl

si

-da

I

xɨi

-xa

you

tsitsε txana xɨitsε

-axa -a -xaxa

pro we (excl) we (incl) you

infl

ĩ dilε -wa -(ha)na

he, she, it, they who, someone, no-one someone, people (indf.sbj) one (impr.sbj)

15 Kwaza

725

As Table 15.3 shows, there is a number distinction, although the third person is ambiguous1 for number. The interrogative pronoun, dilɛ ‘who’ can be used as an indefinite pronoun ‘whoever, someone, no-one’ in certain specific constructions, as shown in (4)−(6). (4)

dilε ya-re apara who eat-int banana ‘Who ate the bananas?’

(5)

dilε-here kui-tara-tse who-intn drink-proc-decl ‘Someone should drink this.’

(6)

dilε kui-he-tse who drink-neg-decl ‘No one drank.’

Unmarked pronouns refer to the subject of the sentence. In object function the accusative case marker -wã is attached, as discussed in Section 4.5. Other local semantic role distinctions such as beneficiary, comitative, etc. are also marked morphologically on the same pronouns. The different applicative morphemes on the verb, discussed in Section 5.5, may signal the presence of specific arguments. The derivation of pronouns in possessive function is discussed in Section 4.7. Person inflection, which is an obligatory part of the verb, displays additional distinctions and is discussed in Section 5.2. Kwaza has no demonstrative pronouns. Demonstrative distinctions are expressed by stative verbal roots that must be further derived by nominalizers or classifiers. The details of that process are further discussed in Section 4.3. The basis of the demonstrative system is person-oriented, distinguishing speaker-anchoring from hearer-anchoring, as shown in (7). (7)

a. Q: ĩ̵-hĩ̵-re this-nmlz-int ‘Is it this one?’

b. A: aɨ-hĩ̵-ki that-nmlz-decl ‘(Yes) it’s that one.’

The root ĩ̵ ‘this’ indicates a location close to the speaker and and the root aɨ ‘that’ indicates a location close to the hearer. In addition, distance distinctions can be expressed by the two prefixes that Kwaza possesses: na- ‘proximate’ and yã- ‘distal’,

1 In this respect, third-person pronouns are like common nouns, which are not inflected for number either. Their number can be characterized as ‘general’, in Corbett’s (2000) sense.

726

Hein van der Voort

which occur exclusively on the demonstrative roots. Furthermore, notions of location or relative physical distance such as ‘here’ and ‘there’ are expressed by the same demonstrative roots in combination with the nominalizer -hĩ̵ plus cislocative instrumental case -ko, implying ‘close by’ or the cislocative suffix -xuko ‘within the same settlement’ or the classifier/directional -ryĩ̵, literally ‘area’ but here implying ‘distant’. In combination with the adverbializing suffix -wɨ ‘time, moment of’ demonstrative roots can be used to express temporal deixis, as in (8). (8)

na-aɨ-wɨ onε-hĩ̵-ki prox-that-time arrive-nmlz-decl ‘He arrived at that time.’

4.2 Nominal number and quantification There is no grammatical nominal number marking, but multiplicity can be expressed by the collective suffix -nahere (see also Section 4.4). Its application to common nouns is optional and produces a plural sense. When applied to a personal name, it includes reference to the person’s family members, group members, etc. When combined with a place name, it refers to the inhabitants of the place. When occurring on a verb root, it has a nominalizing function and refers to a (plural) participant in the event denoted by the verb, as in (9). (9)

okya-nahere hunt-coll ‘the hunters’

Kwaza has a quinary numeral system, based on attributive verbal roots for numbers 1, 2 and 5 and adverbs expressing 3 and 4, as listed in (10). (10) teihĩ̵ akɨhĩ̵ emã elele bwakoye

‘one’ ‘two’ ‘three’ ‘four’ ‘five’

(tei ‘be one’, ‘alone’; -hĩ̵ nmlz) (akɨ ‘be two’, ‘company’; -hĩ̵ nmlz) (emã ‘one more, again’) (elele ‘several, many, very’) (bwa ‘end’, ‘to finish’; -koye clf:hand)

Traditionally, gestures in which the fingers shown in pairs may accompany counting. The fact that the number five requires the classifier for hand, -koye, as in (11), clearly shows a relationship of the system with the fingers of the hand. (11) bwa-koye-tya akɨ-hĩ̵-tse end-clf:hand-man two-nmlz-decl ‘seven’ (Lit. ‘A hand ending, it is two (fingers).’)

15 Kwaza

727

Higher numerals are derived from the above forms. Numerals above 10 are built as rather elaborate sentences and were probably rarely used. Nowadays there is much insecurity about the numerals; numerals above 5 are generally borrowed from Portuguese. Kwaza equivalents of other quantifying expressions are based on attributive verb roots, such as tyo ‘be many’ and nĩ̵ ‘be big, much’.

4.3 Classifiers There is no grammatical gender marking in Kwaza, but it is possible to identify the sex of referents. The nouns tswa ‘man’ and etaɨ ‘woman’ form the basis of the classifiers -xwa ‘man’ and -taɨ ‘woman’, respectively, as in (12) and (13). (12) Toweñẽ ĩ̵nĩ̵-wa-xwa Toweñẽ call-indf.sbj-clf:man ‘that man called Toweñẽ’ (13) mãrε’a-taɨ spirit-clf:woman ‘non-indigenous woman’ In addition, another less productive classifier that refers to males is -rati ‘man’. The Kwaza classifiers represent a typical Amazonian system (Grinevald & Seifart 2004; Seifart & Payne 2007). In Kwaza, classifiers form a large category of around 150 bound morphemes. Their semantics ranges from rather general ones, such as -ri, which refers to flat objects or beings, including discs, paddles, plates, fields and livers, to highly specific ones, such as -tehu ‘spoon’, -nãko ‘bee’ and -tuku ‘tongue’. They tend to be highly productive and have a wide distribution. They can be attached to bare (13) or derived nouns, including possessives (44); bare (11) or derived verb roots of any kind, including attributive (32); and demonstrative (14), numeral (16), and interrogative (15a). Classifiers can be incorporated into verbs and refer to arguments (17) (see also Section 5.6), they can function as anaphors, as in (14) through (17) and (44), and they can be used in a nominalizing function (12, 41). Several grammatical constructions, such as numeral, demonstrative and possessive constructions, have an obligatory position for a classifier, and in case no class can be specified, the nominalizer -hĩ̵ takes its place as a semantically neutral classifier, functioning as a grammatical placeholder, as in (7, 11, 15b). (14) (ɨwɨnwĩ̵) yã-ĩ̵-nwĩ̵ tree dist-this-clf:tree ‘the tree (over) there’

728

Hein van der Voort

(15) a. ti-nwĩ̵ what-clf:tree ‘Which tree?’

vs.

b. ti-hĩ̵ what-nmlz ‘Which one, where?’

(16) deda akɨ-dyaɨ-tse snake two-clf:snake-decl ‘There are two snakes.’ (17) ãwãta-dyaɨ-da-ki see-clf:snake-1sg-decl ‘I’m looking at the snakes.’ Their etymology divides the classifiers in three sets: those which are different from an existing free noun (although possibly related), such as -mãi ‘tooth’ versus mĩki ‘tooth’, and -dyaɨ ‘elongated animate being’ versus deda ‘snake’ (set I); those which are clearly related to an existing free noun, such as -xwa ‘man’ versus tswa ‘man’, and -tehu ‘spoon’ versus dihu ‘spoon’ (set II); and those for which no corresponding free noun exists (set III). The members of this latter set are predominantly but not exclusively body-part classifiers. In the case that an independent noun is required to express the semantic content of a set III classifier, the classifier can be attached to the semantically empty noun-formative root e in order to form an independent noun, carrying the meaning of the classifier, as in (18) and (19). (18) e-kai er-clf:leg ‘(its) lower leg’ (19) e-mũ er-clf:liquid ‘liquid, juice’ Around 30 per cent of the classifiers in Kwaza are composite, that is, they consist of multiple classifiers with lexicalised semantic content, such as -romũtsa ‘wrist’, from -romũ-tsa ‘thigh-hand’; -sikilo ‘shield’, from -siki-lo ‘skin-tube’; -tori ‘corner point’, from -to-ri ‘kernel-flat’; -tsumãi ‘point, hook’, from -tsu-mãi ‘stick-tooth’; -xuri ‘chest’, from -xu-ri ‘bone-flat’, etc. Some classifiers are only partially transparent, containing recognisable and unanalyzable parts, such as -rihi ‘flat seed’, from ‘flat-?’. In all its properties and characteristics described here, the Kwaza classifier system strongly resembles that of several of its neighboring languages. Most of the languages of the southwestern Amazon region being genetically unrelated to one another, the system must have spread areally (van der Voort 2015a, 2018).

15 Kwaza

729

Kwaza has no articles, and no contrast in definiteness is expressed, except, perhaps, by demonstratives (Section 4.1) and the interrogative pronoun dilε ‘who’, which can be used with an indefinite sense (5). There is also an important category of indefinite person markers, which will be discussed in Section 5.2.

4.4 Nominal tense, aspect, and related phenomena Nouns may show tense or aspect distinctions through suffixes that can be attached to verbs and nouns alike. There is a lamentative evaluative suffix -’wĩ̵tε that can refer to animate beings who have passed away or for whom one feels pity, e.g., dyɨ’wĩ̵tɛ ‘(my) regretted/poor/late older brother’. There is an intensifying suffix -tete that emphasizes genuineness, as in kanwã-tete ‘real canoe’. Furthermore, the suffix -ĩ’ĩ, which also is attached both to nouns and verbs, may refer to a remote past event or state, as in (20). (20) xɨi-dɨ-hĩ̵ etohoi-ĩ’ĩ-hĩ̵ tsasi-hĩ̵-tsɨ 2sg.pro-poss-nmlz child-rem.pst-nmlz follow-nmlz-rsg ‘He who was once your son came after you, you fool!’ Since there is no obligatory number marking in Kwaza, and a noun can either represent a single unit or a set of units, the collective -nahere (discussed in Section 4.2) and paucal -rɨ, as in (21), come close to producing nominal aspect in the sense of Rijkhoff (2002). (21) etohoi-rɨ-rai child-pauc-depr ‘those shameless children’ The paucal suffix in (21) relates to a small collective of several children. The deprecative suffix -rai is an evaluative suffix that can be attached to either nouns or verbs (see also van der Voort 2015b). Most tense and aspect distinctions are marked on the verb. Obviously, the highly productive nominalization processes of Kwaza allow also for verbal person, tense and aspect marking to occur inside the derived noun (see Section 5.1). Furthermore, some inflectional person markers carry tense distinctions (see Section 5.2).

4.5 Characteristics of the noun phrase Kwaza has a limited case system that reflects differential object marking. Of the core argument functions, an animate definite direct object often receives the accusative suffix -wã, as in (22).

730

Hein van der Voort

(22) kũtyẽ otsi-ki etohoi-wã Brazil.nut smash-decl child-acc ‘A Brazil nut fell on top of the child.’ Even if the object is cross-referenced on the verb (see Section 5.2), case marking may be required. Plants and plant parts are treated grammatically as inanimate. Accusative case marking of nouns referring to plants or any inanimate entity in object function is ungrammatical, as in (23). (23) akɨ-hĩ̵ emã-tya dai-ki João manga(*-wã) two-nmlz more-man take-decl João mango-acc ‘João took three mangoes.’ The case marking system seems to a certain extent lexically determined. With some verbs, such as tsye ‘close in, grab and kill’, the object is optionally marked and probably has a disambiguating function. With some transitive verbs such as tyari ‘shoot, kill game’ the object is not case-marked, unless it refers to a person. Some verbs that contain classifiers do not allow animate case marking. Some transitivized or causativized verbs only allow case marking of animate indirect objects, as in (24). (24) etohoi waya-ta-xa-re Wãzeri-wã child bring-trns-2sg-int Wãzerip-acc ‘Did you bring the child to Mrs. Wãzerip?’ Also, certain inanimate nouns, such as ‘house’, may receive accusative case marking under some circumstances and furthermore there is uncertainty about the animacy of some beings, such as the moon. The factors that determine accusative case marking are not perfectly understood and some of the uncertainty may be associated with obsolescence. There is a certain interplay between case marking and word order. As explained in Section 6.1, word order is variable, but there are some tendencies and limitations. Apparently, if case marking is optional, it becomes obligatory when the object precedes the verb, as in (25). (25) yerexwa kurakura*(-wã) tsye-ki jaguar chicken-acc close-decl ‘The jaguar grabbed the chicken.’ Kwaza does not have adpositions. Instead, semantic functions of nouns such as location, and instrument are indicated by what I have called oblique case suffixes.

15 Kwaza

731

The following functions are distinguished: locative -na; instrumental and locative -ko; beneficiary -du; and comitative -dɨnĩ̵. These are exemplified in (26)−(29). (26) hɨrikoro ãwĩ̵i-da-hĩ̵-ki kɛiñũ-koro-na monkey see-1sg-nmlz-decl jatobá-clf:arm-loc ‘I saw a monkey on a jatobá tree (Hymenaea courbaril) branch.’ (27) tsoye-ko waɨryĩ̵-ra hand-ins sweep-imp ‘Sweep (the table) with your hand!’ (28) kosa waya-nã-da-ki João-du sun take-fut-1sg-decl João-ben ‘I will take a/the flashlight to João.’ (29) axehĩ̵-da-hĩ̵-ki etaɨ-dɨnĩ̵ find-1sg-nmlz-decl woman-com ‘I found him with a woman.’ Note that the instrumental case can also be used in a cislocative function, as the contrast between (30) and (31) shows. (30) axɨ-na bukwa-xa-tsɨ-tse house-loc stay-2sg-pot-decl ‘You will stay inside (your) house!’ (31) axɨ-ko bukwa-xa-tsɨ-tse house-ins stay-2sg-pot-decl ‘You will stay inside (my) house!’ Due to highly productive nominalization possibilities in Kwaza, nouns and noun phrases can be internally very complex. There is no case agreement between the different parts of complex noun phrases. Case marking occurs only once, on the final constituent of the noun phrase, as in (32). (32) ɨwɨnwĩ̵ kara-koro-tsɨ-hĩ̵-na mouru ũ-ki tree dry-clf:arm-ger-nmlz-loc woodpecker be-decl ‘The woodpecker lives in a dry tree branch.’

4.6 Modification The absence of a grammatical category of adjectives does not impede attributive modification of nouns. Instead of adjectives, nouns are used to modify nouns. A

732

Hein van der Voort

modifying dependent noun is placed in juxtaposition with a modified head noun. Both constituents can be bare nouns (33) or derived nouns based on nouns, verbs or adverbs. (33) koreya ɨwɨnwĩ̵ knife tree ‘wooden machete’ In Kwaza, so-called “adjectival” semantic content is usually expressed by stative verbal roots, such as txi ‘be big’, awɨ ‘be cold’, wai ‘be good, nice’, atsile ‘be heavy’, etc. They can be nominalized and used as a dependent modifying noun in juxtaposition with a head noun, as in (34). (34) axɨ arwa-hĩ̵ house new-nmlz ‘new house’ A small set of adjective-like stative verb roots require an explicit attributive suffix -ĩ̵ before they can be nominalized. They often relate to appearance, such as erere ‘spotted’, tare ‘shine’, and most roots for color, as in (35). (35) hãrokãi mɛ̃ rũi-ĩ̵-hĩ̵ lagoon blue-attr-nmlz ‘blue lagoon’ There are also a number of stative verbs that require a nominalizer or a classifier before anything can be done (including verbal use). The verb arwa ‘new’ (34) is one of them. Some other roots from this set are haka ‘old’, ko ‘empty’, the numerals tei ‘one’ and akɨ ‘two’, and the demonstrative roots ĩ̵ ‘this’ and aɨ ‘that’. There is in principle no limit to the complexity of the verb or verb phrase to be nominalized, and there is no special relative clause construction that is different from the noun-noun modification constructions described above. Nominalized verbs can be internally complex with derivational and inflectional morphology, and they can also be phrases that include arguments of any type, as in (36) and (37). (36) ãwĩ̵i-da-hĩ̵-ki towɨ waraya-xa-hĩ̵ see-1sg-nmlz-decl clearing make-2sg-nmlz ‘I saw the garden that you made.’ (37) servezya kui-tɨ-ta-hĩ̵ mɛ̃ -nã-da-ki yerexwa beer drink-mal-1sg.obj-nmlz beat-fut-1sg-decl jaguar ‘I’m going to beat the dog that drank my beer.’

15 Kwaza

733

In all these constructions, instead of a nominalizer, a classifier can be used, as in (38) and (39) (also cf. 12). (38) hã ki-ĩ̵-mũ water red-attr-clf:liquid ‘red river’ (39) (axɨ) haka-xɨ (house) old-clf:house ‘old house’ Note that the semantic specificity of a classifier makes it easy to omit the head of the attributive construction without loss of referentiality. In (34) a headless construction would have the general meaning ‘new one’, but in (39) the meaning is just as specific as when headless. Of course, the pragmatic context continues to play a role in the interpretation of classifier reference and, consequently, in identifying the absent head. Example (40), based on an adverb, can in principle refer to any other leafshaped object. (40) habui-ĩ̵-xɨ outside-attr-clf:leaf ‘piece of plastic lying outside’ The first constituent of (41), nominalized by a classifier, can be considered a headless relative clause. (41) axɨ-dɨ-xa-tsɨ-kanε ãwĩ̵i-da-ki house-caus-2sg-pot-clf:oblong see-1sg-decl ‘I saw boards that you can make a house with.’ The issue of Kwaza attributive constructions, including possessives, is addressed in detail in van der Voort (2006).

4.7 Possession There is no principled distinction between attributive possessive constructions and the nominal modification constructions described in Section 4.6. Kwaza does not have specific possessive pronouns. Any noun, including the personal pronouns described in Section 4.1, or nominalized verb, can be derived by a possessive suffix -dɨ to create a stem that is obligatorily nominalized, either by a nominalizer or a classifier. The resulting possessor noun can be juxtaposed to a head noun that rep-

734

Hein van der Voort

resents the possessum, as in (42) and (43), or it can function as a headless possessive construction, as in (44) and (45). (42) si-dɨ-hĩ̵ axɨ 1sg.pro-poss-nmlz house ‘my house’ (43) kirihiu-dɨ-hĩ̵ axɨ bird.sp-poss-nmlz house ‘nest of the yellow-rumped cacique (Cacicus cela)’ (44) Tawiwi-dɨ-xɨ Tavivi-poss-clf:house ‘Tavivi’s house’ (45) na-aɨ-hĩ̵-dɨ-hĩ̵ isi-ki prox-that-nmlz-poss-nmlz die-decl ‘That person’s (son) has died.’ These examples show possession marked on the dependent. However, there is a frequently occurring alternative head-marked possessive construction in Kwaza. It involves the suffix -tyate that refers to a third-person possessor, as in (46). (46) kanwã-tyate canoe-3.poss ‘his canoe’ There is no full inflectional paradigm for possession, and this odd head-marked construction only concerns third-person possessors. Kwaza shares this irregularity with the neighboring isolate Aikanã. There is no clear productive grammatical distinction between alienable and inalienable nouns. However, for set III classifiers, which refer predominantly to body parts (see Section 4.3), there are no corresponding independent nouns. They require the empty root e in order to form such nouns. This empty root plus classifier construction may be a vestige of such an inalienable category. Synchronically, it can be regarded as a regularity in the lexicon, since not all set III classifiers refer to body parts, and not all body part nouns contain the empty noun formative root e. The empty root construction is probably an areal feature of the languages of the Guaporé-Mamoré region (see van der Voort 2009a, 2015a). However, it is specific for Kwaza that the empty noun root is, conspicuously enough, homophonous with a verb root e ‘have’. This verb is used both for existential (‘there is’) and external possessive (‘X has’) expressions.

15 Kwaza

735

4.8 Compounding, derivation and non-verbal predication Whereas compounding of verbal roots is common in Kwaza, nominal compounds are very rare and usually lexicalized, as in (47). (47) ãrũi=eke-e tapir=horn-too ‘cow’ There are also some rare compounds of nouns with an uninflected stative verb root (48) or an adverb (49). (48) atxitxi=dara maize=green ‘green maize’ (49) yerexwa=txuhũi jaguar=small ‘(wild) cat’ / ‘little dog’ Reduplication has various temporal and aspectual functions (see Sections 5.1 and 5.4). Reduplication of nouns is very rare. Reduplication of the interrogative pronoun dilɛ ‘who’ is attested: dilɛ~dilɛ ‘who else’. Apart from those that involve classifiers, there are only a few exclusively nounto-noun derivational processes. They involve certain specific suffixes, such as -rata ‘first’, -dɨnãi ‘way of, language of’, -le ‘only, just’. Other suffixes can be attached to any category and derive verbs, nouns, adverbs or any category, such as -e ‘too’; -tete ‘real’; and -rai ‘damned!’, deprecative (see also Section 4.4. and van der Voort 2004). Attachment of the verbal causative suffix -dɨ (followed by person and mood inflections) to nouns results in the meaning ‘to make N’. There are no dedicated noun-to-verb suffixes, except perhaps -tsɨ ‘you fool!’, which expresses resignation. Then again, a very frequent and productive process that I have previously called zero-derivation (van der Voort 2004), turns nouns into verbs without intervening dedicated verbalizing morphology, as in (50). (50) Konã etaɨ-ki Konã woman-decl ‘Konã is a woman.’ Nouns can be verbalized by the attachment of verbal person and mood inflections, resulting in the different meanings ‘to be N’, ‘N to be there’, ‘to want to have N’, ‘to need N’, or ‘to say N’, depending on the pragmatic context. In specific contexts it

736

Hein van der Voort

means ‘to produce’ (e.g., of flowers), or ‘to bear’ (e.g., of fruit). Nouns (including proper names) are verbalized by mood inflections when coordinated. Perhaps zeroverbalization is not a derivational process and should be regarded simply as nonverbal predication. However, the process can be recursive, leading to repeated word class changes, as illustrated by (51a,b). (51) a. yerexwa-xa-he-xwa jaguar-2sg-neg-clf:man ‘You’re not a jaguar.’ (i.e., ‘You shouldn’t eat so much meat.’) b. yo-kane txitxi-dɨ-da-hĩ̵-ki manioc-clf:oblong fry-caus-1sg-nmlz-decl ‘They are pieces of manioc which I fried.’ Example (51a) features a noun that is used as a non-verbal predicate and is then turned into a noun again by a classifier. Example (51b) starts out as a verb, which is nominalized and is then used as a non-verbal predicate.

4.9 Adpositions There are no adpositions in Kwaza. Case suffixes express some basic notions such as locative, instrumental, beneficiary, comitative, etc. (see Section 4.5). More specified locations and directions are expressed by adverbs and directionals (see Section 5.3).

5 The verb phrase 5.1 Tense, aspect, modality, evidentiality (TAME) Tense, aspect, modality and evidentiality (TAME) are optional verbal categories in Kwaza that shade into one another, forming a continuum. TAME categories are expressed mainly by derivational suffixes. Nominalization, compounding, reduplication and recursive inflection can also have TAME effects. Sentential mood is an obligatory verbal inflection. Not everything about TAME in Kwaza is well understood.

5.1.1 Tense The basic tense distinction in Kwaza is future, marked by the suffix -na, versus unmarked non-future. The unmarked verb can have either a present or a past inter-

15 Kwaza

737

pretation, depending on the pragmatic context and the presence of temporal adverbs, as in (52) and (53). (52) awɨ-mũ-ki cold-clf:liquid-decl ‘The water is/was cold.’ (53) lato awɨ-mũ-ki txarwa kike-mũ-ki yesterday cold-clf:liquid-decl first hot-clf:liquid-decl ‘Yesterday the water was cold. Today it is warm.’ The future tense suffix -na can also have a desiderative connotation, as in (54). (54) txahĩ̵ ba-yãhĩ̵-nã-tse path clear-clf:path-fut-decl ‘He wants/is going to clear the road.’ Insertion of the nominalizer -hĩ̵ between person and mood marking can create a sense of recent past, as in (55b). (55) a. xɨi awɨ-hĩ̵-xa-re 2sg.pro cold-nmlz-2sg-int ‘Are you cold?’

b. xɨi awɨ-hĩ̵-xa-hĩ̵-re 2sg.pro cold-nmlz-2sg-nmlz-int ‘Were you cold?’

Past tense use of the nominalizer is usually accompanied by other morphological processes. There is a dedicated past suffix -kɨ of relatively rare occurrence, which occurs in combination with -hĩ̵. In addition, the person marker may be reduplicated, as in (56). (56) kukuihĩ̵-da~da-kɨ-hĩ̵-ki ill-1sg~red-pst-nmlz-decl ‘I was ill.’ The difference between these alternative past constructions is not well understood. For remote past there is another dedicated suffix -ĩ’ĩ. It can be applied to both verbs and nouns and requires an accompanying nominalizer, as in (57). (57) awɨ-mũ-ĩ’ĩ-hĩ̵-ki cold-clf:liquid-rem.pst-nmlz-decl ‘The water was cold.’ (a long time ago) Alternatively, another reduplication construction involving different allomorphs of the person marker expresses remote past, as in (58).

738

Hein van der Voort

(58) xɨi awɨ-hĩ̵-xa~xaɨ-hĩ̵-re? 2sg.pro cold-nmlz-2sg~red-nmlz-int ‘Were you cold?’ (one week up to three months ago) Reduplication of person inflections in order to express tense distinctions is not syllable or mora-based but morpheme-based, and may be unique, since it has not been attested in any other language so far (van der Voort 2003, 2009b).

5.1.2 Modality Several modal meanings are distinguished in Kwaza and are expressed by derivational verbal suffixes such as -te purposive, -heta desiderative, -here conjectural, -rɨdɨ irrealis, -tara procrastinative. Several of these also have tense-like characteristics, such as -tsɨ potential, which can be used to express a less certain future than future tense suffix -na does, as in (141), (142), and (158) below. The frustrative suffix -le can have the modal connotation of unfortunateness or pity, as in (59), but it usually has an aspectual sense of incompleteness or unrealizedness, as in (60): (59) kike-mũ-here-da-le-ki hot-clf:liquid-intn-1sg-frus-decl ‘I thought (the coffee) was hot (but it isn’t).’ (60) isi-nã-xa-le-hĩ̵-ki die-fut-2sg-frus-nmlz-decl ‘You almost died.’ It is conspicuous that modal suffixes often occur in a position between person and mood inflection. Mood inflection, which will be discussed in Section 6.3, is obligatory and always occurs verb-finally, after the obligatory person inflection. Several inflectional mood morphemes have derivational modal uses or variants. For example, the monitory mood suffix -tsi sometimes occurs as a preventive suffix -tsi, before the person marker. The hortative mood -ni has an indirect causative variant -nĩ.2 The volitive mood -mĩ̵ has a volitional variant in a fixed combination with first person marker -da. These modal variants and uses of mood morphemes are regarded as instances of degrammaticalization by van der Voort (2002, 2004, 2009b, 2016), a process in Kwaza that was probably triggered by a quotative construction involving

2 Distinct from the canonical causative -dɨ, which occurs in a different position in the verb and which is neutral with regard to direct or indirect causation. The indirect causative has been referred to as “causational” in previous publications. It usually expresses well-intentioned non-forceful indirect causation.

15 Kwaza

739

recursive person inflection. Examples (61) and (62) show how speech quotation involves two layers of inflections: the inner layer relates to the quoted event and the outer layer to the event of quoting. (61) kukuihĩ̵-da-ki=da-ki ill-1sg-decl=1sg-decl ‘I said I am ill.’ (Lit. ‘I (said) “I am ill.”’) (62) danĩ̵ hɨhɨrwa-a-ni=xa-re still move-1pl.incl-hort=2-int ‘Are we still going for a walk?’ (Lit. ‘Do you still (say,) “Let’s walk!”?’) This has led to a reanalysis of some mood markers, such as hortative -ni, as in (63) and (64). (63) ya wãwĩ̵i-nĩ-xa-re already sleep-caus.indr-2sg-int ‘Did you already make (him) sleep?’ (perhaps lit. ‘Did you already (say), “Sleep!”?’) (64) hadai-nĩ-da-ki hack-caus.indr-1sg-decl ‘I cut myself (by accident).’ The modality suffixes -te purposive, -heta desiderative do not have an inflectional verb-final mood counterpart, but their position in the verb can only be understood if regarded as analogous with quotation, as in (65). (65) tyutyu=oya-da-heta-xa-re Wetsa-na walk=go-1sg-desid-2sg-int Wetsa-loc ‘Do you want to walk to Wetsa?’ (perhaps Lit. ‘Do you (say), “I want to walk to Wetsa”?’)

5.1.3 Aspect Several verbal aspects are expressed grammatically through suffixes and specific constructions. There does not seem to be a clear perfective-imperfective aspectual contrast. As mentioned in Section 5.1.1, the insertion of the nominalizer -hĩ̵ between the person and mood marker can produce a past tense meaning. Often, however, the inserted nominalizer has a progressive meaning, as in (67).

740

Hein van der Voort

(66) atxitxi bε-to-ki maize roast-clf:seed-decl ‘She roasted the maize.’ (67) atxitxi bε-to-hĩ̵-ki maize roast-clf:seed-nmlz-decl ‘She is roasting the maize.’ Depending on the semantics of the verb, reduplication of the verb root or of one of its syllables can have various aspectual effects, such as progressive (ongoing events): ɛhĩ̵ ‘nail’, ɛ~ɛhĩ̵ ‘be nailing’, iterative/repetitive (events repeated cyclically in identical manner): hãidi ‘drip’, hãidi~hãidi ‘be dripping’, frequentative (events that occur frequently during a certain time), distributive (events scattered in time or space), intensive, attenuative (the opposite of intensive). Furthermore, habitual aspect can be expressed through embedded reduplication of the subject person marker, as in (68). (68) itso-na ũi-xa~xa-ta-xa-re hammock-loc lie-2sg~red-dep-2sg-int ‘Do you always sleep in a hammock?’ There are several other ways to express habituality. The combination of a medial clause with matrix verb a ‘exist’ creates a habitual sense, for example, ‘He exists eating fish.’ Another way to express habituality is the combination of a nominalized clause with matrix verb e ‘have’, as in (69). (69) hehe-nĩ̵-kutɨ-nãi e-da-hĩ̵-he-ki cover-refl-clf:head-nmlz have-1sg-nmlz-neg-decl ‘I don’t use a blanket.’ The verb root ĩ’ĩta ‘as always’ and the bound morphemes -ĩ’ĩ’e ‘always’ and -tyaryĩ̵ ‘much’, are frequently used for habitual aspect. Whereas the verb root occurs as the second part of a compound, the bound morphemes are applied between the verb stem and the person marker, as in (70) and (71). (70) ĩ̵hĩ̵ko awe=ĩ’ĩta-tse here rain=always-decl ‘Here it always rains.’ (71) tomã-kore-tyaryĩ̵-da-ki bathe-dir:matutinal-much-1sg-decl ‘I always take a morning bath.’

15 Kwaza

741

Intensive aspect is not only expressed by reduplication, but also by the genuineness morpheme -tete ‘real’, as in (72). (72) awe-tete-tse ĩ̵-hĩ̵-ko rain-real-decl this-nmlz-ins ‘Here it always rains a lot.’ Finally, in addition to reduplication there is also a morpheme -e ‘again’ that expresses repetition, as in (73). (73) bui-e-da-ki axɨ-ko leave-again-1sg-decl house-ins ‘I left the house again.’

5.1.4 Evidentiality Evidentiality is not a clear grammatical category in Kwaza. Person markers can be involved in expressions of evidentiality. In certain situations, the use of the indefinite person marker -(ty)wa ‘someone, people, they’ (see Section 5.2.1) may have an inferential meaning as in (74). (74) kawε kui=bwa-tywa-ki coffee drink=end-indf.sbj-decl ‘(There is no coffee in the pot, so, apparently) they finished the coffee.’ Furthermore, recursive person and mood marking is used to quote speech, which obviously shifts the burden of evidence from the speaker to the person quoted, as in (75).3 (75) yanãdɨ-da-ki-tse hungry-1sg-decl-decl ‘She says she is hungry.’ (Lit. ‘She says, “I’m hungry.”’) The verb-final morpheme -tehere (also -tyehere) ‘it seems that’ can express a sense of deductive conjecture, here called inferential, as in (76).

3 As we can see from (62) and (65), this construction is not used only to literally quote direct speech, but also to report on (presumed) thoughts, intentions, desires, purposes and even inanimate processes, presenting those as fictive interaction (van der Voort 2013, 2016).

742

Hein van der Voort

(76) wɨ-tsɨ-hĩ̵-tyehere storm-ger-nmlz-infr ‘Apparently it has stormed.’ (e.g., there are many branches of trees on the ground) The appellative morpheme -hereyã expresses an appeal to the hearer’s knowledge and can function as a tag question with the meaning ‘Isn’t it?’, as in (77). (77) kaɨte~te-a-hĩ̵-hereyã-re scrape~red-1pl.incl-nmlz-appl-int ‘We clean it (the mortar) on the inside, don’t we?’

5.2 Person Grammatically speaking, the expression of arguments as free constituents is optional. The only obligatory parts of a sentence are the verb root and inflectional suffixes that express subject arguments and mood. Person (subject) and mood inflections form part of coherent paradigms and they canonically occur together at the end of the verb root or stem (i.e., the derived verb), in that order, as in (78). (78) hɨhɨrwa-da-ki walk-1sg-decl ‘I am walking.’ The expression of object person is less regular, and the morphemes have a different distribution. Exceptions to the pattern sketched here include omission of person and mood inflections, omission of the root (i.e., free occurrence of the inflections), and recursive application of the inflections (see Section 5.1.2). Then again, the omission of roots or inflections – called ‘morphological ellipsis’ in van der Voort (2004) – occurs only when they are clearly understood from the pragmatic context. In (79) the second person and interrogative mood morphemes -xa-re are understood, and in (80) the verb root kui ‘drink’ is understood (see also Section 7.5). (79) kawe kui coffee drink ‘Did you have coffee?’ (80) a. Q: kui-nã-xa-re drink-fut-2sg-int ‘Are you going to drink?’

b. A: he-nã-da-ki neg-fut-1sg-decl ‘I’m not.’

15 Kwaza

743

Because of their canonical expression as a tight inflectional cluster, the categories of person and mood are both treated in the present subsection. The insertion of morphemes from a very limited set, such as countersupposition (see Section 6.4), and frustrative and potential (see Section 5.1.2), forms an exception to the adjacency of person and mood inflections. As mentioned in Section 4.1, Kwaza has a category of free personal pronouns and a set of person inflections. Table 15.3 in Section 4.1 compares both paradigms. Similar to pronouns, person inflections contrast for person, number, and in/exclusive. Unlike pronouns, person inflection is an obligatory part of the verb phrase and refers to the subject of the sentence. The default interpretation of absence of person inflection is the third person, or, in other words, third person is zero-marked. An exception to this pattern is due to interaction with mood: in the imperative mood, it is the second person that is zero-marked, as in (81). (81) waɨ-siñwã-ra e-siñwã sweep-clf:yard-imp er-clf:yard ‘Sweep the yard!’ The presence of a free argument constituent such as a pronoun is optional and expresses contrastive or, as in (82), identificational focus. (82) txana bu’ũdwa-a-re 1pl.incl.pro sit-1pl.incl-int ‘Is it us who are sitting?’ (discussing a photo) The object argument can be expressed by inflections, although they do not constitute a clear regular paradigm. Some of the object morphemes also have a limited subject value (i.e., the forms are relevant to particular subject-object combinations), and one of them even has a tense value; these can be considered portmanteaux. Some can only be combined with certain subjects. Table 15.4 shows in the vertical axis the object part of the suffixes in bold type, and in the horizontal axis the subject part in italic type. Note that third person objects are zero, similar to third-person subjects. Furthermore, several object morphemes, including first and second plural, and indefinite, combine with the regular subject inflections. The first-person singular object, however, does not combine with the standard second person subject morpheme, and is therefore ambiguous and requires an overt subject (pro)noun or demonstrative expression, as in (83) and (84). (83) xɨi erewe-ta-ki 2sg.pro write-1sg.obj-decl ‘(It was) you who body-painted me.’

744

Hein van der Voort

Tab. 15.4: Object person inflections combined with subject and future tense. 1sg 1sg.obj 2sg.obj

nĩnã-da

3

-da

2sg

3

ta

ta hata

-xa eteya-xa

1pl.obj

2pl.obj

etelexwada

indf.obj

ya-da

ya-xa

1pl.incl

1pl.excl

nĩnã-a

nĩnã-axa

-a

-axa

eteya etelexwa

etelexwaa

etelexwaaxa

ya

ya-a

ya-axa

2pl

indf.sbj

ta-xa

ta-xwa xu-xwa

-xaxa

-wa

eteyaxa

eteya-wa etelexwawa

ya-xaxa

fut

leya

leya

ya-wa

(84) na-aɨ-hĩ̵ erewe-ta-ki prox-that-nmlz write-1sg.obj-decl ‘(It was) he who body-painted me.’ The second person object is expressed in the most diverse way: with first persons it is -nĩnã-, with third persons it is -hata-, with indefinite subjects it is -xu-, and there is even a future tense second person (singular or plural) object -leya. It replaces the normal future tense morpheme (see Section 5.1.1), as in (85) and (86). (85) si erewe-nĩnã-da-ki 1sg.pro write-2sg.obj-1sg-decl ‘I’m body-painting you.’ (86) si erewe-leya-da-ki 1sg.pro write-2.obj.fut-1sg-decl ‘I’m going to body-paint you (sg or pl).’

5.3 Directionals Kwaza has no adpositions, but instead local case markers and a set of approximately 40 verbal derivational morphemes that specify direction when attached to action, process or motion verb stems like hɨ ‘go’, dai ‘take’, and aru ‘put’. The semantic load of directionals tends to be quite specific, for example -kurwa ‘foot of tree’, -mãte ‘into water’, -kore ‘matutinal, of the morning’, -totε ‘upwards from below’. Various directionals relate to orientation with respect to the forest, hillsides, and rivers, such as -dwatya ‘forest’, -katsa ‘middle of the path’, -kutɨtε ‘beyond, behind, over the hill’, and -nõwẽ ‘upstream’, as in (87).

15 Kwaza

745

(87) hɨ-nõwẽ-da-ki go-dir:upstream-1sg-decl ‘I went upstream.’ (paddling, swimming, etc.) Directionals can be used in combination with a ground, which can be an adverb (88) or a noun, as in (89) and (90), or even a classifier (91). If the ground is a noun, it may bear the general locative case marker -na, whatever the direction. (88) habui aru-tywa-ra outside put-dir:high-imp ‘Put it outside!’ (89) axɨ-na hou-hĩ̵=otye-ritsa-da-ki house-loc take-nmlz=throw-dir:outside-1sg-decl ‘I pull her out of the house.’ (90) lokãtywa-na koreyaro aru-kwate-da-mĩ̵ oven-loc pan put-dir:inside-1sg-vol ‘I’m going to put the pan into the oven.’ (91) aru-kya-siñwã-wa-hĩ̵-lε place-dir:length-cd:yard-indf.sbj-nmlz-cond ‘…when they have put (the mortar) upright in the yard…’ Directionals differ from classifiers in several respects, e.g., with respect to their more limited distribution and their lack of a nominalizing potential. But there are also similarities. Directionals form a large subset of bound grammatical morphemes with very specific lexical-like semantic content. Moreover, they can be turned into independent adverbs when attached to a dummy root, analogous to classifiers, as in (92). (92) a-nõwẽ toma-ra er-dir:upstream bathe-imp ‘Take a bath upstream!’ Note that the adverb formative dummy root a is homophonous with the existential verb root a, which represents another parallel with the classifier system, where the noun formative dummy root e is homophonous with the verb root e ‘have’. Furthermore, there is a small subset of morphemes that are ambiguous between directionals and classifiers, such as -ryĩ̵ ‘area, surroundings, all over, room, place, weather’, -rwε ‘around’, and -katye ‘wood (on top of horizontal wood)’. These morphemes can be used either way.

746

Hein van der Voort

(93) do-katye=du-ryĩ̵-da-ki meza-na leak-cd:wood=all-cd:area-1sg-decl table-loc ‘I spilt water all over the table.’ As with classifiers, most of which probably originate from nouns, directionals may also have a lexical origin, probably originating from verb roots that express associated motion. There are still some verbs that, in second position in a compound, can function as directionals. Verbs such as kwε ‘enter’ in (94) and asa ‘leave, end, do away’ in (95) occur often in that position and suggest a grammaticalization path (see also Section 5.9). This cannot be said of all compound verbs, as in (96) and (97). (94) tsũ=kwε-ki hakidwa-rona-na swim=enter-decl stone-clf:hole-loc ‘He entered the cavern swimming.’ (95) ɨwɨnwĩ̵ dɨ=asa-wa-ki tree cut=leave-indf.sbj-decl ‘Someone cut the log up (and cleared it out of the way).’ (96) aɨ-ryĩ̵ buru=oya-ki that-cd:area pass=go-decl ‘He passed by there.’ (i.e., he did not stop) (97) onε=kui-da-mĩ̵ come=drink-1sg-vol ‘I came to drink.’ ~ ‘I’m going there to drink.’

5.4 Pluractionals Pluractionality, or verbal number, is not a highly productive morphological operation. It is expressed basically in two different ways: through reduplication or through root suppletion. Both full and partial reduplication have various uses, including repetitive, frequentative and distributive, e.g., tyari ‘shoot’ > tya~tyari ‘shoot many’, esa ‘sting’ > e~esa ‘sting frequently/by many (e.g., a swarm of wasps)’, dai ‘take’ > da~dai ‘take various’, and aru ‘jump’ in (98). (98) tei-hĩ̵-kai aru~aru-tse one-nmlz-clf:leg jump~red-decl ‘He hops on one leg.’

15 Kwaza

747

Some cases of reduplication express participant plurality (which is different from argument agreement, which represents nominal number), e.g., aruu ‘cross’ > arurunĩ̵ ‘cross (pl subject)’, bui ‘go outside’ > bubui ‘go outside (pl subject)’, and oya ‘go away’ > o’oya ‘go away (pl subject)’. Participant plurality can also be expressed by verb root suppletion: hɨ ‘go (often sg subject)’ and towε ‘go (pl subject)’; onε ‘arrive (hither, originating from here)’ and bwεnε ‘arrive (pl subject)’; onĩ̵ ‘come, arrive (pl subject)’ and wanĩ̵ ‘come, arrive (pl subject)’; asa ‘end, leave, terminate, obliterate (sg object)’ and (u’)uya ‘end, leave, terminate, obliterate (pl object)’. There are only few suppletive verb roots in Kwaza.

5.5 Valency-changing morphology Kwaza verb roots can be inherently transitive, such as dai ‘take, grab something’, tsɛ̃ i ‘tear something’, or intransitive, such as atsile ‘be heavy’ and tsũ ‘swim’. These verbs do not have any morphological marking that identifies their valency. Also, the inherent valency of many transitive verbs, such as ãwĩ̵i ‘see’ and wadɨ ‘give’, can be decreased without any further morphological adaptation. There are several verbal morphemes that can increase valency. The causative morpheme -dɨ promotes a causer to argument status, demoting the original subject to object status, as in (99)− (101). (99)

hoi-ki sweet-decl ‘It is sweet.’

(100) hoi-dɨ-ki sweet-caus-decl ‘He sweetened it.’ (101) etohoi-wã mĩkau hɛro-dɨ-da-ki child-acc porridge lick-caus-1sg-decl ‘I’m feeding the child porridge.’ The causative morpheme -dɨ may derive from the verb root wadɨ ‘give’. Creissels (2010) suggests that wadɨ ‘give’ in second position in Kwaza compounds can be analysed as a grammaticalized benefactive applicative. The data from Kwaza, as in (102), appear to indicate that such a development may be taking place. (102) towɨ tyutɛ=wadɨ-nĩnã-da-hĩ̵-ki clearing fell=give-2sg.obj-1sg-nmlz-decl ‘I cleared a garden plot for you.’

748

Hein van der Voort

Other valency-increasing morphemes are applicative and introduce additional objects with different semantic functions. The morpheme -tɨ has a detrimental or malefactive meaning, introducing an object’s possessor at the expense of which an event occurs, as in (104). (103) auxwana ya-da-hĩ̵-ki meat eat-1sg-nmlz-decl ‘I ate meat.’ (104) auxwana ya-tɨ-da-hĩ̵-ki meat eat-mal-1sg-nmlz-decl ‘I ate his meat.’ In case an overt malefactive object occurs, it does not get accusative or an oblique case marker, but it is marked as the possessor of the involved direct object, as in (105). (105) atxitxi ba-tɨ-nã-da-ki Mario-dɨ-hĩ̵ maize cut-mal-fut-1sg-decl Mario-poss-nmlz ‘I’m going to harvest maize at Mario’s expense.’ Kwaza has a comitative applicative -ete-, shown in (106). (106) babaitye-ete-nã-da-ki Magarida dance-com-fut-1sg-decl Margarida ‘I will dance with Margarida.’ The meanings of the applicative -ta or -tya are less well defined. The morpheme generally adds an object with respect to which an event occurs, as in (107) through (110). (107) hɨya-ki fall-decl ‘It fell.’ (108) hɨya-ta-ki fall-trns-decl ‘It fell in front of him.’ (109) ĩwã kitse-tya-taxwa-ki idly lie-trns-indf.sbj.1sg.obj-decl ‘They lied to me.’

15 Kwaza

749

(110) hoñe-ta-ki mã-tyate-wã hide.self-trns-decl mother-3.poss-acc ‘They hid from their mother.’ When applied to a transitive verb, it may introduce another object, as in (111). (111) waya-ta-ki bring-trns-decl ‘He is taking it there to him.’ The reflexive is expressed by the morpheme -nĩ̵. It indicates that the subject acts upon itself, as in (112). (112) haɨ-xɨi-kotye-nĩ̵-nã-da-ki exɨi cut-clf:hair-clf:neck-refl-fut-1sg-decl hair ‘I’m going to cut my own hair.’ The reciprocal is expressed by the morpheme -lε, as in (113). (113) Pedru Sabina huruya-lε-ki Pedro Sabina like-recp-decl ‘Pedro and Sabina like each other.’ Kwaza does not have morphology expressing voice. Instead of a passive voice that demotes a subject, the impersonal subject inflection -na / -hana (115) and the indefinite person inflection -wa (116) can be used with a similar effect (see Foley 2007: 429), without, however, promoting the object to subject status. (114) Kwaza-dɨ-nãi pɛ̃ rɛ̃ yã-axa-ki Kwaza-poss-nmlz speak-1pl.excl-decl ‘We are speaking Kwaza’ (115) ĩ̵hĩ̵ko Kwaza-dɨ-nãi pɛ̃ rɛ̃ yã-na-ki here Kwaza-poss-nmlz speak-impr.sbj-decl ‘Here one speaks Kwaza.’ (116) hirini ta-wa-hĩ̵ healer say-indf.sbj-nmlz ‘the one called healer, the one they call healer’ The reflexive, mentioned in Section 5.5, can also have a valency-decreasing effect in case the pragmatic context does not permit the subject to act upon itself. Example (118) is as close as one can get to a passive or middle voice in Kwaza.

750

Hein van der Voort

(117) iritxi-ki cover-decl ‘He closed (put the lid on) the pan.’ (118) iritxi-nĩ̵-ki tie-refl-decl ‘It (e.g., the pan) is closed.’ There is no argument structure altering incorporation in Kwaza. Classifiers, which tend to derive from nouns and have rather lexical content, can be incorporated in the verb root, without changing the argument structure. The classifier -tyukwe ‘pod’ agrees with and further specifies the object in the following example, as in (119). (119) kũtyẽ kã-tyukwe-da-ki Brazil.nut toast-clf:pod-1sg-decl ‘I toasted Brazil nuts.’

5.6 Morpheme ordering Basic morpheme order of the verb is root-derivation-inflection. Table 15.5 shows the different slots. The table shows some order variability. As an example, the negative morpheme -he-, which is normally located at the end of the verb stem, can be inserted between the person and mood marker to express negation of supposition (see Section 6.4). The other alternative positions of derivational morphemes are the result of quotative, reduplicative and nominalization constructions that involve recursive verbalization. Alternative morpheme orders often involve lexicalized idiomatic interpretations, and no clear or consistent effects of scope variation are attested. Perhaps alternative negation positioning can be explained in terms of scope.

Tab: 15.5: Morpheme positions in the verb. root

drv

drv

drv

drv?

drv

infl

drv

drv

infl/drv

root

dir clf attr other

valency

tense modality aspect clf dir

object

neg

subject

neg

tense modality aspect other

mood modality aspect other

15 Kwaza

751

5.7 Syntactic characteristics of the verb phrase The verb phrase obligatorily consists of a verb root and its subject person and mood inflections. The presence of other elements, such as derivational morphemes, overt independent arguments, and adverbs, is optional from a grammatical point of view. Word order appears to be relatively free, although some tendencies dominate. In transitive sentences both VO and OV orders are frequently attested. In intransitive sentences, SV order prevails although VS order is also allowed. Different word orders do not appear to go with different semantic meanings. They probably are, however, associated with different pragmatic meanings.

5.8 Serial verb compounds Compounding of verb roots is very common, unlike nominal or hybrid compounding. In fact, verbal compound constructions tend to refer to a single event and can be considered as serial verb constructions. Inflection occurs only once, on the last member of the compound, as in (120) and (121), and elsewhere. (120) itsosisi dɨ=wanε-ra cord cut=bring-imp ‘Bring me a piece of rope!’ Nevertheless, the first member may be morphologically complex, as in (121). (121) tyari-tɨ=asa-a-ni kill-mal=leave-1pl.incl-hort ‘Let’s kill him off!’ The entire compound construction can be nominalized, as in (122). (122) were=oya-taɨ-nahere revolt=go-clf:woman-coll ‘the runaway women’ Often the compound reflects a sequence of events, in iconic order, as in (120) and (123). (123) terya=hɨya-tse slide.down=fall-decl ‘He slipped and fell.’

752

Hein van der Voort

Compounds may express purpose, as in (124). (124) ãwĩ̵i=ε-ki see=go-decl ‘He went there to see.’ A semantically attributive member of the compound may modify the other member as if it were an adverb, as in (125). (125) be=hε-dɨ-hĩ̵-ta-ki black=paint-caus-nmlz-1sg.obj-decl ‘I’m painted black (with genipap) by someone.’ Certain semantically abstract verb roots have a rather grammatical function when they occur in serial compounds, as if they were used as suffixes (see also Sections 5.3 and 5.5): asa ‘leave’ (terminative); bwa ‘end’ (completive); kwɛ ‘enter’ (illative); hãrã ‘stop’ (terminative); and wadɨ ‘give’ (benefactive), as illustrated by (102) and (126). (126) ui kurye=wadɨ-nĩnã-da-ki tobacco wind=give-2sg.obj-1sg-decl ‘I rolled a cigarette for you.’

6 Simple clauses 6.1 Basic constituent order As mentioned in Section 5.7, basic constituent order is variable. Focus is generally on the first constituent of the sentence. In transitive sentences with two overt independent arguments, the first argument tends to be the subject. Distinctions of pragmatic nature, however, be they related to the speech context or to knowledge of the world, may determine alternative orders. On the basis of general knowledge, it should be clear to any speaker of Kwaza who represents the subject and the object in (127). (127) kurakura tsye-ki yerexwa chicken close-decl jaguar ‘The/a jaguar grabbed the/a chicken.’ The position of oblique arguments is also determined by pragmatic and iconicity factors. The same holds for the position of adverbs, which, although they tend to

15 Kwaza

753

occur at the beginning of the sentence, can be placed anywhere. Finally, the order of noun and attributive modifiers (including demonstratives) is usually NMod although other orders are also attested (see Sections 4.5 and 4.6). The pragmatic conditions and implications of word order variation in Kwaza await further research. Most complex constructions in Kwaza tend to be head-final, except attributive phrases and clauses. Their predominant head-initial position is probably due to their verbal origin, as (128) and (129) suggest. (128) yerexwa kanexu bu-ki jaguar necklace put-decl ‘The dog wears a necklace.’ (129) ãwĩ̵i-da-ki yerexwa kanexu bu-hĩ̵ see-1sg-decl jaguar necklace put-nmlz ‘I saw the/a dog with a necklace.’ The object in (129) can be considered a relative clause, the head of which, yerexwa, is the subject of the nominalized verb buhĩ̵.

6.2 Alignment Kwaza follows a nominative-accusative alignment pattern. The distinctions can be expressed by case marking, often when the object is animate (Section 4.5) and by person marking, with the object marker preceding the subject marker (Section 5.2). Inanimate objects are identified by the pragmatic context and a syntactic tendency to reserve first position in the sentence for subjects.

6.3 Sentential mood In Section 5.1.2, reanalysis, or better, degrammaticalization, of certain inflectional mood markers was discussed. Mood markers are verb-final inflections that can be divided into two general types: matrix moods and subordinate moods. Matrix moods are inflections that express the modal status of the main clause, such as declarative and imperative, whereas subordinate moods determine the mood of subordinate adverbial clauses, such as conditional and concessive (Section 7.1). There are four basic matrix mood categories, as Table 15.6 shows. The table shows two forms for the declarative mood. The mood expressed by -ki could be indicative rather than declarative, because in the third person it (unlike -tse) never occurs in combination with future or potential morphemes, which express irrealis situations. An (ir)realis distinction could perhaps be a historical expla-

754

Hein van der Voort

Tab. 15.6: Mood markers of Kwaza matrix verbs. Declarative

-ki /-tse

Interrogative

-re

Persuasive

Imperative Hortative-Causative Volitive

-ra -ni -mĩ ̵

Prohibitive

Negative Imperative Negative Hortative Monitory

-kɨ -(i)ni -tsi

nation, but as argued in van der Voort (2004: 284–96) and shown in various examples in the present chapter no such distinction are attested in a consistent way. Two mood categories, persuasive and prohibitive, can be further divided into different sub-moods that have a more limited distribution. These sub-moods complement each other in the sense that they all occur with different subject persons. As one might expect, the imperative occurs only with (unmarked) second person subjects, the volitive only with first persons, and so on. Example (130) contains two juxtaposed matrix clauses. (130) wãwĩ̵i-he-kɨ oya-tsi sleep-neg-neg.imp go-admon ‘Don’t sleep (lest) he will leave!’

6.4 Negation Negation is realized by the verbal negative suffix -he that usually attracts main word stress and precedes the inflectional person and mood complex, as in (6), (130), and (131). (131) axe-hĩ̵-he-da-ki ãrũi find-nmlz-neg-1sg-decl tapir ‘I did not see a tapir.’ In case negation concerns a presupposition on the part of the speaker or the hearer, the negative morpheme occurs in between the person and mood markers and does not attract stress, as in (51), (69), (132), and (134). Note the contrast between (131) and (132).

15 Kwaza

755

(132) axe-hĩ̵-da-he-ki ãrũi find-nmlz-1sg-neg-decl tapir ‘I did not see a tapir.’ (but other game, e.g., peccary) In the third person, under standard negation, the declarative mood marker is -tse, as in (133), whereas under negation of presupposition, it is -ki, as in (134). (133) mĩu baɨ-mũ-he-tse chicha sour-clf:liquid-neg-decl ‘The chicha4 isn’t fermented.’ (speaker warns a hearer who intends to drink) (134) mĩu baɨ-mũ-he-ki chicha sour-clf:liquid-neg-decl ‘The chicha isn’t fermented.’ (speaker establishes an unexpected fact) Negation in the declarative and interrogative moods is transparent, either preceding or following the person marker. In other moods, however, things are not that straightforward. In Section 6.3, the special prohibitive moods negative imperative, negative hortative, and monitory were mentioned, contrasting with the three persuasive moods imperative, hortative causative, and volitive.

6.5 Clause-level particles There are various interjection-like elements, such as the hortative particle hĩ̵dɛ ‘let’s (go)!’, the delay particle yale ‘wait (a little)!’, the exclamative kukui ‘wow!’ (apparently derived from the verb root kukui ‘hurt’), deprecative kwɨ ‘how horrible!, yecch!’ and the amazement or surprise particle tsεh ‘why!’. In polysyllabic particles, stress falls on the final syllable. (135) hĩ̵dε ya-a-ni biskoito go.imp eat-1pl.incl-hort biscuit ‘Let’s eat biscuits!’ The adverb harɨkɨ ‘now, then’ can also be used as a particle expressing hesitation, as in (136). (136) awɨ-e-hĩ̵ etohoi awɨ-ki=tywa-ki harɨkɨ etohoi-tete get-again-nmlz child get-decl=indf.sbj-decl now child-real ‘People say she got a child again, uhh, a real child.’

4 Slightly alcoholic traditional drink made of fermented manioc, maize or other produce.

756

Hein van der Voort

7 Clause-linking, information structure, and discourse 7.1 Subordination As mentioned in Section 4.6, there is no principled difference between nouns and nominalized verbs or clauses. Nominalized verbs can function as arguments, as attributive (“relative”) clauses, or as complement clauses.5 Examples (137) and (138) show how both subjects and objects are accessible to “relativization” (see Keenan & Comrie 1977). (137) hẽu-da-ki auxwana ho-hĩ̵ smell-1sg-decl meat rot-nmlz ‘I smelled rotten meat.’ (138) xui owɨ-da-hĩ̵ wadɨ-leya-da-ki bag crochet-1sg-nmlz give-2.obj.fut-1sg-decl ‘I will give you the bag that I crocheted.’ Attributive clauses are the result of nominalization by -hĩ̵ and refer to participants of events. Complement clauses, which refer to the events themselves, may be formed by the adverbial nominalizer -nãi, as in (139). (139) etohoi buru-tsɨ-hĩ̵-ki lote kuro-da-nãi child remember-ger-nmlz-decl door close-1sg-nmlz ‘The child remembered that I had closed the door.’ Most adverbial clauses are not formed through nominalization but by specific subordinate moods expressing notions such as concessive, conditional, additive, and manner. In Section 6.3, matrix clause moods were mentioned. Subordinate moods occur only in complex constructions. Usually, the matrix clause is preceded by the subordinate clause. Example (140) contains a concessive clause. (140) ĩ aru-nĩ̵-tõi-lete ãwĩ̵i-ryĩ̵-he-tse 3sg.pro place-refl-clf:eye-conc see-cd:area-neg-decl ‘Although he uses glasses, he does not see.’

5 Besides being used as a nominalizer, the morpheme -hĩ̵ is also used as a generic classifier (as discussed in Section 4.3), and it has temporal (Section 5.1.1) and aspectual (Section 5.1.3) uses.

15 Kwaza

757

There are several subordinate “moods”, some of which require a particular modality in the main clause, as the examples in (141) and (142) show. (141) xɨi towɨ tyutɛ-xa-kɨwɨ si yo anũ-da-tsɨ-tse 2sg.pro clearing fell-2sg-cond 1sg.pro manioc plant-1sg-pot-decl ‘If you clear the field, I will plant manioc.’ (142) mĩu horonĩ̵hĩ̵-da-le ũkai-da-tsɨ-tse chicha prepare-1sg-prec lie-1sg-pot-decl ‘First I’m going to prepare chicha and then I’ll lie down (in the hammock).’ Complex nominalized clauses can also be used adverbially. Because of the frustrative modality morpheme, (143) can be considered as containing an antithetic adverbial clause. (143) manini mũrɛ̃ dɨ-da-le-hĩ̵ ho-hĩ̵-ki fish contribute-1sg-frus-nmlz rot-nmlz-decl ‘I bought (paid) fish, but it was rotten.’ Example (144) shows the use of the nominalizer ‘time of’, being used to form a temporal adverbial clause. (144) kehĩ̵=bwa-nã-tsɨ-wɨ dɨtena-da-ki burn.up=end-fut-ger-time extinguish-1sg-decl ‘I extinguished the fire when (the house) was about to burn down.’

7.2 Clause chaining and switch-reference Another type of grammatically subordinate clause occurs in chains of medial clauses that do not differ in mood from the matrix clause, although they lack specific matrix mood marking.6 Medial clauses in Kwaza are marked by a “dummy” mood morpheme -ta, which I have called dependent mood. They tend to precede7 the

6 In van der Voort (2004) I called them cosubordinate clauses, following Foley & Van Valin (1984). Later analyses by Foley (2010) and Bickel (2010) show, however, that cosubordination probably does not represent a third category that is intermediate between coordination and subordination. For Kwaza one can argue that medial clauses are both coordinate and subordinate. They can be regarded as syntactically subordinate to the matrix clause and semantically dependent on it for mood. Furthermore, they can occur in chains of multiple medial clauses that are coordinate among one another. 7 That is, in narratives rather than in formally elicited sentences, where the medial clause often follows the matrix clause as an afterthought.

758

Hein van der Voort

matrix clause, which contains the mood inflection that holds for all preceding medial clauses. In principle, the matrix clause can be in any mood, as demonstrated by (145) through (147). (145) butyehĩ̵-a-nãi areta-he-da-ta okya-he-da-ki shoot-1pl.incl-nmlz know-neg-1sg-dep hunt-neg-1sg-decl ‘I can’t hunt because I don’t know how to shoot.’ (146) ε-toto-mã-da-ta barε-nĩ̵-da-mĩ̵ go-dir:upward-dir:water-1sg-dep warm-refl-1sg-vol ‘I’m going to go up on the riverbank to warm myself up.’ (147) tsuhũ-du-xa-ta ohui-nã-xa-re what-ben-2sg-dep play-fut-2sg-int ‘Why are you going to play ball?’ Medial clauses generally express consecutive or overlapping events or explanations of purpose, causation, or reason that contextualize the matrix clause. Clause chains can be very long, containing many medial clauses. In the course of the subsequent clauses, the mood does not change, but the subject may. Two morphological switchreference subsystems account for subject changes with regard to the next clause. The following examples contrast same subject reference with different subject reference. If both different subjects are third persons, the switch is marked by the preinflectional element -dɨ, as in (149). (148) kopo koreyari-na hɨya-hĩ̵-ta kãu-tse cup plate-loc fall-nmlz-dep break-decl ‘The cup fell onto the plate and broke.’ (149) kopo koreyari-na hɨya-hĩ̵-dɨ-ta koreyari kãu-tse cup plate-loc fall-nmlz-ds-dep plate break-decl ‘The cup fell onto the plate and the plate broke.’ If at least one of the different subjects is not a third person, the switch is indicated by an alternative medial clause marker, as in (150) and (151), which I have called “switch-reference (mood)”, -si.8

8 Similar to the dependent mood suffix -ta (note that I inaccurately regarded the subordinate manner clause marker -tya in previous work as an allomorph of -ta), the suffix -si is a purely grammatical element that does not have any mood value of itself, but depends on the matrix clause for its mood content. As it happens, -si does not only indicate that the subject of the next clause is different, but also that its person value is different, and -dɨ always indicates that the different subject of the next clause is a third person (van der Voort 2021).

15 Kwaza

759

(150) atsile-si daiwarya-he-da-ki heavy-swr lift-neg-1sg-decl ‘As it is heavy, I can’t lift it.’ (151) onε-xa-si bui-da-hĩ̵-ki arrive-2sg-swr leave-1sg-nmlz-decl ‘You returned and then I left.’ Traditional narratives may consist of many clauses forming in fact one single sentence, terminated by an elliptic formula in the declarative mood: tywa-rati-ki (indf.sbj-foc-decl) ‘That’s how they (say)’.

7.3 Coordination As shown in the sections on subordination and clause chaining, clauses can be linked by bound morphemes. Several strategies to indicate conjunction and disjunction are expressed by grammatically subordinate constructions. Medial clause constructions can often (but not always) be interpreted as semantically coordinate, as in (152). (152) tyaɨ darahĩ̵-ta tãi-ki papaya green-dep tough-decl ‘The papaya is green and hard.’ In principle, juxtaposed matrix clauses can only be interpreted as separate sentences and not as coordinate clauses. In clause chains, however, multiple medial clauses occur juxtaposed and can be regarded as grammatically coordinate among themselves. Such coordinations are also attested with multiple conditional clauses, as in (153). (153) ti-wɨ-here ɛ-da-lɛ, a-da-lɛ towɨ waraya-da-tsɨ-tse what-time-intn go-1sg-prec exist-1sg-prec clearing make-1sg-pot-decl ‘When I leave, I will live over there and make a clearing.’ The expression of negative coordination often involves ellipsis of person and mood marking, as in (154) and (155). (154) ya-he=kui-he-tse eat-neg=drink-neg-decl ‘He didn’t eat and didn’t drink.’

760

Hein van der Voort

(155) kawe e-he atsuka e-he-da-ki coffee have-neg sugar have-neg-1sg-decl ‘I have neither coffee nor sugar.’ Apart from the morphological means described above, there are three clause-linking particles in Kwaza. One of those is the sequential particle tana ‘(and) then’, which occurs often between chained medial clauses, as in (156). (156) kui-da-ta tana kui-da-ta tana isi-da-ki drink-1sg-dep then drink-1sg-dep then die-1sg-decl ‘I drank and drank (the whole night) until I passed out.’ Adversative coordination (Haspelmath 2007: 2) is expressed by the concessive particle wara ‘but’, which is cliticized to a nominalized clause in a subordinate construction, as in (157). (157) koreyaro kãu-tsɨ-hĩ̵=wara hã hɨ-he-tse pan break-pot-nmlz=but water go-neg-decl ‘The pan has a crack, but it still holds water.’ Disjunction can be expressed by the selective particle hele ‘or’, as in (158). This particle is used quite rarely. (158) hã kui-da-tsɨ-re hele mĩu kui-da-tsɨ-tehere water drink-1sg-pot-int or chicha drink-1sg-pot-infr ‘I’m going to drink either water or chicha.’

7.4 Information structure As a general rule, the first constituent of the sentence is in focus position (6.1). Furthermore, the use of the grammatically optional personal pronouns has a focus function (4.1). Another strategy to create focus is recursive nominalization, as in (159). (159) towɨ-na-ĩ̵-hĩ̵ clearing-loc-attr-nmlz ‘from the garden’ (Lit. ‘something that is from the garden’) Apart from these basic focus strategies, topic and focus are indicated by various derivational elements or clitics that occur on different types of lexemes. There is a grammatical construction by which the speaker expresses a contrast with what the hearer knows or presupposes. It involves the verb-final morpheme -hehĩ̵, which replaces the mood marker, as in (160).

15 Kwaza

761

(160) a. C: tsewe-dɨ-tõi-hata-tsi throw-caus-clf:eye-3.sbj.2.obj-mon ‘Take care that the toad doesn’t spit in your eye!’ b. R: aru-nĩ̵-tõi-da-hehĩ̵ place-refl-clf:eye-1sg-cntr.top ‘But I have glasses!’ The pragmatic contrastive topic clitic =mãyã introduces a new discourse topic that contrasts with the previous topic, as in (161). (161) hayere=mãyã tãlo-he-tsɨ-nãi e-hĩ̵-tsɨ pig.sp=top angry-neg-ger-nmlz have-nmlz-res ‘Now the collared peccary is not very aggressive.’ The contrastive focus morpheme -txitxi, which may be a clitic, was attested with adverbs and nouns, as in (162). (162) Teteru-txitxi mũi-ni Teteru-cntr.foc submerge-hort ‘(I want) Teteru to fetch water! (and not you)’ The verbal morpheme -txetxi expresses surprise and has the forceful sense of ‘really’, as in (163). (163) korewe-txetxi-tɨ-ta-le-re pregnant-really-mal-1sg.obj-frus-int ‘Why! Would she really be pregnant on me?’ The morpheme -rati has an identificational focus function and can applied to nouns, as in (164). (164) tsũhũ-rati-mũ-tete kui-xa-re what-foc-clf:liquid-real drink-2sg-int ‘What kind of chicha are you drinking?’ It can furthermore be applied to verbs, in which case it expresses historical narrative and is inserted between the person and mood markers, as in (165). (165) axɨhi ĩ-wa-rati-ki parica sniff-indf.sbj-foc-decl ‘This way they sniffed parica.’

762

Hein van der Voort

The categorical element -le ‘only, just, nothing but’ is attached to nouns. In narratives it may have a much wider scope, meaning ‘suddenly, exactly’, and introduce a new topic or a new event in the discourse, as in (166). (166) yerexwa-le ti-hĩ̵ tsũhũ-rati-wã nãi-xaxa-re ta jaguar-only what-nmlz what-foc-acc like-2pl-int talk ‘Then (a man called) Jaguar said: “What are you all looking at?”’ Finally, switch-reference marking in clause chains does not always indicate different subjects. In cases where the subject remains the same, as in (167), it signals a different topic. (167) erewe-xɨ-da-si kukui-koye-da-ki write-clf:leaf-1sg-swr hurt-clf:hand-1sg-decl ‘It is because of writing that my finger is hurting.’ Grammatically speaking, one would have expected same subject marker -ta in this example rather than switch-reference -si. Wherever there is subject continuity in spite of switch-reference marking, there is topic discontinuity.

7.5 Discourse strategies Although Kwaza person and mood marking are analyzed as a complex of bound verbal inflectional suffixes that are obligatorily attached to a verb, either they or the verb root can be omitted if the context is such that these are understood. For example, under the proper intonation and in the right context, inflections can be omitted, as in (79) above and in (168). (168) tsuhũ emã what cry ‘Why are you crying?’ Similarly, in the right context, the verb root can be understood, which leaves the inflections free, forming, as it were, an inflectional particle, as in (80) above and in (169). (169) a. Q: kuro-xaxa-re close-2pl-int ‘Did you close the door?’

b. A: axa-ki 1pl.excl-decl ‘We did.’

I have analyzed both types of minimal expressions as morphological ellipsis, and not as real particles or auxiliary verbs, in part because their interpretation is highly

15 Kwaza

763

context dependent and not interpretable in isolation, and especially because consultants say that these expressions are “not complete”. However, some cases of morphological ellipsis occur so frequently in certain grammatical contexts that they may be regarded as (lexicalized) particles. Especially in narratives, inflections often occur by themselves as resumptive particles, to re-establish an interrupted medial clause-chaining structure, as in (170), but also in other monologues, as in (171). (170) nãi-tya-hĩ̵-tadɨ tya ĩwã like-man-nmlz-excl man nothing ‘Ah, that’s what they are doing! (Yes) they lied.’ (171) enãi-tya-ta-tsɨ-tse da-ta areta-nã-da-ki quarrel-trns-1sg.obj-pot-decl 1sg-dep know-fut-1sg-decl ‘He is going to quarrel with me, so I, I’m going to learn.’ When asked for a translation of a freely occurring inflectional morpheme, informants typically give various different possibilities, depending on the situation and the particular inflection, such as ‘and then’, ‘well’, ‘I’, ‘then I’, ‘so I’, ‘she’, ‘well you’, ‘and thereafter’, ‘yes’, etc. Under root ellipsis, derivational morphemes may also remain attached to the inflectional morphemes. As a result, the elliptic expression hehĩ̵si, which is probably a semi-lexicalized combination of -he neg, -hĩ̵ nom, and -si swr, can be used as a tail-head linking particle meaning ‘well then’, as in (172). (172) a. Q: ya ya-xa-re already eat-2sg-int ‘Did you already eat?’

b. A: danĩ̵ ya-he-da-ki still eat-neg-1sg-decl ‘I have not yet eaten.’

c. C: he-hĩ̵-si ya-ra neg-nmlz-swr eat-imp ‘Well then (since (you) didn’t), eat!’

8 Final observations The Kwaza language has several properties that stand out typologically. Morphemebased reduplication of person markers (Section 5.1.1) has not been attested anywhere else. This is very different from recursion of person and mood marking, which functions to express quoted speech as well as fictive interaction (Section 5.1.4), but which has also led to degrammaticalized structures and forms (Section 5.1.2). Several other properties are characteristic for many languages of South America, such as complex morphology, prolific nominalization, lack of grammatical number, a dis-

764

Hein van der Voort

tinction between inclusive and exclusive reference, clause chaining, and a switchreference system. Certain traits are regional and are probably best explained in terms of areal diffusion, such as specific properties of classifying constructions, possessive constructions, and certain formal and lexical traits. The language is still relatively little researched, and some points are not well understood or merit a better description. As a linguistic isolate, Kwaza furthermore continues to represent a challenge to historical linguistics.

9 Acknowledgments It is a pleasure to express my gratitude to my Kwaza friend and teacher Mario Aikanã, with whom I have worked on the Kwaza language for more than 25 years. I am also immensely grateful to the other speakers of the language who have always been willing to be consulted and have contributed substantially to my research. Furthermore, I am hugely indebted to the indigenous communities at the TubarãoLatundê and Rio São Pedro indigenous reserves in Rondônia for their infinite hospitability. I also owe many thanks to the organizations that have funded various projects concerning Kwaza during the years: the Netherlands Organisation for Scientific Research (NWO), under TSL project #300–72–021 and VIDI project #276–70–005; the Volkswagen Foundation (VWS), under DobeS projects #85.611 and #92.740; and the Fundação de Amparo à Pesquisa do Estado do Pará (FAPESPA), under PPDOC project #004/2010; as well as the respective hosting institutions Universiteit Leiden, Universiteit van Amsterdam, Radboud Universiteit Nijmegen, Max Planck Institute for Psycholinguistics (MPI), and the Museu Paraense Emílio Goeldi (MPEG). Finally, I want to thank Pattie Epps, Lev Michael, Zachary O’Hagan and the anonymous reviewers from whose comments and corrections this chapter has benefitted enormously. Responsibility for the contents is mine only.

10 References Bickel, Balthasar. 2010. Capturing particulars and universals in clause linkage: A multivariate analysis. In Bril, Isabelle (ed.), Clause linking and clause hierarchy: Syntax and pragmatics, 51−101. Amsterdam & Philadelphia: John Benjamins. Caspar, Franz. 1975. Die Tuparí: Ein Indianerstamm in Westbrasilien. Berlin & New York: Walter de Gruyter. Corbett, Greville G. 2000. Number. Cambridge: Cambridge University Press. Creissels, Denis. 2010. Benefactive applicative periphrases: A typological approach. In Fernando Zúñiga & Seppo Kittilä (eds.), Benefactives and malefactives: Typological perspectives and case studies, 29−70. Amsterdam & Philadelphia: John Benjamins. Crevels, Mily & Hein van der Voort. 2008. The Guaporé-Mamoré region as a linguistic area. In Muysken, Pieter (ed.), From linguistic areas to areal linguistics, 151−179. Amsterdam & Philadelphia: John Benjamins.

15 Kwaza

765

Foley, William A. 2007. A typology of information packaging in the clause. In Shopen, Timothy (ed.), Language typology and syntactic description, Vol I: Clause structure, 362−446. Cambridge: Cambridge University Press. Foley, William A. 2010. Clause linkage and nexus in Papuan languages. In Bril, Isabelle (ed.), Clause linking and clause hierarchy: Syntax and pragmatics, 27−50. Amsterdam & Philadelphia: John Benjamins. Foley, William A. & Robert D. Van Valin. 1984. Functional syntax and universal grammar (Cambridge Studies in Linguistics 38). Cambridge: Cambridge University Press. Fortescue, Michael. 1994. Polysynthetic morphology. In R. E. Asher & J. M. Y. Simpson (eds.), The encyclopedia of language and linguistics, 2601−2602. Oxford: Pergamon Press. Fortescue, Michael, Marianne Mithun & Nicholas Evans (eds.). 2017. The Oxford handbook of polysynthesis. Oxford: Oxford University Press. Grinevald, Colette & Frank Seifart. 2004. Noun classes in African and Amazonian languages. Linguistic Typology 8(2). 243−285. Haspelmath, Martin. 2007. Coordination. In Shopen, Timothy (ed.), Language typology and syntactic description, Vol II: Complex constructions, 1−51. Cambridge: Cambridge University Press. Keenan, Edward L. & Bernard Comrie. 1977. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8. 63−99. Lévi-Strauss, Claude. 1948. Tribes of the right bank of the Guaporé River. In Steward, Julian H. (ed.), Handbook of South American Indians: Volume 3, the tropical forest tribes, 370−379. New York: Cooper Square Publishers Inc. Maldi, Denise. 1991. O complexo cultural do Marico: Sociedades indígenas dos rios Branco, Colorado e Mequens, afluentes do médio Guaporé. Boletim do Museu Paraense Emílio Goeldi (Antropologia) 7 (2). 209−269. Nordenskiöld, Erland. 1915. Forskningar och Äventyr i Sydamerika. Stockholm: Albert Bonniers Förlag. Rijkhoff, Jan. 2002. The noun phrase. Oxford: Oxford University Press. Seifart, Frank & Doris L. Payne. 2007. Nominal classification in the north west Amazon: Issues in areal diffusion and typological characterization. International Journal of American Linguistics 73(4). 381−387. Snethlage, Emil H. 1939. Musikinstrumente der Indianer des Guaporégebietes. Baessler-Archiv, Beiträge zur Völkerkunde [Beiheft X]. Berlin: Dietrich Reimer (Andrews & Steiner). van der Voort, Hein. 2002. The quotative construction in Kwaza and its (de-)grammaticalisation. In Crevels, Mily, Simon van de Kerke, Sérgio Meira & Hein van der Voort (eds.), Current Studies on South American Languages, 307−328. Leiden: Research School of Asian, African, and Amerindian Studies (CNWS). http://www.etnolinguistica.org/illa:vol3n20 van der Voort, Hein. 2003. Reduplication of person markers in Kwaza. Acta Linguistica Hafniensia 35. 65−94. van der Voort, Hein. 2004. A grammar of Kwaza. Berlin: Mouton de Gruyter. van der Voort, Hein. 2005. Kwaza in a comparative perspective. International Journal of American Linguistics 71(4). 365−412. van der Voort, Hein. 2006. Construções atributivas em Kwazá. Boletim do Museu Paraense Emílio Goeldi (Ciências humanas) 1(1). 87−104. http://www.scielo.br/pdf/bgoeldi/v1n1/v1n1a07.pdf van der Voort, Hein. 2008 [1998]. Kwazá. In Ricardo, Fany Pantaleoni (ed.) Encyclopedia of indigenous peoples in Brazil. São Paulo: Instituto Socioambiental. http:// pib.socioambiental.org/en/povo/kwaza van der Voort, Hein. 2009a. Possessive expressions in the southwestern Amazon. In McGregor, William B. (ed.), The expression of possession, 343−388. Berlin: Mouton de Gruyter.

766

Hein van der Voort

van der Voort, Hein. 2009b. Reduplication and repetition of person markers in Guaporé isolates. Morphology 19(2). 263−286. http://www.springerlink.com/content/1545301353748536/ fulltext.pdf van der Voort, Hein. 2013. Fala fictícia fossilizada: O tempo futuro em Aikanã. Boletim do Museu Paraense Emílio Goeldi (Ciências Humanas) 8(2). 359−377. http://www.scielo.br/pdf/bgoeldi/ v8n2/09.pdf van der Voort, Hein. 2015a. Sistemas de classificação nominal no sudoeste amazônico. MOARA− Revista Eletrônica do Programa de Pós-Graduação em Letras 43(2). 5−22. http:// www.periodicos.​ufpa.br/index.php/moara/article/view/3836/3861 van der Voort, Hein. 2015b. Kwaza. In Grandi, Nicola & Lívia Körtvélyessy (eds.), Edinburgh handbook of evaluative morphology, 606−615. Edinburgh: Edinburgh University Press. van der Voort, Hein. 2016. Recursive inflection and grammaticalized fictive interaction in the southwestern Amazon. In Pascual, Esther & Sergeiy Sandler (eds.), The conversation frame: Forms and functions of fictive interaction, 277−299. Amsterdam & Philadelphia: John Benjamins. van der Voort, Hein. 2018. Development and diffusion of classifier systems in southwestern Amazonia. In McGregor, William B. & Søren Wichmann (eds.), The diachrony of classification systems, 201−240. Amsterdam & Philadelphia: John Benjamins. van der Voort, Hein. 2021. Clause chaining and switch-reference in Aikanã and Kwaza. Boletim do Museu Paraense Emílio Goeldi. Ciências Humanas 16(3). 1–18. https://doi.org/10.1590/21782547-BGOELDI-2021-0077

Raoul Zamponi and Chris Rogers

16 Máku 1 2 3 4 5 6 7 8 9

Classification, demographics, and sociolinguistic background Phonology Word classes and morphological structure The noun phrase The verb phrase Simple clauses Clause-linking, information structure, and discourse Conclusion References

1 Classification, demographics, and sociolinguistic background Máku (maku1246) is an extinct northern Amazonian language1 once spoken by a group of people dwelling in the vicinity of the Auari (or Auaris) River located in the extreme northwest of the state of Roraima, Brazil (Koch-Grünberg 1913: 457, 1917: 48, 170). The recorded history of this language is marked by a steady decline in the speaker population as a consequence of communicable diseases and external social pressures. In 1925, there were probably little more than 50 speakers of the language split among at least three communities: two malocas (‘community houses’) along the upper Uraricoera (Rice 1928: 214, 217), and an encampment along the middle course of the river near the Kulekuleima Rapids (Migliazza 1965: 30). Attacked by the Kasilapai (a Ninam-speaking Yanomami group) and, especially, decimated by European diseases in the 1930s, by 1950 there was only one Máku family, of eight or nine people, left on the lower Uraricoera. By 1969, two speakers of Máku remained, Sinofrônio (Kuluta) Magalhães (approximately 50 years old) and his sister, Maria (approximately 55 years old; Migliazza 1980: 115). At the end of the 1990s, both of

1 Not to be confused with Makú – a non-genealogical grouping of northwest Amazonia that has long been assumed to include the languages Hup, Yuhup, Dâw, and Nadëb (the Naduhupan family), the sisters Kakua and Nukak (the Kakua-Nukak family), and Puinave (an isolate) – nor with Mako (or Maco), a language of the Sáliban family (closely related to Piaroa) spoken by a group located along the Ventuari river and its tributaries (Amazonas, Venezuela).

Raoul Zamponi, Macerata, Italy Chris Rogers, Brigham Young University, USA https://doi.org/10.1515/9783110432732-003

768

Raoul Zamponi and Chris Rogers

16 Máku

769

these speakers had died (Maciel 2014: 8), and the language ceased to be spoken.2 The map on the previous page shows the Máku speaking-area around 1910. The name Máku (recorded as [ˈmaˈko] ~ [ˈmaˈku]3) is a word of Arawakan origin used in the Içana-Vaupés basin to refer to various nomadic populations of this region (Koch-Grünberg 1906: 877; Martins & Martins 1999: 251). It is highly probable that the Máku had contact with these peoples (see below), and it is commonplace to refer to this language as Máku in published resources. According to Sinfrônio, the Máku referred to themselves, among themselves, as jukude-itse ‘person-pl; people’ (pronounced [zoˌkudeiˈtse]) (Ernesto C. Migliazza, p.c.). Máku is a linguistic isolate (Koch-Grünberg 1913: 457; Loukotka 1968: 151; Migliazza 1965: 1; Rodrigues 1986: 95). Greenberg (1987: 93) suggested that it seems to form a group with two other geographically close languages, Sapé (Kaliana) and Arutani (Urutani, Uruak), but no significant evidence has been found supporting this claim and it is largely disregarded by scholars. However, common grammatical features and loanwords suggest contact with speakers of other languages of the region of the Içana-Vaupés basin, such as Arawakan and Tukanoan. For example, numeral classifier suffixes, evidentiality morphemes, two or more past tenses, a distinction between near and remote future, suffixal negation, and serial verb constructions with concordant marking of subject person are shared by many of the languages in that area. The following (probable) loanwords also provide evidence for some form of contact.

Máku loan

gloss

source

inamulu juʔwi lipina

‘stingray sp.’ ‘jaguar’ ‘knife’

Baniwa hinámaɾu ‘stingray’ (cf. also Baniva (of Guainía) inámàɺu) Tariana jáwi or Baniwa dzáwi perhaps from an Eastern Tukanoan language (cf. Tucano diʔpʰĩ ́ ́ [nĩpĩ ŋ́ ã] ‘knife’) and Tuyuca dĩpĩgã

1.1 Previous work and sources Not much data about the Máku language are known to have been collected, and only a small amount of the information that has been collected was ever analyzed and published. The information that was published was either short word lists

2 Maria died about six years before Sinfrônio (Thiago Costa Chacon p.c.). Neither Sinfrônio nor Maria had children who spoke the language (though Maria did have a child that spoke Portuguese; Migliazza 1965: 18). 3 These are probably careful pronunciations (see Section 2.2), recorded respectively by Migliazza (1964g) and Koch-Grünberg (1917: 302), of a name that, very likely, was normally pronounced with the stress on the second syllable (cf. Faria’s (1927) ). The Ninam (Yanomami) called the Máku makuri (see Koch-Grünberg 1917: 200).

770

Raoul Zamponi and Chris Rogers

(Koch-Grünberg 1913: 458, 1928: 317–24; Migliazza 1978), basic profiles of phonology and syntax (Migliazza 1965, 1966), or a reproduction of earlier materials (Aikhenvald & Dixon 1999: 361–362). Iraguacema Lima Maciel (2014) provides the text of a traditional tale with a free translation. Other unpublished materials were produced by João Barbosa de Faria (1927) and Iraguacema Lima Maciel (1991). However, important information on the language was collected independently by three linguists, Ernesto C. Migliazza, Aryon Rodrigues, and Iraguacema Lima Maciel, from 1958 to 1990. This information consists of handwritten field notes and unanalyzed audio recordings of Sinfrônio Magalhães,4 which present a fairly good overview of the grammatical properties of the language. Since these recordings are of a native speaker of Máku, they provide an invaluable database of information that can be used to describe Máku grammar and are used as the primary data throughout this chapter. In light of the situation and rather scarce documentation available on Máku, it is obvious that the corpus of available language data is now closed. This, in turn, affects the types of analysis and/or descriptions of the language that are possible. For example, methods with a high degree of control over the data collection process – such as experimental techniques and/or data elicitation – are either highly restricted or are not possible. The descriptions below represent a rigorous and systematic representation of the data, as it is now understood. However, there are many questions, issues, and generalizations that are left unanswered or untreated. The data for the descriptions below mostly come almost exclusively from the audio recordings, because these represent the most empirically accurate access to the language. The transcriptions of words and sentences were then transferred to a grammatical database using FieldWorks Language Explorer and ELAN.5 This database consists of 1,580 lexical items, 447 example sentences, and 15 texts (naturalistic discourse). Each of the lexical items was organized according to a headword in phonemic representation, glossed in both English and Portuguese, and marked for grammatical category and cross-referenced to the example sentences and texts.

2 Phonology The phonetic description of Máku sounds is difficult since the quality and quantity of the data are less than ideal. However, based on the transcriptions and phonetic analysis of the audio recordings, the following phonemic contrasts are observed.

4 The audio material recorded by Migliazza is available online at the Archive of the Indigenous Languages of Latin America (ailla.utexas.org; PID: ailla:124425); see Migliazza (1958b, 1960b, 1960c, 1960d, 1964b). 5 Our work on Máku was supported by a National Science Foundation (NSF) Documenting Endangered Languages (DEL) grant number 1524606 .

16 Máku

771

Tab. 16.1: Máku consonant inventory.

stop fricative affricate nasal approximant

voiceless voiced voiceless voiceless voiced voiced

labial

alveolar

p b ɸ

t d s ʦ n l

m w

palatal

velar

glottal

k

ʔ h

j

Tab. 16.2: Máku vowel inventory. front

high mid low

p:b p:ɸ ɸ:w w:b m:n t:d t:ʦ s:ʦ j:d ʔ:Ø ʔ:h ɸ:h u:e i:ɨ i:y y:ɨ i:e e:a u:y

central

back

unrounded

rounded

unrounded

rounded

i e

y

ɨ

u

a

tepediba paneʔsia ɸanikɨsia wesia maʦu sukute tene sekepe jakwana buʔte menehe leɸese jukude ekɨnɨ ki lymydia ene tene nu

‘I will stay.’ ‘It is sharp.’ ‘(S)he is thinking.’ ‘It is hot.’ ‘mouse’ ‘(his/her) eye; grass’ ‘I’ ‘(S)he stays.’ ‘De’kwana’ ‘earth’ ‘paca’ ‘fly’ (n.) ‘person’ ‘grease’ ‘this’ ‘(S)he caught.’ ‘you’ (sg) ‘I’ ‘my mother’

tebediba ɸanikɨsia wane besia naʦu jukude ʦene ʦekete dakalu bute weʔe leheniamuʦa kete ekinuba ky lɨnɨkɨʦɨ ine tane ny

‘I will throw.’ ‘(S)he is thinking.’ ‘his/her brother-in-law’ ‘(S)he is throwing.’ ‘grass’ ‘person’ ‘we’ (incl) ‘our (incl) head’ ‘a kind of basket’ ‘river’ ‘(S)he does not sleep.’ ‘(S)he came inside.’ ‘his/her head’ ‘You (sg) die.’ ‘(S)he sucks.’ ‘sand’ ‘louse’ ‘his/her cousin’ ‘It flies.’

Table 16.1 shows the consonant contrasts and Table 16.2 shows the vowel contrasts. These tables are followed by a set of minimal pairs as evidence of these contrasts. While the analysis of Máku phonology is ongoing, we can suggest some of the ways that Máku seems to compare to other languages in the region. First, two seg-

772

Raoul Zamponi and Chris Rogers

ments are cross-linguistically noteworthy: /y/ and /ɸ/. The first is extremely uncommon in Amazonian languages and not found in any of Máku’s neighbors. The second is also uncommon in Amazonia, but an allophone [ɸ] of /p/ seems to have been present in the above-mentioned Sapé and Arutani (see Zamponi forthcoming: Sections 6.1 and 6.2). The relative rarity of this sound in Amazonia and its presence in these two neighboring languages suggest a potential connection between speakers of these languages. Second, there are a number of gaps in the phonemic inventories of both consonants and vowels. Of particular interest is the regular symmetry for four series of consonants (voiced stops, voiceless stops, fricatives, and nasals) at the labial and alveolar places of articulation, but the asymmetrical relationship for consonants at the velar place of articulation. Similarly, the vowel contrast between /y/ and /ɨ/ creates an interesting asymmetry in the inventory for rounding. Third, an acoustic analysis of the vowel tokens in the audio reveals that [o] is both acoustically and statistically distinct from [u]. However, phonologically, [o] and [u] are often in free variation and no clear minimal pairs are known. Consequently, [o] is considered an allophone of /u/ though the complete profile of this vowel phone is still undetermined (see Rogers 2020: 31–32). Lastly, the exact articulation of /j/ is difficult to determine. In some words, it is clearly pronounced as the palatal approximant [j], while in others it is pronounced [z] or [ð]. It seems that [ð] is used between certain vowels, and [z] occurs word-initially.

2.1 Phonological variation 2.1.1 Voicing of /k/ The voiceless velar stop /k/ is sometimes voiced intervocalically in regular speech (found in the texts and some elicited sentences), but remains always unvoiced in slow, careful speech (found in elicitations): /asake/ [ˌasəˈke] ~ [ˌaːsəˈge] ‘today, now’, /mas(a)taka/ [maˌsətəˈka] ~ [ˌmastəˈgə] ‘child’.

2.1.2 Palatalization and vowel reduction to glides When an alveolar or velar consonant is followed by /i/ and an additional vowel, the consonant and /i/ sounds coalesce into a single palatalized sound. Thus, before other vowels the sequences, /ti/, /di/, /li/, /si/, /ʦi/, /ni/, /ki/, are produced as [tʲ], [dʲ], [lʲ], [ʃ], [ʧ], [ɲ], [c], respectively. For example: /ʦiumu/ [ʧuˈmu̞ ] ‘how?’, /niamu/ [ɲaˈmu̞] ‘your (sg) bone’, /lukia/ [ˈluˈca] ‘(S)he is old’, /kudialu/ [ˌkudʲaˈlu̞] ‘canoe’, /keʔlia/ [keʔˈlʲa] ‘day’. Only in careful speech do we (rarely) see evidence for /i/ between /s/, /ʦ/, or /n/ and a vowel different from /i/ in the form of a palatalization of the preceding consonant, though in all cases it is unclear if speakers of Máku

16 Máku

773

would be aware of this trace of the high vowel. This is the case, for example, of the temporary aspectual marker [-ʃa] ~ [-ʃja] that we analyzed as containing of an underlying /i/, that is, -sia (as in /tianjusia/ [ˌtʲanzoˈʃja] ‘I am cold’). In most other cases there is no phonetic evidence to suggest the presence of /i/ following a palatalized consonant. In these cases, we assumed that the vowel is underlyingly present to conform to our phonemic analysis of those words for which an alternation [ʃ] ~ [ʃj], [ʧ] ~ [ʧj], or [ɲ] ~ [nj] was occasionally noted. It must also be observed that when /s/ and /ʦ/ are followed by /i/, and /i/ does not precede another vowel, the two alveolars are realized as palatals, but in this case the front vowel is realized as a separate speech sound: /ʦimala/ [ˌʧimaˈla] ‘arrow’, /sipinu/ [ˌʃipiˈnu̞] ‘(S)he kills’. Unlike the other alveolars, /s/ and /ʦ/ are also palatalized before the high front rounded vowel /y/ and /ʦ/ even before the high central vowel /ɨ/. For example, /sylesia/ [ˌʃyleˈʃa] ‘It is blowing’, /meʦyna/ [ˈmeˈʧyˈna] ‘wind’, /ʦeʦɨʦɨ/ [ˈʦeʧɨˈʧɨ] ‘our (incl) breast’. In /CiV/ sequences different from those indicated above, the vowel /i/ is realized as [j], as in kuʦisibiukudia [kuˌʧiʃiˌbjokuˈdʲa] ‘They hit each other’. The palatal approximant [j] is also the phonetic manifestation of /i/ either word-initially before another vowel or intervocalically: /ia/ [jaː] ‘moon’, /laiamuʦa/ [laˌjamuˈʦa] ‘(S)he made’. In /CuV/ sequences, the vowel /u/ is usually realized as [w]: /kua/ [kwa] ‘that’, /jabui/ [zaˈbwi] ‘near’ (postposition), /maluaka/ [maˌloaˈka] (careful speech) ~ [ˌmalwaˈka] ‘samaúma tree’.

2.1.3 Nasalization Vowels are nasalized when they occur word-finally after a nasal consonant, the glot́ ̞] ‘It burns’, /nyhe/ [nỹˈhɛ͂] tal fricative /h/, or glottal stop /ʔ/, for example: /niʔi/ [nĩ ʔĩ ‘fire’, /dɨʔɨ/ [ˈdɨʔĩ ̵] ‘tapir’. Anticipatory nasalization of vowels is usually weak and not consistent. In the first two examples, this looks very similar to nasal harmony described for many languages in the Amazon region. However, the final example shows that a nasal is not required for nasalization to be produced on a vowel after /ʔ/. This is sometimes called “rhinoglottophilia” and is common in many languages throughout South America (Matisoff 1975; Ohala & Busà 1995).

2.1.4 Centralization of /a/ When the vowel /a/ does not occur in a primary-stressed syllable, it is commonly articulated as [ə]. For example, /salakakɨsia/ [səˌkələˌkakɨˈʃa] ‘(S)he is sitting’, /kalumadana/ [ˈkaːloˌmədaˈna] ‘It is yellow’, /kuditakana/ [ˌkudiˌtəkəˈna] ‘It is small’.

774

Raoul Zamponi and Chris Rogers

When /a/ follows the labial approximant /w/, however, it is often slightly raised and backed to [ɐ] (and gradiently rounded because of the preceding /w/). For example, /waiʦe/ [wɐ̹iˈʦe̞] ‘they’, and /ʦewaʦi/ [ˌʦewɐ̹ˈʧi] ‘their mouth’.

2.1.5 Vowel lowering In word-final position, all vowels except /a/ are usually lowered in the articulatory space. For example, /muʦi/ [ˈmuˈʧi ̞] ‘its meat’, /lemeky/ [ˌlemeˈky]̞ ‘land turtle’, /ekɨnɨ/ [ˈeːkɨˈnɨ ̞] ‘grease’, /ekaʦu/, [ˈekaˈʦu̞ ] ‘horn’, /tene/ [ˈteːˈne̞] 1sg.pro. In wordmedial position, this type of vowel lowering occasionally affects /e/ and /u/, being articulated as [ɛ] and [o], respectively. The lowering of /u/ to [o] is also noted in rare instances in word-initial position, for example, /une/ [oˈne̞] ‘his/her mother’ and /uba/ [oˈba] ‘tree’.

2.2 Syllables and stress Syllable nuclei are the only obligatory component of syllables, and they can be simple or complex. Both monoconsonantal onsets and codas are permitted in the language. Syllables with complex nuclei include diphthongs (/ei/, /ai/, /ui/, /au/) and usually have an onset. One exception is /ain/ ‘(S)he fishes’. The template for the largest possible syllable in Máku is (C)V(V)(C). However, CV syllables are the most common. Stress predominantly correlates with and (high) pitch and is not contrastive. Primary stress is typically placed on the rightmost syllable of a word. Secondary stress appears to alternate every other preceding syllable in words composed of more than two syllables: malakata [maˌləkaˈta] ‘kind of latex’, ʦalaniamuʦa [ˌʦalaˌɲamuˈʦa] ‘they made’. When a word is pronounced carefully or slowly, it can begin with an extra-stressed syllable, or each syllable may become stressed or be produced separately from the others: /kalumana/ [ˈka:lumaˈna] ‘it is while’, /luna/ [ˈloːˈna] ‘rattlesnake’, /lemena/ [ˈleːmeˈna] ‘it is red’. The vowel of the initial syllable of the word may be lengthened, as the examples indicate.

2.3 Phonological rules and processes 2.3.1 Vowel deletion The (rightmost) vowel /e/ in the pronominal markers of possessor (Section 3.1.2), subject (Section 3.5.1), and direct/indirect objects (Section 3.5.2) does not appear before a vowel-initial stem, but does appear before consonant-initial stems. For exam-

16 Máku

775

ple: te-ia (1-arm) > tia ‘my arm’, te-alaka (1.sbj-sit) > talaka ‘I sit’, te-Ø-iun-diba (1.o-3sg.sbj-rem.fut) > tiundiba ‘(S)he will hit me’, te-sibu (1.sbj-speak) ‘I speak’.

2.3.2 Haplology The syllable ke is omitted from the prefixes sEke- 3sg (subject marker) and teke- 1+3 (possessor and subject marker) when these precede a stem beginning in /kV/ or /VkV/. For example, Ø-seke-ke (3.o-3sg.sbj-eat) > seke ‘(S)he eats’, teke-ikidialu (1+3canoe) > te-ikidalu > tikidalu ‘our (incl) canoe’, teke-kinuba-dia (1+3sbj-die-pfv) > tekinubadia ‘We (excl) died’ (but note teke-kai in Table 16.8).

2.3.3 Harmonic vowels Vowel harmony in Máku is restricted to three affixes, the prohibitive -ʔV, negative -ʔA, and the affix sEke 3sg.sbj, when used as a prefix rather than as a suffix. The prohibitive -ʔV (3.5.3) has a vowel that completely harmonizes to the features of the vowel that precede it (which, however, is usually deleted in normal speech, due the syncope rule illustrated below in 2.3.4), for example, ke-a-ʔV (2.sbjeat-proh) > kaʔa ‘Don’t eat (fruit)!’, ke-ki-ʔV (2.sbj-eat-proh) > kekiʔi ‘Don’t eat!’, ke-we-ʔV (2.sbj-sleep-proh) > keweʔe ‘Don’t sleep!’, ke-sipinu-ʔV (2.sbj-kill-proh) > kespinuʔu ‘Don’t kill!’. The negative -ʔA (Section 3.5.4) includes a vowel A that has the same quality as the preceding vowel if it is non-high; if the vowel in the stem is high, then A is represented by /a/, for example, Ø-we-ʔA-bala=ke (3sg.sbj-sleep-neg-neg=foc) > weʔebalake ‘(S)he does not sleep’, te-a-ʔA-bale=ke (1sg.sbj-eat-neg-neg=foc) > taʔabalake ‘I do not eat’, Ø-sipinu-ʔA (3sg.sbj-kill-neg) > sipinuʔa ‘(S)he does not kill’. When the affix sEke- 3sg.sbj is used with an /a/-initial stem, the final vowel is deleted, as discussed in the preceding section. The remaining vowel completely harmonizes with the following /a/ stem vowel. For example, sEke-alaka-kɨsia (3sg.sbj-sit-temp) > sakalakakɨsia ‘(S)he is sitting’, sEke-ne-sia (3sg.sbj-hear-temp) > sekenesia ‘(S)he is hearing’.

2.3.4 Syncope The following deletion processes occur within phonological words regardless of their morphological structure and are optional in that they do not take place in slow, careful speech. The first of two identical vowels in two contiguous CV syllables that do not constitute an independent phonological word is usually deleted in non-

776

Raoul Zamponi and Chris Rogers

elicited/careful speech, for example, ke-ki-ʔi (2.sbj-eat-proh) > kekʔi ‘Don’t eat’, Øsipinu-dia (3sg.sbj-kill-pfv) > spinudia ‘(S)he killed’, ku-seke-dia (see-3sg.sbj-pfv) > kuskedia ‘(S)he saw’, te-teke-niu-dia (1+3.o-1+3.sbj-cut-pfv) > tetkeniudia ‘We (excl) cut ourselves’. Note, however, that, when the two identical vowels are separated by the glottal stop and the second vowel is followed by a suffix which is not the permanent marker -na (Section 3.5.3) or the negativizer -bala (Section 3.5.4), the deleted vowel is the second one, for example, Ø-niʔi-sia (3sg.sbj-burn-temp) > niʔsia ‘It is burning’, Ø-pane-ʔe-sia (3sg.sbj-sharp-neg-temp) > paneʔsia ‘It is dull’, e-kuʦilinuʔu-sia (2.sbj-dirty-pl-temp) > ekuʦilinuʔsia ‘You (pl) are dirty’, Ø-laʔa-ʦa-muʦa (3sg.sbj-there.is.not-firsth-rem.pst) > laʔʦamuʦa ‘There was nothing’. When the suffix -na or -bana follows a sequence /CV1ʔV1/, the deleted vowel is always the first one of the sequence, according to the general rule indicated above. For example, eluke-nuʔu-na (2.sbj-tall-pl-perm) > elukenʔuna ‘You (pl) are tall’, Ø-pane-ʔe-na (3sg.sbj-sharp-neg-perm) > panʔena ‘It is dull’, seke-ke-ʔe-na-bala=ke (3sg.sbj-eatneg-perm-neg=foc) > sekʔenabalake ‘(S)he never eats’.

2.3.5 Apocope Word-final vowels tend to be omitted before either a consonant or another vowel – see examples (19), (22), (32), (47), (56), (70), (86), and (91).

2.3.6 /ʔ/-loss Loss of the glottal stop /ʔ/ is widespread in non-elicited/careful speech, for example, uʔsi ‘other’ > usi, juʔwi ‘jaguar’ > juwi, keʔle ‘sun’ > kele, baʔi ‘lake’ > bai, naʔme ‘water, rain’ > name.

3 Word classes and morphological structure Word classes in Máku include nouns, pronouns, quantifiers, verbs, adverbs, particles, conjunctions, interjections, ideophones,6 some postpositions (though affixal morphology indicates most of the case relations in the language), and, perhaps,

6 Only one ideophone can be recognized for sure in the corpus of Makú: kiɾi. It occurs in a text recorded by Migliazza from Sinfrônio without a translation in Portuguese, and we recognize it as an ideophone based on the fact that it is repeated consecutively eleven times, and it is being phonologically deviant for the presence of the alveolar flap [ɾ]. What this form means exactly remains unclear.

16 Máku

777

some postpositions and demonstrative determiners. There is no separate class of adjectives. In general, the morphology is agglutinating and includes prefixes, suffixes, and infixes. Máku words generally contain only a few morphemes, with the verbal complex permitting the greatest complexity and number of bound morphemes. The language is generally head-marking. In each of the following sections, the morphosyntactic attributes of the word classes with inflectional morphology (nouns, pronouns, quantifiers, and verbs) are exemplified and described.

3.1 Nouns Nouns are quite common in the Máku corpus and are identified through the presence of case marking, possession, number marking, and/or specific derivation strategies. These are discussed in this section.

3.1.1 Case Case is marked on nouns and pronouns through a set of suffixes. The list of case suffixes evident in the data is:7 comitative singular -siky in (1), (76), and (77); comitative plural -dai in (2); dative or allative -le in (74) and (99); terminative -ky in (98); ablative -jeni in (1) and (63); elative or prolative -wani in (3), and (99); locative or instrumental -sa in (4), and (67); inessive -wa in (26), (69), and (93); and temporal -de, referring to locations in time, in (5).8

7 The numbers listed in parentheses after these suffixes indicate where examples can be found later in the chapter. 8 The following abbreviations for the example sources will be used: bm = text “Bucha and Makunaima” (recorded in 1958 by Migliazza); esm = elicited sentences collected by Migliazza (tape-recorded in 1964); esr = elicited sentences collected by Rodrigues (tape-recorded in 1986); fa = Formulário dos vocabulários padrões para estudos comparativos preliminares nas línguas indígenas brasileiras (segunda edição): língua máku (version A) (tape-recorded in 1960 by Migliazza; Migliazza 1960a); fb = Formulário dos vocabulários padrões para estudos comparativos preliminares nas línguas indígenas brasileiras (segunda edição): língua máku (version B) (tape-recorded in 1964 by Migliazza; Migliazza 1964a); m1 = text “The manioc” (version tape-recorded in 1958 by Migliazza); m2 = text “The manioc” (version recorded in 1990 by Maciel); mfn = E. C. Migliazza’s field notes (1958, 1960); mw = text “My wedding” (tape-recorded in 1990 by Maciel); op = text “Other plans” (taperecorded in 1990 by Maciel); ot = text “The Opossum and the Turtle” (tape-recorded in 1960 by Migliazza); q = Questionário padrão para a pesquisa nas línguas indígenas brasileiras: língua máku (tape-recorded in 1964 by Migliazza; Migliazza 1958a); s = text “The stone” (tape-recorded in 1990 by Maciel); uarm = untranslated audio recording from Maria Magalhães (tape-recorded in 1960 by Migliazza); uars = untranslated audio recording from Sinfrônio Magalhães (tape-recorded in 1960 by Migliazza); w = text “The water” (recorded in 1990 by Maciel); wiws = text “When I was small” (tape-recorded in 1990 by Maciel). Note that Migliazza worked with Sinfrônio Magalhães in three

778

Raoul Zamponi and Chris Rogers

(1)

Buʦia inaiamuʦa makunaimasiky wapitejeni. Buʦia Ø-ina-nia-muʦa makunaima-siky wapite-jeni Bucha 3sg.sbj-come-nfirsth-rem.pst Makunaima-com.sg sky-abl ‘Bucha came with Makunaima from the sky.’ (w)

(2)

tene usiiʦedai. tene uʔsi-iʦe-dai 1.pro other-pl-com.pl ‘I with others’ ( fa)

(3)

buʦi seketusia nawani. 9 buʦi sEke-tu-sia naʔme-wani leaf 3sg.sbj -float-temp water-prol ‘The leaf is floating on the water.’ ( fb)

(4) uja ʦabukuna bɨkɨʦisa. uja Ø-Ø-ʦabukuna bɨkɨʦi-sa 3sg.pro 3.o-3sg.sbj-tie-perm rope-ins ‘(S)he ties it with a rope.’ ( fa) (5) tekene tikiteʔbalake desembrude. tekene teke-ite-ʔA-bala=ke desembru-de 1+3.pro 1+3.sbj-go-neg-neg=foc December-temp ‘We will not be travelling in December.’ (op)

3.1.2 Possession Possessive nominal prefixes (shown in Table 16.3) mark person and number of the possessor, for all persons, with an additional inalienable/alienable distinction for the third person singular. Two numbers (singular and plural) are distinguished together with a clusivity contrast in the first-person plural: inclusive (1+2) and exclusive (1+3). An additional suffix -nuʔu is used to express plural number in conjunction with the prefixes e- 2 and ʦe- 1+2. Note that when one of the possessive prefixes, except the monoconsonantal n-, is bound to a vowel-initial stem, the vowel in the marker is dropped, as indicated in Section 2.3.1.

occasions in Boa Esperança farm: in August 1958, in 1960, and in the months of March and April 1964, supported by the Museu Paraense Emilio Goeldi. All his audio recordings, now in the Archive of the Indigenous Languages of Latin America (AILLA), were made in these three years, not only in 1960 as reported by AILLA’s stated metadata. 9 The elative/prolative case suffix is accompanied by a contraction in the noun root naʔme ‘water’ (becoming na). It is unclear whether this is a consequence of the suffix or not.

16 Máku

779

Tab. 16.3: Possessive prefixes. singular 1 2 3

plural te-

inalienable alienable

eØ-/ne-

1+2 1+3

ʦetekeeʦe-

Tab. 16.4: Paradigm for basuku ‘foot’ (inalien.) and ʦimala uba ‘bow’ (alien.), citation forms. sg

pl

1 2 3 1+2 1+3 2 3

te-basuku e-basuku Ø-basuku ʦe-basuku(-nuʔu) teke-basuku e-basuku ʦe-basuku

te-ʦimala uba e-ʦimala uba e-ʦimala uba ʦe-ʦimala uba(-nuʔu) teke-ʦimala uba e-ʦimala uba ʦe-ʦimala uba

As in many other Amazonian languages, nouns in Máku are lexically stipulated as being either inalienably or alienably possessed, though the only evidence of inalienable/alienable distinction is with the third person singular possession: zero prefix for inalienably possessed nouns; prefix e- for alienably possessed nouns. Inalienably possessed nouns are terms that denote relatives and parts of wholes: body parts, parts of plants, and space portions (relational nouns). Kinship terms and most body part terms would seem to be obligatorily possessed; terms denoting those body parts which are common to find separately from the body which they make up (e.g., kute ‘hair, feather’ and ekaʦu ‘horn’), plant parts, and space portions would seem to be optionally possessed. Table 16.4 illustrates examples of the full paradigm for each type of noun.10 Note that the inalienable/alienable distinction is maintained also when the possessor is expressed as a full noun phrase. (6)

malebe netu malebe Ø-netu capuchin 3sg-tail ‘capuchin’s tail’ ( fa)

10 The forms given in Table 16.4 are citation forms that maintain segments that may be lost in casual speech forms. The common speech form of tebasuku, for example, is tebasku, while that of ʦebasukunuʔu is ʦebaskunʔu, with two vowels lacking (see the discussion of syncope in Section 2.3.4).

780 (7)

Raoul Zamponi and Chris Rogers

nelabana ebunʦi nelabana e-bunʦi woman 3sg-pot ‘the woman’s pot’ (q)

3.1.3 Number Nouns are also marked for number (singular vs. plural). Plural nouns are marked with the suffix -iʦe, and singular nouns are unmarked.11 Number marking on nouns is not associated with animacy, as indicated by the following examples: masaiʦe ‘boys’, baskuiʦe ‘his/her feet’. Nouns are (normally) not marked with the plural suffix when numerals or other quantifiers are present; see (12)–(14), (18), and (19). Lastly, the two comitative case markers and verbal cross-referencing make a singular/plural distinction (-siky vs. -dai).

3.1.4 Noun derivation Some agent nouns (8), patient nouns (9), and abstract nouns (10) obtained from verbs by means of the suffix -na (homophonous with the permanent aspect suffix; Section 3.5.3) were recorded. (8)

sibuʔana sibu-ʔA-na speak-neg-nmlz ‘one who does not speak’ (mfn)

(9)

sibuna sibu-na speak-nmlz ‘language, word’ or ‘one who speaks’ (mfn, Koch-Grünberg 1928: 321)

(10) anjuna anju-na be.cold-nmlz ‘ice’ (Lit. ‘that which is cold’) (q)

11 The suffix -iʦe is also found in the plural personal pronouns eniʦe ‘you’ (only recorded by Faria 1927), from ene-iʦe (2.pro-pl), and waiʦe ‘they’, likely the result of a diachronic phonological erosion of the expected sequence uja-iʦe (3sg.pro-pl).

16 Máku

781

Another deverbal nominal suffix, -niu, also appears in one agent noun (11), though we do not understand the difference between -niu and -na. (11) sipinuniu sipinu-niu kill-nmlz ‘hunter’ (Maciel 1991: 46) Only two suffixes that derive nouns from other nouns are clearly attested in Máku corpus:12 -taka diminutive (e.g., mas(a)taka ‘child’ (q) (cf. masa ‘boy’)) and -ʔna bereavement, with the basic meaning ‘one to which his/her [relative] died’ (e.g., nabalekeʔna ‘widow’ (Koch-Grünberg 1928: 320) (cf. nabaleke ‘husband’)).13

3.2 Personal pronouns Personal pronouns have the same person and number distinctions observed above with the possessive morphology, with an additional distinction in the plural forms (1+2+3). They are presented in Table 16.5. The 1+2 pronoun is minimal inclusive referring to the speaker and one or more addressees. The 1+3 pronoun excludes the addressee(s) and is exclusive. Lastly, the 1+2+3 pronoun refers to the speaker, the addressee(s) and everyone else and can be termed “augmented inclusive” (see Cysouw 2003: 217).

Tab. 16.5: Personal pronouns (citation forms). singular

plural

1

tene

1+2 1+3 1+2 + 3

2 3

ene uja

ʦene(nuʔu) tekene tenenuʔu enenuʔu waiʦe

12 The meaning and existence of a third denominal nominal suffix is uncertain. This alleged suffix appears only in the term for the Máku language, maku-ema (mfn). 13 This suffix is a hapax and we cannot, therefore, exclude that it is a mere sequence of the negative suffix -ʔA and the permanent aspect suffix -na (Section 3.5.3) and even that nabalekeʔna is a stative verb (‘She is a widow’) rather than a noun.

782

Raoul Zamponi and Chris Rogers

3.3 Demonstratives There are two demonstrative pronouns in the available Máku material that are used to refer to entities on the discourse stage in relation to the speaker (proximal and distal). When used with three case suffixes – locative -sa, lative -le, and inessive -wa – these two demonstrative pronouns form the base of corresponding demonstrative adverbs. Contrasts in the demonstrative system are represented in Table 16.6.

Tab. 16.6: Demonstratives.

proximal distal

demonstrative determiners

demonstrative pronouns

demonstrative adverbs locative

allative

inessive

ki kua

ki akua

kisa kuasa isa usa

‒ kuale ‒ ule

kisawa kuasawa isawa ‒

There are possibly some other generalizations that can be made about these demonstratives, but the data are either contradictory or missing. For example, the sets of demonstrative determiners and demonstrative pronouns differ exclusively for their distal form: kua always occurs adnominally in the corpus (e.g., kua meʦi ‘that flower’ ( fa)), while akua is the form we always find in a headless context (e.g., tene akuasiky ‘I with that one’ ( fb)). Nevertheless, we also have two adnominal occurrences of akua (e.g., akua malia ‘that knife’ (q) and akua wike ‘that mountain’ ( fb)), suggesting that this form and kua may be variants of a single element that, just like the proximal demonstrative ki, might be used both adnominally and pronominally. The uses of the demonstrative adverbs of the last two lines of Table 16.6 and the semantic parameter that distinguishes the forms beginning in i from those beginning in u are unclear.

3.4 Quantifiers Máku quantifiers are independent words including esuʔu ‘many’, peʔtaka ‘all’, and numerals. The numerals and peʔtaka come immediately after the head they modify, as in (12) and (13). The quantifier esuʔu comes immediately before the head (14). (12) masina batasaba masina baʔtasaba earthworm four ‘four earthworms’ ( fb)

16 Máku

783

Tab. 16.7: Numerals 1 to 5. 1 2 3 4 5 a

nukujabaʔtasynialabaʔtasaba-a ʦisku nukujante See baʔta- ‘two’.

(13) jauku peʔtaka jauku peʔtaka star all ‘all the stars’ (q) (14) esʔu basku mete esuʔu Ø-basuku Ø-mete many 3sg-foot 3sg-finger ‘his/her many toes’ (q) The numerals from one to four, as well as the quantifier peʔtaka, also modify personal pronouns agreeing in person (and optionally in number, in the case of peʔtaka) with their head. (15) tekene tekesyniala telukena. tekene teke-syniala te-luke-na 1+3.pro 1+3-three 1.sbj-tall-perm ‘We (excl) three are tall.’ ( fa)14 (16) enenu epetakanu tiunukɨse! ene-nuʔu e-peʔtaka-nuʔu te-un-nuʔu-kɨse 2.pro-pl 2-all-pl 1.o-hit-pl-imp ‘You all hit me!’ ( fa) The numerals one to five are listed in Table 16.7. Distinct stems exist for the numbers one through four. The numeral five is a phrase meaning ‘one our (incl) hand’, as shown in (17). (17) ʦisku nukujante ʦe-isuku nukuja-nete 1+2-hand one-clf:body.part ‘our (incl) hand; five’ ( fa) 14 The verb should be marked by the prefix teke- ‘1+3’.

784

Raoul Zamponi and Chris Rogers

Numerals higher than five were recorded by Faria (1927) and Koch-Grünberg (1928: 324), and Migliazza (field notes of 1960). Those from six to nine seem to consist of descriptive names for various fingers, like, for example, ʦisku butena ‘our (incl) thumb; six’ (Lit. ‘our big finger’). ‘Ten’ is ʦisku baʔtanete ‘our (incl) two hands’ (Faria, Migliazza).

3.4.1 Classifiers The data exhibits evidence of two numeral classifiers for Máku: -nete, referring to body parts, as in (17) and (18), and -sy, referring to periods of time (19).15 The classifiers are only used in conjunction with the numerals one through four. (18) ʦebasuku baʔtante sukute ʦe-basuku baʔta-nete Ø-sukute 1+2-foot two-clf:body.part 3sg.sbj-full ‘our (incl) two full feet’ (Koch-Grünberg 1928: 324)16 (19) batasabas kelia baʔtasaba-sy keʔl-ia four-clf:time sun-moon ‘four days’ (m)17

3.5 Verbs Verbs in Máku, like in many other Amazonian languages, can be morphologically, syntactically, and semantically divided into two sub-classes: active and stative. Active verbs are transitive or intransitive verbs that generally denote events, actions, and processes. Stative verbs are intransitive verbs that generally denote states and cover many of the cross-linguistic semantic classes of adjectives suggested in Dixon (1982: 16, 34–49, 2006b: 3–5). Active verbs can be negated using either the suffixes -ʔa and -bala; stative verbs are negated using only -ʔa. In addition, most stative verbs are obligatorily marked for either the temporary or permanent aspects (we

15 -nete always occurs syncopated to -nte in the audio recordings. Its full form is suggested by the expressions of the numbers six (, i.e., ʦ-iku baʔta-nete ‘our (incl) two hands’, actually, ‘ten’) and nine (, again ‘our (incl) two hands’) recorded by KochGrünberg. 16 Recorded as the “numeral” twenty. 17 The numeral precedes the head noun in this example, presumably due to an interference of Portuguese.

785

16 Máku

discuss exceptions below), whereas active verbs are not. Both active and stative verbs are inflected for the person and number (gender is not expressed) of the subject argument, tense, aspect, mood, and evidentiality; transitive active verbs are inflected for object agreement as well.

3.5.1 Subject agreement There are different inflectional classes of both active and stative verbs based on the way they are inflected for subject agreement. The active verb inflectional subclasses are: P-verbs (which mark subject through agreement prefixes), I-verbs (which mark subject through agreement infixes), and S-verbs (which mark subject agreement through suffixes). Table 16.8 gives examples of conjugations of each of these subclasses of verbs, as far as the information has been recorded. The stative verb inflectional subclasses are: P-verbs and I-verbs. Table 16.9 gives examples of paradigms for each. The examples given in these two tables show the verbal stems (roots) without TAME markers (see Section 3.5.3 for a discussion of TAME). Lastly, there are separate second person imperative agreement affixes on verbs, as in (20) and (21). With most verbs, however, there is omission of second person marking in (positive) imperatives, as in (22). The imperative construction is mentioned in Section 6.4.

Tab. 16.8: Paradigms of selected active verbs (citation forms). P-verbs

I-verbs

S-verbs

-kai ‘stand’

-ain ‘fish’

kuʦi ‘wash’

dai_(i)na ‘bring’

bukulu‘look for, hunt’

ku‘see’

1sg

te-kai

tain (te-ain)

kuʦi

bukulu-te

ku-te

1+2

ʦain (ʦe-ain) kuʦia

ku-ʦe-nuʔu

1+3

ʦe-kainuʔu teke-kai

daitina (daiina) –

1+2+3

te-kai-nuʔu

tekain (teke-ain) –

2sg

ke-kai

kain (ke-ain)

2pl 3sg

ke-kai-nuʔu kain (ke-ain) Ø-kai ain (Ø-ain)

3pl

ʦe-kai-pu

–

kuʦi

–

bukulu-ʦenuʔu bukulu-teke

kutsinuʔu –

–

–

–

daikina (daiina) – daina (daina) –

–

ku-seke

– bukulu-seke

– ku-seke

– kuʦi kuʦia

ku-teke

bukulu-ʦe-pu kula-Ø

a In the audio recordings, Sinfrônio pronounces the affix ʦe with an /i/ when it is inserted within the root kuʦi ‘wash’. It is unclear whether the vowel harmonization is mandatory in this context.

786

Raoul Zamponi and Chris Rogers

Tab. 16.9: Paradigms of selected stative verbs (citation forms). P-verbs

I-verbs

-leme ‘be red’

-kuduma ‘be good’

inene be afraid’

1sg 1+2 1+3

te-leme ʦe-leme(-nuʔu) teke-leme

inene inene-nuʔu inene

1+2+3 2sg 2pl 3sg 3pl

– e-leme e-leme-nuʔu Ø-leme ʦe-leme-pu

te-kuduma ʦe-kuduma-nuʔu te-kuduma (teke-kuduma) te-kuduma-nuʔu e-kuduma e-kuduma-nuʔu Ø-kuduma –

inene-nuʔu inene inene-nuʔu inene inene

(20) ene kunʦikɨse! ene kuʦi-kɨse 2.pro wash-imp ‘You (sg) wash!’ ( fa) (21) enenu lilanukɨse! ene-nuʔu lila-nuʔu-kɨse 2.pro-pl pull-pl-imp ‘You (pl) pull!’ (fb) As noted in the two tables above, there are two suffixes that can (optionally) be used to indicate a plural subject: -nuʔu and -pu, the latter of which is reserved for thirdperson subjects. Both suffixes occur on both stative and active verbs. The suffix -nuʔu can also be used to mark non-third-person plural object agreement, as in (23).

3.5.2 Object agreement Transitive active verbs also agree in person and number with the direct objects and indirect objects (or “extension-to-core” arguments, following Dixon 2006a: 7) by means of the prefixes reported in Table 16.10. Object agreement prefixes occur before the subject agreement prefixes in Pverbs, as the following examples indicate. (22) ten etkunuba. tene e-te-aku-nuʔu-ba 1.pro 2.o-1.sbj-hit-pl-nr.fut ‘I will hit you (pl).’ (fb)

16 Máku

787

Tab. 16.10: Object (O/E) prefixes. singular

non-singular

1

te-

1+2 1+3 1+2+3

2 3

eØ-

ʦeteketeeØ-

(23) malebe mesyky ʦeskeneiamuʦa. malebe mesyky ʦe-sEke-ne-nia-muʦa capuchin manioc 1+2.e-3sg.sbj-show-nfirsth-rem.pst ‘The capuchin showed us (incl) the manioc.’ (m1 )

3.5.3 Tense, Aspect, Mood, and Evidentiality Máku verbs can be morphologically marked for tense, aspect, mood, and evidentiality (TAME). The Máku tense system is symmetrically centered on the present tense (unmarked), with two future tenses, -ba near future and -diba remote future, and two past tenses, -muʦa remote (or narrative) past and -nka recent (or general) past. Tense markers (especially the past tense markers) are optional when the meanings can be recovered based on discourse context and deictic temporal reference, cf. the absence of tense markers in (26), (94), the first two clauses of (95), and (97). Two evidential distinctions are marked alongside both past tenses, and never with any other temporal reference: -ʦa firsthand experience (firsth) and -nia non-firsthand experience (nfirsth) (usually -ia in spontaneous speech). Specification of evidentiality is quite common for past tense marked verbs, as seen in examples, as in (24)–(26), but does not appear to be required, as indicated by (73). (24) me lamamu tesketunʦanka name tesibana. me lamamu te-sEke-tune-ʦa-nka naʔme my.father early.in.the.morning 1.o-3sg.sbj-wake.up-firsth-rec.pst water te-si-bana 1.sbj-bathe.with-purp ‘My father woke me up in the morning to bathe.’ (wiws) (25) line line kalamana laiamuʦa Buʦia. line line Ø-kalama-na Ø-Ø-la-nia-muʦa Buʦia stone stone 3sg.sbj-white-perm 3.o-3sg.sbj-make-nfirsth-rem.pst Bucha ‘Bucha made the stone, the white stone.’ (s)

788

Raoul Zamponi and Chris Rogers

(26) jaule lemeky peneia kulalwa. jaule lemeky Ø-Ø-pene-nia kulalu-wa opossum land.turtle 3.o-3sg.sbj-put-nfirsth cage-ine ‘The opossum put the turtle into a cage.’ (ot) There are six aspectual suffixes: temporary18 -si(a), permanent (non-temporary) -na, perfective -di(a) (27), imperfective -debena (28), temporary progressive -dikisi(a), and permanent progressive (non-temporary progressive) -dikina.19 (27) ʦimala sipintedia. ʦimala Ø-Ø-sipinte-dia arrow 3.o-3sg.sbj-shoot-pfv ‘(S)he shot an arrow.’ ( fa) (28) uja kyʔte milanedbena. uja kyʔte Ø-Ø-milane-debena 3sg.pro egg 3.o-3sg.sbj-count-ipfv ‘(S)he is counting the eggs.’ ( fa) As previously mentioned, the temporary and permanent suffixes are obligatorily used with most of the verbs of the stative class when these occur both predicatively, in the present tense, and attributively. When a verb belonging to the stative verb class is marked with -si(a), it indicates a temporary or contingent property, as in (30) and (39). When a stative verb is marked with -na, it indicates that the state denoted by the verb is a permanent or inherent property of an individual or object, as in (15), (25), (32), (58), and (71). (29) masa sekenesia. masa sEke-ne-sia boy 3sg.sbj-listen-temp ‘The boy is listening.’ ( fa) (30) bunʦi ʦilisia. bunʦi Ø-ʦili-sia pot 3sg.sbj-dirty-temp ‘The pot is dirty.’ ( fa)

18 The allomorphs -kɨsia and -kɨsi are used only with the verbs -pe ‘stay, live’, -ja ‘lie down’, -laka ‘sit’, -kai ‘stand’, -ɸani ‘know’, and -ɸanimita ‘think’ in the available material. 19 Based on the phonological similarity and the semantic congruity, the latter two suffixes are likely from -diki + -sia and -diki + -na, respectively. However, since there is no evidence for -diki as a separate morpheme in Máku, we include them here as unanalyzed morphemes.

16 Máku

789

(31) ujake weʔenabalake. uja=ke Ø-we-ʔA-na-bala=ke 3sg.pro=foc 3sg.sbj-sleep-neg-perm-neg=foc ‘(S)he never sleeps.’ ( fa) (32) kel ekumasana. keʔle Ø-ekumasa-na sun 3sg.sbj-round-perm ‘The sun is round.’ ( fa) The temporary progressive and permanent progressive suffixes indicate that an event is either occurring during speaking time and is relatively temporary for a time (for -dikisi(a)), or that it is occurring during speaking time and is relatively permanent for a time (for -dikina). (33) uja wedikisia. uja Ø-we-dikisia 3sg.pro 3sg.sbj-sleep-temp.prog ‘(S)he is sleeping.’ ( fa) (34) tekene bukulutekedikina. tekene bukulu-teke-dikina 1+3.pro hunt-1+3.sbj-perm.prog ‘We (excl) are hunting.’ ( fb) Finally, verbs can optionally be marked for five moods by the following suffixes: imperative -kɨse (sometimes simply -kɨ or -se), as in (16) and (20) through (22); hortative -kada (35); prohibitive -ʔV (36); conditional -wake, as in (90) and (91); and purposive -bana, as in (24), (81), (83), (87), and (88). There is no formal marking of the indicative mood, and it is considered the default when there is no overt marking of other moods. (35) ʦalynyhykada! ʦe-alynyhy-kada 1+2.sbj-start-hort ‘Let’s start!’ (mfn) (36) keweʔe! ke-we-ʔV 2.sbj-sleep-proh ‘Don’t sleep!’ ( fa)

790

Raoul Zamponi and Chris Rogers

Interestingly, imperative clauses can also command or prohibit states expressed by stative verbs in Máku, as illustrated by the following two examples. (37) kudumakɨse! kuduma-kɨse good-imp ‘Be good!’ (Migliazza 1966: 25) (38) kekalamaʔa! ke-kalama-ʔV 2.sbj-white-proh ‘Don’t be white!’ (Migliazza 1966: 25)

3.5.4 Negation Stative verbs are negated by means of the suffix -ʔA, as in (39), while active verbs are negated by means of -ʔA plus, optionally, the suffix -bala, as in (40) and (41). See the discussion of vowel harmony in Section 2.3.3 regarding the phonetic realization of the vowel in -ʔA. (39) lipina paneʔsia. lipina Ø-pane-ʔA-sia knife 3sg.sbj-sharp-neg-temp ‘The knife is dull.’ (Lit. ‘not sharp’) ( fa) (40) ujak as weʔbalake. uja=ke asa Ø-we-ʔA-bala=ke 3sg.pro=foc today 3sg.sbj-sleep-neg-neg=foc ‘(S)he does not sleep today.’ ( fb) (41) uja jukude sipinuʔa. uja jukude Ø-Ø-sipinu-ʔA 3sg.pro person 3.o-3sg.sbj-kill-neg ‘(S)he doesn’t kill people.’ ( fa) Adverbs may also be negated using the -ʔA suffix. (42) asaʔa asa-ʔA today-neg ‘not today’ (Migliazza 1966: 24)

16 Máku

791

(43) lamaʔa lama-ʔA tomorrow-neg ‘not tomorrow’ (Migliazza 1966: 24)

3.5.5 Verb derivation The diminutive suffix -taka (Section 3.1.4) also occurs with stative verbs without modifying their word class. The suffix occurs together with the negative suffix -ʔA when added to a verb stem (cf., however, -laʔtaka ( fa) from the verb stem -laʔa ‘few’ which is etymologically from -la-ʔA (make-neg)). (44) kahiʔatakana. Ø-kahi-ʔA-taka-na 3sg.sbj-long-neg-dim-temp ‘It is short.’ ( fa) Other class-maintaining derivational suffixes used with verbs are the following. The intensifier suffix -ʦa is found in one stative verb in our corpus of resources, exemplified in (47). (45) naʔme wateʦasia. naʔme Ø-wate-ʦa-sia water 3sg.sbj-have.a.taste-ints-temp ‘The water is very salty.’ (Lit. ‘has a strong taste’) (esr) The distributive suffix -te, used with both active and stative verbs to indicate that the verbal referent involves a group of people referred to by either the subject or object argument of the verb, is illustrated in (46) and (47). (46) uja niutenudiba epetakanʔu. uja n-Ø-iu-te-uʔnu-diba e-peʔtaka-nuʔu 3sg.pro 2.o-3sg.sbj-hit-distr-pl-rem.fut 2-all-pl ‘(S)he will hit you, you all.’ ( fa) (47) en epetaka elemetenusia. ene e-peʔtaka e-leme-te-nuʔu-sia 2.pro 2-all 2.sbj-red-distr-pl-temp ‘You all are red.’ (fb) The desiderative suffix -ʦimu is shown in (48), and the negative desiderative -tia in (49).

792

Raoul Zamponi and Chris Rogers

(48) palu daikinaʦimusia? palu Ø-daiina-ʦimu-sia banana 3.o-bring-desid-temp ‘Do you want to bring bananas?’ (esr) (49) tene kaɸe temitiasia. tene kaɸe Ø-te-mi-tia-sia 1.pro coffee 3.o-1.sbj-drink-neg.desid-temp ‘I do not want to drink coffee.’ (esr) The undirected motion suffix -liku is exemplified in (50). The suffix -dakana might be a further (attested) directional or motional suffix of Máku. Its semantics, however, is obscure.20 (50) uja patelikusia. uja Ø-pate-liku-sia 3sg.pro 3sg.sbj-walk-undir-temp ‘(S)he is walking around.’ ( fb) There is also a valency-decreasing affix ku in Máku. However, the semantic and morphological characteristics of this affix are largely unclear. It appears to be used with active verbs, most commonly with a reciprocal meaning, in some cases as a prefix preceding the subject agreement prefix (51), in other cases as a suffix directly attached to the root (52), and in still other cases as both a prefix and a suffix, shown in (53). (51) waiʦe kuʦisibiudia. waiʦe ku-ʦe-isibiu-dia| 3pl.pro recp-3pl.sbj-hit-pfv ‘They hit each other.’ ( fb) (52) waiʦe ʦeʦeniukudia. waiʦe CV~ʦe-niu-ku-dia 3pl.pro red~3pl.sbj-cut-recp-pfv ‘They cut each other.’ (fb) (53) waiʦe kuʦisibiukudia. waiʦe ku-ʦe-isibiu-ku-dia 3pl.pro recp-3pl.sbj-hit-recp-pfv ‘They hit each other.’ ( fa) 20 See ikilusa patedakanadia (ikilu-ʔsa Ø-pate-dakana-dia path-loc 3sg.sbj-walk-?-pfv) ‘(S)he walked on the path’ (fa).

16 Máku

793

But it can be also observed in a few sentences with a singular subject without marking the host verb (-sipinu ‘kill’ in all cases) as intransitive, as in (54) and (80).21 (54) ujake dɨʔɨ spinukudia. uja=ke dɨʔɨ Ø-Ø-sipinu-ku-dia 3sg.pro=foc tapir 3.o-sg.sbj-kill-recp(?)-pfv ‘(S)he killed tapirs.’ ( fa)

4 The noun phrase The head of a Máku noun phrase may be modified by other nominal elements or stative verbs. Possessors, demonstratives, and the quantifier esuʔu ‘many’ occur in pre-head position, as in (14), (55)‒(57), and (68), while the remaining quantifiers (including numerals), stative verbs, and relative clauses occur in the post-head position, as in (12), (13), (15), (16), (25), (58), and (59). Note that stative verbs do take cross-referencing prefixes when used attributively, as in (59). Nouns can be occasionally used as nominal modifiers though there is no way to be certain whether these examples show nominal modifiers or parts of a nominal compound.22 Whatever their function, nominal modifiers also follow the head, as in (60). (55) malebe isa malebe Ø-isa capuchin 3sg-liver ‘capuchin’s liver’ (q) (56) en ewaʦinʔu ene e-waʦi-nuʔu 2.pro 2-mouth-pl ‘your (pl) mouth’ ( fa) (57) ki meʦi ki meʦi this flower ‘this flower’ ( fa)

21 Perhaps as in Tariana (Aikhenvald 2003: 264) and Hup (Epps 2011: 319), a transitive verb used with a marker of reciprocity might just imply multiple participants in Máku. The root -sipinu ‘kill’ might, therefore, be used in (54) and (80) together with ku in the meaning ‘kill in a hunting expedition’ or, literally, ‘kill with a lots of people’. 22 Both elements of this construction receive stress.

794

Raoul Zamponi and Chris Rogers

(58) pi kuditakana Ø-pi Ø-kuditaka-na 3sg-nose 3sg.sbj-small-perm ‘small nose’ (q) (59) juʔwi esʔu kudumana iʦe juʔwi esuʔu kuduma-na iʦe dog very good-perm pl ‘very good dogs’ (mifn) (60) bunʦi buʔte bunʦi buʔte pot earth ‘clay pot’ ( fa)

5 The verb phrase Motion verbs can be serialized in Máku. They are postposed to a main verb (from an open class) that determines the transitivity of the whole serial verb construction. (61) masa aliwi spinu patedia. masa aliwi Ø-Ø-sipinu Ø-pate-dia boy alligator 3.o-3sg.sbj-kill 3sg. sbj-go-pfv ‘The boy went to kill the alligator.’ ( fa) Both verbs in a serial construction must agree in person and number with the subject. (62) ʦenenu ʦespinu ʦatenudia. ʦene-nuʔu ʦe-sipinu ʦe-ate-nuʔu-dia 1+2.pro-pl 1+2.sbj-kill 1+2.sbj-go-pl-pfv ‘We (incl) went to hunt.’ (Lit. ‘kill’) (fa) Plurality and TAME, on the other hand, are usually marked only on the second verb. Example (63) shows a serial verb construction in which tense and evidentiality are marked on every verb. (63) maluakajeni mesyky sapaniamuʦa inaniamuʦa. maluaka-jeni mesyky Ø-Ø-sapa-nia-muʦa samaúma-abl manioc 3.o-3sg.sbj-pull.out-nfirsth-rem.pst Ø-ina-nia-muʦa 3sg.sbj-come-nfirsth-rem.pst ‘(S)he (the paca) came to pull out the manioc from the samaúma tree.’ (m2)

16 Máku

795

6 Clauses Independent clauses may be declarative, imperative, interrogative, and exclamatory. Declarative clauses, in turn, can be classified according to their nuclear constituents as verbless, intransitive, transitive, and ditransitive.

6.1 Verbless clauses Verbless clauses are used to indicate an equative, locative, or existential meaning, as in (64) through (69). They have two core constituents: subject and complement, and most often (perhaps always in clauses denoting equation) the subject precedes the complement. Note that some examples of verbless clauses are included with an apparent subset of the available tense, aspect, and evidentiality morphology available for regular verbs, as in (64) and (69). (64) asaiʦeke dɨʔɨ juwi liana jabi menehe jukudeʦamuʦa. asaiʦe=ke dɨʔɨ juʔwi liana jabi jukude-ʦa-muʦa long.ago=foc tapir jaguar curassow.sp. agouti person-firsth-rem.pst ‘Long ago, the tapir, the jaguar, the curassow, and the agouti were people.’ (uars) (65) asake emine patele. asa=ke e-mine patele today=foc 3sg-house glass ‘Today his house is glass (i.e., is invisible, transparent like glass).’24 (bm) (66) enʦeke Maɾia. Ø-enʦe=ke Maɾia 3sg-name=foc Maria ‘Her name is Maria.’ (mfn) (67) sapnawi wapitesa. sapanawi wapite-sa cloud sky-loc ‘The cloud is in the sky.’ ( fa)

23 It is unclear what motivates the use of the firsthand evidential in this clause. Its occurrence might be due to language obsolescence as well as signal general knowledge. The fact that the nonfirsthand evidential is also used in the same text from which this example is from does not come down in favor of the latter interpretation, however. 24 That is, Makunaima’s, a supernatural being.

796

Raoul Zamponi and Chris Rogers

(68) asake esu line. asa=ke esuʔu line today=foc many stone ‘Today there is much stone.’ (s) (69) bunʦiwa ekɨnɨsia. bunʦi-wa ekɨnɨ-sia pot-ine grease-temp ‘There is grease in the pot.’ ( fb)

6.2 Intransitive clauses Intransitive clauses, both with active and stative verbs, most often occur with the subject preceding the predicate, as indicated by (70) and (71). (70) ijan teneke teweba. ijani tene=ke te-we-ba then 1.pro=foc 1.sbj-sleep-nr.fut ‘Then I will sleep.’ (esm) (71) taba butena. taba Ø-bute-na stick 3sg.sbj-big-perm ‘The stick is big.’ (fa)

6.3 Transitive and ditransitive clauses and ordering of obliques Máku does not distinguish transitive and ditransitive clauses morphosyntactically. The basic order of transitive clauses is AOV. (72) Buʦia wike laiamuʦa. Buʦia wike Ø-Ø-la-nia-muʦa Bucha mountain 3.o-3sg.sbj-make-nfirsth-rem.pst ‘Bucha25 made the mountains.’ (s)26 In ditransitive (extended transitive) clauses, the pragmatically neutral position for the indirect object constituent seems to be clause-finally, after the verb.

25 Bucha is Makunaima’s younger brother (see footnote 24). 26 The translation of wike is plural in the sense that he created all mountainous terrain, i.e., the mountains.

16 Máku

797

(73) malebe naba jaʦi sekenemuʦa menehe nabale. malebe Ø-naba jaʦi Ø-seke-ne-muʦa menehe capuchin 3sg-daughter manioc.bread 3.e-3sg.sbj-show-rem.pst paca Ø-naba-le 3sg-daughter-dat ‘The capuchin’s daughter showed the manioc bread to the paca’s daughter.’ (m2) Non-core (i.e., oblique) constituents show flexibility in ordering within the clause. (74) ujake unesiky patenuna. uja=ke Ø-une-siky Ø-pate-nu-na 3sg.pro=foc 3sg-mother-com.sg 3sg.sbj-walk-?-perm ‘(S)he walks (?) with his/her mother.’ ( fa) (75) uja patenusia unesky. uja Ø-pate-nu-sia Ø-une-siky 3sg.pro 3sg.sbj-walk-?-temp 3sg-mother-com.sg ‘(S)he is walking (?) with his/her mother.’ ( fb)

6.4 Imperative, interrogative, and exclamatory clauses Imperative clauses (see Sections 3.5.1 and 3.5.3.) most commonly do not have overt subject arguments, though they do seem to be used for emphasis, as in (16), and (20) through (22). The available data contain mostly constituent (i.e., content) question clauses and only a few instances of polar question clauses.27 Constituent questions involve two things: an interrogative pronoun (e.g., tuʦi ‘who’, ʦini ‘what’, ʦile ‘where’, ʦibani ‘when’, or ʦiumu ‘how’) in clause-initial position and the enclitic particle =ta attached to the interrogative pronoun, as in (76), or to the last phonological word of the clause it is marking, as in (77). (76) ʦinita nisike? ʦini=ta Ø-niʔi-si=ke what=cnt.int 3sg.sbj-burn-temp=foc ‘What is burning?’ (q)

27 One striking feature of the audio-recorded interrogative clauses is that they do not appear to have a specific interrogative intonation. All of them were obtained by elicitation and, perhaps for this reason, they were articulated by Sinfrônio Magalhães without changes in pitch or intensity.

798

Raoul Zamponi and Chris Rogers

(77) ʦini kalanudibata? ʦini Ø-ke-ala-nuʔu-diba=ta what 3.o-2.sbj-do-pl-rem.fut=cnt.int ‘What will you (pl) do?’ (m2) There are only four examples of polar questions in the Máku corpus. They do not show any unique marking that would distinguish them from declarative statements. Examples (48) and (78) show two of the four recorded polar questions. (78) ene kesibu katediba ule? ene ke-sibu ke-ate-diba u-le 2.pro 2.sbj-speak 2.sbj-go-rem.fut there.ana(?)-all ‘Will you (sg) go to speak there?’ (esr) Exclamatory clauses may consist of one noun phrase (ʦe-mesu ‘our (incl) manioc stalks!’ (m1 )) or an interjection, like e e ‘yes’ (mfn), edi ‘that’s right, okay’ ( fa, q), and laʔa ‘no’ (esr, fa, fb, q).

7 Clause-linking, information structure, and discourse This section describes what is known about Máku structure above the level of the simple clause.

7.1 Coordination There is no overt coordinator in Máku; rather coordination is accomplished by phrase or clause juxtaposition. A coreferential noun phrase in two coordinated clauses is often omitted in the second clause (80). (79) aleme mesyky wesyky palu / esu dikiamuʦa. aleme mesyky wesyky palu esuʔu Ø-dikia-muʦa pineapple manioc sweet.potato banana a.lot 3sg.sbj-have(?)-rem.pst ‘Pineapples, manioc, sweet potatoes, (and) bananas. It (the samaúma tree) had a lot.’ (m1 ) (80) ujake dɨ spinukudia aliwi spinukudia. uja=ke dɨʔɨ Ø-Ø-sipinu-ku-dia aliwi 3sg.pro=foc tapir 3.o-3sg.sbj-kill-recp(?)-pfv alligator Ø-sipinu-ku-dia 3sg.sbj-kill-recp(?)-pfv ‘(S)he killed tapirs and ((s)he killed) alligators.’ (fa)

16 Máku

799

7.2 Subordination There are not many examples of subordinate clauses attested in the data, and for the most part those that are available do not have any explicit subordinator. There are examples of complement clauses, relative clauses, and adverbial clauses. Only verbless relative clauses, temporal adverbial clauses, and, likely, adverbial causal clauses are formed with a dedicated subordinator. There is no indication of switchreference marking in Máku. On the other hand, the use of the verbal purposive inflection (see examples (24), (87), and (88)) as a main clause intentional future tense marker (81) may be the outcome of an insubordination process (Evans 2007). (81) ujake sekebana lama. uja=ke sEke-ke-bana lama 3sg.pro=foc 3sg.sbj-eat-purp tomorrow ‘(S)he will eat tomorrow.’ (fa)

7.2.1 Complement clauses Complement clauses appear to function exclusively as an object argument in a transitive clause, and may be unmarked (82), or marked with either the purposive -bana (83) or the nominalization suffix -na (Section 3.1.4), as in (84). These different complementation strategies are conditioned by the verb of the main clause. (82) lamamu kulaniamuʦa tseniuna laʔabalake. lamamu Ø-kula-Ø-nia-muʦa ʦe-niu-na at.dawn 3.o-see-3.sbj-nfirsth-rem.pst 3pl.sbj-cut-nmlz Ø-laʔa-bala=ke 3sg.sbj-there.is.not-neg=foc ‘At dawn, they saw they had not cut (the tree).’ (Lit. ‘There was not their cut.’) (m2) (83) enemu sakamaneiamuʦa mekasa lymybana. Ø-enemu Ø-sEke-amane-nia-muʦa meʔkasa 3sg-brother 3.e-3sg.sbj-order-nfirsth-rem.pst fish Ø-Ø-lymy-bana 3.o-3sg.sbj-catch-purp ‘His brother ordered him to fish.’ (Lit. ‘catch fish’) (uarm) (84) ten bukulu tunisia patena. tene bukulu Ø-te-unisia Ø-pate-na 1.pro ? 3.o-1.sbj-want 3sg.sbj-go-nmlz ‘I want him/her to go.’ (esr)

800

Raoul Zamponi and Chris Rogers

7.2.2 Relative clauses Relative clauses are scantly and ambiguously documented in the data. Apparently, they can be formed in two ways. If the relative clause is a verbal clause, it is formally unmarked, as in (85). (85) tuʦita ki wedikisike? tuʦi=ta ki Ø-we-dikisi=ke who=cnt.int this 3sg.sbj-sleep-temp.prog=foc ‘Who is this who is sleeping?’ (q) If the relative clause is a verbless clause, it seems to be marked – after the final relative clause constituent – with an enclitic relativizer =ni, as in (86) (86) wanaka ʦen ʦeminesani wanaka ʦene ʦe-mine-sa=ni animal 1+2.pro 1+2-house-loc=rel ‘animal that is in our (incl) house’ ( fb)

7.2.3 Adverbial clauses The attested adverbial clauses indicate purpose, conditionality, time, and cause or reason. Purposive clauses have the predicate marked with the purposive mood and may precede (87), be embedded within (88), or, more commonly, follow the main clause (24). (87) ʦikibana laʔʦamuʦa. ʦi-ki-bana Ø-laʔa-ʦa-muʦa 1+2.sbj-eat-purp 3sg.sbj-there.is.not-firsth-rem.pst ‘There was nothing for us (incl) to eat.’ (m1 ) (88) eʦiwa jaʦi ʦikibana lalaiamuʦa. eʦiwa jaʦi ʦe-iki-bana so manioc.bread 3pl.sbj-eat-purp CV~Ø-Ø-la-nia-muʦa red~3.o-3sg.sbj-make-red-nfirsth-rem.pst ‘So (s)he made manioc bread in order for them to eat (it).’ (m1 ) Conditional clauses have the predicate marked for the conditional mood, and they also may precede or follow the main clause, as in (89) and (90).

16 Máku

801

(89) tene tatewake uja inadiba. tene te-ate-wake uja Ø-ina-diba 1.pro 1-go-cond 3sg.pro 3sg.sbj-come-rem.fut ‘If I go, (s)he will come.’ (q) (90) uja juwi spinu bywake. uja juʔwi Ø-Ø-sipinu Ø-Ø-by-wake 3sg.pro jaguar 3.o-3sg.sbj-kill 3.o-3sg.sbj-bite-cond ‘(S)he would kill the jaguar if it bit him/her.’ (fb) The temporal clauses that were recorded are marked by one of these free-form subordinators juʦe ‘when’ (91) or jajani ‘after’ (92). Additionally, there are examples of the predicate being marked with an enclitic sequential marker =jani (93). (91) mastaka juʦe mekas spinuʦanka. masa-taka juʦe meʔkasa Ø-Ø-sipinu-ʦa-nka boy-dim at.the.time fish 3.o-3sg.sbj-kill-firsth-rec.pst ‘When (s)he was a child, (s)he killed fish.’ (fb) (92) kumuka jajani mesu ʦeiamuʦa. Ø-kumuka jajani mesu Ø-ʦe-e-nia-muʦa 3sg.sbj-fall after manioc.stalk 3.o-1+2.sbj-gather-nfirsth-rem.pst ‘After it (the samaúma tree) fell, we (incl) gathered the manioc stalks.’ (m1 ) (93) naʔme liatejani keʔle panaidiba. naʔme Ø-liate=jani keʔle Ø-panai-diba rain 3sg.sbj-finish=seq sun 3sg.sbj-come.out-dist.fut ‘The rain having being over, the sun will come out.’ (Migliazza 1966: 25) The subordinator juʦe indicates that the situation described by the subordinate clause is temporally simultaneous with the action in the main clause. The sequential marker =jani indicates a temporal succession of events. Temporal subordinate clauses are not independently marked for tense, but have the same temporal reference as the main clause. Adverbial clauses of cause or reason are marked with -maka, an element whose morphemic status is unclear, and are postposed to the main clause (94). (94) eʦiwa lemeky simaia jaule kinubamaka. eʦiwa lemeky Ø-sima-nia jaule Ø-kinuba-maka so land.turtle 3sg.sbj-laugh-nfirsth opossum 3sg.sbj-die-reas ‘So the turtle laughed because the opossum died.’ (ot)

802

Raoul Zamponi and Chris Rogers

Lastly, similar to coordinated clauses, when a subordinate clause and a main clause share a co-referential noun phrase, this is (normally) omitted from one of the two clauses, usually whichever occurs second, as shown in (90).

7.3 Information structure and discourse In connected, spontaneous speech, verbal arguments (subject and object) are often elided if the information can be recovered from the previous discourse, as is common cross-linguistically. (95) menehe naisesiky weʔebalake / sekete / lama pateiamutsa. menehe naise-siky Ø-we-ʔA-bala=ke sEke-te paca monkey.sp-com.sg 3sg.sbj-sleep-neg-neg=foc 3sg.sbj-stay.awake lama Ø-pate-nia-mutsa day.after 3sg.sbj-go-nfirsth-rem.pst ‘The paca didn’t sleep with the monkey. She stayed awake. The day after, she went away.’ (w) Placing a constituent in clause-initial position seems to have the effect of bringing it into focus within the discourse. Unfortunately, without necessary grammatical and pragmatic controls, it is impossible to determine the motivation for placing a constituent in clause-initial position. We refer to this as a focus construction without being able to clarify whether it is similar to either focus constructions or to contrastive topic constructions cross-linguistically. (96) uja juwi bydia. uja juʔwi Ø-Ø-by-dia 3sg.pro dog 3.o-3sg.sbj-bite-pfv ‘The dog bit him/her.’ (fa) In addition to placing a constituent clause-initially, the enclitic =ke also has a focus function. It occurs on words of various word classes in the data, but it is particularly common on nouns and pronouns in subject function, as in (5), (31), (40), (54), (64)‒ (66), (74), (76), (80)–(82), (85), and (95). In continuous discourse, the particle eʦiwa ‘so, thus, because of this’, establishes a logical sequence between successive sentences. (97) lamamu lemeky kuske patena / jaule kinubaia / eʦiwa lemeky simaia jaule kinubamaka. lamamu lemeky Ø-ku-seke Ø-pate-na jaule in.the.morning land.turtle 3.o-see-3sg.sbj 3sg.sbj-go-perm opossum

16 Máku

803

Ø-kinuba-nia eʦiwa lemeky Ø-sima-nia jaule 3sg.sbj-die-nfirsth so turtle 3sg.sbj-laugh-nfirsth opossum Ø-kinuba-maka 3sg.sbj-die-reas ‘In the morning, the turtle went to see him (the opossum). The opossum died. So, the turtle laughed because the opossum was dead.’ (ot)28 (98) mene esku siʦiiamuʦa29 / eʦiwa malebe pateiamuʦa maluaka baʦiky. menehe esuku Ø-siʦi-nia-muʦa eʦiwa malebe paca ? 3sg.sbj-cry(?)-nfirsth-rem.pst so capuchin Ø-pate-nia-muʦa maluaka Ø-baʦi-ky 3sg.sbj-go-nfirsth-rem.pst samaúma 3sg-leg-term ‘The paca felt pity and cried (?). So the capuchin went to the foot of a samaúma tree.’ (m1 ) On the other hand, the particle ijani ‘then’ (cf. the sequential marker -jani) links successive sentences without expressing the nature of the relationship between the events described. (99) eminewani pateniamuʦa adiatale / ijani wapitele pateniamuʦa. e-mine-wani Ø-pate-nia-muʦa adiata-le ijani 3sg-house-elat 3sg.sbj-go-nfirsth-rem.pst place.below-all then wapite-le Ø-pate-nia-muʦa sky-all 3sg.sbj-go-nfirsth-rem.pst ‘From his house, he (Makunaima) went downwards. Then, he went up into the sky.’ (bm)

8 Conclusion This chapter aimed at providing a structural overview of Máku, an extinct isolate once spoken in vicinity of the Auari River in northern Brazil, as preserved in its documentary sources. Many of the aspects of the language that were described appear as areal features of northwestern Amazonia, being also attested in languages of different families of the region. These include: (i) possession, either inalienable or alienable, prefixally marked on the possessed noun (Section 3.1.2); (ii) numeral classifiers, although in a very limited number (just two; Section 3.4.1); (iii) a class of stative verbs semantically and morphosyntactically distinguished from the class

28 The final -na in patena is possibly an error for the non-firsthand evidential -nia. 29 [ˌʃiʧiˌjamuˈʦa].

804

Raoul Zamponi and Chris Rogers

of active verb (Section 3.5); (iv) evidential suffixes (Section 3.5.3); (v) verb serialization (Section 5); (vi) core clausal constituent order AOV (Section 6.3); and (vi) simple juxtaposition as a mechanism for coordination of noun phrases and clauses (Section 7.1). On the other hand, there are also some features of Máku that appear uncommon in the whole Amazonian context. These include: (i) the presence of the bilabial voiceless fricative /ɸ/ and the high front rounded vowel /y/ in the phonological inventory (Section 2); (ii) a combination of number and clusivity in the pronominal system that produces three forms for ‘we’ (Section 3.2); (iii) verbal subject person marking that, according to the inflectional class of the verb, is by prefixes, infixes, or suffixes (Section 3.5.1); (iv) a well-developed aspectual system (Section 3.5.3); and (v) different negation strategies for active and stative verbs (Section 3.5.4). The not particularly large corpus of Máku did not permit us to dwell on some aspects of this language that surely deserve more attention. We refer in particular to the demonstrative system (Section 3.3), the semantics of evidentials (Section 3.5.3), the morphological characteristics of an apparent reciprocal affix (Section 3.5.5), serial verbs (Section 5), the formation of polar questions (Section 6.4), and subordination strategies in general (Section 7.2).

9 References Aikhenvald, Alexandra Y. 2003. A grammar of Tariana, from northwest Amazonia. Cambridge: Cambridge University Press. Aikhenvald, Alexandra Y. & R. M. W. Dixon. 1999. Other small families and isolates. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), The Amazonian languages, 341‒383. Cambridge: Cambridge University Press. Boersma, Paul & David Weenink. 2013. Praat: Doing phonetics by computer. http:// www.praat.org/ (accessed 10 July 2021) Cysouw, Michael. 2003. The paradigmatic structure of person marking. Oxford: Oxford University Press. Dixon, R. M. W. 1982. Where have all the adjectives gone? and other essays in semantics and syntax. Berlin & New York: Mouton. Dixon, R. M. W. 2006a. Complement clauses and complementation strategies in typological perspective. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Complementation: A crosslinguistic typology, 1‒48. Oxford: Oxford University Press. Dixon, R. M. W. 2006b. Adjective classes in typological perspective. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Adjective classes: A cross-linguistic typology, 1‒49. Oxford: Oxford University Press. Epps, Patience. 2008. A grammar of Hup. Berlin & New York: Mouton de Gruyter. Evans, Nicholas. 2007. Insubordination and its uses. In Irina Nikolaeva (ed.), Finiteness: Theoretical and empirical foundations, 366‒431. Oxford: Oxford University Press. Faria, João Barbosa de. 1927. Vocabulário da tribo Macú. [Manuscript in Arquivo do Museo do Índio, Rio de Janeiro.] Greenberg, Joseph H. 1987. Language in the Americas. Stanford: Stanford University Press. Koch-Grünberg, Theodor. 1906. Die Makú. Anthropos 1. 877‒906.

16 Máku

805

Koch-Grünberg, Theodor. 1911. Aruak-Sprachen Nordwestbrasilien und der angrenzenden Gebiete. Mitteilungen der anthropologischen Geselschaft 41. 33‒153, 203‒282. Koch-Grünberg, Theodor. 1913. Abschluß meiner Reise durch Nordbrasilien zum Orinoco, mit Berücksichtigung der von mir besuchten Indianerstämme. Zeitschrift für Ethnologie 45. 448‒474. Koch-Grünberg, Theodor. 1917. Von Roroima zum Orinoco. Ergebnisse einer Reise in Nordbrasilien und Venezuela in den Jahren 1911–1913. Vol. 1: Schilderung der Reise. Stuttgart: Stracker und Schröder. Koch-Grünberg, Theodor. 1928. Von Roroima zum Orinoco. Ergebnisse einer Reise in Nordbrasilien und Venezuela in den Jahren 1911–1913. Vol. 4: Sprachen. Stuttgart: Stracker und Schröder. Ladefoged, Peter & Ian Maddieson. 1996. The sounds of the world’s languages. Oxford: Blackwell. Loukotka, Čestmír. 1968. Classification of the South American Indian languages. Johannes Wilbert (ed.). Los Angeles: Latin American Center, University of California. Maciel, Iraguacema Lima. 1991. Alguns aspectos fonológicos e morfológicos da língua máku. Brasília: Universidade de Brasília MA thesis. Maciel, Iraguacema Lima (Iramel Lima). 2014. Mexoko: lenda máku. Belém: Cromos. Martins, Silvana & Valteir Martins. 1999. Makú. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), The Amazonian languages, 251‒268. Cambridge: Cambridge University Press. Matisoff, James A. 1975. Rhinoglottophilia: The mysterious connection between nasality and glottality. In Charles A. Ferguson, Larry M. Hyman & John J. Ohala (eds.), Nasálfest: Papers from a symposium on nasals and nasalization, 265‒288. Stanford: Language Universals Project, Department of Linguistics, Stanford University. Migliazza, Ernesto C. 1958a. Questionário padrão para a pesquisa nas línguas indígenas brasileiras: Língua máku. [Manuscript in Anthropological Division of the Museu Paraense Emílio Goeldi, Belém, Brazil.] Migliazza, Ernesto C. 1958b. 34: Words and sentences. Amazonian Languages Collection of Ernest Migliazza. The Archive of the indigenous Languages of Latin America, ailla.utexas.org. Access: public. PID ailla:124991. Migliazza, Ernesto C. 1960a. Formulário dos vocabulários padrões para estudos comparativos preliminares nas línguas indígenas brasileiras (segunda edição): Língua máku (version A). [Manuscript in Anthropological Division of the Museu Paraense Emílio Goeldi, Belém, Brazil.] Migliazza, Ernesto C. 1960b. 31: National Museum list. Amazonian Languages Collection of Ernest Migliazza. The Archive of the Indigenous Languages of Latin America, ailla.utexas.org. Access: public. PID ailla:124988. Migliazza, Ernesto C. 1960c. 32: National Museum list. Amazonian Languages Collection of Ernest Migliazza. The Archive of the Indigenous Languages of Latin America, ailla.utexas.org. Access: public. PID ailla:124989. Migliazza, Ernesto C. 1960d. 33: National Museum list. Amazonian Languages Collection of Ernest Migliazza. The Archive of the Indigenous Languages of Latin America, ailla.utexas.org. Access: public. PID: ailla:124990. Migliazza, Ernesto C. 1964a. Formulário dos vocabulários padrões para estudos comparativos preliminares nas línguas indígenas brasileiras (segunda edição): Língua máku (version B). [Manuscript in Anthropological Division of the Museu Paraense Emílio Goeldi, Belém, Brazil.] Migliazza, Ernesto C. 1964b. 34: Words and sentences. Amazonian Languages Collection of Ernest Migliazza. The Archive of the Indigenous Languages of Latin America, ailla.utexas.org. Access: public. PID ailla:124991. Migliazza, Ernesto C. 1965. Fonologia máku. Boletim do Museu Paraense Emílio Goeldi (Antropologia) 25. 1‒18. Migliazza, Ernesto C. 1966. Esboço sintático de um corpus da língua máku. Boletim do Museu Paraense Emílio Goeldi (Antropologia) 32. 1‒38.

806

Raoul Zamponi and Chris Rogers

Migliazza, Ernesto C. 1978. Máku, Sape and Uruak languages: Current status and basic lexicon. Anthropological Linguistics 20. 133‒140. Nolan, Francis J. 1983. The phonetic bases of speaker recognition. Cambridge: Cambridge University Press. Ohala, John J. & Maria Grazia Busà. 1995. Nasal loss before voiceless fricatives: A perceptually based sound change. Rivista di Linguistica 7. 125‒144. Rice, A. Hamilton. 1928. The Rio Branco, Uraricuera, and Parima (second part). The Geographical Journal 71. 209–33. Rodrigues, Aryon D. 1986. Linguas brasileiras: Para o conhecimento das línguas indígenas. São Paulo: Edições Loyola. Rogers, Chris. 2021. Máku: A comprehensive grammar. London & New York: Routledge. Zamponi, Raoul. Forthcoming. Extinct lineages and unclassified languages of Greater Amazonia. In Patience Epps & Lev Michael (eds.), Amazonian languages: An international handbook, vol. 3: Smaller language families. Berlin & Boston: de Gruyter Mouton.

Katharina Haude

17 Movima 1 2 3 4 5 6 7 8 9 10 11 12 13

Introduction Phonology Morphology Lexical and functional categories Basic clause structure and grammatical relations Voice, valence, and transitivity Complement and adverbial clauses Negation Relativization Pragmatic effects of constituent order alternations Summary Acknowledgments References

1 Introduction Movima (movi1243) is spoken in and around Santa Ana del Yacuma, a former Jesuit mission in Bolivia’s Beni department with approximately 18,000 inhabitants today. Movima is a linguistic isolate, and nothing is known about the pre-colonial history of the ethnic group by which it is spoken. While “Movima” is also the self-designation of non-natives born in Santa Ana del Yacuma, most indigenous people belong to the lower social class of the town’s population. According to a count conducted by members of the speaker community, in 2012 Movima was spoken by approximately 500 adults. There are no first-language learners of Movima anymore, and despite revitalization initiatives, for example, at primary schools, the language must be considered severely endangered. At present there is no evidence of linguistic phenomena resulting from contact with other indigenous languages. A possible loan from Guaraní might be ro:ya ‘house’ (Guaraní óga); the word charke ‘dried meat’ is originally Quechua (charqi), but it is likely that it was introduced via the regional Spanish charque; and the word pa:ko ‘dog’ is found in other native languages of Bolivia, including in the highlands, but its origin is unclear. The prolonged contact with Spanish, however, has led to a large number of lexical borrowings, which are marked by their phonological treatment and by their behavior in compounding. Also, there seem to be numerous calques from Spanish. For instance, the Movima verb toje:ɬe shows the same polysemy as Spanish pasar ‘pass by/happen’, which is probably no coincidence, and

Katharina Haude, Centre national de la recherche scientifique https://doi.org/10.1515/9783110432732-004

808

Katharina Haude

809

17 Movima

‘not anymore’ is expressed in Movima by jayna kas ‘already not’, parallel to ya no in Spanish. Movima was first investigated systematically by the SIL linguists Robert and Judith Judy (see J. Judy 1965; R. Judy 1965; Judy & Judy 1962, 1967). Colette Grinevald carried out an elicitation session on classifiers in the late nineties (Grinevald 2002). The data on which this chapter is based, stemming from direct elicitation plus a corpus of approximately 30 hours (130,000 words) of annotated spontaneous discourse, were collected by the author during 10 field trips (totaling more than 15 months) between 2001 and 2012. All examples presented below are attested in the corpus, although some have been slightly simplified for the sake of presentation. The present grammatical sketch does not cover all aspects of Movima grammar, and some issues are slightly simplified; it is largely restricted to those morphological and syntactic patterns that are frequent in actual discourse and are central for an understanding of the structure of the language. The reader is referred to Haude (2006) for a more fine-grained picture of the phenomena dealt with (or not dealt with) here, as well as to the separate publications by the author referred to in the corresponding sections.

2 Phonology The five vowel phonemes of Movima are /a, e, i, o, u/. There are no nasal vowels. Contrasts in vowel length are mostly metrically or morphologically based (see below), so no separate set of long vowels is assumed. The consonants are listed in Table 17.1, with differing orthographic symbols – partly based on Spanish orthography – in pointed brackets. Given the regional context, the most remarkable phoneme is the voiceless lateral fricative /ɬ/, which is

Tab. 17.1: Movima consonant phonemes (differing orthographic symbols in < >).

simple plosives labialized plosives implosives affricate fricatives lateral fricative nasals lateral approximant simple vibrant glides glottalized glide

bilabial

alveolar

p

t

ɓ

ɗ

β

s ɬ n l ɾ

palatal

velar

glottal

k kw

ʔ

tʃ͡

m

w

h

j jʔ

810

Katharina Haude

frequently heard in spontaneous speech due to its appearance in several common affixes. There are not many phonologically conditioned alternations. The alveolar nasal /n/ assimilates in place of articulation to right adjacent consonants: before a bilabial consonant, it is pronounced as [m]; before a velar consonant, it is pronounced as [ŋ]. A striking allophony, which contributes to the abundance of glottal(ized) sounds in spoken Movima, involves the voiceless plosives /p/, /t/, and /k/. In coda position, these are realized as glottal(ized) consonants: /k/ is realized as the simple glottal stop [ʔ], which is followed by an echo vowel repeating the preceding vowel at the end of a prosodic phrase (this echo vowel is only represented orthographically when it is actually pronounced); /p/ and /t/, while retaining their place of articu͡ m ] and [tʔ͡ n], lation during the glottal closure, are released nasally and realized as [pʔ respectively. Syllables may be open or closed. Consonant clusters are not allowed, although they may occur in loans. There are some restrictions on the coda position: the bilabial fricative /β/ and the affricate /tʃ͡ / do not occur in coda position; /ɾ/ occurs in coda position only in Spanish loans, where it tends to be realized as [h]; plosives in coda position are realized by their glottalized allophones, described above; and the labialized plosive /kw/ is not found in coda position at all. There are no vowel-initial syllables: what is spelled orthographically as a vowel-initial word is, in fact, preceded by a glottal stop, and the same is true of vowel-initial bound morphemes (except ‘external clitics’, see Section 3.5). Metrical structure, as reflected by the opposition between light (L, i.e., CV) and heavy (H, i.e., CVː or CVC) syllables, is extremely important in Movima word formation, but not yet fully explored. In general, content words are minimally disyllabic and consist of minimally three morae. Stress (not represented orthographically) occurs by default on the penultimate syllable of the word. When the penultimate syllable is open, it is lengthened; in the case of disyllabic words, this lengthening is a way to achieve the three-mora condition. Examples of frequent words that conform to these rules are given in (1). (1)

a. b. c. d. e.

CVː.CV CVC.CVC CVC.CV CVː.CVC CV.CVː.CV

/ˈtoːmi/ /ˈɓajɬim/ /ˈalɾa/ /ˈɓiːhaw/ /koˈɾiːɗi/

‘water’ ‘my garden, field’ ‘my friend’ ‘old’ ‘stick’

Syllables with glottal(ized) coda consonants attract stress. Hence, words ending in /p/, /t/ or /k/, glottalized in coda position, are stressed on the last syllable, for ͡ m ] ‘flea’, /tʃuhat/ [tʃuːˈhatʔ͡ n] ‘motacú (palm tree)’ or example, /kuɗup/ [kuːˈɗupʔ e /meɾek/ [meˈɾeʔ ] ‘big’. The simple glottal stop, [ʔ], can furthermore cause deviations from the lengthening rule: most disyllabic words ending a simple glottal stop have a short initial syllable, for example, /merek/ ‘big’ or /karak/ ‘macaw’.

17 Movima

811

There is a closed class of about seven nouns that are phonologically defective in that they only consist of two light syllables with identical vowels, for example, /koʔo/ ‘tree, wood’ or /βeʔe/ ‘fire’. These nouns all contain a glottal stop, and they have a special form when suffixes or enclitics are added; for instance, the base /koʔo/ ‘tree’ becomes /koj/ in that case. Spanish loans are adapted to the Movima stress rules. A stressed penultimate syllable is lengthened when open: for instance, the loan from Spanish policía ‘police’ is pronounced [poliˈsiːja]. However, when a Spanish word originally carries stress on the antepenultimate syllable, like música ‘music’, the penultimate syllable of the loan is stressed, but remains short, while the originally stressed syllable is lengthened: [muːˈsika]. Furthermore, unlike native words (see Section 3.1), disyllabic Spanish loans with the structure CVːCV retain the long penultimate vowel also when further morphemes are added; for example, [ˈwaːka] ‘cow’ (from Spanish vaca), [waːˈkaːɗi] ‘reins’ (-di ‘clf:long/thin’), [waːkaˈtoːɗa] ‘meat’ (-toda ‘piece’). Other deviations from the stress and lengthening rules, which suggest that long vowels may be synchronically phonemic, are probably due to the lexicalization of morphologically complex words. This can result in minimal pairs distinguished by lengthening, like [ɓaˈloːsi] ‘pink’ vs. [ɓaːˈlosi] (baː- ‘finish’, -losi ‘resin’) ‘finished resin’. Furthermore, different cliticization processes (see Section 3.5) lead to significant stress and lengthening distinctions, as shown in (2). In (2a), the penultimate syllable of the complex unit is stressed and there is no long vowel. In (2b), the antepenultimate syllable of the unit is stressed and lengthened. This is because in (2a), the bound pronoun /us/ is “internally cliticized” (represented by ‘=’; see Section 3.5). Internal cliticization, which attaches a possessor to nouns or the prxm argument to transitive verbs (see Section 5.1), results in a prosodic word that bears no lengthening on either the penultimate syllable of the host or the penultimate syllable of the resulting word. In (2b), the same bound pronoun /us/ is “externally cliticized” (represented by ‘--’). External cliticization, which attaches a pronoun representing s or obv to the predicate (see Section 5.1), has no prosodic effect, and so, the penultimate syllable of the host remains long and stressed. As the meaning contrast between (2a) and (2b) shows, in Movima, variation in the stress and lengthening pattern of a transitive predicate indicates who acts on whom. (2)

a. [ʔajaˈnaʔus] aja-na=us wait-drct=3m.absnt ‘He waits for him/her/it/them.’

b. [ʔaˈjaːnaʔus] aja:-na=Ø--us wait-drct=1sg--3m.absnt ‘I wait for him.’

As far as intonation is concerned, declarative clauses generally bear the pitch accent on the lexical predicate. Marked intonation is the major device for question formation. In these cases, the first syllable of the first word in the sentence, even if nor-

812

Katharina Haude

mally unstressed, receives high pitch, and pitch decreases continuously towards the end of the utterance.

3 Morphology Lexical morphology in Movima is mainly agglutinating. The fusional character of referential elements (Section 4.1), where one morpheme indicates multiple semantic, deictic, and syntactic categories simultaneously, can be attributed to fossilized morphological complexity (Haude 2006: 143). The distinction between derivation and inflection is difficult to make in Movima, since canonical inflectional categories like tense/mood/aspect on verbs or case/ number/gender on nouns do not exist, and there is no agreement morphology. Most morphemes are best considered derivational in that they change the lexical category of a word and/or affect its meaning (e.g., verbalization and nominalization, or valence increasing morphology on verbs, Section 6.2). Verbal direct/inverse marking (Section 6.1) might be considered inflectional since it is fully productive and interacts directly with clausal syntax. Still, given the weak evidence of a clear distinction, it seems preferable to leave the question open and to speak of derivation (and accordingly, of lexical bases rather than stems) everywhere.

3.1 Suffixes and prefixes Suffixes, the most common bound morphemes, create a phonological word that follows the stress and lengthening rules outlined above: stress and length shift towards the right, as shown by the boldface on the stressed syllables in (3). Some endings, for example, /wa/ in (3c), require a ‘linking nasal’ (glossed ‘epenthetic consonant’, ep.c) before a further suffix is added (Haude 2006: 59). (3)

a. iwaːni speak ‘(I/you/X) speak(s).’

b. iwaniː-wa=Ø speak-evnt.nmlz=1sg ‘my speaking’

c. iwani-wa-n-si speak-evnt.nmlz-ep.c-clf:sound ‘way of speaking’ The only real prefix is the oblique marker n(V)-, which attaches to referential elements (Section 4.1). When the referential element is consonant-initial, the prefix receives a vowel that is identical to the first vowel of the base: ni-kinos (oblart.f.ab). Furthermore, word-initial reduplication (Section 3.2) can also be analyzed as prefixation, albeit phonologically underspecified.

17 Movima

813

3.2 Reduplication There are four reduplication processes, all of them regressive (i.e., copy preceding source). They are based on metrical phenomena: prefixing reduplication can be monomoraic (CV~) or bimoraic (CVC~ or CVː~). One reduplication process even involves a full iambic foot (H~, LL~, or LH~), as in for example maj~majni ‘have children’, nono~no:no ‘have animals’, choran~chorankwanto ‘have/wear a hat’. Infixing reduplication is always monomoraic (); it involves copying the last CV-element of the base and inserting it before the source, as in dejal (cook) ‘cook’. Reduplication in Movima covers a large and cross-linguistically unusual range of grammatical functions, listed in Table 17.2. Infixing and bimoraic prefixing reduplication each perform multiple different functions, but disambiguation follows either from the base to which they are applied or from the syntactic environment (Haude 2014a). Only in some cases does reduplication simply serve to satisfy prosodic well-formedness constraints, for example, with monosyllabic bound noun roots, as in di~di-n-kwa (red~grain-ep.c-alien) ‘grain/seed’.

Tab. 17.2: Forms and functions of Movima reduplication (μ= mora; F=iambic foot). reduplicant

base

category indicated

gloss

μ~

Monosyllabic verb root with suffix

direct

drct~

μμ~

Monosyllabic verb root with suffix

inverse

inv~

Monosyllabic verb root without suffix

middle

mid~

Disyllabic verb root or complex base, with suffix

inverse

Disyllabic verb root or complex base

middle

Noun

inalienable possession

Noun

embedded predicate

Noun

predicative possession

poss~

F~

3.3 Infixation Movima has three infixes whose position inside the base is partly morphologically, partly metrically conditioned.

814

Katharina Haude

The first is multiple-event/participant (pluract), which is attached to the root of complex verbal bases: tan-pit (cut-clf:middle) ‘be cut in halves’ → tanka:-pit (cut-pluract-clf:middle) ‘be cut into various pieces’. When this marker is inserted in a synchronically unanalyzable root, it must be analyzed as an infix: iye:ni ‘move’ → iyeni ‘move repeatedly’. The second is direct (drct). It is an allomorph of the direct suffix -na (Section 6.1) and occurs in complex bases after monosyllabic roots of the structure CVC: tan-a-pit=Ø (cut-drct-clf:middle=1sg) ‘I cut (it) in halves’. This affix comes late in word formation, even if, as in this example, it precedes other morphemes in linear order (see also (38) and (39)). As in the case of pluract, when the components of the base cannot be identified synchronically, this element must be analyzed as an infix, as in jomni=Ø (devour=1sg) ‘I devoured it’. When the slot after the first syllable is already occupied by pluract, the direct marker is replaced by the word-final allomorph -na: tan-ka-pit-na=Ø ‘I cut it in various pieces’. The third infix is irrealis (irr), which indicates either participant negation (Section 8.3) or, on verbs, undetermined future. It is inserted after the first iambic foot (LH or H) of the base independently of the base’s internal morphological complexity. It takes the form after vowels, in (4), and after consonants, in (5). As (6) shows, also this marker can be analyzed as a suffix in those cases where it occurs between two identifiable morphemes. On phonologically defective nouns (see above), it is replaced by a reduplicated suffix, shown in (7). (4)

a. aro:so rice ‘rice’

b. aroso rice ‘There is no rice.’

(5)

a. as-na=Ø sit-loc.nmlz=1sg ‘my home’

b. as-na=Ø sit-loc.nmlz=1sg ‘I have no home.’

(6)

a. chi-poj-kay=Ø exit-caus-inv=1sg ‘(They) drive me out.’

b. chi-poj-ak-kay=Ø exit-caus-irr-inv=1sg ‘Nobody drives me out. / May they drive me out.’

(7)

a. ko’o tree,wood ‘tree/wood’

b. ko’-ka:~kak tree;wood-red~irr ‘There is no tree/wood.’

3.4 Compounding and incorporation Compounding and noun incorporation are important word-forming devices in Movima. Here a nominal element is attached to the right of a lexical base. The attached

17 Movima

815

nominal element is rarely a full noun, as in sotak-sema:na (one-week) ‘one week’; more commonly, it is a bound root (like -di ‘grain’), a truncated element (like -mi ‘trnc.water’ from to:mi ‘water’), or a bound element with no corresponding free form (e.g., -waj clf:place; -lomaj clf:time). Multiple compounding with bound elements is fairly productive, for example, bo:ve ‘straw fan’, bove:-mo ‘straw basket’, bove-mo:-ba ‘little round straw basket’. There is a strong tendency towards lexicalization, and many synchronically simple nouns probably originate from compounding. Truncation, that is, the clipping of one part of a noun so that it can be used for incorporation or compounding, usually involves the last syllable of a word, but it is extremely irregular and of limited productivity. There are indications that some apparently truncated elements were originally noun roots which were lexicalized with other elements so that their occurrence in other environments is regarded as truncation. So, for instance, the classificatory bound element -lo ‘liquid’ in charaye:-lo ‘syrup’ (sugarcane-clf:liquid) might be interpreted as a truncation from the noun nonlo ‘milk’; however, the first element in nonlo is identical to the verb root non- ‘suckle’, and, hence, the noun nonlo itself is a compound headed by the classificatory bound element -lo. Furthermore, the truncation of non-final segments is a hint that the source word may have undergone subsequent compounding: the bound element -kos ‘girl’ stems from tolkosya ‘girl’, whose last element -ya, in turn, may be a truncation from kwe:ya ‘woman’. This interpretation is supported by the fact that some speakers still use the word tolkos ‘girl’. There must have been a point in time in which truncation was highly productive, as is also evident from the treatment of some, probably older, Spanish loans. These must be disyllabic when truncated (e.g., -pato from sapa:to ‘shoe’, Spanish zapato), and in the case of originally disyllabic loans, the truncated last syllable is reduplicated: for instance, si:ya ‘chair’ (from Spanish silla) becomes -yaya in constructions like sotak-yaya (one-trnc.chair) ‘one chair’.

3.5 Cliticization Cliticization involves referential elements, that is, determiners, bound pronouns, and demonstratives (Section 4.1). These can be attached to different kinds of hosts and can have syntactic scope over an entire phrase. There are two types of cliticization: “internal” and “external”. Internal cliticization (represented by ‘=’), which marks the prxm argument on transitive predicates and the possessor on nouns, is suffix-like in that it creates a prosodic word with the corresponding penultimate stress (Section 2). In contrast to canonical prosodic words, the penultimate syllable is not lengthened. Internal clitics furthermore require a preceding vowel; on consonant-final hosts, a linking vowel /a/ is attached. The hiatus with vowel-initial enclitics is resolved by a glottal stop. While this process might also be interpreted as suffixation, the cliticization analysis is supported by the fact that it also applies to determiners, which form a syntactic phrase with the subsequent content word. This

816

Katharina Haude

is illustrated in (8), where the article =us is phonologically encliticized to a transitive predicate and at the same time forms a referential phrase (RP) together with the proper noun Ernan. Example (8) also shows the pronoun =y’ɬi internally encliticized to a noun, encoding its possessor. The stressed syllables are marked with an accent here. (8)

jayna jaymoɬ-á=us Ernan us pa:toron-á=y’ɬi dsc call-ep.v=art.m Ernan art.m landlord-ep.v=1excl ‘Then Ernan called our landlord.’

External cliticization (represented by ‘--’) only involves bound pronouns that represent the single argument (s) of an intransitive predicate or the obv argument of a transitive predicate (see Section 5.2). Stress and vowel length remain unaffected, and no preceding vowel is required. The feature that distinguishes external cliticization from both juxtaposition and suffixation is that vowel-initial enclitic elements are resyllabified with a host-final consonant, as shown in (9); in juxtaposition and suffixation, by contrast, vowel-initial morphemes are preceded by a glottal stop. (9)

jo’yaj--us arrive--3m.absnt ‘He arrived.’

/'hoʔ.ja.hus/

Another type of cliticization involves a single consonant that must attach to a preceding vowel. Cases in point are the determining element s and the first-person pronoun ɬ (see Table 17.6), both of which are often neutralized to [h]. They both occur as the final element of the articles (of which they analyzed as an integral part, see Table 17.4) but also on determining demonstratives (Section 4.1) and on the negative copula (Section 8), where they are analyzed as clitics.

4 Lexical and functional categories The most fundamental word class distinction is that between content words, referential elements, and particles. Referential elements (Section 4.1) belong to a formally and functionally clearly definable closed class. Content words (Section 4.2) contrast with particles (Section 4.3) in their morphology.

4.1 Referential elements There are three sets of referential elements: articles, personal pronouns, and demonstratives. Articles (Table 17.3) mark the beginning of a referential phrase. All

17 Movima

817

Tab. 17.3: Movima articles.

presential/generic absential (ab) ceased existence/past (pst)

male sg (m)

female sg (f)

non-human sg (n)

plural/mass (pl)

us kus us

i’nes kinos isnos

as kos os

is kis is(os)

articles have the ending /s/, which can be seen as the element bearing the determining function (it can also occur on demonstrative determiners, see below, and with the negative copula, Section 7). Articles specify the referents of a referential phrase as human male/female versus non-human (“neuter”), singular versus plural/mass, present versus absent, and in existence versus not existing anymore (“past”). The neuter article is furthermore used for nonspecific or derogatory reference to humans. The articles do not mark definiteness. The “presential” and “absential” forms of the article are used for referents that exist at the place of speaking or somewhere else, respectively; these are shown in (10a). The “past” form is used when the referent is absent and does not exist anymore, as in (10b). (10) a. loy iɬ ajlo:maj as / kos pa:ko intn 1sg.pro tell.about art.n / art.n.ab dog ‘I’ll tell you about my (present / absent) dog.’ b. loy iɬ ajlo:maj os pa:ko intn 1sg.pro tell.about art.n.pst dog ‘I’ll tell you about my (former/deceased) dog.’ In past contexts, the past-tense article also occurs in referential phrases referring to entities that may still be in existence but whose existence is considered irrelevant. This is especially common with non-human referents, as in (11), where the ‘machine’ might still have been in existence at the time of speaking even though the narrated events took place many years ago. Here, the article simply marks past tense. (11) yey-na=’ne os ma:kina want-drct=3f.pro art.n.pst sewing.machine ‘She wanted the/a sewing machine.’ In referential phrases referring to non-time-stable and more abstract entities such as points in time or states and events, the absential article acquires a temporal meaning, indicating hodiernal past (see the complement clause in (49)).

818

Katharina Haude

Tab. 17.4: Movima third-person personal pronouns.

presential/generic absential (ab)

human male (m) human female (f)

non-human (n)

plural/mass (pl)

free

bound

free

bound

free

bound

free

bound

u’ko usko

u(’) us

i’ne isne

(i)’ne (i)sne

a’ko asko

a(’) as

i’ko isko

i(’) is

Tab. 17.5: Movima personal first- and second-person pronouns. free

1sg 2sg 1+2 1excl 2pl

inɬa ulkwat i:de iy’ɬi iy’bikweɬ

bound (prxm/possessor)

bound (s)

pre-head

post-head

pre-head

post-head

(i)ɬ – (i)ɬ (i)ɬ –

=Ø =n =n =y’ɬi =nkweɬ

(i)t (i)j (i)ɬ (i)t (i)j

– – – --(i)y’ɬi --(i)y’bi

The set of free and bound third-person personal pronouns is displayed in Table 17.4. Free personal pronouns (glossed pro) typically occur clause-initially, where they function as predicates (Section 9.3). Bound pronouns are encliticized, either internally or externally, to the preceding noun or verb (Section 3.5). The system of personal pronouns referring to speech-act-participants (SAP; Table 17.5) is more heterogeneous (see Haude 2011a). The distinction between free and bound pronouns exists here as well, but there is less morphological similarity between the two categories. There are two sets of bound pronouns, one for marking possessors and the prxm argument of transitive verbs (Section 5.1) and the other for marking the single argument of intransitive predicates (s). Both sets also contain elements that precede the head. These consist of a single consonant, which is either preceded by a dummy vowel /i/ or is phonologically cliticized to a preceding vowelfinal word. None of the pre-head elements are grammatically obligatory. The third set of referential elements contains the demonstratives (Table 17.6). Demonstratives can function as determiners (12), as pronouns (13), and as predicates of existential (14a) and possessive clauses (14b). They convey temporal, aspectual, and evidential information. (12) kodé=s pa:ko dem.n.nstd=det dog ‘that (visible, sitting or lying) dog’

absential

kuwa

kuro’ uso’

past

kulro’

moving away from speaker

non-past

kula’wa

moving towards speaker

kupa kulpa

(prox) (dist)

elevated

kude: kulde:

temporary possession

(prox)

non-standing on ground

kure’ kulre’

kulwa

(prox) (dist)

standing on ground

positional

u:(ru) kul(ru)

human male (m)

elevated distant, or otherwise perceived

(prox) (dist)

close to speaker close to hearer

SAPoriented

Tab. 17.6: Movima demonstratives.

isno’

kino’

kilno’

kila’niwa

kinipa kilnipa

kilniwa

kiniwa

kinede: kilnede:

kine’ kilne’

i:(ni) kil(ni)

human female (f)

oso’

koro’

kolro’

kola’wa

kopa kolpa

kolwa

kowa

kode: kolde:

kore’ kolre’

ay(ru) kal(ru)

non-human (n)

iso’

kiro’

kilro’

kila’wa

kipa kilpa

kilwa

kiwa

kide: kilde:

kire’ kilre’

i:(ri) kil(ri)

plural/mass (pl)

17 Movima

819

820

Katharina Haude

(13) kulro’ joy-cheɬ dem.m.retr go-refl/recp ‘He (in sight, retreating) is leaving.’ (14) a. uso’ us itila:kwa dem.m.pst art.m man ‘There was the/a man.’ b. uso’ us majniwa=us dem.m.pst art.m offspring.of=3m.ab ‘He had a son.’ (Lit. ‘There was his son.’) The functions of referential elements can be illustrated with the structure of referential phrases and the marking of possession. A referential phrase contains, minimally, a determiner and a content word (see also Section 9.2). Possessive marking is realized through internal cliticization to the noun (Section 3.5) of a bound pronoun, as in (15), or of the determiner of the possessor referential phrase, as in (16). (15) us alwaj-a=’ne art.m spouse-ep.v=3f.pro ‘her husband’ (16) us alwaj-a=kine’e=s tolkosya art.m spouse-ep.v=dem.f.std=det girl ‘that (standing) girl’s husband’ As on transitive verbs, the absence of an overt internal enclitic on an obligatorily possessed noun (see Section 4.2) marks the first-person singular possessor, shown in (17). As on verbs, the proclitic first-person marker ɬ is not obligatory here. (17) us alwaj=Ø art.m spouse=1sg.pro ‘my husband’ The way in which referential elements encode the obv argument is illustrated in Section 5.1; predicative personal pronouns are discussed in Section 9.3. All referential elements except bound pronouns can be marked as oblique by the n(V)- prefix, which marks any kind of adjunct, as in (18). (18) n-us alwaj=Ø obl-art.m spouse=1sg.pro ‘with/for/from etc. my husband’

17 Movima

821

4.2 Content words 4.2.1 Nouns and verbs: Morphology Most Movima content words can be identified as either nouns or verbs on morphological grounds (property-denoting words form subclasses of verbs and nouns; see Section 4.2). However, the difference between nouns and verbs is not always overt. Cross-linguistically typical nominal categories like case/number/gender or verbal categories like tense/aspect/mood are either not, or not consistently, morphologically marked; some morphemes are even shared by nouns and verbs, for example, the irrealis infix marking existential negation (Sections 3.3 and 8.3). However, there are morphemes that do depend on the lexical category of the base. For instance, the verbalizing suffix -tik ‘make/do (sth. with) N’ only occurs on nominal bases, such as pokso ‘chicha’ or po:no ‘drum’: pokso-tik ‘make chicha’; pono:-tik ‘play the drums’. Hence, words can usually be tested for their morphological possibilities in order to be classified. While aspect and modality are mainly marked by particles, there are also some affixes that mark these categories on verbs. For instance, there is a suffix -kaɬ, which indicates that an action is going to be carried out immediately. Irrealis can imply an undetermined future when attached to a verbal base: salmo (return) can mean ‘Nobody returns’ and also ‘I’ll be back’. Some verbal and nominal affixes are homophonous, and their interpretation depends on the lexical category of the base. For instance, the suffix -na is an intransitive directional verbalizer on nouns: itwa-na (river-vbz.dir) ‘go to the river’. On bivalent verbal bases, by contrast, -na is a marker of direct voice, and on monovalent verbal bases -na is a locative nominalizer (Section 6.1). Suffixes with the form -kwa can also be found on both nouns and verbs. On nouns, -kwa marks “absolute state”, as in (21b). On verbs, -kwa marks the benefactive (Section 6.2), and -kwa also derives verbs denoting a bodily process, as in joro:-kwa ‘sleep’ or choj-kwa ‘urinate’. Furthermore, infixing reduplication marks inalienable possession on nouns, as in (20b); on verbs, it marks the inverse voice on complex verbal bases (Section 3.2). These meaning differences of homophonous affixes suggest a clear lexical difference between nouns and verbs – or at least between the subclasses of nouns and verbs with which these elements can be productively combined.

4.2.2 Nouns and verbs: Syntax Verbs typically function as predicates and nouns typically occur in referential phrases, and deviating distributions are pragmatically marked (Section 10.1). However, there is no categorical distinction between verbs and nouns on the syntactic level. Basically, all content words can function as predicates, and all content words can

822

Katharina Haude

be combined with a determiner to form a referential phrase. Possessed nouns and proper names, however, are usually not found as main-clause predicates (Section 9.3). Hence, a rough three-way lexical distinction can be made on syntactic grounds between (i) possessed or proper nouns, which have a restricted potential to function as predicates, (ii) bivalent verbs, which freely form transitive predicates, and (iii) unpossessed common nouns and monovalent verbs, which freely form intransitive predicates. A useful criterion to distinguish nominal from verbal predicates is their behavior in embedding (Section 7). Embedded verbal predicates take the suffix -wa, while nominal predicates undergo reduplication. When a noun is combined with the verbal suffix -wa, as in (19), the resulting meaning hints at a zero verbalization that replaces the verbalizer -tik before further suffixes.1 (19) os pokso-Ø-wa=sne art.n.pst chicha-vbz-evnt.nmlz=3 f.absnt ‘her making (of) chicha’

4.2.3 Possession Most nouns can be possessed; they can be directly combined with an internally encliticized person marker (Sections 3.5 and 4.1), for example, roya=n (house =2) ‘your house’. A verb, in contrast, can only be combined with an internal enclitic when it is overtly marked as transitive by either the direct or the inverse affix (see Section 5.2). Hence, as a rule of thumb, a word that can be combined with an internal enclitic without containing a direct or inverse marker is a noun. This criterion identifies, for instance, the words jampa ‘do’ and jankwa ‘say’ as nouns, or at least as non-verbs: jampa=n ‘you do’, Lit. ‘your done (thing)’, jankwa=n ‘you say’, Lit. ‘your said (thing)’. Nouns are either alienable or inalienable. Alienable nouns, like ro:ya ‘house’ or to:mi ‘water’, can occur with and without possessive marking. When denoting inalienably possessed entities, they undergo infixing reduplication (Section 3.2). Examples (20a) and (20b) show the noun ro:ya ‘house’ alienably and inalienably possessed, respectively. (20) a. as roya=us art.n house=3m.absnt ‘his house’

b. as roya=a’ art.n house=3n ‘its house (of the riverboat)’

On inalienable nouns, a zero internal enclitic indexes the first-person singular. This class includes kinship terms, part-of-whole terms, and some terms denoting goods, 1 See Haude (2011b) for a different analysis.

17 Movima

823

like bayɬim=Ø ‘my field’, dokwe=Ø ‘my dress’, no:no=Ø ‘my domestic animal’. Inalienable nouns must be overtly marked if denoting an unpossessed entity. Markers of nonpossession vary depending on the base they attach to, but they are all based, in one way or another, on the complex instrumental suffix -wa-n-ra(-ni) nmlz-epclf(-vbz.inch), for example: bayɬim-wa:nas ‘field’, dokwe-wanra:ni ‘clothes’, and nono-wanra ’domestic animal’. Unpossessed kinship terms take the ending -wawankwa, as in ma’-wawankwa ‘mother (i.e., woman who has children)’. Many part-of-whole terms are phonologically defective roots that consist of two morae only, for example, lora ‘leaf’ (CVCV) in (21a), and they cannot occur without an overt possessive marker. When unpossessed or alienably possessed, these roots receive the absolute state suffix -kwa that marks them as physically detached from the entity whose part they are, as in (21b). This derived form, in turn, can again be marked as inalienably possessed by infixing reduplication, as in (21c), which is a semantically equivalent alternative to (21a). (21) a. lora=as leaf=3n.absnt ‘its (i.e., a plant’s) leaf’

b. lora-n-kwa(=us) leaf-ep.c-alien(=3m.absnt) ‘(his) leaf’ (alienably possessed or unpossessed)

c. lora-n-kwa=as leaf-ep.c-alien=3n.absnt ‘its (i.e., a plant’s) leaf’ Some words that would intuitively be classed as nouns cannot be possessed. These are, for instance, instrumental nouns ending in -ni; see (22). The ending -ni is also found on some intransitive verbs (e.g., iwa:ni ‘speak’), and it marks adjectives (Section 4.2.4) as denoting a process. Therefore, perhaps these supposed nouns originate from intransitive verbs. (22) a. ya:lowe-wanra:-ni drink-ins:ec-vbz.inch ‘beverage’

b. *ya:lowe-wanra-ni=n drink-ins:ec-vbz.inch=2 (Intended: ‘your beverage’)

4.2.4 Adjectives Property-denoting words like merek ‘big’, tochik ‘small’, ja:yaw ‘good, nice’, and color terms, share properties of both verbs and nouns. Like monovalent verbs, property-denoting words cannot be combined with an internal enclitic and are not found with the verbalizing suffix -tik. Like bivalent verbs, in combination with the suffix -na they form a direct transitive predicate: jayaw-na=n ‘you make (it) nice’ (nicedrct​=2). Unlike bivalent verbs, however, the unmarked form does not denote the

824

Katharina Haude

result of an action (Section 6.1). Furthermore, unlike verbs, but like nouns, property-denoting words can undergo reduplication when functioning as embedded predicates, as in (23a). There is, however, an alternative form, which is derived with the suffix -ɬe, like in (23b). This may be taken as evidence that property-denoting words form a separate subclass of nouns, that is, adjectives. (23) a. n-os tochik-a=sne obl-art.n.pst small-ep.v=3 f.absnt ‘when she was small’ (Lit. ‘in her past-being small’) b. n-os tochik-ɬe=sne obl-art.n.pst small-stat.nmlz=3 f.absnt ‘when she was small’ (Lit. ‘in her past-being small’) A further distinctive feature of property-denoting words is that they can only occur as the leftmost (i.e., modifying) element of a compound, as in merek-ro:ya ‘big house’. There is no evidence indicating whether this is compounding or mere juxtaposition. If analyzed as juxtaposition, this would mean that there is a class of content words, adjectives, which can directly precede another content word within the same phrase. Quantifiers like ba:- ‘all’, kaw- ‘much/many’, the four native numeral terms sotak- ‘one’, oy- ‘two’, tas- ‘three’, and oyka- ‘four’, as well as some property-denoting roots such as dit- ‘hard’, mol- ‘unripe’, or pola- ‘new’, require a classifier or other bound nominal element: for example, mol-ba ‘unripe round fruit’ (unripeclf:round). When no further semantic specification is given, they take as a placeholder the semantically neutral bound element -ra, as in mol-ra ‘unripe’.

4.2.5 Content words: summary The three major classes of content words as distinguished by the above criteria are listed in Table 17.7.

Tab. 17.7: Some criteria distinguishing classes of Movima content words.

verbs nouns adjectives

occurs as main predicate

marker if embedded

+ vbz-tik

+ /-na drct, or loc.nmlz

+ internal enclitic

yes yes yes

-wa

, -ɬe

no yes no

yes no yes (rare)

yes, if drct/inv yes (most) no

17 Movima

825

4.3 Particles Movima has at least 40 particles, that is, words that neither belong to the closed class of referential elements nor participate in any of the morphological processes that apply to content words. Particles are generally morphologically invariant and typically only bimoraic (unlike content words). Some, however, have a long and a short form, for example, lajat/lat pre-hodiernal past (glossed ant), to(je)ɬ ‘very’, or nok(o)wa future; some particles can occur either with or without a final syllable -ka, like kwey(ka) hodiernal past, rey(ka) ‘again; you know’, or po:ra(ka) ‘briefly’. Particles cover a large range of functions. They can be subdivided into several groups based on their functional and distributional properties. Coordinating particles, such as che ‘and’, ban ‘but’, or bo ‘because’, occur at the beginning of the adjoined clause. Most other particles can occur anywhere in the sentence, even inside a referential phrase, and many can occur more than once in the same sentence. This includes the ubiquitous discourse particle jayna (33), which is analyzed as marking a discontinuity in the course of events but still requires more research. Particles marking tense, aspect, mood, and evidentiality are not restricted to a particular position. Some particles, for example, loy intentional, loy subordinate negator (Section 8.2), and the detransitivizing marker kaw/kwey (Section 6.3) immediately precede the predicate. Interjections usually occur at sentence margins. Particles are the primary means for encoding tense, aspect, modality, and evidentiality. Tense particles distinguish between remote (kwil), pre-hodiernal (la’ or lajat), and hodiernal (kwey(ka)) past, as well as future (nok(o)wa and loy). Tense particles usually occur only towards the beginning of a text, where they establish the time of the narration; past and non-past tense are further indicated by articles (see Section 4.1). Aspectual categories indicated by particles include chot ‘habitually’ or the durative particles ena’ ‘standing’ (see (89)), da(ya)’ ‘non-standing’, and buka’ ‘moving’. Modal, epistemic, and evidential particles include, for instance, words such as rey ‘you see / as we all know’ (regional Spanish pues) or ɬat, used to attract the addressee’s attention (‘Look!’) or indicate indirect knowledge in nonpersonal narratives. Homophonous particles can be distinguished by their allomorphies and/or their syntactic distribution. For instance, hodiernal past tense kwey has an allomorph kweyka, which the detransitivizer kwey does not have. In contrast, the detransitivizer kwey is a speaker-dependent variant of kaw, and its occurrence is limited to relative clauses and referential phrases (Section 6.3). Another case of homophony is loy subordinate negation (Section 8.2), which only occurs in subordinate constructions and is therefore distinct from the intentional future loy. Finally, hypothetical di’ can be distinguished from relativizer di’ in that the former can occur at the beginning of a basic clause, while the latter has to follow the noun it modifies (Section 9.1). Incidentally, there are several modal particles commencing with di, for instance: didi’ frustrative, dis optative, and disoy counterfactual.

826

Katharina Haude

5 Basic clause structure and grammatical relations 5.1 Argument encoding Movima syntax is predominantly predicate initial, and the language displays all the properties that this feature commonly entails.2 In terms of morphosyntactic ordering, there are no postpositions, and the only non-reduplicative prefix in the language, the oblique marker nV-, can be compared to prepositions in other languages. The order of possessive phrases is possessed-possessor. Relative clauses always follow their head. Incorporated elements follow the verb root. There is neither a copula nor a possessive verb, and the most common main-clause alignment pattern is ergative. Movima also provides evidence for Myhill’s (1985) claim that in verb-initial languages a verb in non-initial position is deverbalized (Section 9.3). Movima shows neither agreement nor case marking. The two arguments of transitive clauses are distinguished by constituency. They can be characterized as “internal” versus “external” to the predicate phrase based on the fact that the former is tightly phonologically attached to the predicate while the latter is less tightly connected, can be left unexpressed, or can be “extracted” from its position (Section 9; see Table 17.8 below). The internal argument represents the event participant that ranks higher in the hierarchies of person (1>2>3), animacy (human > animate > inanimate), and discourse topicality, while the external argument represents the event participant that ranks lower on these hierarchies. (When the animacy and discourse hierarchies are in conflict, animacy tends to be more important, although there are exceptions; see Haude 2014b.) Because of this impact of referential hierarchies (reminiscent of Algonquian obviation systems), the internal argument of a transitive predicate is labeled prxm (proximate), and the external argument is labeled obv (obviative). An illustration of pronominal argument encoding is given in (24). In (24a), the prxm argument is the first-person singular, which is zero-marked, so that the verb (aya:na) has the typical prosodic pattern with stress and length on the penultimate syllable. The externally cliticized pronoun (--us) does not induce or undergo any phonological process. In (24b), prxm is the second-person enclitic =n. Being nonsyllabic, this element does not affect the stress pattern of the host; however, like all internal enclitics, it causes the penultimate syllable of the host to lose its lengthening. The subsequent vowel-initial externally cliticized pronoun --us takes the consonant /n/ as its onset. Finally, example (24c) illustrates internal cliticization of a syllabic pronoun, =us, which causes the main stress to shift to the new penultimate syllable, illustrating the suffix-like character of internal cliticization (albeit without the corresponding penultimate lengthening). The following third-person pronoun is preceded by an element k-, which is glossed obviative since the pronoun encodes the obv argument. 2 See Clemens and Polinsky (2015: 3).

17 Movima

827

Tab. 17.8: Properties of prxm versus obv. prxm

obv

internal cliticization (=): stress shift, epenthetic /a/, pronouns and articles are cliticized obligatory (=Ø is 1sg) higher in referential hierarchy all persons

external cliticization (--): no stress shift, no epenthetic /a/, only pronouns are cliticized, resyllabification not grammatically obligatory lower in referential hierarchy only 2pl and 3rd persons

(24) a. aya:-na=Ø--us wait.for-drct=1sg--3m.absnt ‘I wait for him.’ b. aya-na=n--us wait.for-drct=2--3m.absnt ‘You wait for him.’ c. aya-na=us--k-as wait.for-drct=3m.absnt--obv-3n.absnt ‘He waits for it.’

[ʔa.ˈjaː.na.ʔus]

[ʔa.ˈja.na.nus]

[ʔa.ja.ˈna.ʔus.kas]

First person and second-person singular, being highest in the referential hierarchy, can only be encoded in the internal position. When the two persons interact, only the first person is encoded, while the second person is either understood from the context or expressed by a free pronoun, as in (25). (25) (ulkwat) aya:-na=Ø 2sg.pro wait-drct=1sg ‘I wait for you.’ Table 17.8 sums up the properties of prxm and obv.

5.2 Transitive and intransitive predicates Transitive verbs are always marked as either direct or inverse. These markers indicate which of the two arguments, prxm or obv, is the actor (i.e., agent, causer, experiencer, etc.) and which one is the undergoer (i.e., patient, goal, theme, causee, stimulus, etc.). In this way, the combination of the morphological marking on the verb and the syntactic position of the nominal constituents (Section 5.1) unambiguously indicates the semantic roles of the arguments.

828

Katharina Haude

A direct and an inverse transitive clause, respectively, are contrasted in (26a) and (26b). In both, prxm is expressed by a pronoun and obv by a referential phrase, which is the most common situation (see Haude 2014b). (26) a. jomni=as os ke:so devour=3n.absnt art.n.pst cheese ‘It (the fox) devoured the cheese.’ b. jommi-kay-a=’ne is ka:wup devour-inv-ep.v=3f art.pl mosquito ‘She will be devoured by mosquitos.’ Intransitive predicates can take only one argument (s), which shares the properties of obv, listed in Table 17.8. An example of an intransitive verbal predicate with the argument expressed by a bound pronoun is jo’yaj--us ‘he arrived’ in (9), which also displays the resyllabification of the external enclitic with the host-final consonant. Unpossessed common nouns can function directly as predicates of equational clauses. They behave like intransitive verbal predicates whose argument can, for instance, be expressed by an externally cliticized pronoun, as in (27). (27) rulrul--as jaguar--3n.absnt ‘It is/was the/a jaguar.’ Possessed nouns are rare as main-clause predicates (see Section 9.3); when they function as such, their argument can only be expressed by a referential phrase, as in (28) but not by a bound pronoun, demonstrated in (29). (28) jayna pekato=sne os jeya=sne dsc sin=3f.absnt art.n.pst state.of=3f.absnt ‘Her sin was that she was like that.’ (Lit. ‘Her state was her sin.’) (29) *pa:ko=us--k-as dog=3m.absnt--obv-3n.absnt (Intended: ‘It is his dog.’) Existential, locative, and possessive clauses are formed with a demonstrative pronoun in predicate function, as in (30). (30) kuro’ kus majni=Ø n-as Santakurus dem.m.absnt art.m.absnt offspring=1sg obl-art.n Santa.Cruz ‘I have a son in Santa Cruz.’ (Lit. ‘There is my son in Santa Cruz.’)

17 Movima

829

As will become apparent in Section 9, obv and s share not only formal but also behavioral characteristics, since only obv or s can be relativized. In contrast, there is no construction to which prxm has exclusive access (see Haude 2010a, 2019a). Therefore, if anything is to be analyzed as the “subject” of a transitive clause in Movima, this would be obv. Since the encoding of arguments as either prxm or obv depends on the relative position of their referents in a referential hierarchy, the pattern was analyzed as “hierarchical alignment” in Haude (2009a): the argument with the lower-ranking referent is aligned with s. Note that when semantic roles (a for the most agent-like and o for the most patient-like transitive argument) are taken into account, the direct/inverse contrast results in a split-alignment pattern: the direct construction displays ergative (s=o), and the inverse construction displays accusative alignment (a=s). In addition to the core arguments, a clause can take an unlimited number of oblique-marked constituents, here subsumed under the term “adjuncts”. These include all kinds of circumstantial expressions, like the locative adverbial in (30) or the time adverbial in (31). Example (31) also shows that event participants that are not included in the syntactic argument structure of the verb are also encoded as adjuncts, as is, for example, the case with verbs denoting three-participant events. (31) n-os ima:yoj kayɬe=us--k-as obl-art.n.pst morning give=3m.absnt--obv-3n.absnt n-os charke obl-art.n.pst dried.meat ‘In the morning he gave it (= the dog) (the) dried meat.’

6 Voice, valence, and transitivity There is a fundamental distinction between semantic and syntactic valence in Movima, here termed “valence” and “transitivity”, respectively. Semantically, verbs can be mono-, bi-, or polyvalent, that is, they can denote events involving one, two, or more participants. Whatever their valence, however, all verbs are basically intransitive, and transitive verbs (i.e., verbs taking two syntactic core arguments) must be derived through direct or inverse marking. Ditransitive main-clause predicates do not exist; most verbs denoting three-participant events, for example ‘give’, encode the recipient as their non-actor core argument (see (31), also Haude 2012a). A basic outline of the verbal system is given in Section 6.1. Valence-increasing morphology, that is, morphology deriving verbs that denote events with more than one participant, is presented in Section 6.2. Transitivity-decreasing mechanisms, applicable only to transitive verbs, are described in Section 6.3.

830

Katharina Haude

6.1 Verb classes and voice morphemes Different classes of verbal bases can be distinguished on the basis of their combinatorial possibilities with the voice/valence affixes. The most basic distinction is that between bivalent and monovalent bases. Bivalent bases participate in the direct/ inverse alternation. They either belong to a closed class of about 150 semantically bivalent roots, such as vel- ‘watch’, yey- ‘want’, tikoy- ‘kill’, and sal- ‘look for’, or they contain valence-increasing morphology (Section 6.2). Bivalent bases can be productively combined with the voice morphemes from Table 17.9, which determine the argument structure of a verb and indicate the semantic role of its s/obv argument, as well as, in the case of the direct/inverse markers, that of prxm. Only the direct and inverse markers can derive a transitive verb; they add a prxm argument position, which encodes the actor, in the case of direct, and the undergoer, in the case of inverse marking. Example (32) illustrates different voice affixes on the bivalent root ɬek- ‘kick’. Recall that the transitive voices, direct and inverse, obligatorily take an internal enclitic, and that the absence of an overt form indexes first-person singular. Intransitive verbs are translated into English as non-finite forms. (32) a. b. c. d. e.

ɬek-na=Ø ɬek-kay=Ø ɬek-cheɬ ɬek-e:ɬe ɬek-’i

‘I kick you/him/her/it/them.’ ‘You/he/she/it/they kick(s) me.’ ‘kick oneself/each other’ ‘kick at something/someone; kick around’ ‘be kicked’

However, most bivalent bases cannot be combined with the full range of voice markers; for instance, the middle reduplication is unattested with ɬek- ‘kick’, but it does occur with the root kel- ‘open’: kel~kel ‘open by itself’. The root kel-, in turn, is not found with the agentive suffix -eɬe. The only alternations in which all bivalent bases participate are those between direct, inverse, reflexive/reciprocal, and resultative.

Tab. 17.9: Voice affixes and the role of s/obv. affix

function

s/obv is:

-a-/-na -kay -cheɬ -eɬe

-‘i

direct (drct) inverse (inv) reflexive/reciprocal (refl/recp) agentive (act) middle (mid) resultative (res)

undergoer of transitive verb actor of transitive verb actor + undergoer actor potentially affected actor undergoer of state resulting from externally induced event

17 Movima

831

The resultative voice can be considered the least-marked form of a bivalent base (see Haude 2012b). The resultative ending -’i only occurs on simple bivalent verb roots, which cannot form a prosodic word on their own, like ɬek- in (32). Therefore, this ending may be a phonological dummy: a phoneme /i/ is also found in other prosodically defective environments, for example, on the bound third-person feminine pronouns (Table 17.5), or the bound first-person pronouns (i)ɬ, (i)y’ɬi, (i)y’bi (Table 17.6). Morphologically complex bivalent bases, in contrast, can be unmarked for voice. In that case, they denote a state, usually a resultative one, as in (33). (33) jayna ɬok-poj kis ko’o dsc fall-caus art.pl.absnt tree ‘The trees have already been felled.’ Monovalent bases do not participate in the direct/inverse alternation. The diagnostic for identifying them is that the suffixation of -na, which marks direct voice on bivalent bases, derives a possessed locative noun, as illustrated in (34). (34) a. joy-cheɬ go-refl/recp ‘go’

→

joy-na=Ø go-loc.nmlz=1sg ‘where I go’ (Lit. ‘my go-place’)

b. kay~kay mid~eat ‘eat’

→

kay-na=Ø eat-loc.nmlz=1sg ‘where I eat’ (Lit. ‘my eat-place’)

c. jo’yaj arrive ‘arrive’

→

joyaj-na=Ø arrive-loc.nmlz=1sg ‘where I arrive’ (Lit. ‘my arrive-place’)

Some monovalent roots, like joy- in (34a) and kay- in (34b), take the reflexive or middle marker. This is lexically determined, and these roots are not freely combinable with other voice morphemes. Some monovalent verbs, like jo’yaj in (34c), are monomorphemic, while others belong to restricted classes sharing one particular morpheme, such as -kwa for “bodily processes” (e.g., choj-kwa ‘urinate’, achis-kwa ‘sneeze’, joro:-kwa ‘sleep’) or -a for “sensations” (e.g., bele:k-a ‘be happy’, jilo:k-a ‘feel cold’). Transitivity and the direct/inverse distinction are also reflected by imperative marking. The suffix -ki derives an intransitive imperative verb in (35); -ti derives the transitive direct imperative, in (36); -doj/-dok derives the transitive inverse imperative, shown in (37). As shown in the (b.) examples of (35) to (37), on both the transitive direct and the intransitive imperative the plural is formed with the secondperson plural ending -kweɬ (cf. Table 17.6). In the inverse imperative, surprisingly, a plural actor is encoded by an internally cliticized second-person plural pronoun of

832

Katharina Haude

the s/obv set (=y’bi), (37b); no explanation of this unusual pattern can be given here. A plural undergoer of the inverse imperative verb is formed with an internally cliticized first-person exclusive pronoun, shown in (37c), in line with the general argument encoding pattern of inverse verbs. The combination of actor and undergoer enclitics is not attested on imperatives. (35) a. joy-ki! go-imp.intr ‘Go!’

b. joy-ki-kweɬ! go-imp.intr-2pl ‘Go (pl)!’

(36) a. sal-ti! look.for-imp.drct ‘Look for him/her/it/them!’

b. sal-ti-kweɬ! look.for-imp.drct-3pl ‘Look (pl) for him/her/it/them!’

(37) a. aya-dok! wait-imp.inv ‘Wait for me!’

b. aya-doj-a=y’bi! wait-imp.inv-ep.v=2pl.s/obv ‘Wait (pl) for me!’

c. aya-doj-a=y’ɬi! wait-imp.inv-ep.v=1excl ‘Wait for us!’

6.2 Valence-increasing morphology As stated above, bivalent bases can be derived by overt morphemes such as causatives or applicatives. Recall that the resulting verbs are not necessarily transitive (e.g., (33)): they require direct or inverse marking to function as transitive predicates. The causative suffix -poj on a monovalent root indicates direct causation. Example (38) shows the transitivized, direct counterpart of (33), prxm (=is) representing the actor and obv (kis ko’o), encoded like s in (33), representing the undergoer. The benefactive suffix -kwa is illustrated in (39). In both examples, the direct marker, which marks the verb as transitive, is represented by the base-internal allomorph -a, since the roots are monosyllabic (Section 3.3). (38) ɬok-a-poj-a=is kis ko’o fall-drct-caus-ep.v=3pl.absnt art.pl.absnt tree ‘They felled the trees.’ (39) des-a-kwa=as is pa:ko jump-drct-ben=3n.absnt art.pl dog ‘It jumped for the dogs (in order to catch them).’

17 Movima

833

A very common applicative suffix is the ‘co-participant’ marker -ɬe, which adds an undergoer to the meaning of the verb. For instance, the monovalent verb root jiwa‘come’, combined with this suffix, becomes a bivalent base ‘bring’ ((40a), resultative without voice affixes), which can be transitivized through direct (40b) or inverse (40c) marking. (40) a. jiwa:-ɬe come-coprt ‘be brought’

b. jiwa-ɬe:-na=Ø come-coprt-drct=1sg ‘I brought you/him/her/it/them.’

c. jiwa-ɬe-kay=Ø come-coprt-inv=1sg ‘I was brought by you/him/her/it/them.’ The valence-increasing suffixes can also be added to direct-marked bases, from which they derive semantically trivalent verbs. On monosyllabic roots, the direct marker is then optionally replaced by a reduplicative CV-prefix (see Section 3.2, Table 17.2), while elsewhere, the direct marker -na or is retained. On a direct-marked base, the causative indicates indirect causation. The core arguments represent causer and causee, while the patient is optionally encoded as an adjunct (an additional suffix -as may be added to reverse the causee/patient roles). This is illustrated in (41). The causer is encoded as prxm (=Ø), the causee is obv, which is not overtly expressed here but implied by the direct marking on the verb; the patient is represented by an oblique-marked adjunct phrase. (41) sa~sal-poj-na=Ø ni-kis alwambet drct~look.for-caus-drct=1sg obl-art.pl.absnt paper ‘I’ll have (her) look for the papers.’ Example (42) features another valence-increasing morpheme, the malefactive -bij. Again, the suffix is attached to a direct-marked base. This time, the verb is marked as inverse. Consequently, the actor is obv (again unexpressed in this example) and the maleficiary is encoded as prxm (=y’ɬi); the patient is represented by an adjunct. (42) jayna jomni-bij-kay-a=y’ɬi n-is nono=y’ɬi dsc devour-mal-inv-ep.v=1excl obl-art.pl animal=1excl ‘(It) has already devoured our animals (to our detriment).’ A valency increase is also induced by a process called “modifying incorporation”. The most transparent situation is that where a part-of-whole term is incorporated, which raises the possessor to argument status; see the contrast between (43a) and (43b). The verb can then be transitivized through direct marking, in (43c), which

834

Katharina Haude

adds an actor to the event, or through inverse marking, illustrated in (43d), which swaps the A and O roles of the transitive verb. (43) a. ben-’i kis chorimpa=sne paint-res art.pl.absnt fingernail=3 f.absnt ‘Her fingernails are painted.’ b. ben-chorimpa--sne paint-fingernail--3f.absnt ‘She has painted fingernails.’ (Lit. ‘She is fingernail-painted.’) c. ben-a-chorimpa=Ø--sne paint-drct-fingernail=1sg--3 f.absnt ‘I paint her fingernails.’ (Lit. ‘I fingernail-paint her.’) d. ben-chorimpa:-kay=Ø--isne paint-fingernail-inv=1sg--3 f.absnt ‘She paints my fingernails.’ (Lit. ‘I get fingernail-painted by her.’)

6.3 Detransitivizing processes There are two processes that decrease the syntactic transitivity of a verb: argument incorporation, and the detransitivizing operation with kaw/kwey. Both processes operate on direct-marked verbs, with the effect that the verb becomes intransitive, with prxm promoted to s and obv demoted to an adjunct. Argument incorporation involves the insertion of a noun or a bound nominal element (Section 3.4) into a direct-marked verb. The bound element represents the undergoer, and the now intransitive verb has the actor as its s argument. The undergoer can optionally be additionally expressed as an adjunct. This is particularly common when, as in (44), the incorporated element has a broader meaning, so that the adjunct provides more precise information. Note the difference in cliticization of the bound pronoun (i)y’ɬi in the examples below: it is internally cliticized when representing prxm of the non-incorporating transitive verb, as in (44a), and externally cliticized when representing s of the incorporating, now intransitive verb, as in (44b). (44) a. vel-na=y’ɬi is wa:ka watch-drct=1excl art.pl cow ‘We looked after (the) cattle.’ b. vel-a:-poy--iy’ɬi n-is wa:ka watch-drct-animal--1excl obl-art.pl cow ‘We tended (the) cattle.’ (Lit. ‘We animal-watched at [the] cattle.’)

17 Movima

835

The detransitivizing operation with the particle kaw/kwey (speaker-dependent variation) is restricted to relative constructions, which can only target s/obv (Section 9). The particle is inserted before the predicate, illustrated in (45) with a headed relative clause. Through this operation, prxm becomes s, and obv is demoted to adjunct status. In spontaneous speech, this construction only occurs with directmarked verbs, where it functions as an antipassive; elicitation shows, however, that the operation is also grammatical with inverse verbs, where it functions as a passive. (45) kino’ kinos kwe:ya di’ kwey vel-na dem.f.absnt art.f.absnt woman rel detrns watch-drct n-isko obl-pro.3pl.absnt ‘There is a woman who looks after them.’ Detransitivization is obligatory in relative constructions describing scenarios with a human acting on an inanimate entity. When both participants are equally ranked, detransitivization alternates freely with the inverse construction, as in (68). The detransitivizing operation contrasts with argument incorporation in that the latter is not required by any syntactic condition. Argument incorporation modifies the meaning of the verb, indicating that the action is directed towards a particular type of undergoer; by contrast, the detransitivizing operation has a purely syntactic function.

7 Complement and adverbial clauses Complement and adverbial clauses have the form of a referential phrase containing a nominalized predicate, as illustrated by the temporal adverbial clauses in (46) to (48). Verbal predicates are nominalized with the suffix -wa, illustrated in (46); nominal predicates undergo infixing reduplication, shown in (47) (Sections 3.2 and 4.2). These nominalized forms are marked as possessed, that is, they are combined with an internal enclitic, and the absence of overt marking indexes the first-person singular. Embedded referential predicates, that is, demonstratives, in (48), or personal pronouns (Section 9.3), take the ending -niwa, probably a fossilized combination of the verbalizer -ni and the nominalizer -wa. Referential embedded predicates are unpossessed. (46) n-os joy-wa=Ø obl-art.n.pst go-evnt.nmlz=1sg ‘when I left’ (Lit. ‘at my past-leaving’)

836

Katharina Haude

(47) n-os tolkosya=Ø obl-art.n.pst girl=1sg ‘when I was a girl’ (Lit. ‘at my past-being a girl’) (48) n-as koro’-niwa kos laya:kwa obl-art.n dem.n.absnt-vbz:nmlz art.n.absnt drunk.person ‘when there is a drunk person’ (Lit. ‘at there being a drunk person’) Due to possessive marking, which is identical to prxm encoding, embedded intransitive verbal and nominal predicates show a different argument-encoding pattern than in main clauses (see Section 5.2): their s is encoded like prxm of a transitive predicate. Transitive nominalized predicates, in turn, retain their argument structure. In the complement clause in (49), both arguments of the transitive direct predicate are expressed as in the main clause, that is, the actor as prxm and the undergoer as obv. (49) bo yey-na=us choy rey kos reas want-drct=3m.absnt certainly epist art.n.absnt onaye-na-wa=us--k-isne know-drct-evnt.nmlz=3m.absnt--obl-3 f.absnt ‘Because he wanted to get to know her (Lit. ‘… his knowing her, earlier today’), of course.’ Voice marking is only partly retained on nominalized verbal predicates. Most monovalent voice suffixes are dropped; for instance, the reflexive-marked verb joy-cheɬ ‘go’ becomes joy-wa when nominalized. On verbal bases whose morphological structure would require that the direct voice be formed with -a (Section 3.3), for example, kay-a-poj ‘feed’, even the inverse marker -kay is dropped; see (50). Only through the presence of an overt argument expression, like --k-i’sne in (50), can such a nominalized form be identified as inverse. Otherwise, it could just as well be interpreted as a reflexive, middle, or resultative verb, as these forms are unmarked as well in nominalization. (50) bo os kay-poj-wa=y’ɬi--k-isne reas art.n.pst eat-caus-evnt.nmlz=1excl--obv-3 f.absnt ‘… so that she would feed us’ (Lit. ‘for our past-being fed by her’) Since complement and adverbial clauses are referential phrases denoting states and events, their determiner is always an article from the “non-human” set (see Table 17.4 above). In contrast to its function when referring to concrete entities, the absential form of the article here indicates an additional temporal category: a past

17 Movima

837

tense limited to the day of speaking, as in (49). In this way, embedded clauses show a three-way temporal distinction, illustrated in (51).3 (51) as / kos / os joy-wa=Ø art.n art.n.absnt art.n.pst go-evnt.nmlz=1sg ‘my leaving (now or later/earlier today/before today)’ Different types of complement and adverbial clauses are specified by particles (Section 4.3). For instance, a purposive clause is a complement clause introduced by bo, as in (50), which is a clause coordinator when preceding a main clause, as it is in (49).

8 Negation 8.1 Main-clause negation Negated main clauses consist of a copula ka ‘is not’, to which the determining element =s or =ɬ (for first person) is attached (Section 4.1), and an embedded predicate. The embedded predicate is nominalized in the same way as in complement and adverbial clauses (Section 7). Since there is no article, tense is not specified here. The examples in (52) and (53) show an intransitive clause and a direct transitive clause, respectively, with the affirmative version under (52a) and the negated version under (52b). Recall that argument encoding in affirmative intransitive clauses is optional. (52) a. (it) joy-cheɬ 1intr go-refl/recp ‘(I) went/left.’ b. ka=ɬ joy-wa=Ø neg=det.1sg go-evnt.nmlz=1sg ‘I didn’t go/leave.’ (Lit. ‘my going/leaving was not.’) (53) a. onaye-na=u kinos majniwa=u know-drct=3m art.f.absnt offspring.of=3m ‘He knew his daughter.’ b. ka=s onaye-na-wa=u kinos majniwa=u neg=det know-drct-evnt.nmlz=3m art.f.absnt offspring.of=3m ‘He didn’t know his daughter (Lit. ‘his knowing his daughter was not’).’

3 See Haude 2010b.

838

Katharina Haude

8.2 Other types of predicate negation Other types of negation work slightly differently. After the negators ka’ prohibitive, mo:(ka) ‘not yet’, and the subordinate clause negator loy, only intransitive predicates are nominalized, and they are not marked as possessed, as in (54). Transitive predicates retain their main-clause form, either direct (55) or inverse (56), without nominalization. (54) ka’ iwani:-wa--y’bi! proh speak-evnt.nmlz--2pl.obv/s ‘Don’t talk!’ (55) ka’ sal-na=nkweɬ! proh look.for-drct=2pl ‘Don’t (pl) look for it!’ (56) ka’ rey ij ela:-kay=Ø--iy’bi! proh epist 2.pl.pro abandon-inv=1sg--2pl.obv/s ‘Don’t (pl) leave me alone!’ Embedded clauses are negated with loy, as in (57). Here, the nominalized predicate retains the form it has in an affirmative embedded clause; that is, all predicates are nominalized and, where applicable, possessed. The particle loy also negates relative clauses (Section 9), whose predicate then follows the pattern described above. (57) kaw-yemes as loy joy-wa=y’ɬi much-day art.n intn go-evnt.nmlz=1excl ‘It’s been many days that we haven’t been going.’ (Lit. ‘Many days [is] our not-going.’)

8.3 Existential negation To negate the existence of the participant represented by s/obv, the predicate may be marked by the irrealis infix (Section 3.3). The construction is illustrated with an intransitive verb in (58), a transitive direct verb in (59), and a transitive inverse verb in (60). (58) ka=s tijka:rim neg=det work ‘There is no worker/nobody who works.’

17 Movima

839

(59) ka=s pawa-na=Ø neg=det hear-drct=1sg ‘I don’t hear anything.’ (60) ka=s vel-kay-a=sne neg=det watch-inv-ep.v=3 f.absnt ‘Nobody looks after her.’ On nominal predicates, irrealis marking negates the existence of the noun’s denotee or its particular location. This can be existential negation when the noun is unpossessed, as in (61), and negation of possession when the noun is marked as possessed, like in (62). In (63), the latter case is also illustrated in an embedded clause, negated with loy (Section 8.2). (61) ka=s juyeni neg=det person ‘There is nobody.’ (62) ka=s polata=y’ɬi neg=det money=1excl ‘We don’t have money.’ (63) n-as loy juyeni=a obl-art.n neg.sub person=3n ‘when there is nobody’ (Lit. ‘at it [e.g., the house] having nobody’) Alternatively, existential negation can be expressed with the main-clause negation pattern (Section 8.1) and an embedded demonstrative predicate (cf. examples (14) and (30)).4 (64)

ka=s koro’-niwa kos tijka:rim neg=det dem.absnt.n-vbz:nmlz art.n.absnt work ‘There is nobody who works.’

9 Relativization When an underived content word is preceded by a referential element or by a referential phrase, it is syntactically subordinated to it and functions as a semantic speci-

4 For the verb inside the referential phrase, see Section 9.2.

840

Katharina Haude

fier of the referent. Constructions of this type are described here as relative clauses (RCs), even though they can also be characterized as participles or nominal constructions. There are three types of RCs: externally headed RCs, preceded by a referential phrase and a relativizer (Section 9.1); light-headed RCs, preceded by a determiner (Section 9.2); and headless RCs, which occur after a pronominal predicate (Section 9.3). Relativization is restricted to the s/obv argument, which is, to use a transformational metaphor, “extracted” from its clause-final position and “moved” to a position preceding the predicate. The detransitivizing operation (Section 6.3) is necessary to promote prxm to s so that it can be relativized. A further property shared by all RC types is that they are negated with loy (Section 8.2).

9.1 Headed relative clauses Headed RCs follow the referential phrase they modify and are introduced by the particle di’. They can be restrictive and non-restrictive. Headed RCs are the most productive device for modification in Movima. There is no principled restriction on the placement of the modifying and the modified element: for instance, (65a) and (65b) are both attested, with apparently no difference in meaning. (65) a. is bi:jaw di’ mowi:maj art.pl old rel Movima ‘(the) old Movima’ (Lit. ‘old ones who are Movima’) b. is mowi:maj di’ bi:jaw art.pl Movima rel old ‘(the) old Movima’ (Lit. ‘Movimas who are old’) The main-clause argument that is modified by a relative clause is usually s/obv, which is more often expressed by a referential phrase, as in (66). However, prxm can be modified by a relative clause as well, as in (67). (66) joy-a-ɬe=us is we:ye di’ vel-na=us go-drct-coprt=3m.absnt art.pl ox rel watch-drct=3m.absnt ‘He left with the oxen that he was in charge of.’ (67) jomni=is jokme di’ sereram-mo kis devour=art.3pl bird rel wild-clf:bird art.pl.absnt ba~ba-kwa=is ɬok’im red~fruit-alien=art.pl sujo ‘The wild birds eat (the) fruits of the sujo (tree).’

17 Movima

841

The relativized argument must be the s/obv argument of the RC, that is, s of an intransitive predicate (see (65) and (67)); o of a transitive direct predicate (see (66)); or a of a transitive inverse predicate (see (68)). Inside the RC, the relativized element remains unexpressed. (68) kis senyo:ra di’ vel-kay-a=sne art.pl.absnt lady rel watch-inv-ep.v=3 f.absnt ‘(the) ladies who look after her’ In order to relativize the prxm argument, the detransitivizing operation must be applied; see (45) in Section 6.3 above. The negation of a headed RC with the particle loy is shown in (69) with an intransitive verb, which is nominalized but unpossessed (Section 8.2). (69) is juyeni di’ loy joy-wa n-as lo:los art.pl person rel neg.sub go-evnt.nmlz obl-art.n village ‘(the) people who do not go to the village’

9.2 Light-headed RCs: Referential phrases Any type of content word can occur inside a referential phrase, including verbs (see Haude 2019b). Example (70a) illustrates a referential phrase containing a direct verb, and (70b) shows a referential phrase with an inverse verb. (70) a. kis vel-na=’ne art.pl.absnt watch-drct=3f ‘the (ones/people) she looks after’ b. kis vel-kay-a=’ne art.pl.absnt watch-inv-ep.v=3f ‘the (ones/people) (who) look after her’ The content word of a referential phrase has the same syntactic properties as the predicate of a headed RC. The referent of the referential phrase is the event participant that would be encoded as the s/obv argument of the corresponding mainclause predicate. The verbs in (70a) and (70b) are thus similar to the predicates of the headed RCs in (66) and (68), respectively, except that they are not preceded by an overt relativizer and the head is not a full referential phrase. The determiner, which establishes the reference (Section 4.1), can be considered a “light head” (Citko 2004). As with externally headed RCs, detransitivization is necessary if the referent of the referential phrase is the actor of a direct-marked verb, as in (71). Negation is carried out with loy; see (72).

842

Katharina Haude

(71) is kaw vel-na n-is wa:ka art.pl detrns watch-drct obl-art.3pl cow ‘the (ones) (who) look after the cattle’ (72) is loy iyeni:-wa art.pl neg.sub move-evnt.nmlz ‘the (ones) (who) do not move’ There is no structural difference between a referential phrase containing a verb and a referential phrase containing a noun. Even detransitivization can occur in a referential phrase with a noun, with the effect that the referent is the possessor of the noun’s denotee. Compare the possessed noun in (73a) with the “detransitivized” noun in (73b). (73) a. is majniwa=sne art.pl offspring.of=3 f.absnt ‘her children’ b. kinos kwey majni (n-usko) art.f.absnt detrns offspring obl-pro.3m.absnt ‘(his) mother’ (Lit. ‘the/a [woman who has] [him as] child’) Given the many parallels between nouns and verbs, and in particular, the identical encoding of the possessor and prxm, the effect of the detransitivization of nouns makes sense: a verb (at least inside an referential phrase, but possibly also elsewhere) can be interpreted as characterizing an event participant rather than as denoting an event (see Haude 2009b). The referential phrase with a transitive verb in (70a), for instance, can be paraphrased as ‘her watched ones’, and the detransitivized referential phrase in (71) as something like ‘the ones who have the cows as watched ones’.

9.3 Headless RCs: Content words with fronted pronoun In the third construction that can be characterized in terms of relativization, the content word is preceded by a free pronoun (see Tables 17.5 and 17.7), as in (74). The pronoun is the predicate here (Haude 2018a). This becomes apparent in embedding, as in (75) (Section 7), where it is the pronoun that receives morphological marking, and not the content word. (74) asko yey-na=’ne pro.3n.absnt want-drct=3f ‘That was (what) she wanted.’

17 Movima

843

(75) n-os asko-niwa yey-na=’ne obl-art.n.pst pro.3n.absnt-vbz.inch:nmlz want-drct=3f ‘when that was (what) she wanted’ The content word in this construction behaves exactly like the predicate of a relative clause or like the content word of a referential phrase: to express the prxm argument with a clause-initial free pronoun, detransitivization must be used, illustrated in (76), and the negation of the subordinate predicate is carried out with loy, shown in (77). (76) usko kwey ona-ra:-na pro.3m.absnt detrns know-ec-drct ‘He is (the one/someone who) knows (it).’ (77) u’ko loy iwani:-wa pro.3m neg.sub speak-evnt.nmlz ‘He is (the one/someone who) doesn’t talk.’ As in the other two types of RCs, the content word can also be a noun, as in (78). This results in an equational clause, propositionally equivalent to an intransitive nominal clause (see (27)). The pronominal construction, furthermore, is the only construction in which a possessed noun can function as a lexical predicate with a pronominal argument expression, (79) (cf. (29)). (78) a’ko rulrul pro.3n jaguar ‘It is (the/a) jaguar.’ (79) asko pa:ko=us pro.3n.absnt dog=3m.absnt ‘It is his dog.’

10 Pragmatic effects of constituent order alternations 10.1 Predicate nominals with a verb in the argument referential phrase The unmarked constituent order in Movima is predicate initial, with a verb functioning as the predicate, as in (80a). An alternative, pragmatically marked, construction is created by placing the noun representing the s/obv argument in predicate position and the verb (with the prxm argument, if transitive) inside the s/obv referential phrase, as in (80b) (see also Section 9.2). Open questions are also formed this way:

844

Katharina Haude

the predicate is a question word and the argument referential phrase typically contains a verb, as in (81). (80) a. yey-na=us os to:mi want-drct=3m.absnt art.n.pst water ‘He wanted (the) water.’ b. to:mi os yey-na=us water art.n.pst want-drct=3m.absnt ‘Water (was) (what) he wanted.’ (81) ɬéɬa kos yey-na=n? what.is art.n.absnt want-drct=2 ‘What do you want?’ (Lit. ‘What is the [thing] you want?’)

10.2 Fronted free pronouns The construction with a clause-initial free s/obv pronoun described in Section 9.3 has a pragmatically marked status as well. It establishes as the clausal topic a referent that has been introduced immediately before and that is not the main protagonist of either the preceding or the subsequent discourse. The focus is on the content word, which is prosodically prominent. Syntactically, the construction is a cleft, but this does not correspond to its pragmatic function (Haude 2018b). Its information structure rather suggests that the free pronoun has a copula function, attributing the focus (rheme) to the content word. Also, prxm can be represented by a free pronoun to the left of the predicate. In contrast to the pronominal construction, the pronoun here has a disambiguating function, taking up the main protagonist of the previous discourse after another discourse participant has intervened (Haude 2012c), as in (82). Even though there is not necessarily a pause after the pronoun, this construction can be described as left-dislocation. Firstly, the argument is cross-referenced by the obligatory internal enclitic on the predicate; secondly, unlike the pronominal construction, negation follows the main-clause pattern, indicating that the pronoun is located outside the clause, illustrated in (83). (82) joy-cheɬ is chot komersyante di’ juyeni …, che usko go-refl/recp art.pl hab merchant rel person and pro.3m.absnt jayna, eney, ji:sa-na=us os nego:siyo n-is dsc filler make-drct=3m.absnt art.n.pst deal obl-art.pl juyeni person ‘Merchants (Lit. ‘merchant people’) used to come, and he then traded with the people (i.e., the merchants).’

17 Movima

845

(83) che usko ka=s iwani-wa=us and pro.3m.absnt neg=det speak-evnt.nmlz=3m.absnt ‘And he didn’t speak.’ (Lit. ‘And he, his speaking was not.’)

10.3 Fronted referential phrases A referential phrase can also precede the predicate. This may often be due to influence from Spanish, but in many cases the fronted referential phrase indicates a topic shift, signaling that the referent of the referential phrase will persist in the subsequent discourse. When the fronted referential phrase represents s/obv of the predicate it precedes, it may or may not be repeated by a bound pronoun after the predicate, illustrated in (84) and (85), respectively. (84) che kinos alwaj-a=us la’ tera:ni--sne and art.f.absnt spouse-ep.v=3m ant ill--3f.absnt ‘And his wife, she was ill.’ (85) uɬ alwaj=Ø jayna bi:jaw art.m:1 spouse=1sg dsc old ‘My husband is old already.’ In transitive clauses, either prxm or obv may be represented by a fronted referential phrase; the referent can then sometimes only be identified on the basis of the context, as in (86). Here, it is known from the context that the caciques and commissioners do not distribute the meat to others, but that they received it from the cowherds. Therefore, the two juxtaposed referential phrases represent obv and are cross-referenced by the externally cliticized pronoun --k-is. (86) isos kasi:ki, isos komisa:riyo art.pl.pst cacique art.pl.pst commissioner.sp kayɬe-na=is--k-is n-os give-drct=3pl.absnt--obv-3pl.absnt obl-art.n.pst wa:ka-to:da cow-piece ‘They distributed the meat to the caciques, to the commissioners.’ When the fronted referential phrase corresponds to prxm of a transitive predicate, it must be cross-referenced by the obligatory internal enclitic; see (87). If it corresponds to obv (which is more common), it can be repeated by an external enclitic as in (86), or not, as in (88) and (89). The fronted obv referential phrase then represents either p of a direct-marked verb, like in (86) and (88), or a of an inversemarked verb, as in (89).

846

Katharina Haude

Tab. 17.10: Movima constructions and their functions (arg=argument; pred=predicate; rc=relative clause; tr=transitive). construction

function

[verb]PRED [det+noun]ARG [noun]PRED [det+verb]ARG [pronoun]PRED [verb/noun] RC [pronouni] [verb=bound pronouni]PRED.TR [RP] [verb/noun]PRED

unmarked focus on event participant assertion about new discourse participant takes up a given discourse participant topic shift

(87) us itila:kwa tojeɬ yey-na=us isnos art.m man very want-drct=3m.absnt art.f.pst alwaj-a=us spouse-ep.v=3m.absnt ‘The man, he loved his wife very much.’ (88) che is dichi:ye jemak ja’ joy-a-ɬe=as and art.pl child too just go-drct-coprt=3n.absnt ‘And it (the jaguar) also just took (the) children.’ (Lit. ‘the children it also just took.’) (89) is pa:ko ena’ jayna kamay-kay-a=as art.pl dog dur.std dsc bark-inv-ep.v=3n.absnt ‘The dogs were already barking at it.’ (Lit. ‘the dogs, it was being barked at by [them].’) As with the pronominal left-dislocation in (83), negation is marked as on main clauses and does not include the initial referential phrase; see (90). (90) is ri:ko ka=s rey tokbaycho-wa=i art.pl rich.sp neg=det epist remember-evnt.nmlz=3pl ‘(The) rich don’t remember.’ (Lit. ‘the rich, their remembering is not.’) The different syntactic constructions and their pragmatic statuses can roughly be represented as in Table 17.10.

11 Summary Movima is a linguistic isolate with some typologically remarkable features. In the order of the present paper, one may start by pointing out that metricality, that is,

17 Movima

847

syllable length and weight, plays an important role in morphological processes, so that sometimes the metrical effects of a clitic are the only formal means for distinguishing grammatical relations. Lexical composition involving noun roots, classifierlike elements, or truncated parts of words is an important device for word formation. There is no verbal tense marking, but tense is indicated by different spatio-temporal categories of the referential elements, in particular, the article. Nouns, verbs, and adjectives are near-equivalent syntactically: they are all basically predicates, and reference is established through a determiner. There is neither agreement nor morphological case. The grammatical relations in a transitive clause are distinguished by constituent order (internal versus external to the predicate phrase) and by direct/ inverse marking on the verb. The internal argument position is occupied by the prxm argument, whose referent ranks higher in hierarchies of person, animacy, and topicality. The external argument position is occupied by the obv argument, whose referent ranks lower on these hierarchies. The obv argument has the same properties as the single argument of the intransitive clause, including exclusive access to relativization. Direct marking on a transitive predicate indicates that the prxm argument is the actor and the obv argument the undergoer, while inverse marking indicates the reverse. Given that obv shares properties with s, the direct/inverse alternation leads to an unusual split-alignment pattern, with direct clauses patterning ergatively and inverse clauses patterning accusatively. The syntactic structure of Movima is predicate-initial. Variation in constituent order serves to manipulate information structure. When a predicate occurs after a referential expression, it loses some of its mainclause properties. Therefore, Movima can be regarded as a typical predicate initial language.

12 Acknowledgements I am deeply indebted to the Movima speakers who taught me their language and provided the data that have formed the basis of my research. The detailed comments on previous versions of this paper by the editors and an anonymous reviewer are gratefully acknowledged. All remaining shortcomings are entirely my own responsibility. My research was financially supported by the NWO-Spinoza project at Radboud University Nijmegen (2001–2006), the DOBES initiative of the Volkswagen Foundation (2006–2013), the Deutsche Forschungsgemeinschaft through the EuroBABEL programme of the European Science Foundation (2009–2013), and my research lab Structure et Dynamique des Langues (CNRS UMR8202–INALCO–IRD).

13 References Citko, Barbara. 2004. On headed, headless, and light-headed relatives. Natural Language and Linguistic Theory 22. 95–126.

848

Katharina Haude

Clemens, Lauren Eby & Maria Polinsky. 2015. Verb-initial word orders (primarily in Mayan and Austronesian Languages). In Martin Everaert & Henk van Riemsdijk (eds.), The Wiley Blackwell Companion to Syntax, 2nd edn. Malden: John Wiley & Sons, Inc. Grinevald, Colette. 2002. Nominal Classification in Movima. In Mily Crevels, Simon van de Kerke, Sérgio Meira & Hein van der Voort (eds.), Current Studies on South American Languages, 216–239. Leiden: Research School of Asian, African, and Amerindian Studies (CNWS). Haude, Katharina. 2006. A Grammar of Movima. Nijmegen: Radboud University dissertation. Haude, Katharina. 2009a. Hierarchical alignment in Movima. International Journal of American Linguistics 75(4). 513-–532. Haude, Katharina. 2009b. Reference and predication in Movima. In Patience Epps & Alexandre Arkhipov (eds.), New Challenges in Typology: Transcending the Borders and Refining the Distinctions, 323–342. Berlin: Mouton de Gruyter. Haude, Katharina. 2010a. The intransitive basis of Movima clause structure. In Spike Gildea & Francesc Queixalós (eds.), Ergativity in Amazonia, 285–315. Amsterdam & Philadelphia: Benjamins. Haude, Katharina. 2010b. ‘She kisses her late husband’ = ‘she kissed her husband’: Nominal tense marking in Movima. In Michael Cysouw & Jan Wohlgemuth (eds.), Rara & Rarissima: Documenting the fringes of linguistic diversity, 95–116. Berlin & New York: Mouton de Gruyter. Haude, Katharina. 2011a. Argument encoding in Movima: The local domain. International Journal of American Linguistics 77(4). 559–571. Haude, Katharina. 2011b. Referring to states and events: subordination in Movima. In Rik van Gijn, Katharina Haude & Pieter Muysken (eds.), Subordination in South American Languages, 141–168. Amsterdam & Philadelphia: Benjamins. Haude, Katharina. 2012a. The expression of three-participant events in Movima. Linguistic Discovery 10(3). 80–96. Haude, Katharina. 2012b. Undergoer orientation in Movima. In Gilles Authier & Katharina Haude (eds.), Ergativity, Valency and Voice, 259–288. Berlin & New York: Mouton de Gruyter. Haude, Katharina. 2012c. Saillance inhérente et saillance discursive en movima. Faits de Langues 39. 169–180. Haude, Katharina. 2014a. Reduplication in Movima: A prosodic morphology approach. In Hein van der Voort & Gale Goodwin Gomez. Reduplication in South American Languages, 343–373. Amsterdam: Brill. Haude, Katharina. 2014b. Animacy and inverse voice in Movima: A corpus study. Anthropological Linguistics 56(3–4). 294–314. Haude, Katharina. 2018a. Nonverbal predication in Movima. In Simon Overall, Rosa Vallejos & Spike Gildea (eds.), Nonverbal Predication in South American Languages (Typological Studies in Language 122), 217–244. Amsterdam: Benjamins. Haude, Katharina. 2018b. A topic-marking cleft? Analyzing clause-initial pronouns in Movima. In Evangelia Adamou, Katharina Haude & Martine Vanhove (eds.), Information structure in lesser-described languages: Studies in prosody and syntax (Studies in Language Accompanying Series 199), 217–244. Amsterdam & Philadelphia: Benjamins. Haude, Katharina. 2019a. Grammatical relations in Movima: Alignment beyond semantic roles. In Alena Witzlack-Makarevich & Balthasar Bickel (eds). Argument selectors: New perspectives on grammatical relations (Typological Studies in Language 123), 213–256. Amsterdam & Philadelphia: Benjamins. Haude, Katharina. 2019b. From predication to reference in Movima. In Christine Bonnot, Outi Duvallon & Hélène de Penanros (eds.). Individuation et référence nominale à travers les langues, 53–77. Paris: Editions Lambert-Lucas. Judy, Judith E. 1965. Independent verbs in Movima. Anthropological Linguistics 7(7). 10–15.

17 Movima

849

Judy, Robert A. 1965. Pronoun introducers in Movima clauses. Anthropological Linguistics 7(7). 5–9. Judy, Robert A. & Judith E. Judy. 1962. Movima y castellano. Vocabularios Bolivianos 1. Cochabamba: Instituto Lingüístico de Verano. Judy, Robert A. & Judith E. Judy. 1967. Movima. In Esther Matteson (ed.), Bolivian Indian Grammars 2 (Summer Institute of Linguistics Publications in Linguistics and Related Fields 16), 353–408. Oklahoma: SIL Publications. Myhill, John. 1985. Pragmatic and categorial correlates of VS word order. Lingua 66. 177–200.

Lev Michael, Stephanie Farmer, Gregory Finley, Karina Sullón Acosta, Christine Beier, Alexandra Chanchari Icahuate †, Donalia Icahuate Baneo, and Melchor Sinti Saita

18 Muniche 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Introduction Classification, demographics, and sociolinguistic background Phonology Morphological profile, word-classes, and cross-categorial morphology Nouns and noun phrases Adjectives and numerals Adverbs Verbs and verb phrases Simple clauses Clause linking Information structure Conclusion Acknowledgments References

1 Introduction Muniche (muni1258) is an almost-extinct linguistic isolate that was formerly spoken in the eastern Andean foothills of the Huallaga River basin of central Peruvian Amazonia. There are no longer any fluent speakers of Muniche, but documentation of Muniche was carried out by Gibson (1996) and by the Muniche Language Documentation Project (MLDP; Michael et al. 2009a, 2009b, 2013) with individuals who retained significant, if partial, knowledge of their heritage language.1 The results of these two projects constitute the only substantial documentation and description of

1 The work of the MLDP was funded by an NSF DEL RAPID grant (BCS #0941205) to Lev Michael. MLDP team members included (with their affiliations at the time) Christine Beier (Cabeceras Aid Project), Stephanie Farmer (UC Berkeley), Greg Finley (UC Berkeley), Lev Michael (UC Berkeley), Karina Sullón Acosta (independent scholar), and Michael Roswell (Swarthmore College). Karina Sullón Acosta additionally carried out a pilot project to identify speakers and begin preliminary Lev Michael, University of California, Berkeley Stephanie Farmer, Independent Scholar Gregory Finley, Independent Scholar Karina Sullón Acosta, Universidad Nacional Mayor de San Marcos Christine Beier, University of California, Berkeley Alexandra Chanchari Icahuate †, Munichis, Loreto, Perú Melchor Sinti Saita, Munichis, Loreto, Perú https://doi.org/10.1515/9783110432732-005

852

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

18 Muniche

853

the language, apart from brief wordlists collected by Tessmann (1930), Goodall (1950), and Daggett (1976). Both Gibson’s work and the MLDP were carried out when Muniche was quite obsolescent, and each documentation effort involved a relatively small number of months of fieldwork, meaning that there is much about the language that is unknown and probably irretrievable at this point. Gibson worked with a single fairly fluent speaker, Victoria Huancho Icahuate, who is believed to have died in the 1990s, while the MLDP worked with Alejandrina Chanchari Icahuate (ACI), who learned Muniche as a child in the 1920s or 1930s but spoke the language only intermittently after her teen years; Donalia Icahuate Baneo, who spoke the language with her Muniche-dominant mother until she died in the early 2000s but never learned the language entirely fluently; and Melchor Sinti Saita, another semi-speaker, who learned some of the language from his parents. Fortunately, the work of Gibson and the MLDP were partially complementary in the topics that they focused on, and this chapter aims to synthesize the results of these two efforts, summarizing all that is known about this language.

2 Classification, demographics, and sociolinguistic background 2.1 Language name and classification Muniche is considered a linguistic isolate in all recent classifications (Campbell 1997, Kaufman 1990, Solís 2003, and Wise 1999), although earlier authors proposed a number of genetic relationships with other languages. Loukotka (1968: 154), for example, treated Muniche (which he calls Munichi/Balsapuertiño) as a member of a proposed “Munichi stock”, along with nine other putative languages.2 Loukotka noted, however, that Muniche is the only language of the stock for which any linguistic materials exist, which raises significant questions about the empirical status of the other nine named languages, and consequently, their supposed genetic relationship with Muniche (with the exception of “Otonavi”, which is a variety of Muniche, as discussed below). Beuchat & Rivet (1909: 618) classified Muniche as Kawapanan (Cahuapanan) but since they had no access to Muniche data, this classification is presumably based on the geographic proximity of Muniche to the languages of the Kawapanan family and on historical accounts of interactions be-

fieldwork funded by Cabeceras Aid Project. The majority of MLDP fieldwork was carried out by Stephanie Farmer and Greg Finley. 2 Tabaloso, Chasutino/Cascosoa, Huatama/Otonavi, Lama/Lamista, Suchichi/Suriche, Zapaso, Nindaso, and Nomona.

854

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

tween speakers of Muniche and Kawapanan languages (i.e., Shawi and Shiwilu). These interactions included cases of migrations of Shiwilu groups into the Paranapura River basin, where they intermarried with Muniches (Steward 1948: 607),3 as well as the separate formation of the large mixed Shawi-Muniche community of Balsapuerto. Some authors have suggested that Muniche is affiliated with Tukanoan (Voegelin & Voegelin 1977, Key 1979) or with Macro-Tukanoan (Greenberg 1987), but Gibson (1996: 18) found few, if any, compelling Muniche-Tukanoan cognates, casting significant doubt on a Tukanoan affiliation – let alone affiliations with more dubious groupings such as Macro-Tukanoan. We concur with Gibson’s judgment on the basis of our independent comparison of Muniche and Tukanoan lexical data. Separately, Gibson (1996: 18) observed that the person-marking paradigm of Muniche exhibits striking similarities to that reconstructed by Payne (1991) for Proto-Arawakan, and we have observed that there are a handful of suggestive similarities to Arawakan lexical items. However, the evaluation of the possible Arawakan affiliation of Muniche will require future systematic work. We feel, at this stage, that Muniche is best treated as an isolate. The Muniche people and their language have been known by a number of names. The name “Muniche” dates to the earliest colonial records (e.g., Figueroa [1661] 1986), with the variant “Munichi” appearing, presumably due to Quechua influence, in the 19th century (e.g., Anonymous 1851). The speakers with whom the MLDP worked referred to the language as “Muniche” and to its speakers as “Muniches”. In the colonial literature, Muniches were also called “Otonabes” (orthographic variants: Otonavis and Otonahuis), but this name did not survive into the modern period. Some authors have identified “Paranapura” as an alternate name for Muniche (e.g., Tovar 1961: 181), but this appears to be an error, perhaps owing to the fact that the Paranapura River basin was an important part of historical Muniche territory. Paranapura is, or was, a variety of Shiwilu.4

2.2 Location The Muniche speakers with whom both Gibson and the MLDP worked live, or lived, in the municipality of Munichis, which lies some ten kilometers from the major regional urban center of Yurimaguas (see Map). Colonial records indicate that in the early 1600s, Muniches lived in several settlements located along the line of hills

3 Steward (1948: 607–608) repeated the Kawapanan classification of Muniche, despite the availability of Tessman’s (1930) data at the time that he was writing. 4 Veigl ([1785] 2006: 110) wrote, “The Paranapura are another branch of the Xeberos [i.e., Shiwilus]. They speak the same language with some dialectal variations.” Hervás y Panduro (1784) drew a similar conclusion, listing “Cahuapano” and “Paranapuro” as subdivisions of the lengua matriz “Chayavita” (i.e., Shawi). Both authors thus effectively identified Paranapura/o as Kawapanan.

18 Muniche

855

that stretches southeast from the headwaters of the Paranapura River into the lowerlying areas of the Huallaga River basin (Veigl [1785] 2006: 110). This area lay close to the town of Moyobamba, an important early colonial center, where Muniches were known as “Otonabes” (Figueroa [1661] 1986: 200). Muniches almost certainly lived in the Cachiyacu River basin at this time, where a mixed Shawi-Muniche settlement, Balsapuerto, is identified later in the colonial period, and they probably also lived in the southwestern part of these hills, closer to the town of Lamas (Veigl [1785] 2006: 110), whose residents exploited Muniches for labor (Figueroa [1661] 1986: 204). Figueroa ([1661] 1986) indicates that in 1652, Jesuit missionaries began successful efforts to resettle the Muniche population to the lower reaches of the Paranapura River, putting them within easy reach of the larger mission settlements on the Huallaga River. According to interviews conducted by the MLDP with elderly Muniches, the Muniche population moved from the older site to the current site of Munichis in the 19th century. Tessmann (1930: 303) remarked that in the mid-1920s, he encountered a small group of Muniches who had relocated to the lower Itaya River, near the city of Iquitos, presumably due to the demographic turmoil of the Rubber Boom; and that shift to Spanish and Quechua was already quite advanced in this group.

2.3 Language shift, language contact, and dialectal diversity Interviews carried out with Muniche elders in 2008 and 2009 by the MLDP indicate that the last fully-fluent speakers of Muniche were born between 1915 and 1925 and that the language was moribund by the early 1930s. These interviews reveal that by the early 20th century, language shift was already taking place from Muniche to both Quechua and Spanish, with shift to Quechua dominating over shift to Spanish in the early part of the 20th century. By the 1960s, however, Quechua also became moribund in Munichis, and subsequent generations are monolingual in Spanish. As discussed in detail in Michael et al. (2013: 340–345), Gibson and the MLDP appear to have documented two somewhat different varieties of Muniche, with the variety of Victoria Huancho, the speaker with whom Michael Gibson worked, exhibiting a number of phonological mergers absent in the speech of the speakers with whom the MLDP worked, who preserved a larger number of contrasts. There also appear to be some morphological differences between the two groups of speakers, which we will address where relevant. Historical sources support the idea that there were at least two major Muniche subgroups, which may account for the different phonologies documented by Gibson and the MLDP. Hervás y Panduro (1784: 61) identified two dialects, Muchimo and Otonabe, while Maroni ([1738] 1988: 108) mentioned two formerly geographicallydistinct subgroups, the Muniches and the Otonavis. Similarly, Velasco ([1788–1789] 1981: 547) identified two subgroups: the Churitunas and the Otonavis. It is also

856

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

worth noting that, apart from the main Muniche settlement on the Paranapura, the mixed Muniche-Shawi settlement of Balsapuerto on the Río Cachiyacu may have been a source of contact-induced dialectal diversity, as may have been the influx of Shiwilu speakers into the Muniche population in the early 17th century mentioned by Steward (1948: 607). Contact with Andean and Kawapanan languages has left identifiable traces in both the phonology and the lexicon of Muniche. As discussed by Michael et al. (2013: 338–340), there is segmental evidence that a number of originally-Spanish loanwords present in Muniche were borrowed via a Kawapanan language, and that the presence of some infrequent segments in Muniche, in particular /l/ and /ʃ/, may be due to contact with Kawapanan languages, and perhaps Shiwilu in particular. Muniche also exhibits a number of segments that Michael et al. (2014: 55) identified as distinctively characteristic of the North-Central Andean phonological area which, in addition to those already mentioned, include /ʂ, ts, tʃ, ʈʂ/, suggesting that Andean languages have had a significant impact on Muniche phonology. Finally, it is important to consider the effects of language attrition on the data and its interpretation. None of the speakers with whom the MLDP worked were fully fluent, and all three of them preferred to work together as a group, to be able to count on the help of the others in filling in gaps in their memory. The most challenging aspect of working with these data is their internal variability, which at points makes it difficult to tease apart obsolescence-based variability (see, e.g., Campbell & Muntzel 1989) from actual structural differences which should be accounted for in this description. In this respect, it is important to note that the speech of ACI, who was at least 20 years older than the next oldest speaker, differed from that of the younger speakers in a number of telling ways. For example, while ACI was consistent in treating verbal-subject person markers as second-position clitics (see Section 8.1), the younger speakers sometimes left these person markers attached to verbs, even when not in sentence-initial position. This suggests an obsolescenceinduced re-analysis by these younger speakers who rendered Muniche person-marking clitics similar to Quechua verbal person-marking suffixes, with which all of our consultants were intimately familiar. When it is clear that the data from ACI differ systematically from those of the younger speakers, then, we have followed the analysis suggested by the former data in interpreting the data from the younger speakers, from which much of the variability in the data stems. In this description, we will indicate those points at which variability in the data presented challenges for interpretation and analysis.

3 Phonology The consonant and vowel phoneme inventories of Muniche as documented by the MLDP are given in Tables 18.1 & 18.2. The consonantal inventory is large for an

857

18 Muniche

Tab. 18.1: Muniche consonantal phonemic inventory.

stop affricaste fricative nasal liquid glide

bilab.

alv.

post-alv.

retro.

p

td ts s

tʃ ʃ

ʈʂ ʂ

pal.

vel.

glot.

c

kg

ʔ

ç

placeless

h

m

N〈n〉 l ɾ〈r〉

w

j〈y〉

(w)

Tab. 18.2: Muniche vocalic phonemic inventory.

high mid low

front

central

back

i e

ɨ

u

a

Amazonian language (see Michael et al. 2013),5 and it shows a number of traits typical of North-Central Andean languages, such as the post-alveolar fricative, the entirety of its affricate series, and the presence of two liquids (see Michael et al. (2014: 55) for further details). At the same time, the presence of the high central vowel and mid vowel are characteristic traits of Amazonian languages. The reader is referred to Michael et al. (2013) for evidence regarding the contrastiveness of the phonemes posited here. Note that in this section we employ IPA-based representations, but in subsequent sections, we employ the practical orthographic alternatives given in angle brackets in Table 18.1. Several allophonic alternations are exhibited by these segments. The underspecified nasal assimilates to the place provided by an immediately following stop or affricate; when it is unable to obtain a place of articulation in this way, it surfaces as [n] in onsets and as [ŋ] in codas, following general nasal markedness principles in onsets and codas (Rice 1996). This alternation is exemplified in (1a) and (1b) where, in the former case, the placeless nasal acquires its place of articulation from the following bilabial stop while in the latter, lacking an appropriate segment from which to acquire its place of articulation, the placeless segment surfaces as a velar nasal. Note that /m/ does not exhibit place of articulation assimilation, as evident in (2).

5 The inventory in Table 18.1 differs from that given in Michael et al. (2013) in that the latter dispenses with phonemic glides by deriving them from underlying vowels in response to syllable structure constraints. Here we opt for a more surface-transparent phonological inventory.

858 (1)

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

a. in-pa [iʔmpa] peach.palm-clf:mash ‘peach palm (Bactris gasipaes) fruit mash’6 b. in-sa [iʔŋsa] peach.palm-clf:fluid ‘peach palm (Bactris gasipaes) fruit drink’

(2)

hamte-ɲe [hamteɲe] child-1sg ‘my child’

The contrast between post-alveolar and retroflex affricates is lost before front and high central vowels, with only [tʃ] appearing before /i,e/ and only [ʈʂ] before /ɨ/. Finally, /h/ undergoes fortition to [x] when immediately preceding consonants.

3.1 Metrical structure and stress The maximal syllable in Muniche is C1C2VC3, with onsetless syllables permitted. All consonants are permitted in simple onsets, and most consonant combinations are permitted in complex onsets, with the following restrictions: (1) C1 cannot be voiced; (2) C2 cannot be a fricative; (3) neither C1 nor C2 can be a glottal stop; and (4) C1 and C2 cannot both be fricatives or both be affricates. In Michael et al. (2013: 331–338) it is argued that many consonant clusters in the language, such as those in tpɨ ‘pineapple’, tsma ‘mosquito’, and kta ‘horsefly’, arose diachronically from a process of atonic vowel reduction. Muniche exhibits a trochaic stress system that differs in its direction of footing depending on word class: verbs and adjectives exhibit left-to-right footing, while nouns, adverbs, and interjections exhibit right-to-left footing. The acoustic correlates of stress in Muniche include pitch and, in open syllables, length (see Michael (et al. 2013) for discussion and acoustic evidence). As evident in (3), nouns are iteratively footed by right-aligned disyllabic trochees, with primary stress being leftmost. As evident in (3d)–(3f), Muniche does not permit degenerate feet, except for monosyllabic words. It is also evident from forms like (3g) and (3i) that this stress system is not quantity-sensitive, since the basic trochaic pattern is maintained despite the presence of unbalanced feet with a heavy syllable in rightmost position.

6 All data are from the MLDP corpus unless explicitly identified as from Gibson’s corpus.

18 Muniche

(3)

a. b. c. d. e. f. g. h. i. j.

859

(ˈu.ki) ‘peanut’ (ˈtda.na) ‘White-lipped Peccary, Tayassu pecari’ (ˈtnan.stɨ) ‘woodpecker’ tsi(ˈtsu.tɨ) ‘Giant Cowbird, Molothrus oryzivorus’ ɨm(ˈtɨk.wɨ) ‘chapo, mashed plantain drink’ ʂap(ˈnah.ma) ‘Giant Armadillo, Priodontes maximus’ (ˈu.hum)(ˌcu.me) ‘long ago’ (ˈtʃa.tʃa)(ˌpɨ.ʔɨ) ‘flower’ (ˈdeʔ.cuh)(ˌtɨʔ.ma) ‘llica, hand net’ (ˌˈmɨh.mɨ)(ˌte.ɲe) ‘my wife’

The metrically-productive nature of this stress pattern is evident from comparing forms like (4a) and (4b), where morphological augmentation results in the predicted shift in the stress pattern. (4)

a. ip(ˈse.ɲe) ‘my load’ b. (ˈip.se)(ˌɲet.ma) ‘my loads’

Verbs and adjectives, in contrast, exhibit left-aligned disyllabic trochees, as can be seen in (5c)–(5d) and (5g)–(5h), where forms with an odd number of syllables exhibit initial stress; compare nominal forms with the same number of syllables, as in (3d)–(3f) which exhibit stress in the second syllable from the left. As is the case with nouns, primary stress is leftmost in verbs; degenerate feet are likewise not permitted, as evident in the forms with an odd number of syllables (cited above); and stress is quantity-insensitive, as evident in the unbalanced feet with closed syllables in rightmost foot position in (5c) and (5f). (5)

a. b. c. d. e. f. g. h. i. j.

(ˈtma.ɾa) ‘It is delicious.’ (ˈwec.pa) ‘Ask (for it) (imp).’ (ˈswɨ.nɨt)pa ‘Cut your hair (imp).’ (ˈtʃpɨ.tsa)ʔa ‘drunk’ (ˈtɨ.sta)(ˌmɨ.ɾa) ‘They tore it apart.’ (ˈsi.cat)(ˌpɨ.nɨ) ‘Buy (it) for me (imp).’ (ˈtʂɨ.na)(ˌmɨ.ɾa)nɨ ‘It pricked me.’ (ˈah.ne)(ˌda.tsa)ʔa ‘white’ (ˈmaj.na)(ˌsu.tɨ)(ˌnɨ.pɨ) ‘I will visit you.’ (ˈtu.ja)(ˌsiʔ.tʃu)(ˌma.na) ‘I want to farm.’

860

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

4 Morphological profile, word-classes, and cross-categorial morphology 4.1 Morphological profile and word classes Muniche is an exclusively suffixing language with relatively morphologically-complex verbs and comparatively little nominal morphology. Verbs exhibit causative suffixes; a range of TAM morphology, including a morphological desiderative (Mueller 2014); sentential mood markers; and aspect. Verbs also exhibit the incorporation of lexical bases (see Section 8.7), which probably derive historically from incorporated nouns. The language tends toward a head-marking profile, but unlike many head-marking languages, verbal arguments are not marked on the verb by affixes per se. Instead these are expressed by clitics, with subjects expressed by secondposition clausal clitics. Nouns exhibit number, possessive, and diminutive marking, as well as evidence of a nominal classifier system. The language lacks case markers, and possession relations are head-marked on the possessum in possessive phrases. However, non-core arguments are licensed by postpositional suffixes, a dependent-marking feature. Stative quality concepts are mainly expressed by stative verbs. The roots undergo deverbal adjectival derivation to function attributively, and in addition, the language exhibits a small number of underived basic adjectives. Adverbs are distinguished by not taking any affixal morphology and by their ability to appear both clause-initially and clause-finally.

4.2 Agreement Both verbal argument agreement and nominal possessor agreement are expressed by bound morphemes distinguishing person and number, as seen in Table 18.3. The segmental form of the verbal and nominal agreement morphemes is identical, and it is clear that the verbal agreement markers are enclitics (see Section 8.1). Thus, we also treat the possession markers as clitics. Muniche distinguishes singular and plural in first and second persons, but not in third person (which does not distinguish gender). The first-person inclusive person marker =wɨdɨ appears to be derived, at least historically, from the concatenation of the first-person plural exclusive and second-person plural person markers. The conditioning environments of the allomorphy exhibited by the first-person singular and third-person markers are a challenge to resolve, due to what appears to be considerable obsolescence-induced variability (Campbell & Muntzel 1989). The trends in the data indicate, however, that =na appears following /a/, while =ɲe appears following /e/ and /i/, with =nɨ appearing elsewhere. In accord with the preceding comments about obsolescence-induced variability, however, the reader will note some exceptions to these generalizations in the data in this chapter. The

18 Muniche

861

Tab. 18.3: Person markers. person

singular

plural

1 2 3

=nɨ ~ =ɲe ~ =na =pɨ =ra ~ =ça

=wɨ (excl); =dɨ =ra ~ =ça

=wɨdɨ (incl)

allomorphy exhibited by the third-person marker in the data collected by the MLDP appears intractable at this point, a conclusion also reached by Gibson (1996: 68– 69) for the dataset he studied. As discussed in Section 8.3, in addition to the person markers enumerated in Table 18.3, Muniche exhibits a second-person singular marker, =pa, found in imperative constructions, and a first-person plural inclusive marker =wa, found in hortative constructions. A minimally attested third-person marker, =ya, is also found in prohibitives.

5 Nouns and noun phrases 5.1 Pronouns and demonstratives Muniche free pronouns, given in Table 18.4, are formed from a shared pronominal base, aʔpaʔ, to which the person markers described in Section 8.1 attach.7 Note that there are no third-person pronouns. When not expressed by NPs, third-person arguments are expressed either as verbal or nominal agreement, or omitted entirely and recovered from context. Free pronouns are attested solely as verbal arguments, as in (6), and not as possessors in possessive constructions. (6)

niçta-çu-mɨ=nɨ aʔpaʔ-nɨ çahe-sa drink-desid-pfv=1sg pro-1sg hot-clf:liquid ‘I want to drink a hot beverage.’

An additive suffix -tsdu(ʔ) can intervene between the pronominal base and the person markers, as exemplified in (7). (7)

tʃaʂdi-c=pɨ aʔpa-tsduʔ-pɨ! laugh-irr=2sg pro-add-2sg ‘You also laugh!’

A single prenominal, seemingly proximal, demonstrative, ɨnhane ‘this’ was documented. 7 The predicted first-person inclusive pronoun aʔpaʔwɨdɨ is not attested in our data.

862

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

Tab. 18.4: Muniche pronouns. person

singular

plural

1 2

aʔpaʔnɨ aʔpaʔpɨ

aʔpaʔwɨ (excl) aʔpaʔdɨ

5.2 Possessive marking and possessive phrases The basic possessive construction in Muniche is a head-marking one in which an agreement marker, reflecting person and number features of the possessor, appears on the possessum, as in (8). This person marker may or may not co-occur with a coreferential pronoun or referential noun. (8)

tsapsi=sa=ra mɨhmɨʈʂɨ ʂupiu-ra hamte-ra wash-prog=3 woman clothing-3 child-3 ‘The woman is washing her child’s clothing.’

In general, the order in possessive phrases that contain an overt nominal possessor is possessum-possessor, as in (9). (9)

hu-sa=nɨ tʃapɨ=ra çua pull.out-prog=1sg seed=3 cotton ‘I am pulling out cotton seeds.’

This construction exhibits an alienability distinction, with inalienable nouns directly taking a possessive suffix, as in (10), while alienable nouns take an additional alienable possession suffix -te, as in (11). (10) ɨʔka-mɨ=ra suu-nɨ grow-pfv=3 beard-1sg ‘My beard grew.’ (11) tʃaʂta-mɨ=ra tʃatʃa-te-ɲe grow-pfv=3 chick-aln.poss=1sg ‘My chick grew.’ A second alienable possession strategy is attested for domestic animals, involving the classificatory noun mɨtɨ, to which the possessive suffix attaches instead of to the possessum, as in (12). This type of construction is quite common in Amazonian languages, where certain semantic classes of possessums, especially domestic animals, cannot be directly possessed, but must instead be possessed in this periphrastic manner, using a classificatory noun.

18 Muniche

863

(12) kɨa-mɨ=ra mɨtɨ=nɨ kaka die-pfv=3 domestic.animal=1sg chicken ‘My chicken died.’ Muniche exhibits a distinct strategy for expressing what appears to be exhaustive focus (see also Section 11) on the possessor, involving an element wɨhɨ, which bears a person marker indexing the possessor, as suggested in the free translations given in (13) and (14). (13) mara wɨhɨ=na tʃa-ʂu cop poss.foc=1sg tree-clf:stick ‘It’s my stick.’ (14) e-ce-pa, maya wɨhɨ=nɨ hiɲuʔu! leave-irr=2sg cop poss.foc=1sg dog.dim ‘Leave it alone, it’s my puppy!’

5.3 Other nominal morphology Muniche nominal morphology consists of nominal classifiers, plural marking, possessive marking, nominal diminution processes, and postpositional suffixes. In this section, we describe all but the postpositional suffixes, for which the reader is referred to Section 9.2.

5.3.1 Noun classes and classifiers Muniche does not exhibit gender or any other noun-class system. It does, however, like many Amazonian languages, exhibit bound noun classifiers which specify additional information about the form and mechanical properties of the nominal stems to which they attach, deriving new nominal stems. Consider, for example, the following set of Muniche forms: tʃaçu ‘plant stalk’, tʃahpɨ ‘tree root’, tʃapɨ ‘thick stick’, tʃaʂa ‘raft’, tʃaʂu ‘slender stick’, and tʃatui ‘leaf’, all formed from the bound nominal root tʃa ‘tree’, to which are suffixed classifiers that specify the shape or properties of the referent. Similar evidence for a different set of nominal classifiers is provided by the following forms: inpa ‘pijuayo mash’, insa ‘pijuayo beer’, instɨʔɨ ‘pijuayo fruit bunch’, based on the nominal root in ‘pijuayo (peach palm), Bactris gasipaes’. Due in part to the extreme obsolescence of the language, our knowledge of this classifier system has considerable gaps; in particular, it is not possible at this stage to fully determine the productivity and combinatorial possibilities of the original system. Similarly, the semantics of certain classifiers is not entirely clear. Nevertheless, it is

864

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

possible to identify quite a few classifiers, the majority of which are exemplified below. Examples of stems that seem to consist of nominal roots bearing the classifier are provided for each classifier, although we do not segment the stems, due to uncertainty regarding allomorphy and the synchronic decomposability of certain stems. -hpɨ clf:rope-like: imahpɨ ‘barbasco, Sp. Lonchocarpus’ (cf. ima, ‘fish’);8 tʃahpɨ ‘tree root’; wuhpɨ ‘rope’; pihpɨ ‘central fiber of manioc root’.9 -htɨʔma clf:net: tʃayahtɨʔma ‘jicra, net shoulder bag’; deʔcuhtɨʔma ‘llica, hand fishing net’; wɨtɨhtɨʔma ‘hammock’. -me clf:powder: apuhtume ‘dust’; hpume ‘fariña, manioc meal’ (cf. hpupa ‘manioc mash’); preme ‘ashes’. -na clf:fiber: maʂpɨna ‘chambira fiber’ (cf. maʂpɨ ‘chambira palm, Astrocaryum chambira’) -pa clf:goop: açcupa ‘foam, froth’; hpupa ‘manioc chicha mash’ (cf. hpume ‘manioc meal’); iʔmpa ‘pijuayo chicha mash’; kutpa ‘gruel’, steʔpa ‘pepper and manioc paste’ (cf. stede ‘pepper’); pihpaʔpa ‘mud’. -pɨ clf:tubular: gakipɨ ‘boa’; tʃapɨ ‘thick stick’; tsipɨ ‘sugarcane’; ʈʂaʂpɨ ‘worm’; tʃtɨpɨ ‘intestine’ (cf. tʃtɨ ‘belly’); uʔpɨ ‘tamshi, Heteropsis sp. vine’; wipɨ ‘snake’. -sa clf:liquid: iʔnsa ‘pijuayo chicha drink’; diʔtsasa ‘refresco, sweet drink’ (cf. diʔtsa, ‘sweet’): itesa ‘broth’; maknasa ‘sachapapa chicha drink’ (cf. makna ‘sachapapa, Dioscorea trifida’); tsisa ‘green corn chicha drink’ (cf. tsite ‘green corn’); tʃuhtesa ‘distilled alcohol’ (cf. tʃuhtɨ ‘smoke’, tʃuheka ‘burning brand’). -stɨʔɨ clf:fruit.bunch: maʃpɨnistɨʔɨ ‘chambira palm (Astrocaryum chambira) fruit’; instɨʔɨ ‘pijuayo palm (Bactis gasipaes) fruit’; htawnastɨʔɨ ‘aguaje palm (Mauritia flexuosa) fruit’; sninstɨʔɨ ‘ungurahui palm (Oenocarpus bataua) fruit’; wɨpɨstɨʔɨ ‘garlic’. -ʂu clf:stick: tʃaʂu ‘stick’; kaʂu ‘branch’; paʂu ‘manioc stalk’; snapʂu ‘cane stalk, Gynerium saggitatum’.10 -tʂɨ clf:people: çamyintʂɨ ‘adolescent male’; itʂɨ ‘man’; mɨhmɨtʂɨ ‘woman’; natwatʂɨ ‘young man’; tʃaʂmɨtʂɨ ‘young woman’; tʃaʂpɨtʂɨ ‘old man (affectionate term for spouse)’. -tʂa clf:platform-like: çawɨnʈʂa ‘seat’ (cf. çawɨ ‘sit’); tʃaʈʂa ‘raft’. From the above data, we can infer that Muniche classifiers have a nominalizing function,11 as evident in (15a) and (15b), where classifiers appear to derive a nominal form from a verbal root. 8 The roots of the this plant are long and rope-like and contain a poison used for fishing. 9 As evident by comparison with the other classifier-bearing form paʂu ‘manioc stalk’, this root exhibits a pa~pi alternation. 10 This classifier also seems to appear in the name of several plantain varieties, e.g., waʔpaʂu ‘bellaco plantain’. 11 This is a property of classifiers in a number of Amazonian language families, such as the Tukanoan family (see Vol. 3).

18 Muniche

865

(15) a. diʔtsasa ‘refresco, sweet drink’ (cf. diʔtsa, ‘be sweet’) b. çawɨnʈʂa ‘seat’ (cf. çawɨ ‘sit’)

5.3.2 Nominal number Muniche exhibits a general number versus plural number contrast (Corbett 2000), where unmarked general number forms may be contextually interpreted as either singular or plural, while overtly plural marked forms yield only a plural interpretation. The nominal plural, -tma, is exemplified in (16) and (34). (16) muʔtsi-sa=ra çamyintʂɨ-tma play-prog=3 child-pl ‘The children are playing.’

5.3.3 Diminution Muniche exhibits two distinct mechanisms for forming diminutives, which were each very frequent in the speech of the consultants working with the MLDP. The first simply involves the suffixation of -ɲɨʔɨ to the nominal base, as in (17). (17) a. puʔpi ‘butterfly’ → puʔpiɲɨʔɨ ‘little butterfly’ b. tʃampi ‘bat’ → tʃampiɲɨʔɨ ‘little bat’ c. ʂtutu ‘guinea pig’ → ʂtutuɲɨʔɨ ‘little guinea pig’ The second diminutive formation process involves palatalization of certain consonants, as exemplified in Table 18.5, together with the additional of a final “echo syllable”, which consists of a glottal stop onset and a copy of the preceding vowel. The

Tab. 18.5: Diminutive consonant mutation. mutation

base

k → tʃ n→ɲ t→c ʂ→ç d→j s→ç h→ç

kaka hinu tuʔhana tseʔʂu titipdisti smahmɨ smahmɨ

diminutive form ‘chicken’ ‘dog’ ‘jaguar’ ‘añuje’ ‘bare belly’ ‘carachama’ ‘carachama’

tʃatʃaʔa hiɲuʔu cuʔhaɲaʔa tseʔçuʔu titipyiçciʔi çmaçmɨʔɨ çmaçmɨʔɨ

‘chick’ ‘puppy’ ‘little jaguar’ ‘little añuje’ ‘little bare belly’ ‘little carachama’ ‘little carachama’

866

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

consonant mutations all involve palatalization of one kind or another near the original place of articulation of the obstruent, except for the buccalization of /h/ to /ç/. The diminutive consonant mutation processes are not entirely predictable, in the sense that some segments that are potential targets for mutation processes do not experience mutation. This is evident in Table 18.5 where, for example, /t/ palatalizes to /c/ in ‘tortoise’ but not in ‘bare belly’, and /d/ palatalizes to /j/ in ‘bare belly’ but not in ‘spider’.

5.4 Nominal suffix ordering The three classes of nominal suffixes occur in the order given in (18), as exemplified in (19). Note that the form ipseɲetma ‘my loads’ demonstrates that the possessive person marker precedes the plural suffix, while the form hnananɨ ‘to my house’ demonstrates that postpositional suffixes precede the possessive person marker. (18) N-posp-poss-pl (19) ipsi-sa=na ipse-ɲe-tma hna-na-nɨ carry-prog=1sg load-1sg-pl house-loc-1sg ‘I am carrying my loads to my house.’

5.5 Nominalization Three deverbal nominalizers are relatively well attested in the MLDP data: an agentive nominalizer -u, exemplified in (20); an instrumental nominalizer -cu, exemplified in (21); and a habitual nominalizer -hihe (possibly restricted to non-volitional verbs), exemplified in (22). Note that all attested stems derived with the instrumental nominalizer bear a classifier, suggesting that such stems may be required to bear classifiers. (20) htedi-u steal-nmlz:ag ‘thief’ (21) a. tmidi-cu-ʂu pole.boat-nmlz:ins-clf:slender.rigid ‘tangana, pole for poling boat’ (22) tdɨ-hihe fart-nmlz:hab ‘habitual farter’

b. swɨnɨ-cu-pe cut.hair-nmlz:ins-clf:forked ‘scissors’

18 Muniche

867

Evidence for an object nominalizer, -ʂma, is found in a single high frequency stem, niʂma ‘beverage, manioc beer’, which involves the root ni ‘drink’.

5.6 NP syntax Nominal modifiers generally precede their heads: adjectives, numerals, quantifiers, and demonstratives are all overwhelmingly prenominal, as exemplified in (23) and (24) (also see (26), (25), (29)); the few examples of postnominal modifiers are possibly due to influence from Spanish. (23) ɲa-mɨ=nɨ mɨtsa watu see-pfv=1sg large rat ‘I saw a large rat.’ (24) nu=ra miɲaʃa waipɨʔsa exist=3 a.little rallasca ‘There is a little rallasca (manioc beer with freshly grated manioc added).’

6 Adjectives and numerals Only a few dozen words denoting property concepts were documented by the MLDP, and all but a handful of these behave like verb roots, occupying clause-initial position and taking verbal morphology and verbal person marking when functioning predicatively, as in (25) and (27). These roots require derivation with a deverbal adjectivizer -tsa ~ -tsaʔa to function attributively, as shown in (26) and (28). As evident in (29), the deverbal adjectivizer can also derive stative nominal modifiers from non-stative verbs. (25) pah-ma=ra ipse-ɲe-tma be.a.lot-ipfv=3 load-1sg-pl ‘My loads are numerous.’ (26) ʂa-çu-mɨ=ra pah-tsaʔa maʔçu eat-desid-pfv=3 be.a.lot-adjr deer ‘S/he wants to eat a lot of deer (meat).’ (27) hiʔtsa=ra ni-ʂma bitter=3 drink-nmlz:obj ‘The drink is bitter.’

868

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

(28) ni-ma=na ni-ʂma hiʔtsa-tsaʔa drink-ipfv=1sg drink-nmlz:obj bitter-adjr ‘I drank a bitter drink.’ (29) tʃuʔmta-me=ɲe kɨa-tsaʔa maʔçu found-pfv=1sg die-adjr deer ‘I found a dead deer.’ Attested underived basic adjectives include mɨtsa ‘large’, as in (30), miɲaʃa ‘few’, possibly pɨtsɨ ‘other’, and the set of numerals. Muniche appears to be typical of Amazonian languages in exhibiting a limited number of native numerals, up to four – that is, wɨtsaʔa ‘one’, utspa ‘two’, utsmɨ ‘three’, and tsaʔha ‘four’ – and using borrowed numerals beyond this point. Numerals from five to ten are borrowed from Quechua, and from Spanish beyond ten. (30) ɲa-mɨ=nɨ mɨtsa watu see-pfv=1sg large rat ‘I saw a large rat.’

7 Adverbs Muniche adverbs are a class of non-inflecting free morphemes (though they may bear second-position subject clitics if they appear sentence-initially). The modest number of adverbs documented by the MLDP includes temporal adverbs, spatial locative adverbs, and a few others such as tʃumɨra ‘well’ and tihna ‘far’. Muniche temporal adverbs include ɨʔmtse ‘today, now’, wataʔa ‘tomorrow’, exemplified in (31), uʔhumeɲe ‘long time’, uʔhumcume ‘before’, and uhunata ‘later’. (31) ena-çu=nɨ ʂapumtɨʔha wataʔa make-desid=1sg juane tomorrow ‘I will make my juanes12 tomorrow.’ The MLDP documented four spatial locative adverbs, which we provide here with the original Spanish glosses given by the consultants: naneʔe ‘acá’, exemplified in (32); naneca ‘aquí’, raneʔe ‘allá’, exemplified in (33); raneca ‘allí’.13 We should be 12 Juanes are a traditional dish among Peruvian Amazonian mestizos, and now many Peruvian Amazonian peoples as well, consisting of a ball of seasoned rice steamed in leaves, typically with meat at its center. 13 The precise semantics of the locative adverbs is not entirely clear in the local variety of Spanish, and even for standard varieties Spanish the difference between the locative adverbs ending in /i/ and those ending in /a/ is a matter of scholarly debate. Gomez & Jungbluth (2015), to take one

18 Muniche

869

extremely cautious in interpreting the glosses literally, however. At this point, it is probably not possible to determine the precise spatial semantics of these demonstratives, other than the fact that the first two express a proximal relationship with respect to a relevant deictic center and the latter two, a distal one. (32) ma-me=pɨ aʔpaʔ-pɨ naneʔe! come-pfv=2sg pro-2sg here ‘Come here!’ (33) ca-t=pa raneʔe ɲaʔtʂu! put-irr=2sg there firewood ‘Put the firewood there!’

8 Verbs and verb phrases 8.1 Verbal person markers Muniche verbal agreement markers, enumerated in Section 4.2, are clitics, with subject markers being second-position enclitics and object markers being verbal enclitics. First addressing the second-position subject-agreement clitics, we often find these cliticized to the sentence-initial verb, since Muniche exhibits VSO basic constituent order (Section 9.1), as in (34), (35a), and (36a). If, however, some other element appears sentence-initially and before the verb, such as an adverb, negation morpheme, or interrogative word, this sentence-initial element will serve as the host for subject-agreement markers, as in (35b), (36b), and (37), respectively. (34) ʂusi-sa=ra ima-tma swim-prog=3 fish-pl ‘The fish are swimming.’ (35) a. de-tsa=wɨ converse-fut=1pl.excl ‘We will converse.’

b. wata-wɨ de-tsa tomorrow=1pl.excl converse-fut ‘We will converse tomorrow.’

(36) a. tsa-tsa=nɨ finish-fut=1sg ‘I will finish.’

b. ap=nɨ tsa-tsa neg=1sg finish-fut ‘I will not finish.’

example, suggest the following semantics acá ‘speaker proximal region’, aquí ‘speaker proximal point’, allá ‘speaker distal region’, allí ‘speaker distal point’.

870

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

(37) manca=pɨ wa-sa-mɨ? where=2sg suck-prog-int ‘Where are you sucking?’ Clauses with transitive verbs typically exhibit two verbal agreement markers (unless the object is third person, see below), and if the verb is clause-initial, both will cliticize to it, with the subject marker preceding the object marker, as in (38) and (39), and the second clause in (41). (38) ʂa-mɨ=ra=wɨ wipɨ bite-pfv=3=1pl.excl snake ‘The snake bit us.’ (39) ʈʂɨna-mɨ=ra=nɨ çude stick-pfv=3=1sg thorn ‘The thorn stuck me.’ There is no overt third-person object marker. As a result, in no case does a thirdperson object agreement marker appear following a subject agreement marker (except in imperatives, see Section 8.3). In cases where we would expect this configuration, the third-person agreement marker is instead omitted, as exemplified in (40). Note that a first-person object interpretation is not possible for the the first-person clitics in (40), since that interpretation would require a third-person subject enclitic to precede the first-person enclitic. (40) a. uhume=ɲe ʂa-mɨ tʃupi long.ago=1sg eat-pfv sajino ‘I ate sajino (Pecari tajacu) a long time ago.’ b. ɨta-sa=nɨ ɲimaʔa-nɨ wait.for-prog=1sg husband-1sg ‘I am waiting for my husband.’

c. tʃpɨna-me=ɲe hit-pfv=1sg ‘I hit him/her.’

If a non-verbal element appears in the initial position of clauses which exhibit two verbal argument agreement markers, the subject agreement marker appears in second position, cliticizing to the clause-initial element, while the object agreement marker remains on the verb, as exemplified in the first clause of (41). This behavior is consistent with object agreement markers being verbal enclitics and not secondposition clitics. Note that even in this type of construction, no overt third-person object marker appears. (41) ap=nɨ sica-tʂɨ=pɨ, pa-ma=nɨ=pɨ aʔpaʔ-nɨ neg=1sg sell-caus=2sg, give-ipfv=1sg=2sg pro-1sg ‘I will not sell (it) to you; I am giving (it) to you.’

18 Muniche

871

8.2 Tense, aspect, and pluractionality Muniche exhibits a basic aspectual contrast between perfective aspect, on the one hand, and progressive and imperfective aspects, on the other. Progressive aspect is restricted to positive polarity non-stative verbs, while the imperfective is restricted to stative verbs and negated non-stative verbs. Clauses with present and past temporal reference are unmarked for tense, with temporal reference evidently being inferred from aspect, while clauses with future temporal reference optionally exhibit overt future tense or irrealis marking. The progressive -sa and perfective -mɨ~-me are exemplified in (42a) and (42b) respectively. (42) a. ʈʂaʂti-sa=na laugh-prog=1sg ‘I am laughing.’

b. ʈʂaʂti-mɨ=ra laugh-pfv=3 ‘S/he laughed.’

The allomorphy of the perfective is difficult to untangle, given the advanced obsolescence of the language and the variability it entails, but the tendency evident in the data is that the -me allomorph is found when the root contains a palatal or postalveolar consonant (/c, ç, ɲ, tʃ/), while the -mɨ allomorph is found elsewhere.14 As indicated previously, the aspectual contrast for stative verbs is not between perfective and progressive but instead between perfective and imperfective, -ma. The imperfective is exemplified for stative verb roots in (43a) and (43b), and in (44) we see that if the imperfective appears on active verb stems it yields a habitual meaning, as we would expect of an imperfective. Note that this example also demonstrates that when the imperfective appears word-finally, it surfaces as -maʔa, with an echo vowel. (43) a. tʃuçu-ma=na be.happy-ipfv=1sg ‘I am happy.’

b. upap-ma=ra ima be.rotten-ipfv=3 fish ‘The fish is rotten.’

(44) aʔpaʔ-nɨ tʃaʂmɨʈʂɨ aɲi-maʔa pro-1sg young.woman walk-ipfv ‘I used to walk as a young woman.’

14 There are also a handful of cases where the perfective surfaces as -ma, and in these cases the vowel in the adjacent and/or following syllable is /a/. However, the perfective often appears as -mɨ in these same phonological environments, leading us to suspect that this variant is obsolescenceinduced.

872

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

The imperfective also appears in place of the progressive in negated clauses, as exemplified in (45a) and (45b), and in (97b).15 (45) a. ap=nɨ tuyasi-cu-ma=na neg=1sg weed-desid-ipfv=1sg ‘I don’t want to weed.’

b. ap=ça tru-maʔa Melchor neg=3 tire-ipfv prop.n ‘Melchor doesn’t tire.’

Finally, we observe that the perfective suffix does not co-occur with the interrogative suffix, whose two allomorphs me ~ mɨ overlap in form with those of the perfective. In the interlinear representations in this chapter, we segment words with both perfective and interrogative meanings (i.e., ones that would otherwise bear both the perfective and interrogative suffix), as exhibiting a null perfective and an overt interrogative suffix, as in (46b).16 (46) a. knɨdi-sa-me=pɨ? choke-prog-int=2sg ‘Are you choking?’

b. knɨdi-Ø=me-pɨ? choke-pfv-int=2sg Have you choked?

In addition to the imperfective and perfective grammatical aspect markers, Muniche exhibits a telic lexical aspect marker, -ta, that derives stems with telic lexical aspect from atelic roots. In all clear attestations of this suffix in our corpus, it co-occurs with perfective marking, and in the case of activity verbs, such as the roots ʂa ‘eat’ and ʈʂɨ ‘climb’, speakers’ translations indicate a sense of completion of the activity, as shown in (47) and (48), respectively. (47) ʂa-ta-mɨ=ra eat-tel-pfv=3 ‘S/he already ate.’ (48) ʈʂɨ-ta-mɨ=ra climb-tel-pfv=3 ‘S/he already climbed up.’ With what are probably stative verbs, such as stɨn ‘be asleep’ and tsmei ‘be angry’, a change-of-state meaning obtains from the addition of the telic suffix, as exemplified in (49) and (50), respectively.

15 Arguably the use of the imperfective in (45b) is also motivated by the habitual sense of the utterance. 16 At least two other analyses are possible that are equally valid given the limited data: (1) that such forms exhibit an overt perfective and null interrogative suffix; or (2), that the single overt suffix that appears is a portmanteau perfective interrogative suffix.

18 Muniche

873

(49) stɨn-ta-mɨ=ra be.asleep-tel-pfv=3 ‘S/he fell asleep.’ (50) tsmei-ta-mɨ=ra Gregorio be.angry-tel-pfv=3 prop.n ‘Gregorio got angry.’ Muniche appears to exhibit at least one pluractional, -na,17 evident in the contrast between (51a) and (51b). b. pna-na-me=ɲe insert-pluract-pfv=1sg ‘I inserted it several times.’

(51) a. pna-me=ɲe pitsua-na insert-pfv=1sg needle-1sg ‘I inserted my needle.’

Temporal reference appears, in general, to be achieved by inference based on aspectual marking or overt temporal adverbs, but Muniche does exhibit an overt future tense suffix, -tsa ~ -sa, as in (52). However, even in sentences with very clear future temporal reference, this affix is rare. Note that verbs bearing the future tense suffix do not bear aspectual marking. (52) mara wata=nɨ ni-tsa çuʔpwiʔi cop tomorrow=1sg gather-fut copal ‘Tomorrow I will gather copal.’18 Muniche exhibits a necessity modal suffix, where the precise nature of the modality – whether internal, external, or deontic in nature – is not clear from the extant data. Also unclear is whether the form of the suffix is -tu, as it is given in (53a)– (53c), or -tuʂɨ. If it is the former, then there is a special form of the second-person subject clitic attested in these cases; alternatively, (53a) and (53b) do not bear person clitics and instead exhibit a different modal suffix, perhaps distinguishing the type of necessity modal. (53) a. rara-tu=ʂɨ hurry-nec=2sg ‘You have to hurry.’

b. sa-tu=ʂɨ eat-nec=2sg ‘You have to eat.’

c. decu-tu=pɨ speak-nec=2sg ‘You have to speak.’

17 As a hapax legomenon, we cannot be sure whether this is a suffix -na or simply reduplication of the preceding phonological material. 18 Focus constructions such as this one are discussed in Section 11.

874

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

8.3 Reality status Muniche exhibits a suffix, -c(e) ~ -t(e), that is employed in positive polarity directives (imperatives and hortatives); in some instances of future temporal reference; in conditional constructions; and in certain “prospective” or future-oriented verbal complements. Given this range of functions, we treat it as an irrealis marker. The -c(e) allomorph appears following the front vowels /i/ and /e/, while -t(e) appears elsewhere; the principles governing the elision of the vowel in these suffixes remain unclear at this time. The use of the irrealis suffix in directive constructions is amply exemplified in Section 9.7, to which the reader is referred. Examples of the irrealis suffix appearing in utterances with future temporal reference include (54) and (55). Note that verbs bearing the irrealis suffix do not bear aspectual marking of any kind. (54) nu-t=ra pahtsa hiɲuʔu exist-irr=3 many dog ‘There will be many dogs.’ (55) ap=nɨ mera-te=pɨ neg=1sg forget-irr=2sg ‘I will not forget you.’ The irrealis suffix is employed in precautional utterances that express warnings of possibly imminent harm, as in (56) and (57), the latter being an utterance intended as a warning to someone eating a dry and hard-to-swallow food. (56) wi-pum=pɨ wɨhta, hi-c=pɨ touch-proh=2sg pot burn.self-irr=2sg ‘Don’t touch the pot, you will burn yourself.’ (57) knɨdi-c=pɨ choke-irr=2sg ‘You are going to choke.’ In addition to the above uses in simple sentences, we find the irrealis suffix in the protasis of conditional constructions and in the verbs of clausal complements of nuçta ‘wonder’ and açta ‘want’, as discussed in Section 10. One outstanding question is the fact that there appears to be functional overlap between the future temporal reference functions of the irrealis suffix in examples like (54) and (55), and the future temporal reference of the future tense suffix -tsa ~ -sa, exemplified in (52). While at this point it is probably impossible to determine the difference in their temporal reference functions, one can imagine that the two

18 Muniche

875

suffixes differ in, for example, the presupposed certainty of the realization of the future event.

8.4 Desiderative Muniche exhibits a desiderative suffix, -çu ~ -cu, exemplified in (58)–(61), which indicates that the subject of the verb bearing the suffix desires the outcome denoted by the clause. The default form of the desiderative is -çu, as in (58)–(59), with the allomorph -cu conditioned by a preceding high front vowel, as in (60) and (61). (58) ʂa-çu-me=ɲe iʔteʔsira maçu eat-desid-pfv=1sg soup deer ‘I want to eat deer soup.’ (59) çawɨ-çu=nɨ raneʔe sit-desid-1sg there ‘I want to sit there.’ (60) napsi-cu=nɨ muntapsa fish.with.barbasco-desid=1sg Yanayacu.Creek ‘I want to fish with barbasco in Yanayacu Creek.’ (61) ap=nɨ tuya-si-cu-ma=na neg=1sg weed-intr-desid-ipfv=1sg ‘I don’t want to weed.’ Note that the subject of the desiderative-bearing verb is necessarily both the one desiring that the action be realized and the one who would realize the action. To express a desire that someone else realize an action, it is necessary to use the complement-taking verb açta ‘want’, as in (115). Interestingly, the desiderative suffix yields different, non-desiderative, meanings when attached to verbs that assign non-volitional semantic roles to their subjects, or when the subjects of such verbs are intrinsically incapable of volition. In the former case, avertive or frustrative meanings arise, as in (62) and (63). In the latter case, the desiderative appears to yield a temporal imminence meaning (i.e., an inchoative meaning, in the aspectual sense of this term), as in (64). (62) wɨsɨ-çu-me=ɲe fall-desid-pfv=1sg ‘I almost fell.’

876

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

(63) ap=nɨ stɨna-çu-mɨ=nɨ uʔtsa neg=1sg sleep-desid-pfv=1sg night ‘I couldn’t sleep during the night.’19 (64) aʔi-çu-me=ra rain-desid-pfv=3 ‘It is about about to rain.’

8.5 Associated motion There are several inflected verb forms in the MLDP corpus that suggests that Muniche exhibits an associated motion suffix -tu, one example of which is given in (65). Evidence for an associated meaning for -tu also appears in Gibson’s field notes, as in (66). (65) decu-tu=pɨ! talk-am=2sg ‘Go and talk!’ (66) niʂta-tu=pɨ! drink-am=2sg ‘Come and drink!’ (Gibson n. d.: 2:15) Curiously, the associated motion meanings expressed in the free translations of examples in the MLDP and Gibson corpora, respectively, are the exact opposite of each other, in the sense that the former data attest ‘go and do’ meanings, while the latter attest ‘come and do’ ones. Whether this reflects that -tu is a general associated motion marker, or that the difference is due to dialectal differences, or language attrition, is difficult to say at this point. It is also noteworthy that all the relevant examples in both the MLDP and Gibson corpora are imperative utterances, and that one allomorph of the irrealis suffix, which is used to form the imperative, has a similar phonological shape (-cu, see Section 8.3). Nonetheless, Muniche is solidly in western Amazonia’s ‘associated motion belt’ (Guillaume 2016), making it very plausible, in conjunction with the free translations provided, that -tu is indeed an associated motion suffix.

19 Note that this sentence does not mean ‘I did not want to sleep last night.’

18 Muniche

877

8.6 Valency-changing morphology Muniche valency-changing morphology includes a valency-increasing causative suffix and two valency-deceasing suffixes: a passive and a reciprocal. In addition, it exhibits suffixes that explicitly indicate the valency of verb roots that tend to be cross-linguistically labile. The Muniche causative suffix has at least four allomorphs in the MLDP dataset. Two of these allomorphs appear to be conditioned by the transitivity of the verb stem to which they attach: -tʃa attaches to intransitive stems, as in (67); and -ʂti attaches to transitive ones, as in (68). (67) çawɨ-tʃa-t=pa wɨhtaʔa tʃusɨ=nɨ! sit-caus-irr=2sg pot cooking.fire=1sg ‘Put (Lit. ‘make sit’) the pot on my cooking fire!’ (68) ʂa-ʂti-sa=na kaka eat-caus-prog=1sg chicken ‘I am feeding the chicken.’ A third allomorph of the causative, -sta, is attested on only one verb in the MLDP corpus, given in (69), but is also exemplified in Gibson (1996: 45), attached to stɨ ‘sleep’. In both attested cases, it attaches to verbs denoting non-volitional actions, suggesting that it may be lexically conditioned by verbs with these semantics. (69) hnɨ-sta-mɨ=ra=nɨ itwɨhma forget-caus-pfv=3=1sg rooster ‘The rooster made me forget.’ (i.e., by crowing and driving the thought from the speaker’s mind) A fourth allomorph, -ta, is likewise attested with a single verb stem, given in (70), and it is not possible at present to characterize its distribution. (70) tuhu-ta-me=ɲe break.intr-caus-pfv=1sg ‘I broke it.’ Muniche exhibits two valency-decreasing morphemes: the passive -ʂi, exemplified in (71) and (97b), and the reciprocal -iki, exemplified in (72). The MLDP encountered no examples of constructions reintroducing the agent of the argument as an oblique argument (nor does Gibson (1996) provide any), suggesting that it may not be grammatical to do so.

878

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

(71) tʃi-ʂi-mɨ=ra pasta kill-pass-pfv=3 tapir ‘The tapir was shot.’ (72) tʃpɨn-iki-sa=ra kaka-tma fight-recp-prog=3 chicken-pl ‘The chickens are fighting with each other.’ (adapted from Gibson (1996: 48)) In addition to the valency-changing morphology described above, certain verbs that tend to be labile cross-linguistically (Haspelmath 1993) appear with explicit valencymarking morphology, with the intransitive member of the pair appearing with the suffix -si intr and its transitive counterpart appearing with -pɨ tr (see Gibson (1996: 46) for a discussion of similar data). (73) kɨ-si-sa=na dig-intr-prog=1sg ‘I am digging.’ (74) kɨ-pɨ-sa=na ahiʔhe dig-tr-prog=1sg hole ‘I am digging a hole.’

8.7 Noun incorporation Muniche exhibits an incorporation or incorporation-like process in which elements exhibiting nominal referential properties (or lexical bases; see next paragraph), affix to stative roots, as in (75a) through (75c). (75) a. tsdami-htɨ=na be.swollen-lb:foot=1sg ‘My foot is swollen.’

b. ɨʂɨ-stipi=ɲi be.skinny-lb:bone=1sg ‘I am skinny.’

c. umi-pdisti=nɨ be.bloated-lb:belly=1sg ‘My belly is bloated.’ Significantly, the apparently-incorporating elements are distinct from the nouns with the same meanings that appear in non-incorporation contexts, which suggests that the nominal elements that appear suffixed to stative roots should be considered lexical bases, rather than true incorporated nouns. For example, the non-lexical base counterparts of the forms in (75a) and (75b) are hta ‘foot’ and tʃaʂtipi ‘bone’,

18 Muniche

879

respectively, which are similar but not identical; while the form we obtained for ‘belly’ tʃtɨ bears very little similarity to its corresponding lexical base in (75c). This same generalization holds for the other three lexical bases we obtained: -piʔʂu ‘lb:shank’ (compare uʔʂu ‘shank’), -pɨʔsɨ ‘lb:hand’ (compare htaʔsɨ ‘hand’), and -tuʔsi ‘lb:old.clothes’ (compare ʂupiu ‘clothes’). As with several other processes in Muniche, the obsolescence of the language and its sparse empirical attestation leaves us with only a partial understanding of the phenomenon. These lexical bases also surface in compound nouns, as in (76) and (77), which often seem to have quite idiomatic meanings, as evident in these examples. (76) tʃua-pdisti catfish.sp-lb:belly ‘big belly’ (77) tʃampi-tuʔsi bat-lb:old.clothes ‘tattered clothes’

8.8 Verbalizers A modest number of examples provide evidence for a denominal verbalizer, -di, that derives a verb denoting a customary action involving the nominal base, as in (78a). The onset of the suffix appears to devoice when immediately following voiceless consonants, as in (78b). (78) a. hɲansu-di-sa=na drum-vbz-prog=1sg ‘I am beating a drum.’ b. ɲaʂ-tɨ-sa=na firewood-vbz-prog=1sg ‘I am making firewood (i.e., cutting and gathering it).’

9 Simple clauses In this section we describe the structure of simple clauses, first using declarative clauses to exemplify our descriptive generalizations and then turning to examine the structure of simple interrogative and directive clauses. We also address nonverbal predication clauses and clausal standard negation.

880

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

9.1 Basic constituent order and alignment Simple declarative clauses, in which both arguments are expressed with overt NPs, typically exhibit VSO order, as evident in (79) through (81). (79) tsapsiʔ-mɨ=ra mɨhmɨtʂɨ ʂupiu wash-pfv=3 woman clothes ‘The woman washed the clothes.’ (80) ɲa-mɨ=ra hamte=ɲe ihɲiʔi see-pfv=3 child=1sg ghost ‘My child saw the ghost.’ (81) casi-me=ɲe aʔpaʔ-nɨ tsaʔa plant-pfv=1sg pro-1sg corn ‘I planted corn.’ There is no dependent marking on verbal arguments indicating grammatical relations; instead, grammatical relations are inferred from verbal argument agreement clitics. Verbal argument agreement markers exhibit nominative-accusative alignment in their clausal positional properties: the agreement clitic for S and A arguments is a clausal second-position clitic, while the agreement clitic for O arguments is a verbal enclitic (see Section 8.1).

9.2 Oblique arguments Oblique arguments are licensed by either free prepositions or suffixal postpositions. Spatial relations and path information related to motion predicates are expressed by suffixal postpositions, of which six were documented by the MLDP – no doubt a fraction of those which the language exhibited when it was vital. These include a general locative -na, as in (82) and (83); the goal marker -ta, as in (84); -mna ‘on top of’, as in (85); -rati ‘underneath’, as in (86); naʔta ‘onto’, as in (87) and -namata ‘from’, as in (88). (82) kɨpɨ-sa=na smaʔsu-na dig-tr-prog=1sg sand-loc ‘I am digging in the sand.’ (83) ate-mɨ=ɲɨ ruʔmu-na go-pfv-1sg forest-loc ‘I went to the forest.’

18 Muniche

881

(84) tʃatwɨ pɨjna-t! let’s.go island-goal ‘Let’s go to the island!’ (85) ca-t=pa tʃapi uʔmɲu-mna put-irr=2sg basket rack-on.top.of ‘Put the basket on top of the rack.’ (86) a-sa=ra hinu ipma-rati=ɲɨ go-prog=3 dog bed-underneath=1sg ‘The dog is going under my bed.’ (87) wɨʔsu-me=ɲe pɨmɨ-naʔta fall-pfv=1sg ground-onto ‘I fell onto the ground.’ (88) wɨʔsu-me=ɲe ipma-namata fall-pfv=1sg bed-from ‘I fell from the bed.’ Instrumental and comitative obliques, in contrast, are licensed by free prepositions. The instrumental preposition mara is exemplified in (89) while the comitative preposition nura is illustrated in (90). Note that the instrumental preposition is homophonous with the third-person copula, discussed in Section 9.3, while the comitative one is homophonous with the third-person existential verb, discussed in Section 9.4. (89) tʂɨna-sa=na ima mara iʔʂua spear-prog=1sg fish ins fishing.spear ‘I am spearing fish with a fishing spear.’ (90) lansa-sa=ra Karina nura Melchor dance-prog=3 prop.n com prop.n ‘Karina is dancing with Melchor.’

9.3 Non-verbal predication Non-verbal predication is expressed in Muniche in one of two ways: with the copula ma; or via simple juxtaposition of the copular subject and nominal predicate. As exemplified in (91), the copula appears in clause-initial position, bearing the subject clitic, as one would expect of a verb, and it is used to express both identity predication, as in (91a) and (91c), and locative predication, as in (91b).

882

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

Note that subject may be expressed by a referential noun, as in (91a) and (91b), or solely by the subject clitic, as in (91c). Nominal predication expressed via juxtaposition is exemplified in (92). (91) a. ma=ya wɨcu-ça Lidia cop=3 name-3 Lidia ‘Her name is Lidia.’

b. ma-ra hna-na-ra Demetrio cop=3 house-loc-3 Demetrio ‘Demetrio is in his house.’

c. ma-ra uçi-nɨ cop=3 hair-1sg ‘It is my hair.’ All attested instances of inclusion predication are expressed via the juxtaposition strategy, as in (92). (92) a. aʔpaʔ-nɨ natwa-ʈʂɨ pro-1sg young.person-clf:person ‘I am a young person.’

b. ɲaʔʂu-nɨ kaʔʂu firewood-1sg branch ‘My firewood is a branch.’

Meanings that are expressed by adjectival non-verbal predication in many other languages are expressed by stative verbs in Muniche. If the subject of the predicate corresponds to an inalienable noun, it is typically expressed in our corpus as an incorporated lexical base, as already exemplified in Section 8.7, although this does not appear to be obligatory, given examples like (93). If the subject is alienable, however, it is not incorporated and is either realized as a free NP, as in (94a), or omitted entirely, especially if it is a speech act participant, as in (94b). (93) mɨwɨʂma=ra uçi=ɲe be.long=3 hair=1sg ‘My hair is long.’ (94) a. sua=ça kutʃi be.fat=3 pig ‘The pig is fat.’

b. sua=nɨ be.fat=1sg ‘I am fat.’

In a very small number of cases, adjectival predication is expressed with a derived form that is used for nominal modification and is thus not a verbal form. In this construction, the adjectival predicate immediately precedes the subject and does not bear verbal agreement, as in (95). It is very possible, however, that this construction is a Spanish or Quechua calque in the context of language obsolescence. (95) ahɨneda-tsaʔa tʃaʂmɨʈʂɨ be.white-adjr young.woman ‘The young woman is white.’

18 Muniche

883

9.4 Existentials Existential clauses are formed with the existential verb nu, which appears in verbal clause-initial position, as in (96a) and (96c), except when preceded by negation, as in (96b); consequently, it typically bears the subject person marker. The existential verb is attested with tense and aspect morphology, as in (96b), where it exhibits the imperfective suffix, and (96c), where it bears the irrealis. (96) a. nu=ra miɲaʃa exist=3 a.little ‘There is a little.’

b. ap=ra nu-ma tʃaʂpɨ-te=ɲe neg=3 exist-ipfv husband-aln.poss=1sg ‘I don’t have a husband.’ (Lit. ‘my husband does not exist.’)

c. nu-t=ra pahtsa hiɲuʔu exist-irr=3 many dog.dim ‘There will be many puppies.’

9.5 Negation Standard negation in Muniche is expressed by the preverbal particle ap which, by virtue of its position, typically serves as the host for verbal subject person markers. If the verb is intransitive, the person marker corresponding to its sole argument cliticizes to the negation particle, as in (97a) and the first clause of (97b). If the verb is transitive and has two associated person markers, the subject marker cliticizes to the negation particle, while the object marker appears on the verb, as in (98). (97) a. ap=nɨ pɨhna-ma neg=1sg know-ipfv ‘I don’t know.’

b. ap=ra hna-ʂi-ma’a hinu, puʔʂ-ma-ra neg=3 bathe-pass-ipfv dog smell-ipfv=3 ‘The dog hasn’t been bathed, and it smells.’

(98) ap=nɨ mera-te-pɨ neg=1sg forget-irr=2sg ‘I will not forget you.’ As discussed in Section 8.2, verbs in negated clauses obligatorily bear the imperfective suffix -ma and cannot bear either progressive -sa or perfective -me, meaning that the basic imperfective/perfective contrast found in positive polarity clauses is neutralized in negative polarity ones.

884

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

9.6 Interrogative clauses Muniche exhibits two basic interrogative constructions: a polar interrogative construction, formed with the interrogative suffix -me ~ mɨ ~ -m and rising intonation; and a content interrogative construction which, in addition, exhibits a fronted interrogative word. Polar interrogatives exhibit the same word order as declaratives, with the interrogative suffix appearing immediately following the progressive suffix, when present, and immediately preceding agreement morphology, as exemplified in (99) through (101). These examples also illustrate the three allomorphs of the interrogative suffix. (99)

tʃɲi-sa-me=pɨ ʂupiu-pɨ? mend-prog-int=2sg clothes-2sg ‘Are you mending your clothes?’

(100) u-sa-mɨ-pɨ? cry-prog-int=2sg ‘Are you crying?’ (101) tʃdi-sa-m=pɨ? deceive-prog-int=2sg ‘Are you lying?’ The allomorphy of the interrogative suffix is unpredictable; it may be that the allomorphy of the interrogative is due to obsolescence-driven leveling between the interrogative and the very similar perfective suffix (see next paragraph), or its conditioning factors may similarly have been lost due to obsolescence. Note that polar existential interrogative utterances, as in (102), and polar interrogatives involving certain discourse particles, as in (103), do not appear to bear the interrogative suffix. (102) nu-ra ʂmɨ? exist=3 food? ‘Is there food?’ (103) irawɨʔɨ, itʃaduʔu? sure tag ‘You’re sure, right?’ As indicated above, content interrogatives are formed with fronted, sentence-initial interrogative words, rising intonation, and an interrogative suffix, as in (104) and

18 Muniche

885

(105). The attested interrogative words include tiʔku ‘what’, maɲca ‘where’, and maʔne ‘which’, exemplified in (104), (105), and (106), respectively.20 (104) tiʔku ʂa-Ø-mɨ=pɨ utsɨ-na? what eat-pfv-int=2sg garden-loc ‘What do you eat in the garden?’21 (105) maɲca=pɨ wa-sa-mɨ? where=2sg suck-prog-int ‘Where are you sucking?’ Content interrogatives involving non-verbal predication appear to lack interrogative suffixes, as in (106) and (107). (106) maʔne hamte=pɨ? which child=2sg ‘Which is your son?’ (107) tiʔku=ra? what=3 ‘What is it?’ As discussed in Section 8.2, the interrogative suffix does not co-occur with the overt perfective suffix which is homophonous with the interrogative, although as exemplified there, the imperfective and interrogative do co-occur. In (108), we provide an example of the same distributional generalization holding for a content interrogative construction. (108) maɲca ma-Ø-me=pɨ? where come-pfv-int=2sg ‘Where did you come from?’

9.7 Directives Muniche exhibits two distinct directive constructions: a positive polarity general directive construction and a prohibitive construction. The positive polarity directive 20 Also attested is an interrogative word pe that appears to mean both ‘what’ (which Gibson (1996) also includes with this meaning) and ‘where’. The small number of examples of this interrogative word in the MLDP corpus are profoundly confusing, however, and short of additional fieldwork, it is unclear if it will be possible to understand its function. 21 Note that the subject enclitic does not appear in second position here, i.e., cliticized to the interrogative word, where we would otherwise expect it to be but rather to the verb. In the MLDP

886

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

consists minimally of a verb bearing the irrealis suffix, and it is used for both second-person directives (imperatives), as in (109) and (110), and first-person plural inclusive directives (hortatives), as in (111). The verb in a directive construction bears a second- or first-person subject agreement clitic depending on whether the directive is an imperative or hortative, respectively; the irrealis suffix supplants all other aspectual and modal morphology. If more than one argument is expressed by person clitics, the addressee of the directive is treated as the subject, as in (109) and (110), and the other arguments are treated as objects, as in (110) and (112c), just as one would expect from the corresponding declarative utterances. Recall from Section 8.3 that irrealis-marked verbs do not take aspectual morphology. (109) pna-te=pɨ panantu! stick.in-irr=2sg plantain ‘Stick the plantain in (the fire)!’ (110) tʃi-c=pa=ça pcuʔʂu! shoot-irr=2sg.drtv=3.drtv trompetero.sp ‘Shoot the trompetero (Psophia crepitans)!’ The subject and object clitics of directives differ in some respects from those found in non-directive utterances. The second-person subject clitic exhibits a special form, =pa 2sg.drtv,22 as in (110) and (112a) and (112c), while the first-person plural inclusive exhibits the special form =wa 1pl.incl.drtv, as in (111). Furthermore, while third-person (object) clitics are not permitted to follow subject clitics in nondirective utterances, they can in directives, and take the special forms =ça ~ =ya 3.drtv, as in (110) and (112c). (111) ni-ce=wa niʔʂma! drink-irr=1pl.incl.drtv drink ‘Let’s have a drink!’ (112) a. masna-t=pa! listen-irr=2sg.drtv ‘Listen!’

b. masna-t=pɨ=nɨ! listen-irr=2sg=1sg ‘Listen to me!’

c. masna-t=pa=ya itwɨhma! listen-irr=2sg.drtv=3.drtv rooster ‘Listen to the rooster!’ corpus, this is in fact typical of sentences in which tiʔku ‘what’ appears. Whether this is an obsolescence-induced phenomenon or not is unclear. 22 Imperatives bearing both =pa and =pɨ are attested; it is not clear if the appearance of =pɨ in these contexts is obsolescence-related or not.

18 Muniche

887

Prohibitive constructions are identical to imperative constructions, except that the prohibitive suffix -pum replaces the irrealis suffix, as in (113). (113) a. wi-pum=pa wɨhta! touch-proh=2sg.drtv pot ‘Don’t touch the pot!’

b. hna-pum=pa=ya idɨ! bathe-proh=2sg.drtv=3.drtv water ‘Don‘t pour water on it!’

10 Clause linking Our knowledge of Muniche clause-linking constructions is quite limited because of the rarity with which they were produced by the speakers with whom the MLDP and Gibson worked. With the available data, we describe what little we know at this time about complement clause constructions, purposive constructions, and conditional constructions. The MLDP corpus includes only two instances of complement clauses. In both cases, the verb in the complement clause bears the irrealis suffix and normal verbal agreement. Note that the two complement-taking verbs, nuçta ‘wonder’, in (114), and açta ‘want’, in (115), are verbs whose complements express possible futures, which may account for the appearance of the irrealis in these cases. (114) nuçta-sa=na ɲima=nɨ ninca-te=ra ima wonder-prog=1sg husband=1sg bring-irr=3 fish ‘I’m wondering whether my husband will bring fish.’ (115) açta-me=ɲe sica-t=pɨ=nɨ puʔu ʂnɨrɨ want-pfv=1sg sell-irr=2sg=1sg meat paca ‘I want you to sell me paca meat.’ The one clearly attested purposive construction in the MLDP corpus is one in which the subject of the purposive clause is coreferential with that of the matrix clause. In this construction, the purpose clause bears the same-subject purposive suffix -sɨta,23 as in (116) and (117), and does not bear subject agreement. Discussions with our Muniche consultants suggested that the different-subject causative construction was distinct, but it was not possible to determine exactly what it was. (116) a-cu-sa=nɨ napsɨ-sɨta go-desid-prog=1sg fish.with.barbasco-purp.ss ‘I want to go to fish with barbasco.’ 23 It is possible that -sɨta is in fact a concatenation of two suffixes, as disyllabic verbal suffixes are otherwise not attested. For example, the second syllable may in fact be the telic suffix -ta, discussed in Section 8.2.

888

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

(117) ama-mɨ=ɲe casi-sɨta paʂu go-pfv=1sg plant-purp.ss manioc ‘I went to plant manioc.’ Conditional constructions are characterized by a protasis clause in which the verb bears the irrealis suffix -te, as in (118) and (119). It is not clear that (118) exhibits a true apodosis clause, but (119) does seem to, and we find the verb in this clause is also marked with the irrealis. (118) açta-t=pa mɨɲaʃa ni-ce=pa aʔpaʔ-tsdu=pɨ want-irr=2sg a.little drink-irr=2sg pro-add=2sg ‘If you want to drink a little, take some also.’ (119) açta-te=ra çaʔmte=ɲɨ nucu-te=ra=ɲɨ Yurimawa-na want-irr=3 child=1sg follow-irr=3=1sg Yurimaguas-loc ‘If he wants, my child will follow me to Yurimaguas.’ (Gibson n. d.: 2:65)

11 Information structure Given the limited documentation of Muniche, we cannot say a great deal about information structure in the language. However, there is one attested construction that is very likely a focus construction, in which the copula mara (see Section 9.3) appears in clauses that do not involve non-verbal predication, and that Muniche consultants translated into Spanish with a pitch accent on the Spanish word corresponding to the Muniche word immediately following the copula, as in (52) and (120). In regional Spanish, this type of pitch accent pattern is characteristic of selective focus. Given this translation and the structural resemblance to a cleft, it seems likely that this is a Muniche selective focus construction. (120) mara wata=pɨ tuyasi-sa cop tomorrow=2sg weed-fut ‘Tomorrow you will weed.’ Note that mara does not bear the subject clitic of the lexical verb in these constructions, suggesting that the copula does not belong to the same clause as the remainder of the sentence, making this a bi-clausal structure, as is true of cleft constructions in many languages.

18 Muniche

889

12 Conclusion This chapter has sought to convey as much as is currently known about the phonology and grammar of Muniche. At the time that we are completing this chapter, there are only one or two remaining elderly semi-speakers known to us, meaning that the window of opportunity for learning more about this fascinating language is now extremely narrow. It can be hoped that more work will be carried out with these speakers soon, to the benefit of their descendants and all those concerned with the indigenous languages of South America. Another possibility for gaining further insights into Muniche is presented by Michael Gibson’s unpublished field notes, which contain data that are likely to yield further insights not reflected in either this chapter or in Gibson (1996).

13 Acknowledgements This chapter would not exist without the dedicated efforts of Alejandrina Chanchari Icahuate, Donalia Icahuate Baneo, and Melchor Sinti Saita in sharing their memories of the Muniche language and their generous and cheerful willingness to collaborate with the other members of the MLDP team to document all that we could of their community’s ancestral language. We also wish to thank Josh Homan, who first informed Lev Michael of the presence of living semi-speakers of Muniche in the town of Munichis, thus inspiring the work reported on here. Finally, we wish to thank Michael (Maik) Gibson, who shared copies of his field notes with us, from which we have drawn some examples in this chapter.

14 References Anonymous. 1851. Establecemiento de la Compañía de Jesus en la República del Ecuador en el Año 1851. Quito: Isidoro Miranda. Beuchat, Henri & Paul Rivet. 1909. La famille linguistique Cahuapana. Zeitschrift für Ethnologie 41: 616–634. Campbell, Lyle. 1997. American Indian languages: The historical linguistics of Native America. Oxford: Oxford University Press. Campbell, Lyle & Martha Muntzel. 1989. The structural consequences of language death. In Nancy C. Dorian (ed.), Investigating obsolescence: Studies in language contraction and death, 181–196. Cambridge: Cambridge University Press. Corbett, Greville. 2000. Number. Cambridge: Cambridge University Press. Daggett, J. 1976. Munichi word-list. Información de Campo No. 99 (microfiche). Lima: Instituto Lingüístico de Verano. Figueroa, Francisco de. 1986 [1661]. Informe de las Misiones de El Marañón, Gran Pará ó Río de las Amazonas. Iquitos: IIAP-CETA.

890

Michael, Farmer, Finley, Acosta, Beier, Icahuate, Baneo, and Saita

Gibson, Michael. 1996. El Munichi: Un idioma que se extingue. Pucallpa: Instituto Lingüístico de Verano. Gibson, Michael. no date. Munichi fieldnotes. Manuscript. Gómez, Elena and Konstanz Jungbluth. 2015. European Spanish. In Konstanz Jungbluth and Federica da Milano (eds.), Manual of Deixis in Romance Languages, 240–257. Berlin: De Gruyter. Goodall, Harold. 1950. Munichi word-list. Información Campo No. 99 (microfiche). Lima: Instituto Lingüístico de Verano. Greenberg, Joseph. 1987. Language in the Americas. Stanford: Stanford University Press. Guillaume, Antoine. 2016. Associated motion in South America: Typological and areal perspectives. Linguistic Typology 20(1): 81–177. Haspelmath, Martin. 1993. More on the typology of inchoative / causative alternations. In Bernard Comrie & Maria Polinsky (eds.), Causatives and transitivity, 87–120. Amsterdam: Benjamins. Hervás y Panduro, Lorenzo. 1784. Catalogo delle lingue conosciute e notizia della loro affinità, e diversità. Cesena. Kaufman, Terrence. 1990. Language history in South America: What we know and how to know more. In David L. Payne (ed.), Amazonian Linguistics, 1374. Austin: University of Texas Press. Key, Mary Ritchie. 1979. The grouping of South American Indian languages. Tübingen: Narr. Loukotka, Čestmír. 1968. Classification of South American Indian Languages. Los Angeles: University of California. Maroni, Pablo. 1988 [1738]. Noticias auténticas del famoso río Marañón. Monumenta Amazónica. Iquitos: CETA. Michael, Lev, Will Chang & Tammy Stark. 2014. Exploring Phonological Areality in the CircumAndean Region using a Naive Bayes Classifier. Language Dynamics and Change 4(1): 27–86. Michael, Lev, Christine Beier, Karina Sullón Acosta, Stephanie Farmer, Greg Finley & Michael Roswell. 2009a. Dekyunáwa: Un diccionario de nuestro idioma Muniche. Iquitos: Cabeceras Aid Project. Michael, Lev, Christine Beier, Karina Sullón Acosta, Stephanie Farmer, Gregory Finley & Michael Roswell. 2009b. Una breve descripción del idioma Muniche. Iquitos: Cabeceras Aid Project. Michael, Lev, Stephanie Farmer, Gregory Finley, Christine Beier & Karina Sullón Acosta. 2013. A sketch of Muniche segmental and prosodic phonology. International Journal of American Linguistics 79(3): 307–347. Mueller, Neele. 2014. Language internal and external factors in the development of the desiderative in South American indigenous languages. In Loretta O’Connor & Pieter Muysken (eds.), The native languages of South America: Origins, development, typology, 203–222. Cambridge: Cambridge University Press. Rice, Keren. 1996. Default variability: The coronal-velar relationship. Natural Language and Linguistic Theory 14: 493–543. Solís Fonseca, Gustavo. 2003. Lenguas en la Amazonía peruana. Lima: Programa FORTEPE, Ministerio de Educación. Steward, Julian. 1948. Handbook of South American Indians. vol. 3: The tropical forest tribes. Washington, D.C.: Government Printing Office. Tessmann, Günter. 1930. Die Indianer Nordost-Perus: Grundlegende Forschungen für eine Systematische Kulturkunde. Hamburg: Friederichsen, de Gruyter & Co. Tovar, Antonio. 1961. Catálogos de las Lenguas de América del Sur. Buenos Aires: Editorial Sudamericana. Veigl, Francisco Xavier. 2006 [1785]. Noticias detalladas sobre el estado de la provincia de Maynas en América Meridional hasta el año de 1768. Iquitos: IIAP-CETA. Velasco, Juan de. 1981 [1788–1789]. Historia del Reino de Quito en la América Meridional. Caracas: Biblioteca Ayacucho.

18 Muniche

891

Voegelin, Carl & Florence Voegelin. 1977. Classification and index of the world’s languages. Amsterdam: Elsevier. Wise, Mary Ruth. 1999. Small language families and isolates in Peru. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), The Amazonian languages, 307–340. Cambridge: Cambridge University Press.

Bernat Bardagil

19 Mỹky 1 2 3 4 5 6 7 8 9 10 11 12

Background and history External relationships Internal variation Phonology Morphosyntnactic typology The noun phrase Postpositions The verb phrase Clauses Conclusions Acknowledgements References

1 Background and history Mỹky (ISO: irn, Glottocode: iran1263) is a language isolate spoken in central Brazil, in the western part of the state of Mato Grosso. The Mỹky people used to live in small slash-and-burn agricultural communities in the upper Juruena basin, in the region between the Sangue and Papagaio rivers, with an estimated population of 1000 individuals at the beginning of the 20th century (Pivetta 1993). Not much is known of their history before 1900, when rubber tapper Domingos Antonio Pinto led an assault on a community near the Tapuru Creek. The attack resulted in a massacre of almost the entire village (Rondon 1946: 88) and was followed by a population split between today’s self-named Manoki (also known as Iranxe or Iranche) and Mỹky communities. Manoki and Mỹky elders speak of the two groups as already being differentiated before then. Between 1909 and 1932, small groups of Manoki approached the telegraph workers established near the Papagaio river in order to exchange goods. In the 1930s, Jesuit missionaries started to expand in the region, followed by the arrival of the Inland South American Missionary Union (ISAMU) and the Indian Protection Service (SPI in its Portuguese acronym). These three organizations established missions and contact posts and competed to attract the attention of the Manoki and other Indigenous groups in the region. In the 1940s the second rubber boom once more attracted rubber tappers to the area, which was again followed by violent clashes with the Indigenous populations. Contact with Brazilian society through

Bernat Bardagil, Ghent University https://doi.org/10.1515/9783110432732-006

894

Bernat Bardagil

19 My˜ky

895

these sources caused the spread of a series of epidemics among the Manoki that decimated them (Pivetta 1993). In 1957, ISAMU retreated from the area and SPI delegated contact with native communities to Jesuit missionaries, who became the main channel between the Manoki and Brazilian society. The Manoki community ended up settling in the Jesuit mission of Utiariti, where other Indigenous groups were also brought. At the mission, children lived in a boarding school separated from their families during the week and were physically punished if they spoke their language at all. Adults were also forbidden from not only speaking their language but also performing any rituals or traditional practices. After reaching a population low of 33 individuals in 1960, the Manoki started a demographic recovery in part due to intermarrying with members of Indigenous groups like the Paresi (Arawakan), Rikbaktsa (Macro-Jê), Kayabi (Tupian), and Nambikwara-speaking groups (Bueno 2014). Today, the Manoki live in a protected Indigenous land, the Terra Indígena Iranxe, spanning 206,455 hectares in the municipality of Brasnorte, in western Mato Grosso, and are fighting to complete the demarcation of the adjacent Terra Indígena Manoki, which extends between the Cravari and Sangue rivers. The group that had retreated away from the advancing frontier in 1900, after the Tapuru massacre, spent seven decades actively avoiding contact with outsiders. Constantly on the move, they fled the expansion of Brazilian settlers and ranchers and the attacks from loggers, rubber tappers, and other Indigenous groups, especially the Tapayuna (Jê) and the Rikbaktsa (Macro Jê). By the 1960s they had settled in their current location on the margins of the Papagaio river. Calling themselves mỹky (‘person’), they were contacted in 1971 by missionaries accompanied by Manoki men fluent in the language (Lisboa 1979). The Mỹky group was made up of 23 individuals: 12 men and 11 women. The Manoki and Mỹky celebrated the reunion, and they have maintained a close relationship ever since, solidified by marriages between members of the two groups. Today most of the Mỹky live in a demarcated Indigenous land comprising 47,094 hectares in the area where they were contacted, near the town of Brasnorte. Faced with severe land invasion from neighboring settlers, they are fighting to officially demarcate a larger part of their territory adjacent to the Indigenous land. The 2014 CIMI (Indigenous Missionary Council) census counted 369 Manoki and 128 Mỹky. As a result of schooling, life in the Jesuit mission and extensive intermarriage with members of other Indigenous groups, language loss is severe among the Manoki. Anonby (2009) states that in 2003 most Manoki under 50 were monolingual in Portuguese. This amounted to four speakers of Manoki as a dominant language in 2022, as attested by the author of this chapter. The language is much more vital among the Mỹky, a majority of whom are native speakers, although proficient knowledge of Portuguese is not uncommon. Both communities maintain frequent contacts and consider each other as part of the same people. The Mỹky language was first described in a brief grammar sketch written by Meader (1967), based on the Manoki variety, preceded only by a vocabulary list and

896

Bernat Bardagil

some grammatical notes made by Moura (1957). The language in its Mỹky variety has been the object of study of a doctoral dissertation (Monserrat 2000), later published as a monograph (Monserrat 2010). Both varieties, with a focus on Manoki, have recently been the object of a documentation and description project whose materials are hosted at the Endangered Languages Archive (Bardagil 2019). In 2019 the Manoki community created a language revitalization collective, Watjuho Ja’a. Through this collective, the Manoki have been organizing meetings and intensive schools for the learning of Manoki among younger members of the community, and have also produced language materials (Bardagil, Kajoli & Mampuche 2020).

2 External relationships The Mỹky language is today considered an isolated language (Campbell 2012; Monserrat 2000, 2010), although it has been claimed to be an Arawakan language (see below in this section). I am of the belief that, based on available evidence, Mỹky cannot be genetically related to any attested language and therefore is best considered to be an isolate. Cândido Rondon first learned about a Mỹky-speaking group, today’s Manoki (Iranxe), from the Paresi in 1907. Their language was proposed by the military explorer to be related to the language of their Paresi neighbors (ISO: pab, Glottocode: pare1272), and thus an Arawakan language: “Today they [the Paresi] are divided in four groups, with the following denominations: Uaimaré, Caxiniti, Cozórini and Iranxe. […] From the information given by the Cozárini and Uaimaré, I know that [the Iranxe] speak Ariti – that is, the Paresi language” (Rondon 1946) (my translation). Thus, Rondon classified the evasive Iranxe as a Paresi group. When Kalervo Oberg visited the Utiariti mission in 1949, he met five Manoki there, and he seconded Rondon’s hypothesis that they spoke an Arawakan language (Oberg 1953). As far as we can tell, these are the only first-hand sources that proposed an Arawakan connection for Mỹky. These claims were made with no linguistic evidence available, supported instead by the shared material culture between the Manoki and the Paresi (e.g., the type of hammocks and houses), and also by the fact that the Juruena basin is a region with Arawakan presence. In the 1960s, Pereira (1974) visited the Utiariti mission and collected a list of words from Manoki speakers. He refuted an Arawakan relation for Mỹky, and wrote that he believed it to be “an isolate group, such as the Trumai or the Nambiquara [sic] etc.”, claiming that “the Iranxe and Paresi languages present no connection” (my translation). The same opinion was held by Meader (1967: 9), and further supported by Monserrat (2000, 2010) as a result of her work mostly with the Mỹky group. Table 19.1 summarizes a comparison of Mỹky lexicon for body parts with a reconstruction of Proto-Arawakan (Payne 1991) and with Latundê (Telles 2002), a

19 My˜ky

897

Tab. 19.1: Lexical comparison for Mỹky, Proto-Arawakan and Latundê. Mỹky ‘arm’ ‘back’ ‘foot’ ‘hair’ ‘leg’ ‘mouth’ ‘tooth’

j

mĩm ãkjeʔɨ kjatɨjã məənĩ matɨjri inẽkɨta jaʔa mjuhu

Proto-Arawakan

Latundê

danapha/wahku dinpy kyhtiba iti/si kawa numa/wajana/khanaky ahȼe

nuʔ takon jṵ na̰jn-kih-te tu ju wḭ

Nambikwaran language. This illustrative table is representative of the lack of putative cognates between Mỹky and its neighboring Arawakan and Nambikwaran linguistic neighbors. Due to its current geographic location, Mỹky is also considered a part of a large multilingual region with its center located further west in the Brazilian state of Rondônia and in Bolivia: the Guaporé-Mamoré linguistic area (Crevels 2002; Crevels & van der Voort 2008). This view proposes that the Guaporé-Mamoré area is a “residual zone,” defined as a region where linguistic diversity has accumulated as the languages spoken by politically and economically more powerful communities moved in, pushing aside the previous communities and their languages (Nichols 1992). Indeed, the area west of the current location of the Mỹky and Manoki is a region of high linguistic diversity in Amazonia, with approximately 50 different languages pertaining to at least 18 different genealogical units, if we consider language families and language isolates as separate instances. Seven language families are attested in the area: Arawakan, Chapacuran, Macro-Jê, Nambikwaran, Panoan, Tacanan, and Tupian. Besides these, 11 isolate languages are also spoken in that region: Aikanã, Kanoê, Kwazá, Canichana, Movima, Cayubaba, Itonama, Mosetén-Chimane, Leko, Yurakaré, and also Mỹky itself with a presence on the eastern margin of this area (Crevels & van der Voort 2008). Even though the evidence for considering the Guaporé-Mamoré a linguistic area has been called into question (Aikhenvald 2012), it is evident that the GuaporéMamoré area presents an unusually large number of language isolates, many of which share characteristics with Mỹky. The proposal was also examined by Muysken et al. (2014), who concluded that there are good arguments for areal diffusion in the region. We also find a few areal Wanderwörter among languages of the region that cannot be attributed to genetic inheritance. The word for giant armadillo, harurai in Kwaza (isolate), malola in Paresi and Enawene-Nawe (Arawakan), and mulula in Sabanê (Nambikwaran), is malula in Mỹky. There is also a Wanderwort for ‘star’ in several languages of the region: warəwarə in Arikapú (Jabutian); wirəwirə in Djeoromitxi (Jabutian); warɨwarɨ in Kanoê (isolate); waruwaru in Mekens and Makuráp (Tupian) and in Kwazá (isolate); and even wara in Aymara (Crevels & van der Voort

898

Bernat Bardagil

2008: 168), interestingly similar to the Mỹky word werapu ‘moon,’ containing the negative suffix -pu. Still, a potential genetic connection between Mỹky and other Guaporé-Mamoré languages, beyond the above mentioned Arawakan and Nambikwaran families, is not supported after examining the Mỹky correspondences for basic lexical items of the languages of the Guaporé-Mamoré region. Among the proposed Guaporé-Mamoré grammatical traits we find nasal harmony, evidentials, directionals, verbal number, nominative-accusative alignment, a preference for prefixation, and complex verbal morphology (Crevels & van der Voort 2008: 171). Admittedly, as argued by Muysken et al. (2014), these are not unusual properties in Amazonian languages. While these characteristics are also observed in Mỹky, some others are clearly absent – the proposed preference for prefixation is instead a clear preference for suffixation in Mỹky. However, the geographical situation of the language, the fact that it is an isolate, and some of the shared properties on lexical and grammatical levels make the presence of Mỹky on the eastern fringes of the Guaporé-Mamoré area consistent with the characteristics of this multilingual region of southern Amazonia. Future evidence might yet turn up information on the geographical location of the speakers of Mỹky before the 20th century, which might shed light on the prehistory of the Guaporé-Mamoré area and its adjacent languages.

3 Internal variation Today, two distinct Indigenous communities speak different varieties of the Mỹky language: the Manoki and the Mỹky. Even though their recent history, their dialect, and their current locations are different, in many ways they consider each other a part of the same people. They visit each other frequently, and the two groups take part together in social and ritual activities. There were some marriages between Mỹky and Manoki shortly after the former were contacted in 1971, and they take place to this day. The most apparent dialectal difference is that the Manoki present an additional /l/ consonant segment that the Mỹky variety does not have, as in the term for giant armadillo: /malula/ among the Manoki, but /mauwa/ among the Mỹky (Section 4.1.1). Loss of /l/ in the Mỹky variety triggered some phonological processes that further distinguish the two varieties, and it also had an impact on the form of some morphemes, both nominal (Section 6.3) and verbal (Section 8). The realization of certain phonemes also differs. The Manoki realize /w/ very frequently as [v], while among the Mỹky [w] is the most common allophone. Among the Mỹky, it is not uncommon for /ɨ/ to be fronted and appear as [i], which is rarely observed among the Manoki. There are also sociolinguistic effects that play out differently in the two communities. The younger Mỹky speak with a reduced verbal inflection, often not realizing

19 My˜ky

899

certain categories morphologically, such as sameness morphology (see Section 8.2). When paired with the phonological differences between the two varieties, this often results in superficially quite different realizations of the same morphosyntactic content, as illustrated in (1), where (1a) presents the form in the Manoki variant and (1b) is the form produced by young Mỹky speakers. (1)

a. [tolopaːnĩ] /to-lo-pa-nĩ/ leave-fut-1sg-sm ‘I’m leaving.’

b. [toːpa] /to-o-pa/ leave-fut-1sg ‘I’m leaving.’

There are also non-linguistic differences. Some myths and traditional narratives present significant variations, and there are also differences in the inventory of spiritual beings with whom the Manoki and Mỹky interact not only during rituals but also in their everyday life (Lopes 2015). The two Mỹky-speaking communities that exist today, the Manoki and the Mỹky, were most likely already differentiated to a significant extent by the year 1900, and the linguistic differences observed today were not provoked by a 70-year separation. What is a direct consequence of the separation, however, is how contact and integration with neocolonial Brazilian society took place. The Manoki were exposed to a lengthy period of contact with the Portuguese language, intensified by the conditions of strict linguistic imposition during the period that they spent in the Utiariti mission. The Mỹky came into contact with speakers of Portuguese decades later, in the 1970s, and since then they have never lived as a community outside of their village near the Papagaio river. The 20th century history of the two groups has resulted in a different language contact situation for each of them, producing for instance separate neology processes. The term used to refer to dogs among the Manoki is kasolu, a borrowing from Portuguese cachorro, while the Mỹky adopted a common Amazonian strategy of semantic widening of the term for jaguar, junai. Today’s Manoki and Mỹky are to all appearences the survivors of autonomous groups, each of them with multiple villages, that traditionally kept a close relationship. Manoki and Mỹky elders speak of the two groups as already being differentiated before their separation in the early 20th century. In fact, there is evidence in oral history of other Mỹky-speaking groups, like the Akomia or ‘Toucan People’ (Bueno 2014: 58, Lopes 2015: 50). It is very likely that several regional clusters of villages of Mỹky-speaking people inhabited the region between the Sangue and Papagaio rivers during the 19th century and that they presented dialectal differences. Supporting this hypothesis, Moura (1957: 160) points out that “every group has a different pronunciation of words” (my translation) when describing the different Manoki groups that were brought together at the Utiariti mission, which suggests that even the Manoki group might have originated as several groups that spoke different varieties of Mỹky.

900

Bernat Bardagil

4 Phonology This section contains a description of Mỹky segmental and suprasegmental phonology. The better known divergences between the two varieties of the language, those spoken by the Manoki and the Mỹky, mainly concern phonology. Where relevant, they are pointed out.

4.1 Phonemic inventory 4.1.1 Consonants Mỹky has a total of 19 consonants, with the absence of one of them, lateral /l/, in the non-Manoki variety (Table 19.2). While the consonant inventories of Eastern Amazonian languages are typically quite small, Mỹky has an especially reduced series of nasal stops, lacking a palatal /ɲ/ and a velar /ŋ/. Also atypical is the presence of a series of phonemically contrastive palatalized oral stops. Tab. 19.2: Mỹky consonants. labial oral stops nasal stops fricatives approximants

j

pp m mj w wj

alveolar

postalveolar

palatal

j

tt n nj s rl

velar j

kk ʃ

glottal ʔ h

j

Oral stops contrast in four points of articulation – labial, alveolar, velar, and glottal – while nasal stops are limited to labial and alveolar contrasts. Mỹky also has a series of fricatives, including a voiceless alveolar, a voiceless postalveolar, and a voiceless glottal. There are four approximants, comprising two glides and two liquids, and an alveolar trill, plus an alveolar lateral in the Manoki variety. Table 19.3 provides minimal pairs to support the phonemic contrasts for Mỹky consonants. A glottal stop and fricative are included in the phonemic inventory, although Monserrat (2010) notes that their distribution is restricted to intervocalic positions. Specifically, the optional occurrence of /h/ suggests a sensitivity to the prosodic structure. It is particularly productive in morpheme boundaries, both in /VV/ and /VC/ sequences, which are realized as [VhV] and [VhC] respectively, as in (2). (2)

a. [irehopa] /ire opa/ sun high ‘mid-morning’

b. [ʃahtakora] /ʃa-tako-ra/ arrive-dir-1sg ‘I returned.’

19 My˜ky

901

Tab. 19.3: Minimal pairs for consonant contrasts. contrast

exemplars

t:k p:k pj:tj m:n p:m t:n l:r n:r s:ʃ ʔ:h ʃ:h j:w j:r r:w

tohu pɨri ipja pamã panã to kula juna sapa məʔɨ miʃi jamã məjã iri

‘go’ ‘basket’ ‘smoke (v.)’ ‘take’ ‘cry (v.)’ ‘go’ ‘hen’ ‘jaguar’ ‘bake’ ‘smoke (n.)’ ‘stomach’ ‘small’ ‘father’ ‘bird’

kohu kiri itja panã manã no kura jura ʃapa məhɨ wamã wamã mərã iwi

‘scrape (v.)’ ‘kick (v.)’ ‘burn (v.)’ ‘cry (v.)’ ‘kill (v.)’ non-gradual ‘Paresí’ ‘butterfly’ ‘wing’ ‘fruit’ ‘hunger’ ‘hunger’ ‘hip’ ‘howler monkey’

The labial approximant /w/ is often realized as [v], especially in the Manoki variety (3). (3)

a. [keva] /kewa/ ‘non-Indigenous’

b. [pɨvɨ] /pɨwɨ/ ‘knife’

Mỹky also presents a series of palatalized stops that contrast phonologically with oral and nasal non-palatalized stops (Table 19.4). Tab. 19.4: (Near) minimal pairs for palatalized contrasts. contrast j

p:p t:tj k:kj m:mj n:nj w:wj

exemplars paihi tõpɨ kapɨ matɨ nakaha kewa

‘laugh (v.)’ ‘uncle’ ‘cover’ ‘go’ ‘entangle’ ‘non-Indigenous’

pjaiʔi tjõpi kjapɨ mjatɨ njakəɨ kẽwjã

‘tame’ ‘anus’ ‘close’ ‘scrape (v.)’ ‘testicles’ ‘coati’

The alveolar lateral approximant is only present in the variety of Mỹky spoken by the Manoki. In words that present /l/ in the Manoki variety, the phoneme is entirely missing in the speech of the Mỹky, demonstrated in (4), with one exception. In a context where a morpheme contains /l/ in a nasal environment in Manoki, it corresponds to /n/ in the Mỹky variety (5). An /l/ between two identical vowel segments or between two high vowels in Manoki resulted in a long vowel in Mỹky, as in (6), and vowels assimilate in backness, shown in (7) and (8).

902

Bernat Bardagil

(4)

a. /wakala/ ‘egret’

manoki

b. /wakaː/ ‘egret’

mỸky

(5)

a. /tĩ̵lĩ̵/ ‘to cook’

manoki

b. /tĩ̵nĩ̵/ ‘to cook’

mỸky

(6)

a. /kalatali/ ‘butterfly’

manoki

b. /kaːtai/ ‘butterfly’

mỸky

(7)

a. /karɨli/ ‘anaconda’

manoki

b. /karɨ/ ‘anaconda’

mỸky

(8)

a. /mãlẽna/ ‘beautiful’

manoki

b. /mããna/ ‘beautiful’

mỸky

It is clear that, in the Mỹky spoken by the Manoki, /l/ is a consonantal phoneme. This is supported by the observation of minimal pairs, such as the roots /kula/ ‘hen’ and /kura/ ‘Paresí’. As Meader (1967: 43) indicates, /l/ occurs only wordmedially. It does however occur morpheme-initially, as in the future tense suffix -lo or the progressive suffix -lỹpa.

4.1.2 Vowels Mỹky has an especially large vowel inventory, totaling 26 contrastive vocalic phonemes (Table 19.5). The vowel system is structured according to three levels of vowel height and three levels of backness (Table 19.6) and comprises one long and one short series of both oral and nasal vowels. Tab. 19.5: Mỹky vowels. oral vowels

high mid low

nasal vowels

front

central

back

front

central

back

i iː ɛ ɛː

ɨ ɨː ə əː a aː

u uː ɔ ɔː

ĩ ĩː ɛ̃ ɛ̃ː

ĩ̵ ĩ̵ː ə̃ ə̃ː

ũ ũː ɔ̃ ɔ̃ː

Vowel phonemes contrast in nasality, and nasal segments trigger nasal harmony in neighboring segments (Section 4.2). The low dimension is neutralized on nasal vowels. Vowel length is contrastive in Mỹky, as described by both Meader (1967) and Monserrat (2010). Long vowels are often realized as falling diphthongs (/mə:pu/ [məjpu] ‘bone’). Table 19.7 presents minimal pairs for length.

19 My˜ky

903

Tab. 19.6: Minimal pairs for vowel quality. contrast

exemplars

i:ɨ i:e ɨ:u ɨ:ə ə:o ə:a u:o

iti aʔi pɨri kaɨʔi pəku mərohu muhu

‘ant’ ‘to knit straw’ ‘basket’ ‘to pull’ ‘to climb’ ‘ripe’ ‘rain’

itɨ aʔe puri kaəʔɨ poku marohu mohu

‘plain’ ‘wasp sp.’ ‘to smell’ ‘to open’ ‘bow’ ‘dawn’ ‘bullet ant’

Tab. 19.7: (Near) minimal pairs for vowel length. contrast

exemplars

i : ɨː e : eː ɨ : ɨː ə : əː a : aː u : uː o : oː

inĩ jeryli tatɨ ətaɨ aku aku kowi

‘house’ ‘stomach ache’ ‘pequi’ ‘catfish’ ‘cicada’ ‘cicada’ ‘fish sp.’

i:nĩ jeːwali atɨː əːtaɨ aːku akuː koːwi

‘louse’ ‘male namesake’ ‘waterfall’ ‘to cry’ ‘echo’ ‘scorpion’ ‘opossum’

4.1.3 Palatalization Phonological palatalization is triggered by a preceding high front vowel or a palatal approximant, as in (9). This process affects all consonants except /h/, /ʔ/, /j/, and /ʃ/. In the case of /s/, the palatalized allophone is realized as [ʃ], shown in (10). (9)

C → Cj / i, j _

(10) a. [mĩmjã] /mimã/ ‘hand’

b. [uipja] /uipa/ ‘to hit’

c. [ʃiʃua] /ʃi-sua/ iter-clean ‘to re-clean’

In Mỹky, palatalization also signals nominal modification, as in (11). This includes possession, which is realized as palatalization of the root-initial consonant of the possessum, illustrated in (12) and further discussed in Section 6.4. (11) a. [kulapa] /kulapa/ ‘child’

b. [nãmɨjkjulapa] /nãmyi kulapa/ ‘woman child (little girl)’

904

Bernat Bardagil

(12) a. [poku] ‘bow’ [are pjoku] ‘my bow’ b. [matɨ] ‘head’ [are mjatɨ] ‘my head’

4.2 Suprasegmental phonology Mỹky displays a set of phonologically distinctive suprasegmental features, comprising nasality and tone. In Mỹky, contrastive nasality can spread from a nasal segment to adjacent oral segments (Piggott 2003; Walker 2011). Nasal harmony has been observed in several Amazonian languages – see Wetzels (2009) for Maxacalí (MacroJê) and Singerman (2016) for Tuparí (Tupian). This is illustrated in (13) for the Manoki variety of Mỹky. Spreading of nasality appears to operate in a right-to-left direction from an inherently nasal segment. As seen in (13a), the segments to the right of /n/ do not become nasalized. Voiceless consonants appear to block nasality from spreading, as in (14). (13) a. [jũnali] /junali/ ‘jaguar’

b. [mãnã] /marã/ ‘to carry’

(14) a. [’ʃikɔkɛ̃] /’ʃikɔkɛ̃/ ‘uncle’

b. [pa’ʃɔ̃pa] /pa’ʃõpa/ ‘cough (v.)’

Mỹky presents evidence for partial loss of nasality on a nasal stop adjacent to an oral vowel; see (15). Postoralization of nasal stops before oral vowels is observed in only a reduced number of Mỹky speakers but is attested as a general rule in the Manoki variety (Meader 1967: 44). ͡ uhu] (15) a. [mb /muhu/ ‘rain’

͡ ] b. [malẽnta /malena/ ‘beautiful’

Turning now to tone, little is known about the tonal system of either of the two varieties of the language. As briefly described in Monserrat (2010: 6) Mỹky presents ˆ/. As two level tones, high /o´/ and low /–/, as well as one falling tone, high-low /o the minimal pairs in (16) illustrate, the tone-bearing unit in Mỹky appears to be a prosodic unit such as the mora. Available data suggest that the high-low tone is restricted to long vowels, demonstrated in (17).

19 My˜ky

(16) a. /jamã/ ( – – ) ‘deer’ c. /mjuhu/ ( a´ a´ ) ‘mother’

905

b. /jamã/ ( a´ – ) ‘small’ d. /mjuhu/ ( – a´ ) ‘tooth’

(17) a. /mĩ ̵kɨ/ ( – – ) ‘person’

ˆ–) b. /mĩ ̵:kɨ/ ( o ‘night’

4.3 Prosodic phonology The basic syllable structure of Mỹky is (C)V(C). Any consonant of the language can appear in onset position, although /ʔ/, /h/, and /r/ are restricted to word-internal positions; see (18). Syllables with a long vowel nucleus can also have codas, as in (18c). Word-internal syllables, just like word-initial syllables, do not require an onset, demonstrated in (18d). (18) a. i.tju.ku V.CV.CV ‘to sleep’ c. waaj.mjaa.ta.si CV:C.CV:.CV.CV ‘house gecko’

b. bi.hĩ.na CV.CV.CV ‘sister’ d. u. a.wa V.V.CV ‘tucum (palm tree sp.)’

The coda position is restricted to glides /j/ and /w/, illustrated in (19), with most words ending in an open syllable. (19) a. kãj.ti CVC.CV ‘frog sp.’

b. jaw.ka.i CVC.CV.V ‘proper name’

In compounds and phrases that produce a (C)VCV+CV environment, the morpheme boundary often triggers loss of the vowel preceding it, which results in a VC.CV sequence with an otherwise illegitimate coda, such as the oral stops in (20). This has been observed exclusively when the two words (e.g., noun+noun, noun+adjective, adjective+adverb) are polysyllabic. (20) a. [mjat.ka.lo] CVC.CV.CV /mjatɨ kalo/ cold very ‘very cold.’

b. [ku.rat.ka.nã] CV.CVC.CV.CV /kuratu kanã/ corn water ‘corn juice’

906

Bernat Bardagil

4.4 Orthography The current incarnation of the orthographic system for Mỹky consonants uses 18 letters (including Manoki /l/), five digraphs, and four diacritics. Consonants are represented with the characters shown in Table 19.8. Tab. 19.8: Mỹky orthography for consonants. /p/

/kj/ /nj/ /r/

/pj/ /ʔ/ /s/ /l/

/t/ /m/ /ʃ/ /j/

/tj/ /mj/ /h/

/k/ /n/ /w/

The spelling of short oral vowels is given in Table 19.9. The mid-central vowel /ə/ has been traditionally represented with a macron below (a̱), although the more accessible spelling with a diaeresis (ä) has become more widespread in recent years. Nasality is represented with a tilde on a vowel (/Ṽ/ = Ṽ). Tone is not represented on the spelling. However, the Manoki use the acute accent to mark stressed syllables. Tab. 19.9: Mỹky orthography for vowels. /i/ /e/

/ ɨ/ /ə/ /a/

/u/ /o/

Length is represented in two ways. The first system is spelling a vowel twice to represent long vowels, a spelling choice that has become more common in recent times (/Vː/ = VV, and /Ṽː/ = ṼṼ). This is used especially in the Manoki community. The second system, favored in printed materials published in recent decades by organizations working with the two communities, works differently on oral vowels and nasal vowels: an acute accent represents a long oral vowel (/Vː/ = V́), and a circumflex accent represents a long nasal vowel (/Ṽː/ = V̂). Throughout this chapter I use the first system, with doubled vowels, when presenting Mỹky data orthographically.

5 Morphosyntactic typology Mỹky is a largely synthetic, head-marking language. Both verbs and nouns are inflected for several categories, detailed in the following sections. Mỹky verbs are almost exclusively suffixing, presenting one prefix and up to twelve suffixes (Section 8).

19 My˜ky

907

Mỹky grammar presents a strong tendency towards head-finality. While the constituent order in the sentence is generally verb-final, all possible combinations between subject, object, and verb are possible. Example (21) demonstrates SOV order, and (22) illustrates VO order. Rather than standalone pronouns, dropped participant phrases are expressed in the verbal morphology. (21) kooxi jotamũ jotxi-mta o-maa-karee-ju opossum worm hunt-fut go-infr-pst-3sg ‘The opossum went out to hunt worms.’ (Monserrat 2010: 163) (22) näänamã aa-na-määpi manã-i now see-hab-1pl river-m ‘Now we know the river.’ (Monserrat 2010: 147) All Mỹky adpositions are postpositions (Section 7), since the adpositional head follows its complement, as shown with the instrumental in (23) and the locative in (24). (23) a kapy uipa-ru stick ins hit-3sg.past ‘He hit it with a stick.’ (24) jãntai-nỹ tje boat-def loc ‘In the boat.’ As a polysynthetic language, sentences in Mỹky consist minimally of the verbal word with its inflectional morphology. There is common null anaphora of argument phrases, as in (25), both for the subject and the object. (25) anã-sa-mĩpapju-rã-ra, päätõ listen-2obj-3obj.pl-3refl.pl-1sg grandmother.voc ‘I listen to you with them, grandmother.’ (Monserrat 2010: 150)

6 The noun phrase In Mỹky a category of nouns can be defined based on the inflectional profile which, like that of the verb, is quite synthetic. Noun phrases are formed minimally of nouns

908

Bernat Bardagil

or pronouns combined with other nouns (26), adjectives (27), numerals (28), quantifiers (29), and determiners (30). (26) [mia kulapa] simã-ỹpa-Ø-rã man child vomit-prog-3sg-nsm ‘The little boy is vomiting.’ (Monserrat 2010: 26) (27) [kataty jeewja] kewa-ka-mỹ’ĩ ankaa-nãtã macaw red non.Indigenous-poss-manioc eat-decl ‘The red macaw ate a cookie.’ (28) [werapu nũmã] xa-tako-lo-pa-sã moon two come-dir-fut-1sg-nsm ‘I will come back in two moons.’ (29) Juru [mamã jepte] kyjaky-Ø Juru clothes lots own-3sg ‘Juru owns lots of clothes.’ (30) [anỹ inĩ ] jamã-Ø-nĩ this house small-3sg-sm ‘This house is small.’ Noun phrases can bear morphology indicating definiteness, gender, case, diminutive, and several other inflectional suffixes. The distinctive inflectional categories of noun phrases are described below. Nouns are coordinated by juxtaposition, with no dedicated morphological mediation, as in (31). (31) Haware’ana kumãtã kuratu tolũna-Ø Mariana bean corn sow-3sg ‘Mariana sowed beans and corn.’ (Meader 1967: 28) In Mỹky, noun phrases present a fixed determiner-noun-modifier order (Monserrat 2010: 168), exemplified in (32). (32) are [jepte taty käja-nỹ ] ãnka-lopa fsg lots pequi tasty-def eat-fut.1sg.sbj ‘I will eat a lot of tasty pequi.’ As seen in the example above, definiteness morphology (Section 6.2) does not attach exclusively on the noun heading the noun phrase, but instead it is very com-

19 My˜ky

909

mon for it to be attached to the rightmost element in the noun phrase, that being the adjective käja ‘tasty’ in (32). It is likely that definiteness morphology is an instance of phrasal boundary morphology, specifically one that appears on the right boundary of the noun phrase.

6.1 Person and number In Mỹky, three persons are active in the grammar: first, second, and third. Personal pronouns, which co-occur with person markers on the verb, are specified for three number values: singular, dual, and plural. Singular forms are monomorphemic. Plural forms include an optional -lĩ morph that most likely encodes plurality. The same morphological base used in plural pronouns is used to derive the dual forms, with the addition of an obligatory -tã dual morpheme. The Mỹky pronouns are given in Table 19.10 for the Manoki variety. Tab. 19.10: Mỹky pronouns.

1 2 3

singular

dual

plural

are’y sẽĩ anỹ

jããtã sããtã nããtã

jããlĩ sããlĩ nããlĩ

The first-person pronoun is morphologically complex: are-’y. It carries a suffix -’y, maybe connected to definiteness (Section 6.2) or gender (Section 6.3). This suffix is present most of the times when the pronoun appears in isolation, although it is not obligatory; see (33). It is dropped when it is part of a complex phrase, as in possessive contructions (Section 6.4) or postpositional phrases (Section 7), in which cases are appears in its bare form, as seen in (34) and (35). (33) are(’y) xa-tako-ra-nĩ 1sg.pro arrive-dir-1sg-nsm ‘I came back.’ (34) are kulapa itjuku-Ø 1pl.pro child sleep-3sg ‘My child fell asleep.’ (35) a-ra-mĩ are kinã born-1sg-3sg 1sg.pro ben ‘[The child] was born to me.’ (Monserrat 2010: 67)

910

Bernat Bardagil

Unlike pronouns, lexical noun phrases never carry number morphology, even though they can be specified with quantity modifiers like jepte ‘a lot of.’ This is true of both notionally count and mass nouns, as in (36) and (37), respectively. (36) pää-nỹ jepte awa-pa-määmĩ mojamã forest-def lots take-refl-3pl peccary ‘They caught lots of peccaries for themselves in the forest.’ (Monserrat 2010: 147) (37) jañ a-nã jepte suta-na-määpi cotton-indf lots gather-hab-1pl ‘We gathered a lot of cotton.’ (Monserrat 2010: 159) Given the polysynthetic nature of the Mỹky verb complex, where arguments are cross-referenced with agreeing morphemes, number is expressed in two sites. First, as seen in Table 19.10, the value of number exponence is visible on personal pronouns. Second, for the core arguments, namely subjects and objects, number is also reflected on the corresponding cross-reference morpheme on the verb. This is illustrated in (38). (38) a. are moxe-pa-ra 1sg.pro wash-refl-1sg ‘I washed myself.’

ãnka-määxu-nã b. sããlĩ 2pl.pro eat-2pl.pst-decl ‘You all ate.’

Table 19.11 presents the subject cross-reference markers for the Manoki variety. The three number categories of Mỹky – singular, dual, and plural – are expressed by subject cross-reference morphology. In the formation of the full paradigm they interact not only with person but also evidentiality distinctions (Section 8.3). As can be seen, not all persons or reality status and tense categories reflect the same richness of evidentiality information. First person always lacks evidentiality information, and so does irrealis. In the two realis tenses, second person is always specified for evidentiality, and third person, only in realis past tense. These observations can be represented on a markedness continuum with the generalizations in (39). (39) a. Person: 2nd > 3rd > 1st b. Tense: r-past > r-present > irr The placement of subject cross-reference is exemplified in (40) for intransitive and transitive verbs. (40) a. päja-ro-ra-sã wake-tel.sg-1sg-nsm ‘I woke up.’

b. are ãnka-mijulepa-Ø-ra-nĩ 1sg eat-again-3sg.obj-1sg-sm ‘I’ll eat once more.’

19 My˜ky

911

Tab. 19.11: Mỹky subject cross-reference (Monserrat 2010: 39). singular

dual

plural

par. I: realis present 1st 2nd witness 2nd nonwitness 3rd

ra/ja/ta si hmĩ mĩ~mỹ / Ø

je meexi meehmĩ meemĩ/mee

re/te määxi määhmĩ määmĩ/mää

par. II: realis past 1st 2nd witness 2nd nonwitness 3rd witness 3rd nonwitness

pa xu hmjũ ru~ju mũ

pe meexu meehmjũ meeju meemjũ

pä/määpi määxu määhmjũ määju määmjũ

par. III: irrealis 1st 2nd 3rd

pa pa pa / Ø

pe meepja/mee meepja/mee

pä määpja/mää määpja/mää

Objects are also cross-referenced in the Mỹky verbal complex. As can be seen in Table 19.12, person distinctions are highly reduced in non-singular numbers. For number exponence, the bare pju form is used by speakers of the Mỹky variety for both dual and plural (Monserrat 2010: 67), while the speakers of the Manoki variety favor the augmented pju-based form (e.g., xĩpju) for the plural. Tab. 19.12: Mỹky object cross-reference (Monserrat 2010: 61). singular

dual

plural

par. I 1st 2nd 3rd refl

ra sa Ø pa

tpju pju pju tpja

(t)pju (h)pju pju papju

par. II 1st 2nd 3rd refl

ra xĩ mĩ pa

tpju pju pju papju

nĩ/(t)pju (xĩ)pju (mĩ)pju mĩpapju

It is not clear on what aspect of the grammar the distinction between the two paradigms is based. Unlike subject cross-reference, the choice of paradigm for object suffixes does not appear to correlate with tense, reality status, or evidentiality. The same verbal root can take suffixes from the two paradigms, as seen in (41).

912

Bernat Bardagil

(41) a. ka’a-nỹ jamã-Ø-Ø-nã arrow-def give-3obj-3sbj-decl ‘He gave the arrow to him.’

b. jamã-mĩ-ne-nãtã give-3obj-1pl-decl ‘We gave it to him.’

Mỹky second person object cross-reference is exemplified in (42). These are the same markers used to cross-reference a dative or benefactive argument, independently of the two series, as in (43). (42) kanka-lỹ-sa-ra-nã bite-vis-2sg-obj-1sg.sbj-nsm ‘I bit you.’ (43) poku jamã-sa-lo-pa-nĩ bow give-2sg.obj-fut-1sg.sbj-sm ‘I’ll give you a bow.’

6.2 Definiteness The Mỹky noun phrase presents morphological marking of definiteness. There is a definite-indefinite pair of suffixes on nominal roots, which include deictic pronouns. The definite suffix -nỹ, see (44) through (46), specifies or defines a nominal expression, like determiners do in many languages of the world. As illustrated by (46), definiteness morphology attaches to the right of the noun phrase, rather than to the noun. (44) kuwytakä kyjaky-nỹ itjuku-o-Ø-ti-jã? peanut owner-def sleep-go-3sg-q-nsm ‘Has the owner of the peanuts gone to sleep?’ (Monserrat 2010: 174) (45) are upa-nỹ-kĩnã xĩp-jo-hu 1sg.pro snake-def-obl fear-psych.impf-ident ‘I’m afraid of the snake.’ (Monserrat 2010: 120) (46) ka’a ma-nỹ jamã-hpju-Ø-nã arrow big-def give-2pl.obj-3sg-decl ‘He gave you the big arrow.’ The indefinite suffix nã, in complementary distribution with the definite suffix, signals an indefinite or generic interpretation, as in (47) and (48).

19 My˜ky

913

(47) paka mänã-mĩka-nã-hã bath water-dirty-indf-ident ‘They bathe in murky water.’ (Monserrat 2010: 155) (48) jãnta-nã akatako-pä car-indf return-1pl.pst ‘We went back by car.’ (Monserrat 2010: 156) Both the definite and indefinite suffixes alternate with a null affix. Thus, noun phrases also can be morphologically unspecified for their definiteness value, illustrated in (49). (49) jamã-pju-si ka’a give-obj.pl-2sg arrow ‘You gave arrows to them.’

6.3 Gender Mỹky nouns present three gender classes: masculine, feminine, and unmarked. Gender is expressed morphologically by a series of suffixes that attach to nouns. Two suffixes, -li and -xi, mark masculine nouns, while another two suffixes, -lu and -si, mark feminine nouns. A third category of nouns are unmarked for gender. However, the -lV gender suffixes present rampant allophony, especially in the masculine, triggered for the most part by the loss of the /l/ phoneme in the non-Manoki variety of Mỹky; refer to (50). (50) a. Masculine: -xi, -li~-i~-y~-wy b. Feminine: -si, -lu~-u Loss of the /l/ phoneme in the variant spoken by the Mỹky triggered a series of phonological phenomena involving the length and quality of the surrounding vowels, as well as consonant assimilation processes (Section 4.1.1). For gender suffixes with /l/, this most likely resulted in a series of allomorphs, as illustrated in (50). A simple version of the alternation is illustrated in (51), with the Manoki (left) and Mỹky (right) variants of the same gender suffix. (51) a. jeewa-li ~ jeewa-i ‘male namesake’ b. jeewa-lu ~ jeewa-u ‘female namesake’ The masculine suffixes, in (50a), appear on men’s proper names like Xinũ-li or Tamũ-xi and on male kinship terms like kjap-xi ‘younger brother’ or mypy-y (mypy-

914

Bernat Bardagil

li) ‘older brother’. They also appear arbitrarily on both animate and inanimate nouns like py-wy ‘knife’, kaata-i (kalata-li) ‘butterfly’, or moo-xi ‘peccary’. (52) piwi-jee-ta-nĩ moo-xi-nỹ follow-1du-emph-sm peccary-m-def ‘Let’s follow the peccary.’ (Monserrat 2010: 38) The feminine suffixes, shown in (50b), appear with women’s proper names like Kamũ-u (Kamũ-lu) or Kamẽnpja-si; on female kinship terms like määnã-si ‘sister-inlaw’ or jeewa-u (jeewa-lu) ‘female with the same name’; and also in non-human animate and inanimate nouns like kata-u (kata-lu) ‘mushroom’, alo-u (alo-lu) ‘rice’, or watapa-si ‘pigeon sp.’. (53) alo-lu ãnkaa-ki-nĩ! rice-f eat-imp-sm ‘Eat rice!’ Finally, there are nouns with neither suffix, like upa ‘snake’ in (54), including nouns that refer to humans such as tikãta ‘chief’ or nãmyj ‘woman’. (54) apa upa ura-Ø here snake fierce-3sg ‘Here there is a poisonous snake.’ (Monserrat 2010: 47) The distinction between the suffixes that encode the same gender category is not clear. The same proper name can carry both suffixes: Juru-lu or Juru-si for women, and Nãpu-li or Nãpu-xi for men. This fluidity is also observed in some common nouns: the term for a small nightjar species appears both as xiro-si and xiro-lu.

6.4 Possession In Mỹky, possession, belonging, or dependence are expressed by means of two constructions: juxtaposition, and a dedicated postposition kjany~kaney. As a head-final language, in the possessive construction the possessor precedes the possessum, as in (55). (55) sẽĩ nãmyj weeruku-Ø 2sg.pro woman sit-3sg ‘Your wife sat down.’

19 My˜ky

915

Inherent possession is prototypically expressed by means of juxtaposition. This is the relation reserved for body parts and kinship and also for possession of specific inanimate entities such as bows or hammocks. (56) are nãripjiu-n kĩ wala-ra-sã 1sg.pro husband-def dat speak-1sg-nsm ‘I spoke to my husband.’ (57) are jã, palu ää-piju-si-ti? 1sg.pro father cará to.fish-obj.pl-2sg-int ‘My father, did you fish cará?’ (Monserrat 2010: 25) Inherent possession via juxtaposition triggers a phonological alternation at the onset of the first syllable of the possessum. In our corpus, these are limited to a palatalization of oral stops (e.g., /p/→/pj/) and a /m(j)/→/kj/ change, as seen in (58) through (60). (58) a. papyj hammock

b. maromỹ are pjapyj jamã-sa-lo-pa-ra morning 1sg.pro hammock.poss give-2.vis-fut-1sg-nsm ‘Tomorrow I’ll give you my hammock.’

(59) a. mjuli skin

b. jũna kjuli jaguar skin.poss ‘jaguar skin’

(60) a. maty head

b. opyri kjaty ãnka-lo-pa-nĩ tapir head.poss eat-fut-1sg-sm ‘I’ll eat tapir head.’

The /m(j)/→/kj/ alternation for the possessum, which has been observed mainly with body parts, does not seem to be an entirely phonologically conditioned process. Quasi-homophones like mjuhu ‘tooth’ and mju’u ‘mother’ behave differently in that respect, with only ‘tooth’ undergoing the alternation, as seen in (61) and (62). (61) a. mjuhu tooth

b. are kjuhu 1sg.pro tooth.poss ‘my tooth’

(62) a. mju’u mother

b. are mju’u 1sg.pro mother ‘my mother’

916

Bernat Bardagil

The constituency of the predicate in possessive constructions is reflected morphophonologically via the aforementioned process. For a sentence like ‘I have two hammocks’, seen in (63), when numã ‘two’ is the predicate the noun papyj ‘hammock’ forms a possessive noun phrase with first person pronoun are, appearing in its possessed form pjapyj in (63a). By contrast, when papyj numã ‘two hammocks’ is the predicate, the first-person pronoun are’y appears with its isolated form (see Section 6.1), and the word-initial alternation in possessive constructions for papyj ‘hammock’ is not triggered, as in (63b). Square brackets are used to indicate constituency, and the contrived English translations attempt to mirror the structure of the Mỹky sentences. (63) a. [[are pjapyj ] numã] 1sg.pro hammock.poss two ‘My hammocks are two.’

numã]] b. [ are’y [ papyj 1sg.pro hammoch two ‘Mine are two hammocks.’

The second possessive strategy is mediated by the possessive postposition kany ~ kjany ~ kaney, as in (64), which indicates that its object is the possessor of the noun that heads the possessive noun phrase. Non-contiguous possession is allowed, as illustrated in (65), with the possessed noun miatapa ‘fish’ not adjacent to the possessor Xinũxi kany. (64) are kaney kaso-lu nakata 1sg.pro poss dog-f white ‘I own the white dog.’ (65) miatapa pyri-n kje Xinũxi kany fish basket-def loc Xinũxi poss ‘The fish in the basket are Xinũxi’s.’ (Monserrat 2010: 143)

6.5 Classifiers Mỹky classifiers are noun-like morphemes that combine with other nouns with a function that specifies some property of the referent, generally a physical property. The noun classifiers inventoried for Mỹky are mostly used with parts of the body, parts of plants, and shapes. Table 19.13 is a non-exhaustive list of classifiers as described in Monserrat (2010: 166). Mỹky classifiers show great combinatorial productivity. The ti classifier, used to indicate that something is at the back, is exemplified in (66). (66) a. ti kyypju behind.clf end ‘tail’

mã kimã b. ti behind.clf excrement path ‘anus’

19 My˜ky

917

c. mokje ti kjepu neck behind.clf bone ‘nape’

Tab. 19.13: Mỹky classifiers. classifier

meaning

ti tapa kipu tu~tutu jawa

behind flat edge, tip round thick and flat

7 Postpositions In Mỹky, all adpositions are transitive postpositions. As such, they appear linearized to the right of the postpositional object, as shown in (67). (67) [object P]PP Postpositions combine with noun phrases headed by either a noun or a pronoun. Nominalized verbal predicates (Section 8.6) might be able to act as objects of postpositions, but this is not attested. Postpositions combine with inflected noun phrases; in particular, it is very usual to find definiteness morphology couched between the nominal root and the postposition. This is illustrated in (68), where definite -nỹ is phonologically reduced to [n]. (68) a. wala-maa-karee-ju piapaky-n kĩnã speak-infr-pst-3sg bee-def dat ‘He spoke to the bees.’ (Monserrat 2010: 165) b. a kjaã-n kjehy upa-nỹ wood dry-def loc snake-def ‘The snake is on the dry branch.’ (Monserrat 2010: 122) The postpositions in my corpus are listed with their approximate meanings in Table 19.14.

918

Bernat Bardagil

Tab. 19.14: Mỹky postpositions. postposition

meaning

pjaha kỹjanã kinã kapy kjehy ~ tjehy kany ~ kaney

allative comitative dative instrumental locative possessive

Some Mỹky postpositions are further exemplified in (69). (69) a. mia o-karee-Ø päi-n pjaha i-’i jori-mỹ man go-pst-3sg forest-def all honey-m find-3sg ‘A man went to the forest to find honey.’ (Monserrat 2010: 178) b. are’y to-lo-pa-ni ̃ sẽĩ kỹjãnã 1sg.pro leave-fut-1sg-sm 2sg.pro com ‘I will leave together with you.’

8 The verb phrase The Mỹky verb is realized as a polysynthetic verbal word. Mỹky verbs exhibit a high number of inflectional categories, spanning the 12 suffixal slots on the verb complex and the one iterative prefix xi- (Table 19.15). The inflectional categories exponed by suffixation include agreement with core arguments, aspect, tense, evidentiality, mood, and uncertainty. In addition, dedicated suffixes express negation, sameness indexicality, impersonality, reflexivity, nominalization, and interrogation. There are two morphologically defined classes of verbs: active and stative. Stative verbs share a series of features with the class of adjectives (Monserrat 2010). Mỹky presents a three-way inflectional distinction resulting from the intersection of two verbal categories: tense (past-present) and mood (realis-irrealis). The inflectional morphology that indexes subjects on the predicate head presents different paradigms for realis present and realis past. There are also different paradigms for irrealis: a near future, and a future encompassing also interrogative, imperative, and some evidentiality, as described in Section 6.1. Not all categories are obligatorily present via a morphological exponent on the verb. Person is obligatorily marked on the verb and, with subject cross-reference, reality status and tense are also expressed. There is generational variation in the complexity of the verbal complex, with younger speakers of the Mỹky variety using

19 My˜ky

919

Tab. 19.15: The Mỹky verb complex, adapted from Monserrat & Dixon (2003). position

slot

function

Prefix Verb root Suffixes

 1  –  1  2  3  4  5  6  7  8  9 10 11 12

iterative single or serial aspect object, reflexive identificative evidentiality/negation habitual (personal, impersonal) tense subject + visual/non-visual mood epistemic modality identificative sameness indexicality clause type

certain slots less frequently, such as sameness indexicality and discourse markers, than older speakers do. For the Manoki variety, with no speakers under the age of 60, it is not clear whether age plays a role regarding the richness of morphological exponence on the verb.

8.1 Aspect This section discusses the morphological realization of aspectual reference in Mỹky. Besides the four general aspects reflected in Mỹky verbs, the language also presents an imperfective that combines exclusively with psychological verbs; see Table 19.16. Tab. 19.16: Mỹky aspects. suffix

aspect

ku~ka tu lỹ(pa) jo~jaka

perfective imperfective progressive psychological imperfective

The relationship between the time of utterance and the time of a past event corresponds in Mỹky to the aspectual categories of perfective and imperfective, both of which are encoded in the verbal morphology. Perfective -ku (singular) and -ka (plural) denote a completed event, begun and culminated, as in (70) and (71).

920

Bernat Bardagil

(70) Juru siwu-ku-mỹ papyj-nỹ Juru dry-pfv.sg-3sg.prs hammock-def ‘Juru has dried the hammock.’ (Monserrat 2010: 92) (71) mja-ro-ka-määxi ready-tel-pfv.pl-2pl ‘You guys are ready.’ (Monserrat 2010: 43) Imperfective -tu denotes an event or process that is visualized not as completed but as an ongoing situation, shown in (72) and (73). (72) watjuho apu-tu-maa-karee-ju all fly-ipfv-pst-3sg ‘They were all flying.’ (Monserrat 2010: 92) (73) xãn-tu-te lie.down-ipfv-1pl ‘We were lying down.’ (Monserrat 2010: 93) Mỹky also presents an imperfective aspect that combines exclusively with stative verbs that express a psychological or mental state or event, listed and exemplified in (74) and (75). It is marked with -jo (singular) and -jaka (plural). (74) ãnka kã-jo-ja eat want-psych.ipfv.sg-1sg ‘I’m wanting to eat.’ (75) xĩpy-jaka-ỹ-re be.afraid-psych.ipfv.pl-prog-1pl ‘We’re afraid.’ (Monserrat 2010: 88) Additionally, a progressive aspect is indexed with the suffix -(l)ỹ~-(l)ỹpa, denoting a regular, ongoing activity, as in (76). (76) ximã-lỹpa-Ø-rã ulapa vomit-prog-3sg-nsm child ‘The child is vomiting.’ (Monserrat 2010: 29)

19 My˜ky

921

8.2 Sameness indexicality Mỹky presents morphological indexicality of sameness, which encompasses age, sex, and cosmological category. A dedicated slot in the verb complex, corresponding to slot 11 in Table 19.15, indicates sameness or mismatch between the speech event participants. Sex indexicality is a grammatical phenomenon often realized in what is usually known as male and female speech. As opposed to grammatical gender, which can be defined as a classification of nouns into at least two classes, based on semantic properties, with consequences on the morphological exponence of the different noun classes, sex indexicality (also called “gender indexicality”) is the exponence of a speech event participant’s sex (Silverstein 1985: 223, Rose 2015). When a speaker of Mỹky addresses a person of a different sex, a morpheme in the verb complex indicates that the addressee’s sex is different from that of the speaker. The morpheme indicating sex mismatch between speaker and addressee, -sã~-xã, is the same for male and female speakers, as in (77). Example (77a) demonstrates either a man speaking to a man, or a woman speaking to a woman; (77b) illustrates a man speaking to a woman, or the reverse. (77) a. poku ko-pa-rã-meemĩ-Ø bow scrape-refl-m-3du-sm ‘They scraped a bow for themselves.’ (Monserrat 2010: 72) b. poku ko-pa-rã-meemĩ-xã bow scrape-refl-m-3du-nsm ‘They scraped a bow for themselves.’ To the extent that sex or gender indexicality specifies the sex of the participants of the speech act, rather than the gender values of grammatical participants in the clause, this system is orthogonal to that of grammatical gender, also present in Mỹky (Section 6.3). The morphological indices of sameness are used, besides encoding the sex of the interlocutors, to indicate whether speaker and addressee belong to the same age and cosmological category, as well as certain kinship relations. When speaking to an addressee of the same age group as the speaker, be it an adult or a child, the (a) forms in (78)–(79) are used. Conversely, an adult speaking to a child will use the (b) forms, which are also used when a human is speaking to a spirit or when a spirit addresses a human. (78) a. alamỹ mata-raa-Ø-nã banana eat-1sg-sm-decl ‘I ate banana.’

b. alamỹ mata-raa-sã-nã banana eat-1sg-nsm-decl ‘I ate banana.’

922

Bernat Bardagil

(79) a. to-lo-pa-nĩ leave-fut-1sg-sm ‘I will leave.’

b. to-lo-pa-sã leave-fut-1sg-nsm ‘I will leave.’

For certain in-law kinship relations between men and women, the sameness form in (78a) is used (e.g., nĩ) when otherwise the mismatch form would be used. This indicates that the addressee and the speaker are not just one man and one woman but that they are related for instance as brother- and sister-in-law.

8.3 Evidentiality Two of the suffixes in the Mỹky verbal word are described as carrying evidential semantics. These correspond to the suffix in slot 4, with negation and some evidential values, and suffix 7, with a visual/nonvisual distinction (Monserrat & Dixon 2003). Slot 4 can host four suffixes, which are omitted in absence of the encoded values. The four suffixes are: -(r)äära negation, -maka reportative (“someone said that X”), -aka inferred (“it appears that X”), and -hé ~ -étiro speculative (“it is likely that X”). Thus, in this system negation occurs in a paradigmatic opposition with reported and inferred evidentiality (80) through (82). (80) alỹ-maka-Ø die-rpt-3sg ‘It is said that he just died.’ (81) käja= kao-aka-Ø sweet= lots-infr-3sg ‘It appears to be delicious.’ (Monserrat 2010: 60) (82) maromỹ mänã-eetiro-o-Ø morning rain-spcl-fut-3sg ‘It will probably rain tomorrow.’ (83) ypy-lee-nee-pja eat.meat-neg-hab-1sg ‘I don’t eat meat.’ (Monserrat 2010: 56) Slot 7 corresponds to three paradigms of subject suffixes (Section 6.1): two of them for declarative mood (paradigm i for present and recent past, shown in (84), and

19 My˜ky

923

paradigm ii for remote past, as in (85)), and one paradigm for irrealis – paradigm iii for future, interrogative, and imperative, demonstrated in (86). Paradigms i and ii encode a distinction between visual (the speaker sees it) and nonvisual (the speaker does not see it) categories, as can be seen in (84) and (85). (84) jamã-pju-si ka’a give-3pl.obj-2sg.sbj.prs.vis arrow ‘You have given them an arrow (I saw it).’ (Monserrat 2010: 43) (85) jontje xa-taka-meehmjũ-nã yesterday return-dir-2du.sbj.pst.nvis-decl ‘You two returned yesterday (I didn’t see it).’ (Monserrat 2010: 49) (86) kỹntapy kuraatu-nỹ o-o-pa one corn-def plant-fut-1sg.irr ‘I will plant the corn by myself.’ (Monserrat 2010: 172) Slot 7 presents an impoverished exponence of visual evidentiality for non-second persons. That is, only in the second-person subject cross-reference paradigm is the visual/nonvisual distinction across the board. For first person, the distinction is never indicated. For third person, the distinction is only kept in the remote past paradigm. The paradigm for visual/nonvisual evidentiality of second person subject cross-reference in slot 7 is repeated from Table 19.11 in Table 19.17. Tab. 19.17: Second person subject agreement (Monserrat & Dixon 2003: 239). par. I

par. II

subject

visual

nonvisual

visual

nonvisual

2sg 2du 2pl

-si -meexi -määxi

-hmĩ -meehmĩ -määhmĩ

-xu -meexu -määxu

-hmjũ -meehmjũ -määhmjũ

The sentences below illustrate the morphological distinction in the paradigms in Table 19.17 between visual and nonvisual evidentiality categories, in (87) and (88), respectively. (87) jamã-pju-si ka give-3pl.obj-2sg.sbj+prs+vis arrow ‘You gave them an arrow (the speaker saw it).’ (M&D: 239)

924

Bernat Bardagil

(88) matosi manã-Ø-meehmĩ monkey kill-3sg.obj-2du.sbj+prs+nvis ‘You two killed a monkey (the speaker didn’t see it).’ (M&D: 239) Mỹky thus presents two different evidentiality systems, one for visual/nonvisual specifications and a separate one for reported or inferred evidentiality specifications.

8.4 Other verbal inflectional categories The semantic scope of some semantic categories that are indexed in the Mỹky polysynthetic verb is not completely clear in the current analysis of the language, and different verbal categories are brought under the umbrella of different aspects in Monserrat (2010: 7). Suffixes -so singular and -saka plural indicate a state that is presented as completed, whole. The Mỹky completive is mostly encountered with stative verbs, as in (89a), but it can also occur with active verbs, as in (89b). (89) a. tyypy-saka-re fat-compl.pl-1pl ‘We are completely fat.’ (Monserrat 2010: 90)

b. ãnka-so-tu-hmĩ eat-compl-res-2sg ‘You ate it completely.’ (Monserrat 2010: 90)

Iterative -tja indicates repetition of the action denoted by the verb (90, 91). (90) ãnka-tja-lo-pa eat-iter-fut-1sg ‘I will eat again.’ (91) inỹ-so-tja-ku-ti? louse-compl-iter-pfv-2.int ‘Are you full of lice again?’ (Monserrat 2010: 97) There is a grammatical category that marks noun phrases and also the verb complex, the identificative. The function of the identificative is unclear, but the term is borrowed from Monserrat (2010). A series of nominal suffixes previously thrown together into the identificative category (Monserrat 2010: 115) are here analysed separately as nominal morphology indexing gender (§ 6.3).

19 My˜ky

925

Tab. 19.18: Other inflectional categories of the verb. suffix

category

so~saka tja hu, ri nãtã~ntã~nã

completive iterative identificative declarative

Active wara lipi~lika ti, tamã äty~lety owy~olee

uninterrupted permansive frustrative non-intentional durative

Stative ro~raka to

change of state completive

(92) are upa-nỹ-kĩnã xĩp-jo-hu 1sg snake-def-dat be.afraid-stat.imp-ident ‘I’m afraid of the snake.’ (Monserrat 2010: 120) (93) taka-jawu-ri-ky know-wrong-ident-nmlz ‘The one who knows wrong.’ (Monserrat 2010: 104) It could be imagined that the identificative verbal suffix may index the gender of an argument, in what would be a consistent parallel to the revised view of a subset of suffixes previously considered identificatives as actually indexing gender. There is, however, no evidence for such connection. As seen in (94) and (95), the identificative -hu is present both in the case of a masculine argument, the man in (94), and of a feminine argument, Juru in (95). (94) mija jee-jo-hu-Ø kao man happy-psych.ipfv.sg-ident-3sg very ‘The man is very happy.’ (95) Juru xĩp-jo-hu Juru be.afraid-stat.imp-ident ‘Juru is afraid.’ (Monserrat 2010: 120)

926

Bernat Bardagil

The suffix -nãtã~​-ntã~​-nã, in the very last slot of the verbal complex, is often present in declarative sentences, but absent from interrogative, imperative or exclamatory sentences as seen in (96) and (97). It is thus hypothesized that this morpheme marks declarative sentence type. (96)

poku jamã-pju-Ø-nã bow give-2pl.obj-3sg-decl ‘He gave you the bow.’

(97)

nãmyj to-mũ-nãtã woman leave-3pst-decl ‘The woman went away.’

8.4.1 Active verbs The uninterrupted suffix -wara indicates an ongoing activity with a conscious intent of the agent for the action not to be interrupted, as in (98) and (99). (98)

anỹ o-kare-mta-wara-Ø ã 3sg.pro go-hunt-nr.fut-unint-3sg there ‘He will go hunt non-stop over there.’ (Monserrat 2010: 82)

(99)

ãnka-wara-ra wat-unint-1sg ‘I eat without stopping.’ (Monserrat 2010: 82)

The permansive suffixes -(l)ipi singular and -(l)ika plural indicate a frequently occurring activity, as in (100). (100) mãkanã-si pase-ipi-maa-karee-ju tinamou-f sing-perm-nvis-pst-3sg ‘The tinamou was always singing.’ (Monserrat 2010: 83) Frustrative -ti and -tamã indicate that an event did not occur, in spite of an intent for the desired outcome to take place. These are exemplified in (101) and (102). Frustratives are a common category in languages of Amazonia, as described by Overall (2017).

19 My˜ky

927

(101) weetamã kãnka-ỹ-ti-Ø-maa-karee-ju centipede bite-prog-frust-3obj-infr-pst-3sg ‘The centipede almost bit him, they say.’ (Monserrat 2010: 125) (102) pamã-tamã-Ø-maa-karee-ju karyy catch-frust-3obj-infr-pst-3sg anaconda ‘The anaconda almost caught him, they say.’ (Monserrat 2010: 84) The non-intentional suffix -äty ~ -lety is used to indicate an activity that takes place in a haphazard way, without an existing goal, seen in (103) and (104). (103) miatapa kare-äty-Ø mänã-n-kjehy fish walk-nint-3sg water-def-loc ‘The fish goes here and there in the river.’ (Monserrat 2010: 85) (104) pase-lety kare-na-määpi sing-nint walk-hab-1pl ‘We kept singing as we walked.’ (Monserrat 2010: 85) Durative -o(wy)~-o(li) is used for a persistent activity whose attainment spans over a period of time, sometimes with an implication of an excessive amount of time, as (105) and (106) show. (105) Mãty-xi-nỹ kare-owy-mỹ Mãty-m-def walk-dur-3sg ‘It takes Mãtyxi time to walk.’ (Monserrat 2010: 86) (106) tãĩ-owy-mỹ write-dur-3sg ‘He takes some time to write.’ (Monserrat 2010: 86)

8.4.2 Stative verbs Mỹky also presents a category, indexed by the suffixes -ro singular and -raka plural, that attaches exclusively to verbal roots with no internal gradiency, seen

928

Bernat Bardagil

in (107) and (108). It is most likely a morpheme that denotes a change of state operation from atelic to telic verbs. (107) taka-ro-ra-nĩ know-tel.sg-1sg-sm ‘I’m learning.’ (Monserrat 2010: 89) (108) päja-raka-mä wake-tel.pl-3pl ‘They are awoken.’ (Monserrat 2010: 88) There is also a completive category that combines exclusively with stative verbs. It is expressed with -to, which is homophonous with the verb to ‘to go.’ It indicates that a state or process has reached its completion and is usually translated with adverbial expressions such as ‘completely’ or ‘entirely’ in Portuguese; see (109). (109) papyj ma kyty-to-mỹ hammock big black-stat.compl-3sg ‘The big hammock is completely dirty.’ (Monserrat 2010: 90)

8.5 Verb serialization Like many Amazonian languages, Mỹky presents verb serialization. In these constructions, two or more verbal roots are juxtaposed, as in (110). The serialization of verbs results in what is accomplished in non-serializing languages by different strategies, such as use of adpositions or adverbs. (110) kataty-nỹ apuku-to-Ø macaw-def fly-go-3sg ‘The macaw flew away.’ In Mỹky serial verbs, two or more verbal roots are juxtaposed, and only the last one carries verbal inflection, as in (111) and (112). (111) kuakua uta-xa-ja-Ø chicken catch-come-1sg-3sg ‘I brought the chicken.’ (Monserrat 2010: 143)

19 My˜ky

929

(112) ki-mjã sẽĩ itjuku-aa-ti-nĩ? what-q 2sg.pro sleep-see-2sg-sm ‘What did you dream about?’ Serial verb constructions are highly productive in Mỹky, being quite frequent in spontaneous speech as observed in our corpus. Even though the amount of verbs in a series is most commonly two, series consisting of more verbs are not uncommon; see (113). (113) to-tã-awa-xa-ki-nĩ go-fill-carry-come-imp-sm ‘Go get water.’ (Monserrat 2010: 131)

8.6 Nominalization Mỹky exhibits a single nominalization mechanism, albeit a highly productive one. The morpheme -ky acts as an agent-oriented nominalizer, deriving nouns from verbal roots. Nouns thus derived can take at least some of the inflectional morphology available to inflected verbs, such as the plural object morpheme in (114). (114) Jamãxi tã-taka-pju-ky Jamãxi write-know-obj.pl-nmlz ‘Jamãxi is a professor.’ (Monserrat 2010: 104) As seen in the sentences in (115), nouns formed via nominalization can function as predicate heads. This is in fact a frequent strategy, especially among younger Mỹky speakers (Monserrat 2010: 55). (115) a. patãnka ãnka-ky ajawai, apewi maäjanã ãnka-ky monkey eat-nmlz leaf flower also eat-nmlz ‘The monkey is a leaf-eater and a flower-eater.’ (Monserrat 2010: 55) b. uapa jepte panã-ky-nee-mjũ child lots cry-nmlz-hab-3sg.prs ‘Those children are big criers.’ (Monserrat 2010: 55)

930

Bernat Bardagil

9 Clauses Mỹky clauses consist minimally of the verb complex, with null anaphora of argument phrases. Despite a preference for SOV constituent orders, other combinations are not infrequent. Inflectional categories that operate at the level of the clause are case, not as extensive morphologically as other categories but nonetheless present with an accusative alignment, and negation, with a dedicated morpheme in the verb complex. These are described in the present section, as well as other clause-level phenomena such as interrogation, coordination, and subordination.

9.1 Case Mỹky has no dedicated inflectional morphology for grammatical case. However, there is a strategy for marking accusative case on core arguments. In Mỹky, the internal argument of transitive verbs can sometimes be case marked with a diminutive morpheme -nti~-si~-xi, which doubles as the accusative case morpheme for lexical noun phrases. In such cases, it carries no diminutive semantics. In intransitive verbs, the single argument is unmarked for case, as in (116a) and so is the external argument of transitive verbs, shown in (116b) and (116c). The case marking alignment of Mỹky is thus accusative, with a cross-linguistically common pattern of marked accusative. (116) a. kaso-lu itjuku-mỹ dog-f sleep-3sg ‘The dog slept.’

b. kaso-lu kãnka-lỹ-nã moojamã-nti dog-f bite-vis-decl peccary-dim ‘The dog bit the peccary.’

c. moojamã kãnka-lỹ-nã kaso-lu-nti peccary bite-vis-decl dog-f-dim ‘The peccary bit the dog.’ Given the evidence available, the Mỹky diminutive suffix appears to be in a process of grammaticalization toward a morphological exponent of accusative case. However, its use as a diminutive is still productive, demonstrated in (117). (117) a. opyri ‘tapir’ c. opytxi ‘tapir’s cub’

b. ini ‘house’ d. inxi ‘small house’

It is possible that the optionality of morphological marking of accusative constituents in Mỹky serves a function of disambiguating between third-person arguments in the depicted action. The usual lack of accusative morphology in sentences that depict an action for which the role of the arguments is not ambiguous, such as the killing event in (118), supports this notion.

19 My˜ky

931

(118) juna-li ulipa-na opyjri jaguar-m kill-3sg tapir ‘The jaguar killed the tapir.’ Personal pronouns are not marked for case, although the cross-reference morphology in the verb complex has a dedicated slot for objects (slot 2) and one for subjects of both transitive and intransitive verbs (slot 7). The verbal cross-reference system of arguments is thus oriented to a nominative-accusative alignment.

9.2 Vocative Besides the marking of core cases, Mỹky has a vocative that is used on a noun when naming an addressee. The vocative form is derived by nasalization of the last vocalic segment of the word and lowering of high vowels to mid vowels, presented in (119). (119) V → Ṽ[–high] Note that the lack of a low dimension in the nasal inventory results in raising of /a/ to [ə̃] when nasalized (Section 4.1.2). Nominal morphology, such as gender or definiteness suffixes, are very often absent in the vocative. Some examples of Mỹky vocative case are given in Table 19.18. Tab. 19.18: Mỹky vocative. base are jeewa-li Tapura Nãpu-li juna-li

vocative ‘my namesake’ (proper name) (proper name) ‘jaguar/dog’

Are jeewa-lẽ! Tapurã! Nãpõ! junã!

9.3 Negation In Mỹky, negative predication is expressed with the suffix -pu. This suffix, when added to a noun, negates the presence or existence of the entity denoted by the noun. It is also used more generally to negate non-verbal predication, as in (120). (120) a. are patajãnta-pu 1sg car-neg ‘I’m car-less.’

b. opyjri jamã-pu manã-Ø-pa tapir small-neg kill-3sg.obj-1sg ‘I killed a big tapir.’

Mỹky exhibits three values of clausal negation, all of them expressed morphologically by suffixes in the same slot on the verb complex. The negative suffix appears

932

Bernat Bardagil

in slot 4, where it appears in complementary distribution with one of the two evidential categories in Mỹky, namely the one for inferred and reported evidentiality (Section 8.3). Standard negation is expressed morphologically with one verbal suffix, -(r)äära, with the Manoki variant -lee(ra), shown in (121). (121) a. xa-räära-mĩ come-neg-3sg ‘He didn’t come.’

b. aa-räära-namää-kja-määpi mänã-i-nỹ see-neg-hab.pl-rem.past-1pl river-m-def ‘We didn’t know the river.’ (Monserrat 2010: 148)

c. pali-leera-pa-nãtã laugh-neg-1sg-decl ‘I didn’t laugh.’ (Monserrat 2010: 62) The Mỹky verb has a morpheme for future-oriented negation, -ääto. In these clauses, negation takes scope over the future reality of an event, illustrated in (122). (122) are to-paka-ääto-pa 1sg go-bathe-neg.fut-1sg ‘I’m not going to bathe anymore.’ (Monserrat 2010: 61) Yet a third negative morpheme, -ää, targets the habituality exponed in slot 5 and negates it, as in (123). (123) taka-pu-so-ää-na-määpi know-neg-cpl-neg-hab-1pl ‘We don’t usually forget.’

9.4 Interrogation In Mỹky, polar interrogatives are distinguished from declarative clauses, other than intonation, by three interrogative suffixes in the verb complex. The interrogative suffixes vary with the tense of the clause: -ti serves as the present interrogative, -ty indicates past tense, and -wy (-li in the Manoki variety) is used with future tense. The three interrogative suffixes are exemplified with the sentences in (124) through (126). (124) päja-ro-ti? awake-tel.sg-prs.int ‘Are you awake?’

19 My˜ky

933

(125) ãnka-meepa-Ø-ty? eat-2du-pst.int ‘Did the two of you eat?’ (Monserrat 2010: 33) (126) are-kĩ-Ø-wy? 1sg-dat-3sg-fut.int ‘Is it going to be for me?’ (Monserrat 2010: 33) Content interrogatives are formed with a different mechanism. Mỹky has a question particle -mã~-mjã that can combine with certain words to form interrogative phrases, seen in (127). The same morpheme can also form polar questions, as in (128). (127) ääpa-mã o-si-ti? where-int go-2sg-prs.int ‘Where are you going?’ (Monserrat 2010: 86) (128) joo, karee-mjã anã-sa-re-jã, päätõ? hey walk-int listen-2obj-1pl-nsm grandmother.voc ‘Hey, are you walking for us to listen to you, grandmother?’ (Monserrat 2010: 86) Finally, Mỹky also has question words that form interrogative clauses; see (129). They appear together with the -mã ~ -mjã interrogative morpheme. It is not clear whether fronting of question words is obligatory, but it appears to be highly frequent. (129) panỹ-mjã xipiwi-sa-o-Ø? who-int accompany-2obj-fut-3sg ‘Who will accompany you?’ (130) kikja-mã mama pãp-pju-na-mää-ti? how-int clothes buy-obj.pl-hab-2pl-prs.int ‘How did you buy clothes?’ (Monserrat 2010: 36) Question words can also appear as the head of a finite predicate, projecting a verbal complex in the place usually occupied by verbal roots, illustrated in (131). (131) a. ki-mijã-mjã-nĩ? what-name-prs.int-sm ‘What’s the name?’

934

Bernat Bardagil

b. kimã-mã-né-ti-nĩ, kokẽ? what-int-hab.prs.2sg-prs.int-sm brother.in.law.voc ‘What are you doing, brother-in-law?’ (Monserrat 2010: 177)

9.5 Coordination and subordination In Mỹky syntax, juxtaposition serves a double function as both a paratactic and a hypotactic mechanism. Coordinate clauses are formed by juxtaposition, as seen in (132). (132) nỹhỹ ulipa-Ø-nãtã upa-nỹ ky-lỹ-Ø-nãtã 3sg kill-3sg-decl snake-def burn-prog-3sg-decl ‘He killed the snake and burnt it.’ (Monserrat 2010: 135) There are two strategies for the formation of subordinate clauses: juxtaposition and subordinating suffixes. Juxtaposition is the most common mechanism for the formation of subordinate clauses, as illustrated in (133). (133) Brasnorte-n kje to [määnĩ kjamã kõpra-ra-kĩ] Brasnorte-def loc go foot clothing buy-1obj-imp ‘Go to Brasnorte to buy me shoes.’ (Monserrat 2010: 140) There is also a series of suffixes that form subordinate clauses. These suffixes specify concessive, causative, and final semantics. A concessive subordinate clause is given in (134). (134) [mu-manã-Ø-päm] to-ra rain-water-3sg-conc leave-1sg ‘Even though it’s raining, I’m leaving.’ (Monserrat 2010: 144) The suffix -jỹ is used to form a causative subordinate clause, as in (135). (135) Kamũnũ waa-pju-Ø [ äänã aka-uta-re-Ø-jỹ ] Kamũnũ speak-1pl.obj-3sg firewood go-take-1pl-3sg-caus ‘Kamunũ told us to go get firewood.’ (Monserrat 2010: 142)

19 My˜ky

935

A purpose subordinate clause formed with -kje is shown in (136). (136) Tapura kataty uta-mỹ [tikipu opu-kje] Tapura macaw take-3sg tail pull-purp ‘Tapura caught a macaw to pull [the feathers from] its tail.’ (Monserrat 2010: 141) In this hypotactic construction, subordinate clauses can be inflected for tense relative to the tense of the main clause. In a sentence like (137), the temporal reference in the causative subordinate clause is future even though that event might have taken place at the time of utterance. (137) waa-ra-si [to-o-pa-jỹ] speak-1sg.obj-2sg leave-fut-1sg-caus ‘You told me to leave.’ (Monserrat 2010: 142)

10 Conclusions In this chapter I have strived to present our current understanding of Mỹky. As a language isolate, its properties are the single surviving instance of a unique genealogical linguistic unit. Much about the language is still unknown, but fortunately the existence of ongoing research and community-driven language maintenance initiatives means that our knowledge of Mỹky will yet progress beyond what is reflected on these pages. Mỹky presents itself as a polysynthetic language with both head and dependent marking, and it is extremely rich and complex in its TAME categories. The phonology of Mỹky, with a sizeable segmental inventory, becomes especially engaging not only because of its phonological processes and suprasegmental dimensions (vowel lengthening, palatalization, nasality, tone) but also because of the two stages before and after the loss of /l/ that can be observed synchronically in the two attested varieties. Like virtually every Amazonian population, the speakers of Mỹky have faced considerable challenges up to this very day. Loss of significant portions of their population, their territory, and their culture have severly affected the Mỹky language and the lives of the people that speak it. But both the Manoki and the Mỹky refuse to be shaped by what outsiders have tried to take from them by force. They are determined to not only survive but to thrive as a people by being mỹky, being human on their own terms.

936

Bernat Bardagil

11 Acknowledgements I want to express my gratitude to the Manoki and Mỹky communities for welcoming me with enormous generosity. I am especially grateful to my teachers of Manoki: Regina Jalapyjtasi, Luíz Tamũxi, Angélica Kamũnũ, Manoel Kanũxi, Alípio Xinũli, and Domitila Nãlu, who were also lovely informants during my fieldwork in Manoki land. I am also thankful to my jeewali and co-founder of the Watjuho Ja’a Collective, Edivaldo Lourival Mampuche Nãpuxi, as well as Edinei Kanũxi, Marta Tipjusi, Marina Kamũlu, Cláudio Tamũxi, Dário Kajoli, Maria Ilda Tipjusi, and the chief of the village of Cravari, Lourenço Janãxi. In the Mỹky village of Japuíra I thank Tupy Mỹky, Tapau Mỹky, Marikjau Mỹky, Nilo Jolasi Mỹky, Ujepai Mỹky, Typju Mỹky, Janãxi Mỹky, Mãtyi Mỹky, and Tipjuu Mỹky, as well as Elizabeth Amarante Rondon and Geraldo Abdias Lopes. I am grateful to OPAN for their logistical support, especially to Paulo Eberhardt and Marcos Ramires. Special thanks also to André Lopes for his help and intellectual stimulation. I am indebted to the editors of this volume, Pattie Epps and Lev Michael, for their support and feedback on the writing of this chapter, as well as Paige Erin Wheeler for her editing work. Several people provided me with feedback at various stages throughout the writing of this chapter: Myriam Lapierre, Zachary O’Hagan, Adam Singerman, and Hein van der Voort. Finally, I am especially thankful to Ruth Monserrat, who stood up to the titanic task of describing the Mỹky language and, with her work, made a lasting contribution to linguistics and to the Indigenous people who speak Mỹky.

12 References Aikhenvald, Alexandra. 2012. The languages of the Amazon. Oxford: Oxford University Press. Anonby, Stan. 2009. A Report on the Irantxe and Myky. In SIL Electronic Survey Reports. Dallas: SIL International. Bardagil, Bernat. 2019. Documenting Manoki (Mỹky), an isolate of Brazilian Amazonia. London: Endangered Languages Archive. Bardagil, Bernat, Cledson Dário Kajoli & Edivaldo Lourival Mampuche. 2020. Watjuho Ja’a: reclamation of language and land in the Manoki nation. Cadernos de Linguística 1(3). 1–15. Bueno, Ana Cecilia Venci. 2014. Fios de memórias: um estudo sobre parentesco e história a partir da construção da genealogia manoki (irantxe). University of São Paulo PhD dissertation. Campbell, Lyle. 2012. Classification of the indigenous languages of South America. In Veronica Grondona & Lyle Campbell (eds.), The Indigenous Languages of South America: A Comprehensive Guide. Berlin: De Gruyter. Crevels, Mily. 2002. Itonama o Sihnipadara, Lengua no Clasificada de la Amazonía Boliviana, vol. 16. (Estudios de Lingüística). Alicante: University of Alicante. Crevels, Mily & Hein van der Voort. 2008. The Guaporé-Mamoré region as a linguistic area. In Pieter Muysken (ed.), From Linguistic Areas to Areal Linguistics, 151–180. John Benjamins. Lisboa, Thomáz de Aquino. 1979. Entre os índios Munku: a resistência de um povo. São Paulo: Edições Loyola.

19 My˜ky

937

Lopes, André. 2015. O vídeo como ibirapema. A apropriação dos recursos audiovisuais pelos Manoki e seus discursos sobre a história. São Paulo: University of São Paulo MA thesis. Meader, Robert E. 1967. Iránxe: notas gramaticais e lista vocabular. Rio de Janeiro: Museu Nacional. Monserrat, Ruth. 2000. A língua do povo Mỹky. Federal University of Rio de Janeiro PhD dissertation. Monserrat, Ruth. 2010. A língua do povo Mỹky. Campinas: Editora Curt Nimuendajú. Monserrat, Ruth & Robert M. W. Dixon. 2003. Evidentiality in Myky. In Alexandra Y. Aikhenvald & Robert M. W. Dixon (eds.), Studies in Evidentiality. Amsterdam: John Benjamins. Moura, José de. 1957. Os Iranche, contribuição para o estudo etnológico da tribo (Pesquisas). Porto Alegre: Instituto Anchietano de Pesquisas. Muysken, Pieter, Harald Hammarström, Joshua Birchall, Rik van Gijn, Olga Krasnoukhova & Neele Müller. 2014. Linguistic Areas, bottom up or top down? The case of the Guapore-Mamore region. In Bernard Comrie & Lucía Golluscio (eds.), Language Contact and Documentation, vol. 1, 201–233. Berlin: De Gruyter Mouton. Nichols, Johanna. 1992. Linguistic Diversity in Space and Time. University of Chicago Press. Oberg, Kalervo. 1953. Indian tribes of northern Mato Grosso, Brazil. Washington: United States Government Printing Office. Overall, Simon. 2017. A typology of frustrative marking in amazonian languages. In Alexandra Aikhenvald & Robert M. W. Dixon (eds.), The Cambridge Handbook of Linguistic Typology, 477–512. Cambridge: Cambridge University Press. Payne, Thomas E. 1991. A classification of Maipuran (Arawakan) languages based on shared lexical retentions. In Desmond C. Derbyshire & Geoffrey K. Pullum (eds.), Handbook of Amazonian Languages 3, 355–499. Berlin: Mouton de Gruyter. Pereira, Adalberto de Holanda. 1974. Lendas dos índios Iranxe. Porto Alegre: Instituto Anchietano de Pesquisas. Piggott, Glyne. 2003. Theoretical implications of segment neutrality in nasal harmony. Phonology 20(3). 375–424. Pivetta, Darci Luiz. 1993. Processo de ocupação das dilatadas chapadas da Amazônia Meridional: Iranxe – educação etnocida e desterritorialização. Cuiabá: Federal University of Mato Grosso (UFMT) MA thesis. Rondon, Cândido M. da Silva. 1946. Conferências realizadas em 1910 no Rio de Janeiro e São Paulo (Comissão de Linhas Telegráphicas Estratégicas de Matto Grosso ao Amazonas). Rio de Janeiro: Imprensa Nacional. Rose, Françoise. 2015. On male and female speech and more: categorical gender indexicality in indigenous South American languages. International Journal of American Linguistics 81(4). 495–537. Silverstein, Michael. 1985. Language and the culture of gender: at the intersection of structure, usage, and ideology. In Elizabeth Mertz & Richard J. Parmentier (ed.), Semiotic Mediation, 219–259. Orlando: Academic Press. Singerman, Adam Roth. 2016. Nasal harmony and phonotactic well-formedness in Tupari. International Journal of American Linguistics 82(4). 453–485. Telles, Stella. 2002. Fonologia e gramática Latundê/Lakondê. Amsterdam: Free University of Amsterdam PhD dissertation. Walker, Rachel. 2011. Nasal Harmony. In The Blackwell Companion to Phonology, 1838–1865. Oxford: Blackwell Publishing. Wetzels, W. Leo. 2009. Nasal harmony and the representation of nasality in Maxacalí. In Andrea Calabrese & W. Leo Wetzels (eds.), Loan Phonology, 241–270. John Benjamins.

Zachary O’Hagan

20 Omurano 1 2 3 4 5 6 7 8 9

Introduction Omurano history Phonology Grammar Lexical comparison with neighboring languages Comparison with Tessmann Conclusion Acknowledgments References

1 Introduction Omurano (omur1241) is a language isolate formerly spoken on the upper Nucuray and Urituyacu rivers in the Loreto region of Peru (see Map). As of 2013, approximately 40 words and 15 short phrases were remembered collectively by fewer than 10 individuals born and/or resident, with one exception, on the Urituyacu, all of whom speak Urarina as their language of daily life. These rememberers were born between the early 1940s and the 1980s, and, among them, the closest consanguineal relative that was a dominant Omurano speaker was a grandparent. (In the case of one rememberer, the relevant dominant Omurano speaker was a mother-in-law, whom the rememberer never met – see Section 2.2.) In the precolonial period, linguistic neighbors of Omurano included Urarina, Kandozi, and Taushiro, all isolates. This chapter is based on linguistic, historical, and ethnographic data collected by the author on two exploratory field trips to the Urituyacu in 2011 and 2013,1 as well as on a wordlist of approximately 250 items recorded by Tessmann (1930: 444–458), the sole prior documentation of the language. Notably, Tessmann (ibid.: 444) baselessly asserted that Omuranos were known by the Jesuits as Roamainas, which seems to derive from a misunderstanding of Hervás y Panduro (1784: 61) and the Jesuit notion of a lengua matriz, by which the latter claims that Omurano and Roamaina are related to Maina (Kandozi). This information was supplied by Fr. Juan de Velasco (see below), whose linguistic classifications exhibit numerous errors. While Tessmann’s assertion influenced later scholars (e.g., Villarejo 1953), I do not assume

1 The materials, including field notes and audio recordings, that derive from these trips are publicly accessible online through the California Language Archive: http://dx.doi.org/doi:10.7297/ X29K488M. Zachary O’Hagan, Department of Linguistics, University of California, Berkeley https://doi.org/10.1515/9783110432732-007

940

Zachary O’Hagan

20 Omurano

941

it here, noting that contemporary Jesuits understood Roamaina to be related to Pinche, that is, Taushiro (e.g., Maroni 1988[1738]: 265).

2 Omurano history 2.1 The eighteenth and nineteenth centuries In 1638, the Society of Jesus began missionary activities in lowland Peru in an area that corresponds roughly to the modern-day region of Loreto, and which was known as the Provincia de Maynas. Neither the term Omurano or its orthographic variants are mentioned in the fairly extensive Jesuit records until 1743, in a letter written by Fr. Juan Magnin (1701–1753) (Jouanen 1943: 737) describing the first expedition that explored the Nucuray River (transcribed in Cipolletti 2017: 207–225).2 This expedition was led by Pedro de Ortega y Murga in July 1742, at the behest of a Roamaina man named Gonzalo, who had recently been banished from the mission of Santo Tomé de Andoas (Pastaza River) for inciting the flight of a few dozen Roamainas from that mission some three years earlier.3 During their flight, this group of Roamainas was massacred by Mainas from the headwaters of the Nucuray. Only Gonzalo escaped, his own son having been taken captive. Arriving in Borja, in short, Gonzalo seems to have used Ortega y Murga to exact revenge on the Mainas that killed his people. Ortega y Murga’s expedition, which came together in full at the mouth of the Huallaga River, consisted of nine viracochas, 10 Kandozis (Mainas), 30 Shawis (Cahuapanas), 50 Shiwilus (Jeberos), and Gonzalo with a Simigae companion. With the capture of Kandozi prisoners near an abandoned village some four days’ travel from the mouth of the Nucuray, the expedition was told eight additional Kandozi chiefs (Sp. caciques) upriver, as well as four chiefs of Umuranas [sic]. The expedition never had direct contact with Omuranos. It is evident that, in the years following 1742, while the Jesuits came to understand the extent of Omurano territory better, they were less certain of their relationship to other ethnolinguistic groups. Fr. Pietro Esquini (b. 1717) (Jouanen 1943: 731), for example, in his 1760 numeración of the Maynas missions, states in the context of describing the mission of San Xavier de Urarinas (Marañón River), “There are between the headwaters of the Nucuray and Lorito Yaco the Umuranas.” Two years later, Fr. Franz Xavier Veigl (1723–1798) (Jouanen 1943: 749), in his numeración, states in the context of San José de Pinches (Pastaza River) that “toward the Marañón there is an unknown nation that the Roamainas and Urarinas are familiar with.

2 I thank Joshua Homan for making me aware of the transcription of Magnin’s letter by Cipolletti. 3 This group may correspond to the 20 families discovered on the Capirona in 1737 (Maroni 1988[1738]: 266).

942

Zachary O’Hagan

It is assumed they are Mainas.”4 Regarding San Xavier de Urarinas, he states, like Esquini, that Omuranos live between the headwaters of the Nucuray and Urituyacu, adding that they are a subgroup (Sp. parcialidad) of Mainas. While Ortega y Murga’s expedition makes it clear that Omuranos lived in relatively close proximity to Mainas, linguistic evidence presented in Section 5 makes it clear that Omurano is not related to modern-day Kandozi. Fr. Juan de Velasco (1919: 252, translation mine), based on information supplied to him by exiled Jesuits in Italy in the 1770s and 1780s, reports that in 1767 Fr. Antonio Segundo del Castillo (1736–1781) (Jouanen 1943: 728) “conquers the savage Umuranas.” Hervás y Panduro (1784: 59), with information supplied by Velasco, mentions a “new community” (Sp. pueblo nuevo) of Omuranos. Elsewhere Velasco refers to this settlement as Humuranas de Uritoyaco [sic] (1842[1789]: 230). It seems that Jesuits had only just begun to have a sustained presence on the Urituyacu at the time of their expulsion, not yet having established a permanent mission. Jouanen (1943: 536), an authorative Jesuit historian, for example, does not list a mission on the Urituyacu at the time of the expulsion. Subsequent mentions of Omuranos are relatively infrequent. Requena y Herrera (1994[1785]: 660) mentions the presence of some “Uritos” (presumably Omuranos) in the Urarinas mission. And Vásquez Caicedo (1905[1843]: 270, translation mine), speaking of the Urituyacu, claims that “on this [river] are the heathen Umuranas [sic] at more than 500.”

2.2 The twentieth century The Urituyacu basin of the early 20th century was a zone of intense interaction between Omuranos, Urarinas, Taushiros, Kandozis, and non-indigenous outsiders, particularly as part of a system of debt peonage that originated during the Rubber Boom. By this time, the Omurano population had been severely reduced, and those who remained had begun to intermarry with the more numerous Urarinas, who had begun migrating to the Urituyacu as early as the 19th century. As a result, Urarina is today the dominant language of the Urituyacu, and individuals exhibit different degrees of Omurano ancestry. Locally, however, Omurano ancestry is denigrated. As a consequence, the reconstruction of regional history relies on stitching together various accounts that differ in detail and that are often obscured by pejorative stereotypes of Omuranos, including their lack of clothes, salt, and cooked meat, as well as their fear of outsiders. The first published mention of Omuranos from this period is Tejedor (1927: 66), who enumerates only twenty one Muranas [sic], fourteen men and seven women.

4 These numeraciones are unpublished manuscripts 1279 and 1296, respectively, in the Archivo Colonial de la Compañía de Jesús, held by the Biblioteca Ecuatoriana Aurelio Espinosa Pólit in Quito.

20 Omurano

943

Similarly, Tessmann, who visited the Urituyacu in January 1925, enumerated thirteen men, eight women, and three children (1930: 444). Tessmann relates that Omuranos lived across from the mouth of the Sungaroyacu [sic] – a left-bank tributary known today as the Yanayacu – which is consistent with oral history. He goes on to add that Omuranos were in conflict with Kandozis, who occupied the headwaters of the Bohorceyacu [sic] – perhaps the modern-day right-bank tributaries Huacana or Pintuyacu – and with Pinches, that is, Taushiros, who occupied the headwaters of the abovementioned Sungaroyacu. The latter observation is consistent with that of Tejedor (ibid.), who claims that Pinches occupied the mouth of the Yanayacu, but it is inconsistent with what is known of Taushiro territory of the early twentieth century (see O’Hagan, this volume). Finally, Villarejo (1943: 101), whose information seems to have come from his own travels, lists eight families resident on the Urituyacu.5 Prior to 1925, the lower Urituyacu was controlled by the patrón José Dosantos Gonzales [sic], who occupied an outpost known as Omuranas (Tessmann 1930: ibid.), near modern-day Cafetal. The mixed Urarina-Omurano families that resided near this outpost appear to have been located west of the Urituyacu on the streams and lakes that reach toward the headwaters of the Patuyacu River, which was formerly used as a route connecting the Urituyacu with Santo Tomás, an important settlement on the Marañón (see below). The mixed nature of these families is exemplified by an ancestor of many Urituyacu residents, Bautista Macusi Ahuite (d. c1979): it is said that Macusi’s father was Urarina, while his mother was an Omurano from an upriver settlement near the mouth of the Yanayacu, and that he was born near Cafetal. Urarina appears to have been the dominant language in this downriver region, with children receiving Omurano input from their mothers. More generally, there appear to have been marriage practices that involved downriver men, who experienced significant contact with non-indigenous outsiders and Urarinas, taking women of isolated upriver settlements as spouses. By 1945, the lower Urituyacu was controlled by patrón Rafael Zubiate (c1895–1974), and no Omuranos remained in the Yanayacu region.6 It is likely that they succumbed rapidly to measles in the years immediately preceding 1945, a point at which outsiders appear to have encroached to a greater degree than before into the upper Urituyacu.

5 Between 1943 and the second edition of his book (Villarejo 1953: 149), this author seems to have gotten access to Tessmann (1930), as he adds information about Omuranos that seem to be gleaned from Tessmann, for example, changing the number of families, mentioning the same tributaries of the Urituyacu that Tessmann does, and perpetuating Tessmann’s unsubstantiated assertion that Omuranos are Roamainas. 6 Some today speculate that isolated Omurano settlements may have persisted upriver of a waterfall that separates the headwaters of the Urituyacu from all extant communities. In fact Omuranos are rumored to have been the guardians of a fifteen-meter-high golden waterfall in this region which disappeared when exclusively Omurano settlements disappeared, perhaps in the late 1930s or early 1940s.

944

Zachary O’Hagan

This encroachment is exemplified by the Valencia Paima family, which immigrated to the region in 1945 and which in 1949 was the victim of a Kandozi raid at Collpayo, a site located between the Yanayacu and San Juan de Abejaico. The Valencias fled downriver to Triunfo, where they ultimately seized Kandozi weapons and destroyed them in the river. Either by 1949 or shortly thereafter, a small group of Kandozis settled at the former Omurano site at the mouth of the Yanayacu and became involved in the local system of debt peonage. Unfortunately for Omuranos, the presence of both Kandozis and non-indigenous outsiders in the upper Urituyacu only led to further strife, in particular in the form of a massacre (probably one of many) that took place across from Triunfo circa 1953. By this time, Zubiate had been replaced at Triunfo by patrón Andrés Meléndez, who had previously been located on the Nucuray, and who, upon his departure, had kidnapped a Kandozi woman. As revenge, Kandozis raided the mixed Urarina-Omurano settlement at dawn, unmooring canoes that could be used for escape, murdering some twenty to forty individuals, and retrieving the Kandozi woman. During their retreat along the Huacana, the Kandozis kidnapped two unsuspecting women whose husband was away in the forest, and they have never been seen again. In response to these killings, a moment of extreme and well known social upheaval on the Urituyacu, many Urarina-Omurano families fled to Santo Tomás (now Monterrico), where many of their descendants still reside. Based on a brief exploratory trip to the Urituyacu conducted by Gerhard Fast of SIL International, Ribeiro and Wise (1978: 151) claim that all Omuranos perished in epidemics of measles and whooping cough between 1953 and 1958. It is clear, based on the decimation of the upriver Omurano population before 1945, that disease had riddled the Urituyacu throughout the 20th century, and it continued to do so through the 1970s. Probably the most devastating of recent outbreaks occurred in 1976, when measles was again brought to the Urituyacu from Iquitos. It is important to highlight, however, that disease did not actually eradicate Omurano speakers, as exemplified by individuals like Macusi, who died after this period, and that fluent Omurano speakers survived probably until the late 1990s or early 2000s.

3 Phonology 3.1 Segmental inventory Omurano exhibits ten consonant phonemes (Table 20.1). No fricative or velar consonants are present in the attested data. Omurano exhibits five phonemic vowel qualities (Table 20.2), with nasal counterparts attested only for the high front vowel. Vowels in the initial syllable of disyllabic roots uttered in isolation are often lengthened, but length is not analyzed as phonemic because it disappears when the accusative suffix -ta is added to these roots.

20 Omurano

945

Tab. 20.1: Omurano Consonants.

stop nasal affricate flap

bilabial

alveolar

p m

t n

b

alveopalatal

palatal

ɲ ʧ

ɾ

lateral glide

l j

Tab. 20.2: Omurano Vowels. front high mid low

central

i, ĩ e

back u o

a

3.2 Phonotactics Attested syllable shapes are V, CV, and CVV, with the most frequent being CV. Onsetless syllables are attested root-initially and -internally, but root-internally the only attested onsetless syllable is /a/. CVV syllables are restricted to the only attested diphthong, /au/. Two, three, and four-syllable roots are attested, but no one-syllable roots, suggesting a disyllabic minimum word requirement.

3.3 Postlexical phonological processes Vowel hiatus resolution is attested between the accusative suffix -ta and a following vowel-initial verb root. Two strategies to resolve this hiatus are attested: one involves the deletion of the vowel of the accusative suffix altogether; the other involves the raising of /a/ to /e/, a sort of coalescence (see footnote 7). The voiced bilabial stop /b/ spirantizes to [β] before /e/.

3.4 Prosody Omurano exhibits two surface tones, high and low. Attested tonal patterns per syllable count in roots are summarized in Table 20.3. Given the small amount of data on which this chapter is based, it is not possible to provide a thorough, empirically grounded analysis of Omurano tone. However, some tentative claims can be made. First, it seems that some roots are not specified

946

Zachary O’Hagan

Tab. 20.3: Attested Tonal Patterns in Roots. syllables

tonal patterns

σσ σσσ σσσσ

LH, HL7 LLH, LHL, HLL LLLH, LHLH

for tone. Given the absence of an attested all-low root, I consider such roots to be assigned a default final high tone. This is evident in the behavior of nouns that receive a single final high tone in isolation, and also a single final high tone when suffixed with the accusative suffix -ta (e.g., [pùɾìmá], [pùɾìmàtá] ‘agouti’). It follows, then, that nouns that in isolation receive a high tone on a non-final syllable must be specified for high tone (e.g., /napíʧu/ → [nàpíʧù]). The behavior of tone at morpheme boundaries is only visible in the accusative suffix and the first-person possessive prefix nàùɾú-. That nàùɾú- is high-toned follows from its behavior in combination with the toneless root /opi/ ‘canoe’, where ‘my canoe’ surfaces as [nàwɾǔːpì], with a long vowel and a rising tone. This suggests that a high tone may spread one syllable rightward at a morpheme boundary onto a toneless syllable, and from there all remaining syllables receive low tone.8 Omurano tone, however, also exhibits stress-like properties, most notably the reduction of vowels in low-toned syllables (e.g., [ɾáwǝ̀nǝ̀] ‘child’). Unfortunately little more can be said of Omurano prosody based on extant data.

4 Grammar Given the random nature of attested short phrases of Omurano, only a small number of grammatical domains of the language can be described. These are word order and grammatical relations (Section 4.1), imperatives (Section 4.2), and nominal possession (Section 4.3).

4.1 Word order, grammatical relations, and person marking Basic word order of transitive clauses appears to be AOV (1),9 and verbs do not seem to be crossreferenced for the person of their arguments. A single suffix -ne appears

7 HL is attested only once, in the likely loan pápà ‘father’. 8 On this view the surface form [nàwɾǔ:pì] must involve two adjacent vowels /u/. 9 One token of OVA is attested, but this is interpreted as information-structurally marked. It is worth noting that this is the basic word order of Urarina (Olawsky 2005).

20 Omurano

947

on all attested verb stems regardless of the person of the subject or object, and its meaning is unknown. Verbal arguments are expressed, then, solely via either independent pronouns or nouns, with verb phrases lacking an overt subject argument being interpreted with second-person subjects (1c).10 (1)

a. [màɾàʧí pùɾìmàtá pètèné] maɾaʧi puɾima-ta pete-ne dog agouti-acc kill-? ‘The dog killed an agouti.’ b. [ɲùɾú pùɾìmàtá pètèné] ɲuɾu puɾima-ta pete-ne 1.pro agouti-acc kill-? ‘I killed an agouti.’ ~ ‘I am going to kill an agouti.’ c. [bùrú, ʧànètá pètèné] buɾu ʧane-ta pete-ne? old.man collared.peccary-acc kill-? ‘Old man, did you kill a collared peccary?’

Unlike the subjects of transitive and intransitive verbs, the direct object of a transitive verb is always marked with the suffix -ta, regardless of its information-structural status, which suggests that it is a marker of accusative case, and by extension that Omurano is a nominative-accusative language.11

4.2 Imperatives Imperative instantiations of transitive verbs differ from their declarative counterparts only in that they lack an overt subject argument.12 (2)

́ [màmàtá ījàné] mama-ta ĩja-ne! fish-acc eat-? ‘Eat the fish!’

10 In this way such phrases are structurally identical to imperatives. 11 Four variants of a single intransitive sentence consisting of the verb ‘be good’ are attested, with both VS and SV orders, the difference between which is not understood. 12 The interpretation of (2) is of a singular participant. It is not known whether an interpretation of plural participants is possible.

948

Zachary O’Hagan

4.3 Nominal Possession Nominal possession is attested in two instances of a first-person possessor, with two distinct nominal prefixes, ɲùɾù- and nàùɾú-. The former is attested in conjunction with the word pápà ‘father’, the latter with the word òpí ‘paddle’ (see Section 3.4). That ɲùɾù- is not an instance of the independent pronoun is evidenced by the fact that no high tone surfaces on the second syllable (cf. [ɲùɾùpápà]). The distribution between these two prefixes is not further understood; perhaps it reflects an alienability distinction.

5 Lexical comparison with neighboring languages Omurano has been claimed to be part of a number of genealogical groupings. Among others, Tovar (1961: 151) classified it with Pinche (i.e., Taushiro); Ribeiro and Wise (1978: 151) suggest that it might be Zaparoan, as does Wise (1985: 208); Kaufman (1994: 63) classified it with both Taushiro and Kandozi. However, Loukotka (1968: 157) considers it an isolate, as does Carvalho (2013: 108). The purpose of this section is to list the majority of lexical data collected on Omurano alongside corresponding data from Taushiro, Urarina, Kandozi, and Achuar, in order to argue that Omurano is indeed best considered an isolate. These data are summarized in Table 20.4.13 Three potential loans involving Urarina are boldfaced. Note that the p:h correspondence in the word for ‘blowgun’ suggests that modern-day Urarina h may derive via debuccalization of *p. Tab. 20.4: Lexical Comparison with Neighboring Languages. meaning

Taushiro

young man

Omurano

Urarina

Kandozi

Achuar

ɾáwànà

enamanaː

kanugaasi

nátsa

14

lazy person

ʧàtàné

wamiŋkaanuɾi

náki

scabied person

wìʧúmà

ʃaapi (sarna)

mámu (sarna)

iŋkɾisa

ìŋkis

white person

nìjí

nàʧúɾì

old man

jèjú

bùɾú

biːna

water

wéì

ʧùá

akaʉ

weámɾau (viejo) kuŋku

júmi

13 The sources of the lexical data in this table are as follows: Taushiro from the author’s field notes (see O’Hagan, this volume) and Alicea Ortiz (1975); Urarina from Olawsky (2006) and Epps (n. d.); Kandozi from Overall (p.c.) and Tuggy (1966); Achuar from Fast et al. (1996). 14 The form ʧàtánè is also attested.

20 Omurano

Tab. 20.4 (continued) meaning

Taushiro

Omurano

Urarina

Kandozi

Achuar

manioc beer

ahjãnehoke

ʧùá

baɽʉe

kapuʂi

hamánʧ

agouti

wì ̵ntí ̵

pùɾìmá

mami

punʧuwa

kãjū́k

paca

àjáwà

jàpú

itɕ͡ a

maʂaaʂi

káʃai

collared peccary hùjóntò

ʧàné

ubana

kaʂuuma

jaŋkipík

white-lipped peccary

tàjá

àné

ɽaːna

waŋkaana

páki

dog

wànántà

màɾàʧí15

ɽeːmae

tumuuʂi

jãwā́ã

spider monkey

àhú

bàbàné

alau

ʧuupa

wáʃi

howler monkey

wàʔná

màɾìàbé

ɽuɽu

ʂanʧiiʂi

jakúm

woolly monkey

àhúntù

lùné

aɽauata

ʧuɾu

ʧuː

deer

ùʔwéwì

àlámàɾé

ukwaːe

manʧani

hápa

16

tapir

xèhí

làùtùmé

aɽãla

pamaɾa

pamá

currasow

éìntì

tátànà

ataɽi

maʃu

máʃu

wàhìnó

nàpíʧù

kʉeːɽi

waʧuɾu

kúju

Spix’s guan

tèntá

ʧàùɾí

enʉɽi

kaɾuntsi

aúnts

fish

éìnà

màmá

ate

kajupʧi

namák

manioc

àh ā́

jùné

laːnu

kaʂinʂi

máma

guan sp.17 18

j

w

plantain

àntá

pùɾá

f anaɽa

paɾantama

pánːtam

ayahuasca

ànùʔwí ̵

ìjùné

iɲunu ~ kw aiɽi

ʂuɾuupʂi

natém

canoe

tì ̵ní ̵ntì ̵

òpí

enanihja

kanu

kánu

túnìʧà

kiha

pitʂi

kawín

paddle blowgun

ànètá

pìʧàná

hicɕ͡ ana

ʃuŋkanaaʂi

uːm

1sg

úì

nàùɾú

kanʉ

nu, nuwa

wi

15 16 17 18

The form màɾáʧì is also attested. The form làùtùmá is also attested. This term refers to the Blue-throated piping guan (Pipile cumanensis), or Spanish pava. The scientific name is Penelope jacquacu, or Spanish pucacunga.

949

950

Zachary O’Hagan

6 Comparison with Tessmann Günter Tessmann (1884–1969) was a German naturalist who in the mid-1920s traveled throughout lowland Peru, documenting the cultural practices of dozens of indigenous groups, and, to a lesser extent, compiling (sometimes lengthy) wordlists of the languages they spoke. This work culminated in the publication of Die Indianer Nordost-Perus, which to this day remains the most extensive documentation of northeast Peruvian Amazonian languages from the first half of the 20th century. However, the orthographic conventions that Tessmann used to transcribe Omurano words, and the words of all other languages he documented, are highly opaque. The purpose of this section is not to provide a complete exegesis of Tessmann’s orthographic conventions, but rather to provide evidence that the language documented in 1925 and 2013 is the same one, and to make some disparate observations that bear on the analysis and history of the Omurano language as well as Omuranos’ interactions with Urarinas. Table 20.5 lists all forms collected by the author that are also attested in Tessmann (1930). Tessmann’s orthographic representations are preserved exactly. In the rightmost column, Tessmann’s original German gloss is provided, which at times differs slightly from that given in the leftmost meaning column; however, the meanings are always closely related, and the instances in which they differ suggest a likely misunderstanding on the part of Tessmann. Tab. 20.5: Comparison with Tessmann (1930). meaning

2011/2013

Tessmann German gloss

young man

ɾáwànà

old man

bùɾú

water, manioc beer

ʧùá

tšu’á19 Maniokmassato

agouti

pùɾìmá

pṓ̱lē̱mă Aguti

paca

jàpú

yapṓ Paka

collared peccary

ʧàné

́ kjīan Halsbandpekari

white-lipped peccary

àné

aá ͜ n Weißbartpekari

dog

màɾàʧí

maldží Hund

spider monkey

bàbàné

bā́bani Spinnenaffe

white capuchin

bàbànèʧùmàné20

(ni)ðàwana Knabe bū́ṛo alt

báwani ðnán Heller Kapuzineraffe

19 See also the form ‘water’. 20 Compare this with Tessmann’s form ‘full moon’. This suggests a root /ʧuma/ relating to whiteness, although this is contradicted by his form ‘white’.

20 Omurano

951

Tab. 20.5 (continued) meaning

2011/2013

brown capuchin

bàbànèmíànè

howler monkey

màɾìàbé

woolly monkey

lùné

deer

àlámàɾé

tapir

làùtùmé

currasow 21

tátànà

Tessmann German gloss báwani mien Dunkler Kapuzineraffe imáṛiawue Brüllaffe ðṓnŭ̱ Wollaffe aa ðamáli Reh lautṓma Tapir tatn̄́la Hokko ́ napītšo Penelope jacutinga [sic]

guan sp.

nàpíʧù

fish

màmá

manioc

jùné

n̄njúna Maniok

plantain

pùɾá

púra Plante

ayahuasca

ìjùné

ídn Liane

canoe

òpí

ópi Kanu

be tasty22

ìmà(né) ~ èmà(né)

eat

23 ̀ ĩjà(né)

kill

pètè(né)

mā́ma Fisch

nimā́ hübsch njiyaní Speise napē̱tḗipṅani töten

Most generally, the place of prosodic prominence indicated by the acute accent in Tessmann’s representation is striking in comparison to the distribution of high tone, namely in that the two are frequently on different syllables. It was noted above that pápà ‘father’ is the only two-syllable word with a HL tone pattern. However, in Tessmann, eight two-syllable words show prominence on the first syllable,24 yet today all of these words show a final high tone. Furthermore, four of these eight words today end in a vowel /e/, when in Tessmann those same words either end in a different vowel (e.g., ‘woolly monkey’) or no vowel at all (e.g., ‘white-lipped peccary’). This raises the possibility that the final high-toned /e/ of these words is a productive morpheme with an as-yet unknown meaning that might have been absent in the citation forms provided to Tessmann. Indeed at least one form collected exhibits two variants across two different speakers, làùtùmé ~ làùtùmá ‘tapir’, and

21 See footnote 16. 22 This root probably denotes the general property ‘be good’. 23 Recall that -ne appears to be an analyzable suffix. 24 These are ‘old man’, ‘collared peccary’, ‘white-lipped peccary’, ‘woolly monkey’, ‘fish’, ‘plantain’, ‘ayahuasca’, and ‘canoe’.

952

Zachary O’Hagan

it is the final vowel in the latter of these two variants that appears in Tessmann. However, the pervasiveness of LH tone patterns in the collected data goes beyond these words that show vowel alternations (cf. ‘plantain’) and is perhaps an influence from Urarina, in which citation forms often exhibit a final high tone (Olawsky 2006: 121). Tessmann’s corresponds to modern /ɾ/ and /l/; to /ɾ/; to /ɾ/; and to /l/ and /ɾ/. It is possible that Tessmann heard something akin to the Urarina retroflex flap /ɽ/, although this is not present in modern-day Omurano, and that this is represented by the relatively exotic use of . It is not clear what distinction he was attempting to draw in contrasting with , as both correspond to modernday /ɾ/. And the confusion of modern-day /l/ and /ɾ/ represented in his is attested elsewhere in his work. For example, in his data on Omagua (Tupí-Guaraní), which does not exhibit the phone [l], Tessmann represents many instances of /ɾ/ as . Modern-day /ʧ/ corresponds to , , and ́ in Tessmann. The latter may provide slim evidence that a possible *k was in some environments undergoing palatalization, which would go towards explaining the complete absence of /k/ in the collected data. In fact, the grapheme only occurs three times in all of Tessmann’s list of approximately 250 items. Similarly, occurs once in the word-initial sequence , and in the word-final sequence . Other than that, no grapheme likely corresponding to a fricative is attested in Tessmann, which is in line with the absence of fricatives in the collected data. Finally, there are no graphemes in Tessmann that plausibly denote phonemes not attested in the collected data generally, spanning both vowels and consonants, which lends further credence to the comprehensiveness of the inventories provided in Section 3.1.

7 Conclusion This chapter has described what is currently known about the phonology, grammar, and lexicon of Omurano based on fieldwork with rememberers. Unless other rememberers are located, this is likely all that will be known about Omurano. I have confirmed that Omurano is the same language as the one documented by Günter Tessmann in 1925. Lexical comparison with nearby languages confirms its status as an isolate: there are only two candidate borrowings with Urarina, and, since Urarina is also an isolate, the direction of borrowing is unclear (though the fact that Urarina exhibits two words for ‘ayahuasca’ suggests that at least one of the borrowings is from Omurano into Urarina). In terms of phonology and grammar, I have shown 10 phonemic consonants and 6 phonemic vowels (5 qualities). Syllable structure is (C)V(V). I have illustrated several surface tone patterns, suggesting preliminarily that only some nouns are specified for (high) tone, noting that patterns of vowel reduction hint also at a

20 Omurano

953

stress-like system. Basic word order is AOV, verbs are simple and not crossreferenced for person, and alignment is nominative-accusative with an accusative suffix -ta. Imperatives show no special marking of the verb. Nominal possession consists of apparent prefixes attaching to the possessed noun. A major direction for future research would be to apply some of the phonemegrapheme correspondences discussed to Tessmann’s considerably longer wordlist in order to reconstruct more reliable (though inherently hypothetical) representations of other lexical items. This could significantly expand our lexical knowledge of Omurano beyond the roots recovered in this work.

8 Acknowledgements I thank Msgr. Miguel Ángel Cadenas Cardo and Fr. Manuel Berjón Martínez, who provided invaluable logistical support and intellectual stimulation during my time on the Urituyacu. I am indebted to those who shared their knowledge of Omurano with me: Rafael “Rafico” Inuma Macusi, Simón Inuma Manizari, Jorge Macusi Nuribe, †José Manuel Macusi Nuribe, and Juan “Pihuicho” Macusi Nuribe. Numerous individuals in Iquitos and on the Urituyacu and its environs welcomed me into their homes and told me about their lives and the history of their families and the region: Darío Cariajano Yahuarcani, Sonia Caritimari Huansi, Rafael Huaya, Mamerto Inuma López, Ventura Inuma Vela, Juan Macusi Inuma, Juana Macusi Inuma, Juana Macusi Nuribe, Luis “Lucho” Macusi Nuribe, Rolinson Macusi Inuma, Rosalía Macusi Nuribe, Salvador Macusi Ahuite, Francisco “Pancho” Murayari Macusi, Inés Murayari Inuma, Carmela Padilla, Guillermo Reátegui Trigoso, Jaime Ríos Tuesta, Melita Tangoa Capcha, Carlos Valencia Otejón, Catalino Valencia Paima †, César Vásquez, and Gilter Yuyarima Tapullima. I have benefited from conversations with Martin Kohlberger and Luis Miguel Rojas Berscia on the history of the region, and Jorge Rosés Labrada graciously gave comments on an earlier draft, as did an anonymous reviewer and the editors. Around the time that the final draft of this chapter was completed (March 2021), I began a project with Joshua Homan and Emanuele Fabiano on the (linguistic) history of the Urituyacu River and its environs. I have subsequently benefited from extensive conversations with them in (June 2022) and out of the field, although they are not responsible for the claims made here.

9 References Alicea Ortiz, Neftalí. 1975. Vocabulario taushiro. Lima: Instituto Lingüístico de Verano. Carvalho, Fernando Orphão de. 2013. On Záparoan as a valid genetic unity: Preliminary correspondences and the status of Omurano. Revista Brasileira de Lingüística Antropológica 5(1). 91–116.

954

Zachary O’Hagan

Cipolletti, María Susana. 2017. Sociedades indígenas de la alta Amazonía: Fortunas y adversidades (siglos XVII–XX). Quito: Abya-Yala. Epps, Patience. South American languages. In Claire Bowern, Patience Epps, Jane Hill & Patrick McConvell (eds.), Languages of hunter-gatherers and their neighbors: Database. http:// huntergatherer.la.utexas.edu. Accessed November 2015. Fast, Gerhard, Ruby Fast & Daniel Fast (compilers). 1996. Diccionario achuar-shiwiar castellano. Lima: Ministerio de Educación; Instituto Lingüístico de Verano. Figueroa, Francisco de. 1986[1661]. Informe de las misiones en el Marañón, Gran Pará o Río de las Amazonas. In Informes de jesuitas en el Amazonas, 143–309. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía. Hervás y Panduro, Lorenzo. 1784. Catalogo delle lingue conosciute e notizia della loro affinità, e diversità. Cesena: Gregorio Biasini. Jouanen, José. 1941. Historia de la Compañía de Jesús en la antigua Provincia de Quitio, vol. 1. Quito: Editorial Ecuatoriana. Jouanen, José. 1943. Historia de la Compañía de Jesús en la antigua Provincia de Quitio, vol. 2. Quito: Editorial Ecuatoriana. Kaufman, Terrence. 1994. The native languages of South America. In Christopher Mosley & R. E. Asher (eds.), Atlas of the world’s languages, 46–76. London & New York: Routledge. Loukotka, Čestmír. 1968. Classification of South American Indian languages. Los Angeles: University of California Press. Maroni, Pablo. 1988[1738]. Noticias auténticas del famoso río Marañón. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía. Olawsky, Knut J. 2005. Urarina – Evidence for OVS Constituent Order. In Boban Arsenijevic, Noureddine Elouazizi, Frank Landsbergen & Martin Salzmann (eds.), Leiden Papers in Linguistics, vol. 2.2., 43–68. Olawsky, Knut J. 2006. A Grammar of Urarina. Berlin & New York: Mouton de Gruyter. Requena y Herrera, Francisco de. 1994[1785]. Descripción del Gobierno de Maynas y las misiones en él establecidas. In Pilar Ponce Leiva (ed.), Relaciones histórico-geográficas de la Audiencia de Quito siglo XVI–XIX, vol. 2., 658–700. Ribeiro, Darcy & Mary Ruth Wise. 1978. Los grupos étnicos de la Amazonía peruana. Lima: Instituto Lingüístico de Verano. Riva Herrera, Martín de. 2003[1656]. Traslado de los auto fechos por el General Don Martín de la Riva Herrera en razón de la fundación y población de la ciudad de Santander de la Nueva Montaña. In Fernando Santos Granero (ed.), La conquista de los motilones, tabalosos, maynas y jíbaros, 238–251. Iquitos: Centro de Estudios Teológicos de la Amazonía. Santos, Thomas. 1986[1684]. Relación que da el padre Thomas Santos de la conquista y entrada que por el rio del Tigre hizo á cuatro naciones que son los asouinatoas, los pinches, los cenicientos y habitoas. In Informes de jesuitas en el Amazonas, 341–355. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía. Tejedor, Senén F. 1927. Breve reseña histórica de la misión agustina de San León del Amazonas, Perú. El Escorial: Imprenta del Real Monasterio de El Escorial. Tessmann, Günter. 1930. Die Indianer Nordost-Perus: grundlegende Forschungen für eine systematische Kulturkunde. Hamburg: Friederischen, de Gruyter & Co. Tovar, Antonio. 1961. Catálogo de las lenguas de América del Sur. Buenos Aires: Editorial Sudamericana. Tuggy, John. 1966. Vocabulario candoshi de Loreto. Lima: Instituto Lingüístico de Verano. Vásquez Caicedo, Pedro Pablo. 1905[1843]. Descripción de la provincia de Maynas por el gobernador general de las Misiones. In Carlos Larrabure i Correa (ed.), Colección de leyes, decretos, resoluciones i otros documentos oficiales referentes al Departamento de Loreto, vol. 5, 263–275. Lima: Imprenta de La Opinión Nacional.

20 Omurano

955

Veigl, Francisco Xavier. 2006[1798]. Noticias detalladas sobre el estado de la Provincia de Maynas en América meridional hasta el año de 1768. Iquitos: Centro de Estudios Teológicos de la Amazonía. Velasco, Juan de. 1842[1789]. Historia del reino de Quito en la América meridional, vol. 3. Quito: Imprenta de Gobierno, por Juan Campuzano. Velasco, Juan de. 1919. Historia moderna del Reyno de Quito y crónica de la Provincia de la Compañía de Jesús del mismo Reyno. Boletín de la Sociedad Ecuatoriana de Estudios Históricos Americanos 2(5): 260–280. Villarejo, Avencio. 1943. Así es la selva: Estudio geográfico y etnográfico de la Provincia de Bajo Amazonas. Lima: Compañía de Impresiones y Publicidad. Villarejo, Avencio. 1953. Así es la selva: Estudio monográfico de la Amazonía nororiental del Perú. Lima: Sanmarti y Cía. Wise, Mary Ruth. 1985. Indigenous languages of lowland Peru: History and current status. In Harriet Manelis Klein & Louisa R. Stark (eds.), South American Indian languages: Retrospect and prospect, 194–221. Austin: University of Texas Press.

Raiane Salles

21 Pirahã (Apáitisí) 1 2 3 4 5 6 7 8 9 10

Introduction Phonology Word classes and morphological structure The noun phrase The verb phrase Simple clauses Clause-linking, information structure, and discourse Conclusion Acknowledgements References

1 Introduction Pirahã (pira1253) is spoken by over 700 people in ten different villages spread across the Maici River in the municipality of Humaitá, Amazonas, Brazil.1 Speakers call themselves Hiáitihí, and their language is called Apáitisí; however, Pirahã is the name recognized by the Brazilian government and widely used in the literature.2 In the interest of accessibility, this chapter uses the term Pirahã, but it is nevertheless important to recognize the name used by these people to refer to their own language. The community mostly subsists on fishing, although they also rely on hunting and slash-and-burn agriculture. They grow manioc (for flour), cashew, banana, and mango, harvest honey, and collect Brazil nuts, the latter being an important trade product with Brazilian merchants.3 There is a general consensus that Pirahã belongs to the Muran family, but the status of its extinct members as languages or dialects is unclear.4 Loukotka (1968) presents Mura as a small, isolated stock consisting of the languages Murá (mura1273), Bohurá, Pirahá, and Yaháhi.5 Heinrichs (1964) includes Torá, which is

1 One of the villages is actually located in the Ipixuna River within Parintintin territory. The number of villages is constantly expanding because the Pirahã are semi-nomads, and new villages are created when a family clan migrates to another territory. 2 Note that the language does not have /r/ in its phonetic inventory (see Section 2.1). 3 For more information on Pirahã ethnography, see Gonçalves (2001). 4 Amazonas (1852) speculates that the Mura are originally from Peru. For more information on the Mura people, see Amoroso (2009) 5 It is possible that Bohurá does not even refer to a different language than Mura, given that this word is used for auto-identification by the Mura people (see Nimuendajú 1948). Raiane Salles, University of British Columbia https://doi.org/10.1515/9783110432732-008

958

Raiane Salles

21 Pirahã (Apáitisí)

959

Tab. 21.1: Comparison between Mura words and Pirahã words. mura

pirahã

translation

peissí kutjúhi itokúi auí aurí bahúis tihyhí arais itíhi itai uái bê ái aitáhus kwoabís pitaissa abaái aihyahá huisí kaãnhê arí aupís kái babihí

piiʔi koihí itohí aooí aogí bahóisi tihíihi ágaísi tíhi tai hoái bií aí aitáhoi koabaipí pitáipí abí ahoái hisó kaháíʔái agí aáopí kaáo baábi(hi)

long short big male foreigner female foreigner wild pig louse flour tobaco leaf fire blood bone sleep die drink stay say sun moon cold feisty far bad

now widely accepted as Chapacuran (see Birchall et al. 2016). D. Everett (2005) speculates that Matanawi (mata1275), which Loukotka had previously considered an isolate, is also related to Pirahã. The most widely accepted contemporary classification, though, is that Mura is an isolate language and Pirahã is the last surviving of its numerous dialects (e.g., Nimuendajú 1948; Aikhenvald 2012). See Table 21.1 for a comparison between Mura words documented by Hanke (1950) and Pirahã words, which convincingly shows that Pirahã should be classified as Muran. The UNESCO Atlas of the World’s Languages in Danger (Moseley 2010) classifies Pirahã as vulnerable. Although it is spoken by all generations, the small population and the growing influence of Portuguese in the Pirahã’s contact with other groups suggest a relatively high risk of endangerment (see Rodrigues & Oliveira (1977) for evidence that the acculturation process in the region has been ongoing at least since the 1960s). Aspects of the language were described by a number of missionaries from the Summer Institute of Linguistics. Some of their contributions are the documentation of the phonetic system (Heinrichs 1964), the manuscript of a Pirahã-Portuguese dic-

960

Raiane Salles

tionary (K. Everett 1987), the documentation of verbal suffixes (Sheldon 1988), and a grammar sketch (D. Everett 1986). Pirahã then came to wider attention due to claims that the language lacks a number of properties, such as numbers, color terms, tense, and recursion (e.g., D. Everett 2005; see also Nevins et al. 2009a, 2009b; D. Everett 2009; Salles 2015; Futrell et al. 2016; Amaral et al. 2018; Rodrigues et al. 2018; and Sauerland 2010, 2018 for debate on the issue). Some more recent works are Sândalo and Abaurre (2010) and Silva (2014) on phonetic and phonological aspects of the language; Salles (2015) on recursion within possessive constructions; Salles and Matthewson (2016) and Salles (2018) on the semantics of bare nouns; and Sadlier-Brown et al. (2022) on a genderlect distinction. This chapter brings together the data and analysis provided by these works with new fieldwork data. The new data presented here were collected by the author over five different fieldtrips (September 2013, October–December 2014, July 2016, August 2017, and April 2019).6 This chapter describes only the spoken modality; it does not address the whistled and hummed modes observed by D. Everett (1985).

2 Phonology 2.1 Phonemic inventory Pirahã has three vowel phonemes /i, a, o/, listed in Table 21.2 – free variants are also represented in square brackets – and ten consonant phonemes, listed in Table 21.3. (Near) minimal pairs are presented in (1).

Tab. 21.2: Pirahã vowel phonemes. front high mid low

central

i [ɪ] [e] [ɛ]

back [u] o

a

6 All examples in this chapter belong to my corpus unless otherwise noted. For the purposes of consistency, all examples have been adapted to the orthography and glosses used here. On this note: previous works such as D. Everett (1986) represent the glottal stop with an , but I use the IPA symbol [ʔ] to avoid confusion with the voiceless velar fricative. Low tone is unmarked and high tone is marked with the acute accent .

21 Pirahã (Apáitisí)

961

Tab. 21.3: Pirahã consonant phonemes.

plosive fricative approximants

(1)

a. b. c. d. e. f.

/pee/ /beé/ /baj/ /taj/ /kaj/ /ɡáj/

bilabial

alveolar

p

t s

b

palatal

w

‘water’ ‘blood’ ‘many’ ‘leaf’ ‘daughter’ ‘say’

velar

glottal

k

ʔ h

ɡ

j

g. h. i. j. k. l.

/háj/ /saj/ /ʔajó/ /ʔawe/ /báʔá/ /basá/

‘future’ nominalizer ‘yes’ ‘male foreigner’ ‘good’ ‘apple’

2.2 Processes and variability Pirahã exhibits several different phonological processes, with varying degrees of predictability. This section discusses predictable palatalization and epenthesis, as well as nasalization, free consonant variation, and genderlect differences. Palatalization of /t/ occurs predictably before the front vowel /i/ and its allophones [i, ɪ, e, ɛ] (see Heinrichs 1964; D. Everett 1986). See (2) for the rule and (3) for some examples. (2)

Rule: /t/ → tʃ/_i

(3)

a. /tíoí/ [tʃí.ˈwé] ‘rubber’ b. /kabatíóɡí/ [ka.ba.ˈtʃɪ ́o.ɡɪ ́] ‘meat’

Glottal stop insertion before word-initial vowels also occurs predictably (Heinrichs 1964).7 Rule and examples are given in (4) and (5). (4)

Rule: /Ø/ → ʔ/ #_V

(5)

a. /íboɡi/ [ʔɪ ́.bo.ˈɡɪ] ‘milk’ b. /aáháihi/ [ʔaá.háj.ˈhɪ] ‘sugar’

7 Also according to Heinrichs (1964), voiceless consonants are predictably longer than voiced consonants word-medially.

962

Raiane Salles

Tab. 21.4: Words illustrating phonemes in free variation in Pirahã. [b]

[m]

gloss

[ɡ]

booɡɪ baáɡɪsó bɪɪ ́ɪsaj

mooɡɪ maáɡɪsó mɪɪ ́ɪsaj

proper name ɡahiáo ‘many’ ɡɪ ́aj ‘red’ ɡájhɪ

[n]

gloss

[k]

[ʔ]

gloss

nahiáo nɪ ́aj náihɪ

‘airplane’ ‘two’ dem

kosé kahapé kobiáj

ʔosé ʔahapé ʔobiáj

‘eye’ ‘go’ ‘white’

Sândalo and Abaurre (2010) claim that in Pirahã the neutral position of the velum is lowered, making all segments nasal except for obstruents, which are [-nasal]. According to them, vowels are oralized when an obstruent causes orality spreading. However, my data do not support this conclusion; I observe examples of oral vowels following nasal stops, as in (6a), as well as nasal vowels following oral stops, shown in (6b).8 Though more work is required on nasalization, authors agree that nasality is not distinctive in Pirahã (see also Salomon 2020). (6)

a. [meʔõésaj] ‘soap’

́ ]̃́ b. [kahaajɡẽṍɪ ‘bamboo’

Free variation is very common in Pirahã. The most common cases are [b]~[m], [ɡ]~[n], and [k]~[ʔ], all word-initially.9 See Table 21.4 for examples. Word-internal alternation of [ɡ]~[ʔ] is also possible, as shown in (7). (7)

[kapɪɪɡakaaɡakátɪ ́] ~ [kapɪɪʔakaaʔakátɪ ́] kapiiga-kaagak-átí! paper-mark.make-imp ‘Study!’

Heinrichs (1964) and D. Everett (1979, 1982) also noted variation between [b]~[ʙ] and [ɡ]~[ǐ ̜].10 Although not observed in my fieldwork, anecdotal observations by my consultants suggest that older Pirahã generations used to have [ʙ] and [ǐ ̜] in their inventories.11 Finally, a genderlect distinction has been described in which /s/ is categorically pronounced as [h] by women before /i/ and optionally elsewhere (D. Everett 1986;

8 I thank my research assistant Isabel Salomon for pointing out these data. 9 There is one instance of intervocalic [m] in the word [na.ma.ʔáí] ‘saracura’ (see Silva 2014). It was not yet clear if [b] is also in free variation with [m] in this case. 10 The symbol ǐ ̜ was proposed by D. Everett (1979, 1982) to represent the “voiced, lateralized apicalalveolar/sublaminal-labial double flap with egressive lung air” (D. Everett 1982: 94). 11 D. Everett (1979, 1982) argues that these sounds are only produced around people with whom the speaker is familiar. Keren Madora (p.c., July 2017) still observes speakers producing those sounds.

21 Pirahã (Apáitisí)

963

Silva 2014). However, a more recent quantitative variationist analysis shows that both [s] and [h] are used by men and women in all pre-vocalic environments, though higher usage of [s] is observed for men while women more often use [h] (Sadlier-Brown et al. 2022). The authors suggest that a potential sound change, led by women, is in progress.

2.3 Phonotactics Pirahã syllables have a (C)V(V)(G) shape, where G is a glide (/j,w/).12 This representation also accounts for long vowels (VV), which are bimoraic (D. Everett 1979). Examples of each syllabic pattern are given in (8). (8)

a. VV.CV [aa.ɡɪ ́] ‘coati’

b. CVV.CVVG [naa.hɪáw] ‘plane’

c. CVV.CVG [ɡɪo.páj] ‘dog’

Apart from glides, no other consonant can function as a coda, except the possible exception of the animal classifier ʔís, which is pronounced as [ʔɪ ́] when preceding a consonant-initial word and [ʔɪ ́s] when followed by a vowel-initial word. There are no complex onsets, and the sequence */ki/ is not attested (see also Heinrichs 1964; D. Everett 1979; K. Everett 1998).

2.4 Prosody: Tone, stress, root, or word minimality constraints 2.4.1 Tone According to Silva (2014), Pirahã has four phonemic tones: high, low, rising, and falling – that is, two level tones and two contour tones. He argues that the contour tones can only appear in heavy syllables, that is, (C)VV.13 However, further investigation regarding the contour tones is needed. The tone bearing unit (TBU) in Pirahã is the mora. As stated previously, VV syllables are bimoraic and therefore have two TBUs, each with its own level tone. The apparent contour tones arise when a hightoned mora is followed by a low toned one, or vice versa.

12 See Heinrichs (1964), D. Everett (1979, 1986), D. Everett & K. Everett (1984), K. Everett (1998), and Silva (2014) for different approaches to syllable structure in Pirahã. 13 Previous analyses have variously described three phonemic tones (high, mid, and low) (Heinrichs 1964; Sheldon 1974); two phonemic tones (high and low) (D. Everett 1979); four phonetic level tones (high, mid, low, and “low-lowered”) (D. Everett 1986); and four phonetic tones, two level (high, low) and two contour (low-high, high-low) (K. Everett 1998).

964

Raiane Salles

Sheldon (1974) described six morphophonemic processes which operate in noun-modifier and noun-verb sequences, all of which are related to tonology. They are: tone extension, vowel metathesis, syllable deletion, vowel replacement, and tone replacement. These processes are not further described here; more work on Pirahã tonology is needed.

2.4.2 Stress Stress is non-iterative, and syllable weight is relevant for primary stress assignment (D. Everett & K. Everett 1984; D. Everett 1985). These authors propose the syllable weight hierarchy in (9) and state that stress will fall on the heaviest syllable among the three final syllables of a word. If two or more syllables in this window have the same weight, stress falls on the rightmost. In (9), represents any voiced consonant. (9)

CVV>ZVV>VV>CV>ZV

Crucially, in this analysis the onset also contributes to a syllable’s weight – voiceless onsets make heavier syllables than voiced ones – a typological rarity found in only four out of 127 languages in a genetically and geographically balanced survey (M. Gordon, 2006; see also M. Gordon (2005) for a phonetic study of three of those onset-sensitive systems, including Pirahã). Finally, the acoustic correlates of stress in Pirahã are amplitude and duration, not pitch (K. Everett 1998). Secondary stress has not yet been described in Pirahã.

3 Word classes and morphological structure The language tends towards an agglutinative structure, and it prefers suffixes. Tense, aspect, mood, and evidentiality (TAME) marking is realized through verbal suffixes, while arguments are realized as independent words and not marked on the verb, as in (10). (10) ti kapí kaaipí-haí 1.pro coffee make-fut ‘I will make coffee.’ Locative -ó, as in (11), and instrumental -oa, in (12), are oblique case markers, also appearing as nominal suffixes.

21 Pirahã (Apáitisí)

965

(11) hi moitóhi-ó kahápi-haí 3.pro boat-loc go-fut ‘He will go by boat.’ (12) tí apití táob-áhá tagasága-oa 1.pro arm cut-decl machete-ins ‘(He) cut my arm with a machete.’ Nominalizers are also suffixes, seen in (13). (13) ipoihi kosaaga ʔaoí ʔahoai-sai woman not.know foreigner speak-nmlz ‘The woman doesn’t know the foreigner’s language.’

3.1 Pronominal elements and noun classes Pronominal elements may be used in all argument positions. For example, take ʔi, the third-person feminine pronoun. In (14) it is the subject of a transitive verb, in (15) the object, and in (16) the subject of an intransitive verb. (14) ʔi ʔí ʔobí-koí soioʔáagahai 3f.pro 3.inan roll-ass sewing.thread ‘She rolled the sewing thread.’ (15) ti ʔi ʔogí-bai-koí 1.pro 3f.pro want-lot-ass ‘I like her.’ (16) ʔi ʔaitáhoí-koí 3f.pro sleep-ass ‘She has fallen asleep.’ Pronouns are also used in possessive constructions, see (17), and in constructions where they co-occur with the noun they refer to, as in (18) (see also Section 7). (17) ʔi máosái biíisai ʔáag-ahá 3f cloth red cop-decl ‘Her dress is red.’ (18) ipóihí ʔí máosái ʔog-abagai woman 3f.pro cloth want-frus ‘The woman, she would like some/the cloth.’

966

Raiane Salles

Tab. 21.5: Pirahã pronominal elements. person

singular

first second third third feminine third. clf:animal third inanimate third.clf:plant third.clf:liquid

ti gí(ʔai)/ní(ʔai) hi ʔi ʔí(s) ʔí ʔá pii

plural (optionally marked)

+ hoíhiaagáo or pauc + ‘all/every’

ʔogiáagáo ‘all/every’

These pronominal elements can function as both bound and free forms. Compare, for instance, (19) and (20). In (19), ʔá= 3.clf:plant is phonologically attached to the verb ʔobi ‘to roll’. The first syllable of the verb is dropped so it can host ʔá 3.clf:plant. (19) ʔá=bi-koí pogáihiái 3.clf:plant=roll-ass banana ‘The banana rolled.’ However, (20) shows the same pronominal form occurring freely. See Table 21.5 for a summary of pronominal elements. (20) ʔá ʔobi-koí ʔípoai 3.clf:plant roll-ass mango ‘The mango rolled.’ Nearby and closely related groups (e.g., the Mura) experienced widespread language shift, first to Língua Geral (a Tupian contact language spoken in the Amazonian region between the end of the 17th century and the beginning of the 20th century), and later Portuguese (Nimuendajú 1948). On the other hand, for the Pirahã during that same period, “their knowledge of Portuguese and of the Língua Geral never exceeded a dozen words” (Nimuendajú 1948: 267). Therefore, we might expect few borrowings in Pirahã. However, Thomason and D. Everett (2001) demonstrate a close match between the phonological forms of Pirahã pronominal elements with those from Língua Geral and Tenharim (another Tupian language in contact with Pirahã), suggesting an historical connection. They discuss the first and second person and the third-person masculine, feminine, and non-human forms. Nevertheless, the system presents several other third-person forms which are not the result of borrowing and express noun classes. These seem to be short forms from generic

21 Pirahã (Apáitisí)

967

nouns: ʔís 3.clf:animal, for example, comes from the generic ʔísigihí ‘animal’, and ʔá 3.clf:plant comes from ʔáí ‘plant’.14

3.2 Bare nouns and demonstratives Pirahã does not have articles. Nouns are bare in both indefinite and definite contexts, as in (21a) and (21b), respectively, and in any argument position, as in (22) (Salles 2018, Salles in prep; Salles & Matthewson 2016). (21) Prompt: a storyboard in which the narrator character enters a house and sees a woman and a man. a. ti ʔi=áob-áhá ipóihi igihi piai 1.pro 3f.pro=see-decl woman man too ‘I’ve seen a woman and a man.’ b. ti hi=áho-áhá igihí 1.pro 3.pro=speak-decl man ‘I’ve spoken with the man.’ (22) Context: The speaker sees (a) a picture where a monkey is eating an anaconda, and (b) an anaconda is eating a monkey. a. pigiá ʔís hi kohoaip-áhá páohoahai monkey 3.clf:animal 3 eat-decl anaconda ‘The monkey is eating the anaconda.’ b. páohoahai ʔís hi kohoaip-áhá pigiá anaconda 3.clf:animal 3.pro eat-decl monkey ‘The anaconda is eating the monkey.’ As for the demonstrative system, D. Everett (1986) documented a distal form, gáihi ‘that’, and a proximal form, gíisai ‘this’. However, my fieldwork indicates that speakers use gáihi regardless of proximity (and gíisai was not observed in volunteered sentences nor accepted in grammaticality judgments). The word has a general deictic function, often accompanied by pointing gestures. Though typologically unusual, demonstrative systems with only one term have been documented in Supyire, Dyirbal, colloquial Czech, and some dialects of German (see Dixon 2003). The demonstrative in Pirahã is placed after the noun, as in (23). (23) ipóihí gáihi tí ibáisi ʔáag-áhá woman dem 1.pro spouse cop-decl ‘This woman is my wife.’ 14 Though noun classes are not discussed in previous works, K. Everett (1987) described those pronominal elements as classifiers in her unpublished dictionary.

968

Raiane Salles

3.3 Nouns Words such as tiobáhai ‘child’ and ʔipoai ‘mango’ in (24), and tahoasí ‘beach’ in (25), perform a referential propositional function and denote objects: a person, a thing, and a place, respectively. (24) tiobáhai ʔá hoíhio ʔo-áhá ʔipoai child 3.clf:plant pauc hold-decl mango ‘The child is holding two mangoes.’ (25) ti ʔaitáhoí tahoasí 1.pro sleep beach ‘I will sleep on the beach.’ When used for functions other than reference, these words are treated differently. As possessors of nouns, that is, modifiers, they precede the nouns they modify, as in (26), whereas canonical modifiers, such as koíhí ‘small’, occur after the noun they modify, as in (27). These words cannot be used predicatively without a copula, as shown in (28). Based on criteria defined by Croft (1991, 2001), these facts suggest that Pirahã has a distinct class of nouns. (26) tiobáhái máosái koíhí child cloth small ‘The child’s clothes are small.’ (27) Migíʔoí ʔi báagáobí-koí máigi koíhí Migíʔoí 3f bring-ass fishing.hook small ‘Migíʔoí brought small fishing hooks.’ (28) Posogiái tiobáhai *(ʔáag-ahá) Posogiái child cop-decl ‘Posogiái is a child.’ Compounding is a fairly productive noun-formation process in Pirahã. Combinations of noun+noun, as in (29a), noun+modifier, as in (29b), and noun+verb, as in (26c), are all attested. (29) a. ʔapai + foot b. pi + thorn

toii handle

→

ʔapaitoii ‘ladder’

gáía crooked

→

pigáía

‘scissors’

21 Pirahã (Apáitisí)

c. kao + ogiái → mouth be.big (D. Everett 1986: 322)

kaogiái

969

‘bass (fish)’

3.4 Verbs Pirahã also has a class of words that perform the function of predication and (roughly) denote actions, that is, verbs. In (30), the word ʔabópai ‘return’ makes a predicate whose argument is the speaker. The word ʔabópai ‘return’ also hosts the TAME suffix; along with their function as predicates, verbs in Pirahã are distinguished as the only word class which can host TAME suffixes. (30) Context: Speaker is saying goodbye and announcing that he will come back another day ti ʔabópai-hái ahoahiai 1.pro return-fut another.day ‘I’ll return another day.’ Just as nouns require a copula to perform a non-canonical predicative function, verbs require derivational morphology to appear in atypical, non-predicative functions. For instance, in (31), -sai nominalizer is required to nominalize the verb gái ‘to say’ so that it can appear in a possessive (nominal) phrase. (31) hi gái-sái ti pioáhai ʔog-abagai 3.pro say-nmlz 1.pro soda want-frus ‘His saying is: “I want guarana.”’

3.5 Modifiers Pirahã has modifiers which could be identified as adverbs, such as the temporal expressions in (32), as well as those whose semantics are often associated with noun modifiers (adjectives), such as the stative verbs used for property concepts in (33) through (35). (32) a. b. c. d. e.

ahoapió ~ ahoahiái piʔí ~ piʔái ʔíga soʔóá hoa kahaiʔaíi

‘another day’ ‘now’ ‘already’ ‘day’ + numeral ‘x days’ ‘moon’ + numeral ‘x moons’

970

Raiane Salles

f. piiáiso g. piibigaíso h. ahoigío i. ahoái j. hisó k. hisóogiái l. hibigíbagáʔáiso m. ʔahoakohoaihio n. pi (D. Everett 2005: 22)

‘low water’ ‘high water’ ‘afternoon’ ‘night’ ‘during the day’ ‘noon’ ‘sunset/rise’ ‘early morning, before sunrise’ ‘water’ + numeral ‘x years ago/in x years’

(33) Verão ʔagaoa ʔigapópáp-áhá táió ʔagaoa ʔaasí Verão canoe bring-decl then canoe be.new ‘Verão then brought a new canoe/new canoes.’ (34) ʔí ʔaasí-koí 3f.pro be.new-ass ‘She is very young.’ (35) a. ʔahoasai ‘(be) blue/green/yellow’ b. miísai ‘(be) red/yellow’

c.

kobiai ‘(be) white’

d.

miíopáiʔai ~ kopáiʔai ‘(be) black’

D. Everett (2005) analyzes the color terms in (35) as phrasal expressions, given their apparent morphological complexity. These terms all contain -ai, which appears to be the same as copular -ai, shown in (36). (36) kobi-ai see-cop ‘It sees.’ (adapted from D. Everett 2005: 627) Since other adjectival notions are expressed through stative verbs, as illustrated in (33) and (34), it is indeed possible that these color terms were at one point morphologically complex stative verb phrases which remain etymologically transparent.

21 Pirahã (Apáitisí)

971

3.6 Approximate numerals and quantifiers Quantities are expressed by three numerals: hói(híío) ‘one’, hoí(hio) paucal, and baágisó ‘many’.15 These words are used for approximation; that is, they refer to an approximate amount and not an exact quantification, which is a cross-linguistically rare phenomenon (see P. Gordon 2004 and C. Everett & Madora 2012 for experimental work on Pirahã numerosity). As for word order, these (approximate) numerals, like other noun modifiers, follow the noun they modify, as in (37) through (39). (37) ti máigi hóihíío ʔog-abagai 1.pro fishing.hook one want-frus ‘I would like one fishing hook.’ (38) ti máigi hoíhio ʔog-abagai 1.pro fishing.hook pauc want-frus ‘I would like some fishing hooks.’ (39) ti máigi baagisó ʔog-abagai 1.pro fishing.hook many want-frus ‘I would like many fishing hooks.’ These approximate numerals also interact with other types of noun quantification. The baágiso ‘many’ can also be used as a quantifier. The universal quantifier is ʔogiáagao ‘all’, illustrated in (40), and it may be modified by paucal hoíhio to indicate paucal maximality, as in (41). As indicated in Table 21.5, these quantifiers may also be used to pluralize pronominal elements. (40) ipoihi ʔi ʔogiáagáo ʔáag-áhá ʔagaoa ʔáasí koʔóʔió woman 3f.pro all cop-decl canoe be.new inside ‘All the women are/every woman is inside a new canoe.’ (41) hi hoisai hi hoíhiaagáo ʔa=hoáipí-koí páohói 3.pro son 3.pro all.pauc 3.clf:plant=eat-ass bread ‘His two sons ate bread.’ Words such as koíhi ‘small/little’ may be analyzed as either a modifier (as in (42)) or a quantifier (as in (43)); both occur in the same syntactic position (after the noun they modify).

15 D. Everett (2005) argues that hói(híío) means ‘small amount’, instead of one.

972

Raiane Salles

(42) tiobáhai koíhi-aag-á child small-cop-decl ‘The child is small.’ (43) ti kapíti ʔog-abagai kapíti koíhi 1.pro coffee.powder want-frus coffee.powder small ‘I would like a little bit of coffee powder.’

3.7 Postpositions Location and directionality are expressed by nominal suffixes or by postpositional free forms which can be modified by these suffixes. Location is usually expressed by locative -ó, as in (44), but it can also be marked by -ʔai ‘in’, as in (45), which is homophonous with the copula ʔai ‘be’. (44) hi kaái-ó ʔáag-áhá 3.pro house-loc cop-decl ‘He is in the house.’ (45) ti ʔabi-koí Pitiá-ʔai 1.pro live-ass Piquiá-at ‘I live at Piquiá.’ So far, a single directional, ingressive -ʔio, has been documented. Ingressive -ʔio does not occur independently of -ó locative, as demonstrated in (46). (46) hi kahapí kaái-ó-ʔio 3.pro go house-loc-ingr ‘He entered (Lit. ‘went into’) his house.’ Postpositional free forms are derived from nouns modified by locative -ó. For example, see kohoa-ó ‘side-loc’ in (47), ʔapo-ó ‘head-loc’ in (48), and ko-ó ‘stomachloc’ in (49). (47) ʔagí ʔao ʔaag-á pii kohoaó path poss cop-decl river beside ‘There is a path beside the river.’ (K. Everett 1987: 54) (48) kapiigaitóí tábo ʔapó ʔáag-áhá pencil table on cop-decl ‘There’s a pencil on the table.’

21 Pirahã (Apáitisí)

973

(49) giopái ʔagaoa koó ʔáag-áhá dog canoe in cop-decl ‘The dog is in the canoe.’

4 The noun phrase Noun phrases are head-initial, excepting possessive constructions; all modifiers appear after the noun that they modify. Though there are various strategies for expressing number in Pirahã, number is not grammatically marked on the noun, as seen in (50) – where ʔagaoa ‘canoe’ can refer to singular and plural canoes – and nouns are not marked with other inflectional categories, agreement markers, nominal tense, or aspect markers. (50) Context: Storyboard elicitation; the same sentence was provided by the consultant when Verão brought only one canoe and when he brought multiple canoes. Verão ʔagaoa ʔigapópáp-áhá táió ʔagaoa ʔaasí Verão canoe bring-exist then canoe new ‘Verão then brings a new canoe/new canoes.’ As shown in Section 3.1, gender or noun class is encoded in third-person pronominal elements; possible values are masculine, feminine, animal, plant, inanimate, or liquid. This gender/class marking surfaces only in the pronominal system and is not indicated on modifiers. Pronominal forms are used in argument position and in possessive constructions. The word order in possessive constructions is possessor-possessum. Possessive chains also follow that order, as in (51) (Salles 2015). (51) [ Migíʔoí [ (ʔí) ibáisi [ (hi) máiʔi máosái]]] kopáiʔái ʔao Migíʔoí 3f.pro spouse 3.pro mother cloth black poss ʔáag-á cop-decl ‘Migíʔoí’s husband’s mother has black clothes.’ There is no alienability distinction, and kinship terms can have their possessor omitted, as in (52). (52) máiʔi ʔi ʔapaitai koobiái mother 3f.pro hair white ‘My mother’s hair is white.’

974

Raiane Salles

As for nominal quantification, the pronominal forms used for human persons in this system can be pluralized with the quantifiers hoíhiáagáo ‘pauc+all’ and ʔogiáagáo ‘all’ (see Table 21.5). These quantifiers can also modify noun phrases, as in (53)– (54), as can the numerals discussed in Section 3.6. (53) ipóihi hoíhi-áagáo máosái ʔog-abagái woman pauc-all cloth want-frus ‘These three women want cloth.’ (54) ti Hiaitíhí ʔogiáagáo ʔogí-baá-kóí 1.pro Pirahã all want-lot-ass ‘I like all the Pirahãs so much.’ Hoíhiáagáo yields different readings than koíhí, which is used to indicate ‘small/a little/a few’ as in (55) and (56); hoíhiáagáo has a universal reading, picking the maximal set of entities in a given domain with paucal units, as in (57). (55) ti kapíti koíhi ʔog-abagai 1.pro coffee.powder small want-frus ‘I want a little/some coffee powder.’ (56) ti máigi koíhi ʔog-abagai 1.pro fishing.hook small want-frus ‘I want (the) small hook(s).’ (57) ti máigi ʔi hoíhi-áagáo ʔog-abagai 1.pro fishing.hook 3inan.pro pauc-all want-frus ‘I want both fishing hooks.’ As shown in (55)–(57), like all other modifiers, quantifiers are placed to the right of the noun they modify (see Section 3.5).

5 The verb phrase Verbal morphology in Pirahã is relatively sparse compared to many other Amazonian languages, though it does robustly utilize compounding to derive complex verb bases (Section 5.2). Verbal arguments may be marked by pronominal clitics, which, unlike other categories, appear before the verb base (Section 5.1). Verbs are also marked for a variety of TAME categories with suffixes, though some aspectual categories can be unmarked, and some morphemes can mark more than one category, with interpretation dependent on context (Section 5.3). Pirahã does not have mor-

21 Pirahã (Apáitisí)

975

phological valence-adjusting strategies, relying instead on a variety of other strategies to achieve such meanings as causation or reciprocity that are morphologically marked in some languages (Section 5.4).

5.1 Agreement Subjects and objects may be doubled by pronominal clitics that precede the verb, as in (58). Often, arguments are represented by cataphoric pronominals preverbally with the full noun form appearing post-verbally. (58) ti (pogáihiái) ʔá=ohoáipi-koí (pogáihiái) 1.pro banana 3.clf:plant=eat-ass banana ‘I ate/was eating a/the banana.’ In place of a pronominal, repeaters may be used, such as tioí in (59) and boitó in (60). These may be analyzable as cases of emergent noun classes, since the current noun class pronominals clearly emerged from generic nouns (see Section 3.1). (59) ti tioí ʔog-abagai tioíʔai 1.pro rubber want-frus fishing.line ‘I want fishing line.’ (60) ko Hoaboái boitó hiagabog-átí boitóhoi voc Hoaboái motor turn.on-imp motor ‘Hoaboái, turn on the electric generator.’

5.2 Complex bases Like it is for nouns, compounding is also a productive strategy for verbs, many of which are understood to be combinations of verbs, nouns, and directionals resulting in morphologically complex bases, as in (61) (see K. Everett 1987). (61) a. hoá-abóí fire-out ‘turn off’

b. hoá-aipí fire-in ‘turn on’

c. kapiiga-kaga-kai paper-mark-make ‘study’

5.3 Tense, aspect, mood, evidentiality Uninflected verbs end in i and may be used in subordinate clauses, such as itaipí ‘drink’, which is the complement of the intentional verb ʔogabagai ‘want’ in (62).

976

Raiane Salles

(62) ti pi itaipí ʔog-abagai 1.pro water drink want-frus ‘I want to drink water.’ Uninflected verbs also appear in matrix clauses with habitual situations (63), past temporal reference (64)–(65), and future temporal reference (66). (63) Kaoáibigi aitáhoagí Kaoáibogi sleep ‘Kaoáibogí (a kind of spirit) sleeps.’ (64) ʔaoí soʔoá kahápí foreigner already go ‘The foreigners already left.’ (65) ti soʔoá kahápí 1.pro already go ‘I’m going to leave.’ (66) hi koabáipí soʔógiai 3.pro die long.ago ‘He died a long time ago.’ Though uninflected forms can be used in future contexts, future temporal reference is more often overtly marked by the suffix -hai future, as in (67). (67) ti kahápi-hai ʔahoahiái 1.pro go-fut another.day ‘I will leave tomorrow.’ Propositions with non-future temporal reference bear the suffix -á(há) declarative as in (68), or the suffix -koí assertive, as in (69). Assertive -koí reinforces the assertive speech act, whereas -á(há) declarative is a marker of the basic assertive illocution (see Hengeveld (2004) on the difference between basic illocution and illocutionary modification). Both suffixes are in complementary distribution with interrogative -hí, which suggests that these three suffixes together occupy a single mood slot in the verbal template. (68) ti ʔá=ohoáip-áhá páohói 1.pro 3.clf:plant=eat-decl cracker ‘I am eating a cracker(s).’

21 Pirahã (Apáitisí)

977

(69) ti ʔá=ohoáipí-koí páohói 1.pro 3.clf:plant=eat-ass cracker ‘I am eating a cracker(s).’ The minimal pair in (68)–(69) demonstrates a common strategy; when asked to repeat a sentence previously volunteered with -a(há), in (68), the speaker repeats it using -koí, in (69). Notice also that -koí can reinforce an adjectival notion, as in (34), repeated here as (70), and also in (77). (70) ʔí ʔaasí-koí 3f.pro be.new-ass ‘She is very young.’ Both -á(há) and -koí may be used in situations with different temporal and aspectual values, such as recent past events (71)–(72); remote past events (73)–(74); progressive events (68)–(69); habituals (75)–(76); and states, as in (70), (77), and (78). (71) tiobáhai hi ʔis ʔibáobí-koí mahóigatói child 3.pro 3.clf:animal kill-ass wild.pig ‘The child killed the wild pig.’ (72) tiobáhai hi ʔis ʔibáob-áhá mahóigatói child 3.pro 3.clf:animal kill-decl wild.pig ‘The child killed the wild pig.’ (73) ti máiʔi koabáipí-koí soʔogiái 1.pro parent die-ass long.ago ‘My mom died a long time ago.’ (74) ti máiʔi koabáip-áhá soʔogiái 1.pro parent die-decl long.ago ‘My mom died a long time ago.’ (75) hiabáasí hi=ahoái-koí human 3.pro=speak-ass ‘Humans speak.’ (76) kaoáibogí kohoaip-áhá ʔítiisi kaoaibogí eat-decl fish ‘Kaoáibogí (spirits) eats fish.’ (77) máohoipáí ʔís=aáopí-koí jaguar 3.clf:animal=be.feisty-ass ‘Jaguars are very feisty.’

978

Raiane Salles

(78) ti itaigop-áhá 1.pro hot-decl ‘I’m hot.’ Other moods marked on the verb are imperative -áti, as in (79), interrogative -hí, as in (80), and frustrative -abagai, as in (81). (79) pi itaip-áti! water drink-imp ‘Drink the water!’ (80) nía kao ʔis=ibáobí-hí mahóisi? 2.pro int 3.clf:animal=kill-int wild.pig ‘Did you kill the wild pig?’ (81) ti ʔís ʔibáob-ábagai 1.pro 3.clf:animal kill-frus ‘I almost killed it.’ Aspectual suffixes may appear in combination with mood marking, though some aspectual notions are not morphologically marked. For example, uninflected verbs can be used either imperfectively, as in (63), or perfectively, as in (64). Other aspectual notions, however, are morphologically marked, such as regressive -ta in (82), used to indicate the repetition of a situation, and continuative -iig in (83)–(84), used to indicate that not only a situation is in progress but that it is still ongoing. (82) pi hoíhio ti ʔabópai-tá-há water pauc 1.pro return-reg-decl ‘I will come again in two years.’ (83) Context: speaker is announcing to others that someone (not the speaker) is still eating and therefore cannot come now. ʔi kohoáip-iig-áhá 3f.pro eat-cont-decl ‘She is still eating.’ (84) Context: speaker is promising to stay in the village and not to go elsewhere while her husband is away. ti ʔab-iig-áhá 1.pro stay-cont-decl ‘I will keep staying here.’

21 Pirahã (Apáitisí)

979

According to D. Everett (1986), there are three evidentiality and epistemic markers in Pirahã: -híai hearsay, as in (85), and -sibiga and -saʔáiʔi deduction, as in (86); and -saʔáiʔi inferential. (85) gahió hi ʔabaipi-sai-híai píʔái ʔíga airplane 3.pro sit-nmlz-hrs now really ‘The airplane is landing right now (according to others).’ (D. Everett 1986: 298) (86) Kaogiái ʔís=ibáboí-sibiga Kaogiái 3.clf:animal=kill-infr ‘Kaogiái must be fishing.’ (D. Everett, 1986: 298) Inferential -saʔáiʔí is used when the speaker infers something from the evidence, as in (87). Since no context is provided for (86), no conclusions can be drawn from D. Everett’s English translation regarding the difference between -sibiga deduction and -saʔáiʔí inferential. Both are abbreviated here as infr. (87) Context: a boy fell on the fire and the speaker thinks he will die (given the evidence of his state). ko tiobáhai gíʔa=oabáipí-saʔáiʔi voc child 2.pro=die-infr ‘Child, I think you will die.’

5.4 Valence-adjusting phenomena There do not seem to be valence-adjusting morphemes in Pirahã. For instance, reflexive constructions are achieved via coreferential pronouns in subject and object position, as in (88); that is, there is no special semantic, morphological, or syntactic strategy which expresses reflexivity. The third-person pronoun hi in (89) is ambiguous between a coreferential interpretation, in which case the construction would be reflexive, and a non-coreferential interpretation, yielding a reciprocal interpretation (see D. Everett 1986). (88) ti ti ʔibáob-áhá 1.pro 1.pro hit-decl ‘I hit myself.’ (D. Everett 1986: 216) (89) hi hi ʔibáob-áhá 3.pro 3.pro hit-decl ‘He hit himself’ or ‘They hit each other.’ (D. Everett 1986: 217)

980

Raiane Salles

Middle constructions are achieved lexically. In (90), the verb boitopi ‘take down’ has a valence of two, with the agent ti 1.pro and the patient ʔií ‘tree’ expressed in argument positions. (90) ti ʔií boitopi-hai 1.pro tree take.down-fut ‘I will fell the tree.’ In (91), since only the patient ʔií ‘tree’ is relevant, a verb with a valence of one is used: ʔáobi ‘to fall’. (91) ʔií ʔi ʔáob-áhá tree 3.pro fall-decl ‘A tree has fallen.’ The verb ʔibíibi ‘order, cause to do’ is a lexical causative; see (92). (92) ti giʔai ʔibíibí-sog-abagai kahiaí kai-sai 1.pro 2.pro order-desid-frus basket make-nmlz ‘I want to order you to make a basket.’ (D. Everett 1986: 220) According to D. Everett (1986), the only morphological valence-adjusting operation in Pirahã is nominalization of a verb, which decreases a verb’s valency. However, notice that the agent tiobáhai ‘children’ is still present in (93), though as the possessor of the nominalized clause. Semantically and syntactically, the valence of the nominalized verb is still two. Further research is needed to determine whether nominalized verbs may also be used in instances where one of the arguments is absent entirely. (93) tiobáhai hóoí ai-sai ʔabahíoʔoi child bow make-nmlz incorrect ‘Children’s bow making is incorrect.’ (D. Everett 1986: 270)

6 Simple clauses 6.1 Basic constituent order and alignment Grammatical relations are expressed through a canonical AOV/SV word order. However, as noted previously, arguments may be cataphorically represented by prever-

21 Pirahã (Apáitisí)

981

bal pronominal forms, with the nouns they refer to appearing post-verbally, as in (94).16 (94) tiobáhai hi ʔis ʔibáobí-koí mahóigatói child 3.pro 3.clf:animal kill-ass wild.pig ‘The child killed the wild pig.’ There is no grammatical case marking on nouns or pronouns, and the roles of subject, agent, and object are determined by clausal word order, with agents before objects and both A and S appearing preverbally. As discussed in Section 3.1, pronominal elements also have the same forms for subjects, agents, and patients, resulting in neutral alignment in the pronominal system as well as for full NPs.

6.2 Sentence types As mentioned in Section 5, declarative sentences are either marked with assertive -koí or declarative -a(há); in addition, declaratives have previously been described as using rising intonation (D. Everett 1986). Imperatives are marked with the verbal suffix -áti, interrogatives with -hí, and frustratives with -abagai. This section also addresses negation. Interrogatives are divided into polar questions and content questions. Polar questions may rely solely on intonation (according to D. Everett (1986), they have a different rising contour than declaratives), but there are also morphological strategies. Interrogative suffixes include -hí (discussed in Section 5), -hoáʔai – used for questions dealing with existence/possession, as in (95) – and the free form kao, shown in (96). (95) hi ʔáo-hoaʔái tiobáhai? 3.pro poss-int child ‘Does he have children?’ (D. Everett 1986: 237) (96) níai kao ʔis=ibáobí-hí mahóisi? 2.pro int 3.clf:animal=kill-int wild.pig ‘Did you kill the wild pig?’ Most content questions are formed by a noun or pronoun followed by the interrogative element go, which in turn can be modified by locative -ó, forming ‘where’, as

16 It is not possible to say yet whether these nouns are appositional or functioning as arguments of the main clause. As noted by the editors, the former is quite common in Amazonian languages.

982

Raiane Salles

in (97), or by another interrogative element gíiso forming ‘how many, much; when; why’, as in (98)–(99). (97)

Hiaahoái hi go-ó? Hiaahoái 3.pro int-loc ‘Where’s Hiaahoái?’

(98)

kapiigaitói go gíiso ʔai-hí? pencil int int cop-int ‘How many pencils are there?’

(99)

gíai go gíiso kahápí? 2.pro int int go ‘When are you leaving?’

There is also a form which does not use interrogative go: kaoí ‘who, whose’, as in (100) and (101). (100) go gíiso ʔi ʔogi-hiab-á? int int 3inan.pro want-neg-decl ‘Why don’t you want it?’ (101) kaoí ʔigi-ai-hí? who assc-cop-int ‘Who is that?’ (D. Everett 1986: 240) The word ʔígi associative in (101)–(102) is also sometimes used in interrogative constructions in conjunction with the copula ʔai which is modified by interrogative -hí. It is not yet clear when this construction appears. The interrogative word is positioned in situ in the clause, following a cross-linguistic tendency for verb-final languages (see Dryer 2007). (102) kaoí kaáí ʔigiai-hí? who house assc-int ‘Whose house is that?’ (D. Everett 1986: 240) Clauses are negated with the suffix -hiab, as in (103). (103) ti ʔá ʔogí-hiab-ahá 1.pro 3.clf:plant want-neg-decl ‘I don’t want it.’

21 Pirahã (Apáitisí)

983

There is also a free form, kaba negative, which may function either as a negative polarity item or as a negative existential copula, as in (104). (104) ʔís kaba-ahá 3.clf:animal neg-decl ‘It (the meat) is over.’

6.3 Copulas Pirahã makes use of several different copula verbs, each of which has a sightly different distribution. The copula ʔáagá has the largest number of functions, appearing in locative predications, existentials, equative constructions, and possessive predications. The other copulas, ʔai and ʔiiga, are much more restricted in function. Copula ʔáagá appears in locative predications, as in (105), and in existential assertions, as in (106). (105) Context: answer to the question ‘where is the manioc?’, with goó ‘where’. ʔá kaʔa ʔoó hi ʔáag-á máasí 3.clf:plant basket inside 3.pro cop-decl manioc ‘It is inside the basket, the manioc.’ (106) Context: answer to the question ‘how many spiders are there?’, with go gíiso ‘how many’. ʔi baágisó ʔáag-áhá kopó koó ʔí ʔai 3.pro a.lot cop-decl glass inside 3.pro cop ‘There are a lot of them (spiders), in the glass they are.’ In many cases, constructions with ʔáagá are ambiguous between these locative and existential interpretations, as seen in (107). (107) moísai ʔáag-áhá bee cop-decl ‘Here is a bee/There is a bee here.’ Along with locative expressions and existentials, ʔáagá also appears in equative and predicative contexts, as in (108) and (109), respectively. (108) Hiahoai kasiti ʔáag-áhá Hiahoai chief cop-decl ‘Hiahoai is the chief.’

984

Raiane Salles

(109) óóí ʔís ʔahoasai ʔáag-áhá óóí turtle 3.clf:animal yellow cop-decl turtle ‘The turtle is yellow.’ Possessive predication is expressed by the conjunction of a light verb ʔao ‘have’ and the copula ʔáagá, as in (110). (110) ti tioi miíisai hoihio ʔao ʔáag-á 1.pro ball red pauc poss cop-decl ‘I have two red balls.’ The existential may also be expressed by a ʔao ʔáagá construction, as in (111). (111) kopóhoi ʔao ʔáag-á ahoahái tábo apó cup poss cop-decl blue table on ‘There is a blue cup on the table.’ The decision to treat ʔao as a light verb instead of an auxiliary is based on the fact that it never receives TAME markers; it always appears in its bare form. TAME markers are instead attached to ʔáagá, as in (112) (see Evans (2000) for cross-linguistic examples of light verbs). (112) kapí ʔao ʔáag-iig-áhá coffee poss cop-cont-decl ‘There is still coffee.’ A noun may also be placed between the light verb and the copula, as in (113). (113) ti ʔao hoísai ʔaagáhá 1.pro poss son cop-decl ‘I have a son.’ Another copula is ʔai, which, like ʔáagá, is used for locative predications, as in (106) and (114), and in existential contexts with the modifier -bá ‘many’, as in (115). (114) Kaao Maitá ʔai Kaao Humaitá cop ‘Kaao is in Humaitá.’ (115) kaáí ʔai-bá house cop-many ‘There are many houses (there).’

21 Pirahã (Apáitisí)

985

Finally, ʔiiga is used to indicate a temporary state, as in (116) (D. Everett 1986: 203). It also appears in the expression piai ʔiiga, meaning ‘now’. (116) ti baábi ʔiig-á 1.pro bad cop-decl ‘I am sick.’ (D. Everett 1986: 204)

6.4 Particles Pirahã uses a number of different particles, the best-understood of which is vocative ko. The language also includes some particles which function as modifiers to verbal arguments and some discourse markers or interjections. The vocative particle in Pirahã, ko, is used before a (proper) noun to call out to people, as in (117). (117) ko Migíʔoí, ti maohái ʔog-abagai voc Migíʔoí 1.pro cloth want-frus ‘Hey Migíʔoí, I want cloth.’ D. Everett (1986) describes ʔapaí ‘first’, in (118), and the synonyms gaaba, in (119), and tiohióxio ‘next’ as participant-oriented (as opposed to event-oriented) temporal particles that mark the order in which participants will be involved in a situation. However, they are better understood simply as modifiers. (118) ti ʔapaí ʔopí-ta-áhá 1.pro first go-reg-decl ‘I will go first.’ (D. Everett 1986: 305) (119) gíʔai ʔapaí ti gaaba 2.pro first 1.pro next ‘You first, I (will do it) next.’ (D. Everett 1986: 305) Finally, there are several discourse particles (e.g., ʔaaió ‘right, good’, ʔiigiái ‘okay, enough’, ʔai ‘so, yeah’), as well as interjections, such as ʔáí, used to express surprise in face of unexpected situations, and ʔm (a forcefully ejected glottal stop with closed lips), used to express discontent.

986

Raiane Salles

7 Clause-linking, information structure, and discourse There are multiple strategies for clause-linking in Pirahã. Relative clauses still need further research, but they seem to be externally-headed and either post-nominally embedded, as in (120), or adjoined, as in (121). (120) [ipóihí [tábo ʔapó ʔabaipi sigiái-hi]] ti ibais-áag-áhá woman bench on sit same-int 1.pro wife-cop-decl ‘The woman (who) is the same sitting on the bench is my wife.’ (121) ti baosápisí ʔog-abagai [giai go-ó baosápisí bigaobí ʔai sigíai] 1.pro hammock want-frus 2.pro int-loc hammock show be same ‘I wanted the hammock which is the same you showed me.’ (adapted from D. Everett 1986: 276) Three different elements seem important for relativization: the word sigiai ‘same’, interrogative -hi, and the question word go. Their exact function is not yet clear.17 In (119), sigiai appears at the end of the relative clause and is modified by the interrogative -hi; in (120), the interrogative -hi is not used, but the question word go shows up in the relative clause – apparently as a relative pronoun – along with sigiai ‘same’. Finally, Prado (2018) discusses cases of nominal modification by multiple adjectives where the first adjective seems to be a reduced relative clause, given the presence of copula ʔáaga, as in (122). (122) ʔaʔai [ʔa-opái-ʔaag-áhá] koíhi stone 3-black-cop-decl small ‘The stone (which) is black is small.’ It is worth noting that the strategies for relativization observed here contradict D. Everett’s (2005, 2009) claims that Pirahã has no syntactic embedding, since this issue has been at the center of the recursion debate (e.g., see Hauser et al. 2002; Fitch et al. 2005; Pinker & Jackendoff 2005; Jackendoff & Pinker 2005; Nevins et al. 2009a, 2009b; Sauerland 2010, 2018; Salles 2015; Futrell et al. 2016; Amaral et al. 2018; and Rodrigues et al. 2018). Along with the relative clauses in examples (120)– (122), Pirahã also exhibits possessor embedding (Salles 2015), embedding in prepositional phrases (Amaral et al. 2018), complementation in obligatory control senten-

17 D. Everett (1986) raised the possibility that interrogative -hi and sigiai ‘same’ are complementizers.

21 Pirahã (Apáitisí)

987

ces (Rodrigues et al. 2018), and subordination in reported speech (Sauerland 2018); the latter two will be discussed in this section. At least two complementation strategies have been identified in Pirahã: nonfinite and nominalized complements, the latter of which is a common strategy for complementation in Amazonian languages. For instance, non-finite clauses occur as complements to desiderative predicates with the verb ʔogai ‘want’, as in (123). The verb of the complement clause can also be nominalized, as in (124). Notice that the complement clause may precede the matrix verb or be postposed to it, as in (123) and (124), respectively.18 (123) ti pi itaipí ʔog-abagai 1.pro water drink want-frus ‘I’d like to drink water.’ (124) ti ʔog-abagai pi itaipi-sai 1.pro want-desid water drink-nmlz ‘I’d like to drink water.’ Purposive clauses may also use either a non-finite complement, as in (125), or a nominalized complement, illustrated in (126). In (125) the non-finite complements may also be postposed to the matrix verb. (125) hi bagia-á-ʔio hi ʔao ʔagaoa kobai 3.pro come-decl-ingr 3.pro poss canoe see ‘He is coming to show his canoe.’ (adapted from D. Everett 1986: 265) (126) ti ʔi ʔiga-áti ʔagaoa kaiti-sai 1.pro 3inan.pro take-imp canoe bore-nmlz ‘I must take it to bore the canoes.’ (adapted from D. Everett 1986: 265)

18 Often there is phonological insertion of an /s/ between the complement verb and the desiderative verb, leading some to analyze sentences such as (i) as simple clauses where the verb is modified by a desiderative suffix -sog (see D. Everett 1986 and Sheldon 1988; for a discussion of -sog, see Rodrigues et al. 2018). Even if these structures are analyzed in this way, they would co-occur with other complementation strategies in the language (see Sauerland 2010, 2018 for an analysis of the usage of the nominalizer for complementation in Pirahã). (i) ti hi ʔobái-sog-abagai tiobáhai 1.pro 3.pro see-desid-frus child ‘I want to see a child.’

988

Raiane Salles

Nominalization can also be used to mark the verb in a conditional construction, as in (127). (127) pii boi-hiabi-sai tí ʔahapi water fall-neg-nmlz 1.pro go ‘If it doesn’t rain, I will go.’ (adapted from D. Everett 1986: 264) Reported speech also uses nominalization, but of the matrix verb; the verb gái ‘say’ appears in possessive constructions to introduce a direct quote, shown in (128). (128) hí gái-sai tí pioáhai ʔogái 1.pro say-nmlz 1.pro soda want ‘His saying is: “I want soda”.’ Sauerland (2019) presents evidence that the direct quote acts as a complement clause in such constructions. He conducted a comprehension study where consultants heard a speaker producing a false statement followed by another speaker reporting this false statement in a structure such as (128). When asked for felicity judgments, consultants often correctly classified the first speaker’s statement as false but the report as true. A coordinated or paratactic construction would not render the reported statement as true, so Sauerland concludes that there is embedding; that is, what is true is that the first speaker said the sentence, not the sentence itself. Further evidence that this is an embedded construction comes from the fact that they can stack, as in (129), where the speech of a third speaker is embedded under the speech of a second one. (129) Maaga hi gái-sai Hiahoai hi gái-sai tí pioáhai ʔogái Maaga 3.pro say-nmlz Hiahoai 3.pro say-nmlz 1.pro soda want ‘Maaga says that Hiahoai says “I want soda”.’ (Lit. ‘Maaga’s saying is: “Hiahoai’s saying is: ‘I want soda’”.’) Adverbial clauses are marked by suffixation or free adverbials. Temporal adverbial clauses are marked with the sequentials -so and -áo, demonstrated in (130) and (131).19 These suffixes are phonologically-conditioned allomorphs: -áo is used after consonants (D. Everett 1986: 263). (130) mái ʔi kahápi-so tiobáhai hi gái-sai ti ʔabi-hái parent 3f.pro go-seq child 3.pro say-nmlz 1.pro stay-fut ‘When the mother left the child said: “I will stay”.’

19 When a verb is marked with a sequential, it indicates that another situation will unfold subsequently.

21 Pirahã (Apáitisí)

989

(131) ti ʔaboopai-ta-h-áo pi ʔigi-hai pioaisai 1.pro return-reg-decl-seq 3.clf:liquid bring-fut soda ‘When I come back again, I will bring soda.’ Finally, in reason and result utterances, the reason clause is followed by the adverbial taio ‘because, therefore, then’ and precedes the result clause, as in (132). (132) ʔií ʔi ʔáob-áhá taio ti ʔi(baisi) kahápí-hiab-ahá tree 3inan.pro fall-decl because 1.pro spouse go-neg-decl ‘Because a tree had fallen my husband couldn’t go.’ Turning to coordination, the major strategy for indicating clausal conjunction is shown in (133), where the first clause is paratactically juxtaposed to the second one, which ends with the associative particle píai ‘too’.20 In (133), the verb is elided in the second clause of the conjunct, which is common in Pirahã. (133) ti iíhoí ʔog-abagai ipóihi piai 1.pro beads want-frus woman too ‘I want beads. My wife (wants beads) too.’ Disjunction is expressed through plain juxtaposition of the two clauses, as in (134). (134) tiʔisi ʔibáboi-haí hai ti ʔoi kahápi-hái hai fish fish-fut dub 1.pro woods go-fut dub ‘I will go fishing (or perhaps) I (might) go hunting.’ (D. Everett 1986: 226) A preliminary description of information structure in Pirahã made by D. Everett (1986) shows that topics seem to occur both rightward, as in (135), and leftward, as in (136), and there is also a tendency to repeat topics both leftward and rightward, illustrated in (137). (135) pii-gió-ʔio ʔigóá-ʔai ʔagaoa-ʔai water-down-ingr travel-cop canoe-ins ‘By canoe, (he) travels downriver.’ (D. Everett 1986: 228) (136) tíihí hi bigí káob-áhá Brazil.nut 3.pro ground fall-decl ‘The Brazil nut, it fell to the ground.’ (D. Everett 1986: 228) 20 By ‘paratactically juxtaposed’ I mean that there is a prosodic break.

990

Raiane Salles

(137) Paigi hi ʔob-áaʔáí Paigí Paigí 3.pro see-much Paigí ‘Paigí, he knows a lot, Paigí.’ (D. Everett 1986: 228) Example (136) may not really demonstrate topicalization; as discussed in Section 3.1, arguments of the clause may be doubled by a clitic, in which case tíihí ‘Brazil nut’ would be in subject position, not topicalized. This ambiguity might be resolved prosodically; D. Everett (1986) reports that topics are marked by intonation and a pause between the topic and the remainder of the sentence. It is not clear yet what constraints are imposed on such constructions, or if they have different pragmatic implications. For instance, a rightward topicalization that D. Everett (1986) describes as a clarification topic could be more precisely described as an afterthought; see (135). D. Everett also shows cases in which both arguments are right dislocated, arguing that the leftmost is always the subject and the rightmost the object, as in (138). (138) ʔohoáop-á taío páʔaihi ʔigagí eat-decl because chicken pepper ‘Chickens eat peppers (for) food, therefore.’ (D. Everett, 1986: 203) Pirahã uses a participant-oriented particle ʔagía to mark focus. It is used for focus on a particular participant when other alternatives are present in the discourse, as in (139). (139) Context: I am distributing fishing line and some men are organizing who comes next. Hoaaboái ʔagía Hoaaboái foc ‘Hoaboai, now you.’ Based on its ability to stand alone and function deictically (as second person), shown in (140), one might argue that ʔagía acts as a pronominal element. However, ʔagía may co-occur with a third-person pronominal element, focusing the participant it refers to, as in (141), which suggests that it does not function pronominally in this case. It is not yet clear whether this particle can be used for focus in general, or if it is restricted to participant-oriented functions. (140) Context: men are playing a game. When someone forgets it is their turn, the other players alert them by saying: ʔagía! foc ‘It’s you!’

21 Pirahã (Apáitisí)

991

(141) hi ʔagía gái-sai ti baábaop-á 3.pro foc say-nmlz 1.pro sick-decl ‘He (the one we are talking about) said, “I am sick”.’ (D. Everett 1986: 306)

8 Conclusion Pirahã is a Muran language spoken by a small community that has had more than 200 years of contact with foreigners. Fortunately, their language remains healthy, though still vulnerable. This chapter represents past and recent documentation efforts, presenting state-of-the-art knowledge of Pirahã based on the author’s fieldwork and the work of several other researchers. Pirahã is a tonal language with a small sound inventory: three vowels /i, a, o/ and ten consonants /p, b, t, k, g, ʔ, s, h, w, j/. Interesting phonological aspects include the genderlect distinction between [s] and [h] and the fact that the onset of a syllable contributes to its weight. The basic constituent order is SOV, with neutral alignment in the pronominal system and verbal agreement. Along with cross-linguistically common word classes such as verbs and nouns, there are also some interesting classes, such as a modifier class that can be divided into words that can function as both adjectives and adverbs, and words that function as only one or the other. There is also a class of numerals that denote approximate quantities – something which is cross-linguistically rare – and a class of semantically-transparent postpositions. Nouns do not inflect for gender or number and are divided into at least six noun classes marked by pronominal elements. These pronominal elements are indeterminate between free forms and clitics, suggesting that a grammatical shift to an agreement system might be in process. Of additional interest is the fact that noun phrases rarely have determiners: Pirahã is articleless and has a single demonstrative which performs a general deictic function. Compounding is a very productive strategy in Pirahã, and it is used both for noun and verb formation. Verbs may be inflected through suffixation for tense, aspect, mood, and evidentiality. Pirahã has been claimed to lack numbers, color terms, embedded clauses, and tense (see D. Everett 2005). However, this chapter demonstrates that there is a numeral system, unusual though it may be (encoding approximate quantification); there are color terms; there are several examples of embedding; and there is at least one, future-oriented temporal marker. For those interested in continuing research on Pirahã, this chapter highlights many aspects of the language that would benefit from further research, such as tonology; the modal system and its relationship with temporal, aspectual, and evidential notions; the modifier class; the approximate numeral system; noun classes; the phenomenon of clitic doubling; and information structure. I hope to have in-

992

Raiane Salles

spired future researchers to take part in the documentation of the last surviving language of this family.

9 Acknowledgements I thank the collaboration of the Pirahã, especially to Hiahoai Pirahã, Iaisoai Pirahã, Ioá Pirahã, Opisi Pirahã, Kobio Pirahã, and Toibaiti Pirahã. I thank my advisors Lisa Matthewson, Martina Wiltschko, Hotze Rullmann, Cilene Rodrigues, and Marcel den Dikken; my fieldwork friends Glauber Romling, José Augusto Diarrói, Raimunda dos Santos, Josélia Araújo, Renata Leite, Ayrthon Breder, Sara Campos, Natália Cordeiro, Keren Madora, Elenir Pardo and Netão; my colleagues in the Sound Change in Pirahã project, Emily Sadlier-Brown and Isabel Salomon; and several funding agencies that supported this work: PUC-Rio, SJC Fellowship, and Jacobs Fund. Finally, I thank the editors and anonymous reviewers of this volume for their insightful feedback, which certainly improved the quality of this chapter.

10 References Aikhenvald, Alexandra. 2012. The Languages of the Amazon. Oxford: Oxford University Press. Amaral, Luiz, Marcus Maia, Cilene Rodrigues, Tom Roeper, Filomena Sândalo & Glauber R. Silva. 2018. Self-embedded recursive postpositional phrases in Pirahã: A pilot study. In Luiz Amaral, Marcus Maia, Andrew Nevins & Tom Roeper (eds.), Recursion across domains, 279–295. Cambridge: Cambridge University Press. Amazonas, Lourenço da S. A. de. 1852. Diccionario topographico historico, descritivo da Comarca do Alto Amazonas. Recife: Typ. Comercial de M. Henriques. Amoroso, Marta. 2009. Mura. Povos indígenas no Brasil. Instituto Socioambiental. https:// pib.socioambiental.org/pt/Povo:Mura #Localiza.C3.A7.C3.A3o (28 April 2019.) Birchall, Joshua, Michael Dunn & Simon J. Greenhill. 2016. A combined comparative and phylogenetic analysis of the Chapacuran language family. International Journal of American Linguistics 82(3). 255–284. Croft, William. 1991. Syntactic categories and grammatical relations: The cognitive organization of information. Chicago: University of Chicago Press. Croft, William. 2001. Radical construction grammar: Syntactic theory in typological perspective. Oxford: Oxford University Press. Dehaene, Stanislas, Ghislaine Dehaene-Lambertz & Laurent Cohen. 1998. Abstract representation of numbers in the animal and human brain. Trends in Neuroscience 21. 355–361. Dehaene, Stanislas, Elizabeth Spelke, Philippe Pinel, Ritta Stanescu & Susanna Tsivkin. 1999. Sources of mathematical thinking: behavioral and brain-imaging evidence. Science 284. 979–974. Dixon, R. M. W. 2003. Demonstratives: A cross-linguistic typology. Studies in Language 27(1). 61–112. Dryer, Matthew. 2007. Word Order. In Timothy Shopen (ed.), Language typology and syntactic description, vol. 1, 2nd edn. Cambridge: Cambridge University Press.

21 Pirahã (Apáitisí)

993

Evans, Nicholas. 2000. Word classes in the world’s languages. In Geert Booij, Christian Lehmann & Joachim Mugdan (eds.), Morphology: A handbook on inflection and word formation, 708–732. Berlin: Mouton de Gruyter. Everett, Caleb & Keren Madora. 2012. Quantity recognition among speakers of an anumeric language. Cognitive Science 36(1). 130–134. Everett, Daniel. 1979. Aspectos da fonologia do Pirahã. Campinas: Universidade Estadual de Campinas MA thesis. Everett, Daniel. 1982. Phonetic rarities in Pirahã. Journal of the International Phonetics Association 12. 94–96. Everett, Daniel. 1985. Syllable weight, sloppy phonemes, and channels in Pirahã discourse. Berkeley Linguistics Society 11. 408–416. Everett, Daniel. 1986. Pirahã. In Desmond C. Derbyshire & Geoffrey K. Pullum (eds.), Handbook of Amazonian languages. Vol. 1, 200–325. Berlin: Mouton de Gruyter. Everett, Daniel. 2005. Cultural constraints on grammar and cognition in Pirahã: Another look at the design features of human language. Current Anthropology 46. 621–646. Everett, Daniel. 2009. Pirahã culture and grammar: A response to some criticisms. Language 85(2). 405–442. Everett, Daniel & Keren M. Everett. 1984. Syllable onsets and stress placement in Pirahã. In Mark Cobler, Susannah Mackaye & Michael T. Wescoat (eds.). West Coast Conference on Formal Linguistics III, 105–117. Stanford: Stanford University. Everett, Keren M. 1987. Dictionary (Pirahã-Portuguese). Everett, Keren. M. 1998. The acoustic correlates of stress in Pirahã. The Journal of Amazonian Languages 1(2). 104–162. Fitch, Tecumseh, Marc D. Hauser & Noam Chomsky. 2005. The evolution of the language faculty: Clarifications and implications. Cognition 97(2). 179–210. Futrell, Richard, Laura Stearns, Daniel L. Everett, Steven T. Piatandosi & Edward Gibson. 2016. A corpus investigation of syntactic embedding in Pirahã. PLOS One 11(3). https://doi.org/ 10.1371/journal.pone.0145289. Gallistel, Charles R. & Rochel Gelman. 1992. Preverbal and verbal counting and computation. Cognition 44. 43–74. Gonçalves, Marco A. 2001. O mundo inacabado: Ação e criação em uma cosmologia amazônica: Etnografia Pirahã. Rio de Janeiro: Editora da UFRJ. Gordon, Matthew. 2005. A perceptually-driven account of onset-sensitive stress. Natural Language and Linguistic Theory 23(3). 595–653. Gordon, Matthew. 2006. Syllable weight: Phonetics, phonology, typology. New York: Routledge. Gordon, Peter. 2004. Numerical cognition without words: Evidence from Amazonia. Science 306. 496–499. Hanke, Wanda. 1950. Vocabulário e idioma mura dos índios mura do rio Manicoré. Arquivos: coletânea de documentos para a história da Amazônia 3. Vol. 12, 3–8. Manaus: Associação Comercial do Amazonas. Hauser, Marc D., Noam Chomsky & Tecumseh Fitch. 2002. The Faculty of Language: What is it, who has it, and how did it evolve? Science 298. 1569–1579. Heinrichs, Arlo. 1964. Os fonemas de Mura-Pirahã. Boletim do Museu Paraense Emílio Goeldi 21. Hengeveld, Kees. 2004. Illocution, mood and modality. In Geert Booij, Christian Lehmann & Joachim Mugdan (eds.), Morphology: A handbook on inflection and word formation, vol. 2, 1190–1201. Berlin: Mouton de Gruyter. Jackendoff, Ray & Steven Pinker. 2005. The nature of the language faculty and its implications for evolution of language (Reply to Fitch, Hauser, and Chomsky). Cognition 97(2). 211–225. Loukotka, Čestmír. 1968. Classification of South American Indian languages. California: University of California Los Angeles.

994

Raiane Salles

Moseley, Christopher (ed.). 2010. Atlas of the World’s Languages in Danger. 3rd edn. Paris: UNESCO Publishing. Online version: http://www.unesco.org/culture/en/endangered​ languages/atlas Nevins, Andrew, David Pesetsky & Cilene Rodrigues. 2009a. Pirahã Exceptionality: A Reassessment. Language 85(2). 355–404. Nevins, Andrew, David Pesetsky & Cilene Rodrigues. 2009b. Evidence and Argumentation: A Reply to Everett. Language 85(3). 671–681. Nimuendaju, Curt. 1948. The Mura and Pirahã. In Julian H. Steward (ed.). Handbook of South American Indians. Vol. 3: The Tropical Forest Tribes, 255–269. Washington, D. C.: United States Government Printing Office. Pinker, Steven & Ray Jackendoff. 2005. The faculty of language: what is special about it? Cognition 95(2). 201–236. Prado, Frederico. 2018. Reassessing multiple adjectival modification in Pirahã. Campinas: Universidade Estadual de Campinas undergraduate thesis. Rodrigues, Ivelise & Adélia E. Oliveira. 1977. Alguns aspectos da ergologia Mura-Pirahã. Boletim do Museu Paraense Emílio Goeldi 65. Rodrigues, Cilene; Raiane Salles & Filomena Sândalo. 2018. Word order in obligatory control: Evidence for self-embedding in Pirahã. In Luiz Amaral, Marcus Maia, Andrew Nevins & Tom Roeper (eds.), Recursion across domains, 111–126. Cambridge: Cambridge University Press. Sadlier-Brown, Emily; Raiane Salles & Isabel Salomon. 2022. Exploring variation & change in a small-scale indigenous society: the case of (s) in Pirahã. Linguistics Vanguard. Manuscript. Salles, Raiane. 2015. Understanding recursion and looking for self-embedding in Pirahã: The case of possessive constructions. Rio de Janeiro: Pontifícia Universidade Católica MA thesis. Salles, Raiane. 2018. Bare nouns in Pirahã: A radically [+arg,-pred] language? In Kimberly Johnson & Alex Goebel (eds.), Semantics of Under-represented Languages in the Americas (SULA) 10. Toronto: University of Toronto. Salles, Raiane. in prep. Why you so obsessed with D: Argumenthood with other functional heads in Carioca BrP and Pirahã. Vancouver: The University of British Columbia dissertation. Salles, Raiane & Lisa Matthewson. 2016. Existentials and (in)definiteness in Pirahã. In Thuy Bui & Rudmila-Rodica Ivan (eds.), Semantics of Under-represented Languages in the Americas (SULA) 9, 127–140. Amherst, MA: Graduate Linguistics Student Association, University of Massachusetts. Salomon, Isabel. 2020. Nasal contrastiveness in Pirahã. Poster presented at the Language Sciences Undergraduate Research Conference. Vancouver: University of British Columbia. February 7–8, 2020. Sândalo, Filomena & Maria B. Abaurre. 2010. Orality spreading in Pirahã. Línguas Indígenas Americanas (LIAMES) 10. 7–19. Sauerland, Uli. 2010. Experimental evidence for complex syntax in Pirahã. Tromso: Universitätsbibliothek Johann Christian Senckenberg. Lingbuzz: lingbuzz/001095. (29 April 2019.) Sauerland, Uli. 2018. False speech reports in Pirahã: A comprehension experiment. In Luiz Amaral, Marcus Maia, Andrew Nevins & Tom Roeper (eds.), Recursion across domains, 21–34. Cambridge: Cambridge University Press. Sheldon, Steven N. 1974. Some morphophonemic and tone perturbation rules in Múra-Pirahã. International Journal of American Linguistics 40(3). 279–282. Sheldon, Steven N. 1988. Os sufixos verbais Mura-Pirahã. Série Linguística 9. 146–175. Silva, Glauber R. 2014. A feature geometric analysis of Pirahã phonology and tonology (Mura). Revista LinguíStica 10(2). 1–20. Thomason, Sarah G. & Daniel L. Everett. 2001. Pronoun borrowing. Berkeley Linguistics Society (BLS) 27. 301–315.

Zachary O’Hagan

22 Taushiro 1 2 3 4 5 6 7 8 9 10 11

Introduction Taushiro history and culture Segmental and prosodic phonology The noun phrase The verbal word (simple clauses) Other verbal clause types Combining clauses Nonverbal clauses Conclusion Acknowledgments References

1 Introduction Taushiro (taus1253) is a language isolate traditionally spoken on the rightbank, western tributaries of the middle Tigre River in the Loreto region of modern-day Peru. As of 1971, the 18 Taushiros who resided in this region formed two groups, the Atontu’tua and Einontu’tua (Alicea Ortiz 1976: 1, 6). Since 1971, all except one man, Amadeo García García, have passed away, and Amadeo is considered by himself and outsiders to be the last fluent speaker of the language. In the precolonial period its linguistic neighbors included Zaparoan languages to the west, north, and east, and Roamaina and Zapa to the west and south (Grohs 1974, Santos 1986[1684]). Given the scarcity of data, this chapter describes Taushiro history, culture, and language as thoroughly as possible. The linguistic description focuses on main clause morphosyntax, which is presently best understood. It is based on linguistic, historical, and ethnographic data collected from Amadeo García García during one week in Iquitos in June 2015, and, to a lesser extent, on previously published material, especially Alicea Ortiz (1975a, b, c, 1976).1 The representations and analysis are preliminary. 1 Alicea Ortiz’s work was conducted under the auspices of SIL International. The initial period, on which published linguistic materials are based, lasted from April 1974 to August 1976 (Alicea Ortiz 1976: 3–4). The work was conducted both on the Aucayacu River among the Atontu’tua and at the SIL center in Yarinacocha. According to Mary Ruth Wise (p.c., Jun. 20, 2015), sound recordings from this period were archived with the Peruvian Ministry of Education in 1976, but their whereabouts are currently unknown. In 1979, Alicea Ortiz attempted to locate additional speakers on the Ecuadorian side of the Tigre basin, to no avail (Alicea Ortiz, p.c., Oct. 31, 2015). Regular work continued through 1980, then annual visits were made through 1989, before which portions of the Old and New Testaments were translated into Taushiro and recorded on cassette (Ayala et al. 2006: 161); Zachary O’Hagan, Department of Linguistics, University of California, Berkeley https://doi.org/10.1515/9783110432732-009

996

Zachary O’Hagan

22 Taushiro

997

While Alicea Ortiz’s contribution to the study of Taushiro is pathbreaking, examples are not segmented, the specificity of glosses varies, and many grammatical phenomena are incidentally exemplified in sections not devoted to them, making it a difficult corpus to interpret. More substantively, as Alicea Ortiz herself at times states, the description is often inconclusive (and in some cases contradictory). As a result, I have opted to prioritize my own data, noting divergences from Alicea Ortiz where relevant. Because my data is in turn limited, one effect is that many areas of Taushiro grammar go undescribed in this chapter. I have opted for this as opposed to reintroducing more inconclusive description of Taushiro into the literature. There are some exceptions to this, where I felt a grammatical category was too central to ignore; in these cases I am explicit about how Alicea Ortiz’s work can shed light on things and about what requires further research. Overall, if nothing is said about a particular area of Taushiro grammar in this chapter, then we do not yet understand the language well enough to say anything about it with confidence.

2 Taushiro history and culture I begin with 20th century history before summarizing basic aspects of material culture and then discussing the Jesuit era. Unless cited otherwise, information in this section comes from my personal discussions with Amadeo García. Approximate dates are calibrated relative to his assumed age at the time of particular events.

2.1 The twentieth century Traditional Taushiro social organization involved patrilineal clans. As of 1941, at the outbreak of the Peruvian-Ecuadorian War, two such clans remained, distributed across two right-bank tributaries of the middle Tigre River: the atontuʔtuaʔ on the Huanganayacu River (Taushiro. uwaʔxanoʔ), specifically on a left-bank stream called Ava Blanca (T. axaʔjanoʔ), and the enontuʔtuaʔ in the extreme headwaters of the Aguaruna River (T. iijaʔ). At this time, Peruvian and Ecuadorian forces undertook military campaigns on the Huanganayacu, and the atontuʔtuaʔ fled to a leftbank tributary of the Aguaruna called Gómez Caño (T. nijoʔnoʔ). Prior to this permanent migration, there were probably seasonal migrations between the Ava Blanca and Gómez Caño: according to Amadeo García, the latter is considered the site where all Taushiros originated and spread from. In the early 1950s there were three large houses in Gómez Caño with some thirty residents; Amadeo García was born there circa 1949.

these were 64 pages of the former and some passages from the latter (Wise, p.c., Oct. 4, 2017). Lastly, Alicea Ortiz’s correct first name is Nectalí, not Neftalí as in SIL publications.

998

Zachary O’Hagan

In the first half of the twentieth century, Taushiros were severely affected by disease, including smallpox, measles, whooping cough, and the common cold (Alicea Ortiz 1976: 6). They additionally suffered from raids conducted by outsiders resident on the Tigre River as early as the 1920s, as reported by Tessmann (1930: 583, translation mine) below. Such raids are also mentioned by Alicea Ortiz (1976: 7, translation mine): “in the past mestizos are said to have taken some members of the group to other places and they never returned.” According to a newspaper article from Iquitos (“El Eco”, September 22, 1926), a small number of Pinches [read: Taushiros] on the Huangana were caught unawares in the forest and brought to the Tigre, where they were given away and later released to their own group. A consequence of this must be that they fell upon and killed several balata-tappers [a latex] who had ventured further into the interior.

Circa 1957, the atontuʔtuaʔ jointly settled with the enontuʔtuaʔ on the Aucayacu River (T. awanoʔ), another left-bank tributary of the Aguaruna, at the behest of Cesario Hualinga, a Chicham (Jivaroan) patrón (labor boss) who extracted Cuoma macrocarpa (Sp. leche caspi, Taushiro héʔnìwà) and Manilkara bidentata (Sp. balata) from the region. Hualinga took two Taushiro women as wives, both likely born in the 1930s. He is said to have given only clothes and alcohol in exchange for work. Circa 1964, patrón Felipe Vásquez came into the region. He was in the business of leche caspi, peccary and ocelot hides, game meat, and monkeys. In return, he introduced shotguns (ànètá) into Taushiro communities, as well as radios, machetes (hòhwá), and axes (àxí̵ ). Around 1970, oil exploration was begun in the neighboring Corrientes basin, and a pipeline was built northeastward through Gómez Caño to Bolognesi on the Tigre (in exchange for manioc flour, sugar, and milk). In 1971, SIL International missionary-linguist Daniel Velie visited the Taushiros, and subsequently Nectalí Alicea Ortiz (SIL) was sent to carry out linguistic and Bible translation work with Amadeo García (Ayala et al. 2006: 160–161; Casey 2017) – also see footnote 1. At this time, most atontuʔtuaʔ were monolingual, while the enontuʔtuaʔ were “much more bilingual and [preferred] Spanish, except for women and one monolingual elder man” (Alicea Ortiz 1976: 8, translation mine). This difference probably points to a longer history of contact with Spanish-speaking patrones in the Aguaruna basin than in the Huanganayacu basin. Also in 1971, another patrón removed all enontuʔtuaʔ to work on the Lejía River, a left-bank tributary of the Tigre outside Taushiro territory, severing regular contact between the two groups until some came to reside years later in the regional center of Intuto. As of 1976, there were six adult atontuʔtuaʔ, six adult enontuʔtuaʔ, and six enontuʔtuaʔ-mestizo children (ibid.: 6). These figures are probably overly conservative.

22 Taushiro

999

2.2 Material culture Taushiros constructed houses directly on the ground from (Sp.) irapay palm (àhàã́) thatch, with four foundation poles and holes in the upper reaches to allow smoke from interior cooking fires to escape. The interior was illuminated with torches of copal (Sp.), a resin. Taushiros wove hammocks (ùné), mats (ànòwàhì̵nákì̵ ), net bags (òɕìtòã́), and sieves (àwí̵xà), but not mosquito nets. Taushiros only distilled peach palm (àwànté) as an alcoholic beverage, and not manioc (àhjã́), although they cultivated the latter. This beverage was initially stored in wooden troughs and covered with plantain leaves to permit fermentation. Other cultigens included sweet potatoes (ùnùhìtú), plantains (àntá), bananas (hóʔtì), maize (hwáʔ ntù), peach palm, guava (àníhì), and caimito (kànã́nì). Knives (hòhwá) were made from peach palm wood and axes (àxí̵ ) from stone. Taushiros hunted with spears (àníwì) – tipped with agouti teeth (Alicea Ortiz 1976: 2) – clubs, blowguns, and disguised holes in the ground serving as traps. They fished with spears, line (ànòjá hòòkí̵ ) and hook (kwã́) – line made from chambira palm (òtúà) fiber, and hooks from tinamou bones – barbasco (hèʔɕã́), and traps (àwàtóʔ). According to Alicea Ortiz (1976: 2), women’s hair was cut with piranha teeth. Bows were not used, nor were canoes; riverine travel was by raft. Food was cooked in basic clay pots (wàjùntùnàjá), and cooking techniques included boiling, roasting in plantain leaves, and smoking. Frying was not practiced, nor was toasted manioc flour produced. There was no elaborated pottery tradition, but artistic traditions such as music included flutes (àhúhwí̵ ) and drums (nùʔwìcɕ͡ í), and body painting with genipapo (háʔwì̵) and annatto (àwàʔá) was practiced. Men wore a tunic-like garment (ànòwá) that reached to the knees, and women wore a skirt only, both woven from chambira, which was also used to weave net bags. Sieves were woven from ungurahui palm (àjùxámà). Taushiro shamans were capable of curing and cursing by invoking helper spirits and consuming ayahuasca (ànùʔwí̵ ) – the proper use of which was revealed by the spirit of an armadillo in Taushiro mythology – ojé (àɕṍwà), and toé (xìwànú), but they did not use rattles, shamanic stones, or ladders. The deceased were wrapped in a hammock and buried directly in the ground, with their head facing toward the sun. A fire was subsequently lit on top of the burial mound, the house of the deceased was burned, and all nearby residents moved to a new location.

2.3 Subgroups and ethnonyms in the Jesuit era The ethnonym Taushiro is a Hispanicization of a Chicham word that in some varieties is tawíʃuɾ ‘non-Chicham indigenous person’ (Tuntiak Katan Jua, p.c. 2015). It is first attested in Villarejo (1943: 100) and likely dates to the arrival of Chichamspeaking patrones to the region. In the Jesuit era (1638–1767), modern-day Taushiros

1000

Zachary O’Hagan

were designated by a number of ethnonyms. The most reliably identifiable Taushiro ethnonyms are those that end in the sequence , an early representation of -tuaʔ, which surfaces on all attested Taushiro clan names. These include Asarunatoa and Avitoa. Other ethnonyms from this period that definitively refer to Taushiro speakers are Atio, Pinche, and Uchupa Auca/Ceniciento.2 Araza and Pava probably refer to Taushiro speakers. It can be deduced from the itineraries of colonial era expeditions that the Asarunatoas, and possibly the Cenicientos, lived in the modern-day Aguaruna basin, while the Pinches and Avitoas, and possibly the Atios, lived in the Huanganayacu basin. The Asarunatoas, Avitoas, and Atios were first contacted on the Huanganayacu River during the 1656 expedition of the Spaniard Martín de la Riva Herrera. In January 1657, Salvador Velasques de Medrano, a member of the expedition, wrote about this initial encounter in the following terms (293–294, translation mine). … taking with him more than eighty Indians of the said Conchas and Ruamaynas, and at the end of nine days we arrived at the said Province of Avitoas […] and asking after the people that there were in those provinces, they said that there were some, and that they were all already armed against us […] And arriving at some houses we found them deserted, and, asking about the cause, the prisoners said that they had withdrawn, so that they could all join together upriver […] and they said how upriver of the house of a leader [Sp. cacique] named Taraguaz there were many people withdrawn, waiting, because they wanted to know how we fought […]

He indicated potentially numerous Taushiro-speaking clans present in the region (ibid., translation and emphasis mine). […] they gave up and turned over their spears and shields, clubs and other weapons that they use, and the said general entered the said house embracing the said leader, who seemed to be of the highest authority that the province had, who said that he wanted to be friends, and that they would remove all of his people in a peaceful way, as indeed he did […] and having word that there nearby on another branch of the river there was another nation that they called Attios […] the said leader Taraigua [sic] for whom they have great respect they made peace, and the Açoronatoas [sic] did the same, that they are all neighbors of one another, twentysix leaders were found and seen coming out […]

In June 1684, Father Tomás Santos, resident priest at the mission Santos Ángeles de Roamainas on the Pastaza River, set out in search of the hitherto uncontacted Pinches, first traversing the Pastaza-Tigre interfluvium to what he referred to as the

2 Santos (1986[1684]: 353–354, translation mine, emphasis in original) explains these two ethnonyms in the following terms: “another nation that in the language of the Inca is named Uchupa auca, which means ‘the ashen ones’; and it is because the skin of these people is very dark, the color of ash.” Also note that Father Franz Veigl (2006[1798]: 114), generally a quite accurate source for the colonial era history of the region, specifically indicates that the Pinches, Pavas, Arazas, and Uchpas (cf. Uchupa Aucas) were related.

22 Taushiro

1001

Capirona River (in fact probably the modern-day Copalyacu), then following the Corrientes River, and then proceeding upstream along the Tigre proper (Santos 1986[1684]). He was accompanied by 27 Roamainas and eight Pavas. On July 2 they reached the Asarunatoas on the Aguaruna, being visited while there by the Cenicientos, who had not previously been contacted. He then traveled, on the advice of the Asarunatoas, to contact the Pinches, along with the Avitoas, on the Huanganayacu.3 Following Santos’ expedition, the Pinches and their immediate neighbors appear to have been left alone for some fifteen years. By the late 1690s, however, Jesuit missionary efforts in their territory had intensified, as summarized by Maroni (1988[1738]: 265–266, boldface emphasis mine). [Speaking of Pinches, Pavas, and Arazas …] They had become friends a long time before together with the Roamainas and Zapas, but, as communication had ceased, it was necessary to send someone to speak to them again. Who opened the door for that was the Roamaina leader Damian […] This was who in the year 1698 led to his territory Father Nicolás Durango, with the aim of inviting them to come out and settle on the Pastaza. They refused, being used to living in the middle of the forest, and thus in their own territory two populated places began, the one called San Joseph de los Pinches, and the other Santa Maria de la Asumpcion de los Pavas. Father Superior Gaspar Vidal also having entered there, in the year 1700, he found that there were some five hundred fighting-age men [Sp. indios de lanza], whom he invited and insisted that they come out to the Pastaza to settle, if they wanted to have missionaries that would attend to them regularly; but his invitation did not have an effect until the year 1708, in which, Father Pedro de Campos being in charge of that mission, he entered again into their territory and finally persuaded the Pinches and Pavas to leave their territory and join with the Roamainas on the banks of the Pastaza. […] with this move they sickened and many died. The father seeing this, the ones who were still living, he decided to settle them a ways into the forest on the banks of a beautiful stream, where they live at present in the company of some Roamainas and Arazas, but in very small number. Shortly after the founding of this new village, that today is called San Joseph de los Pinches, Father Pedro having been called to the college of Quito, there succeeded him in the year 1713 Father Juan Zaldarriaga, in whose time, instigated by the Devil, the governor of the Roamainas Ignacio Ratihí, withdrew to the forest with three concubine sisters [i.e., co-wives], and led as many Roamainas as Pinches to flee from the village. The miserable blind man persevered in his apostasy until the year 1721, in which, having encountered him incidentally in the forest very regretful of his many offenses; but the Roamainas and Pinches have still not been able to be recovered.

3 The Pavas were contacted sometime between the writings of Father Francisco Figueroa in 1661, in which they are not mentioned, and the 1684 expedition of Father Tomás Santos, when they appear to have been used as translators with the Asarunatoas on the Aguaruna (see below). Although their linguistic affiliation is less certain, they appear to have lived on streams that drained into the Corrientes at approximately the same latitude as the Aguaruna and Huanganayacu (Maroni 1988[1738]: 265). The Arazas do not appear to have entered into sustained contact with outsiders until a 1731 expedition of Andoas, a Zaparoan people, removed them from a right-bank tributary of the Tigre apparently upriver of the mouth of the Huanganayacu (Maroni 1988[1738]: 110; de Zárate 1988[1735]: 402).

1002

Zachary O’Hagan

Although Maroni never missionized in the Pastaza or Tigre basins, his writings are important to Taushiro history for a number of reasons. First, we glean that he understood the Pinches, Pavas, and Arazas to be linguistically related to the Roamainas.4 Second, we get some sense of the size of the Pinche-Pava-Araza population circa 1700: assuming an average nuclear family size of five individuals, they would have numbered some 2,500 individuals, and with elderly individuals, perhaps some 3,000. It seems that disease had already exhibited a significant impact on these groups (e.g., Velasques de Medrano 2003[1657]: 294), and assuming an initial mortality rate of 25 %, we might expect a pre-contact population on the order of some 4,000 individuals, which suggests a fairly dense settlement pattern in the Corrientes-Tigre interfluvium. Third, we learn that Pinches fled the Jesuit-controlled missions as early as the mid 1710s, and since it is known that San José de los Pinches never counted more than 200 individuals as residents (Grohs 1974: 86), we can be fairly certain that, even with the advent of disease, a sizeable Taushiro-speaking population remained in this region throughout the eighteenth century. By 1735, probable Taushiro speakers at San José de los Pinches did not number more than approximately 150, probably the result of a combination of disease and escape (de Zárate 1988[1735]: 402): there were 46 families totaling some 200 individuals, approximately 50 of whom were Arazas (ibid.). It is noteworthy that the Asarunatoas and Avitoas, crucial to linking the groups visited by de la Riva Herrera in 1656 with the modern-day Taushiros, are not mentioned in major Jesuit works of the region dating from the following century, and that the Atios are not mentioned after the 1650s. This suggests either that Jesuits had ceased identifying Taushiro-speaking groups by clan name or that these clans had been so reduced by the eighteenth century that the Jesuits were simply no longer aware of them. In contrast, at various points during the eighteenth century, some Jesuits linked the Pinches with Roamainas and Zapas. By 1767 the Jesuits appear to have settled on considering at least three other groups to be related to the Pinches: the Arazas, Pavas, and Uchpas, so called ramas de esta tribu ‘branches of this tribe’ (Veigl 2006[1798]: 114). Given the relatively precise locations known of the Asarunatoas, Atios, Avitoas, Cenicientos, and Pinches in comparison with twentieth century evidence, as well as linguistic evidence in the form of , we can conclude that these groups were almost certainly Taushiro-speaking. The Arazas were probably Taushiro-speaking, evidently also occupying a stream that drained into the Tigre at roughly the same latitude as the first four groups mentioned here. The linguistic affiliation of the

4 Since Tessmann (1930: 444), Roamaina has been assumed to be the same language as Omurano (see O’Hagan (this volume) for a rejection of that unsubstantiated claim). While the Roamainas accompanying Santos were clearly able to communicate with Taushiros, it is unclear whether this is because they spoke the same or a related language, on the one hand, or because they were bilingual, on the other. For the time being, Roamaina is best considered unclassified.

1003

22 Taushiro

Pavas is slightly more in doubt, although Jesuit hypothesis and their possible role as translators with the Asarunatoas in the Santos expedition make it also probable that they were Taushiro-speaking. In 1767 Carlos III, King of Spain, expelled the Jesuits from all Spanish territories. Veigl’s writings, which date from this time, are the last before a lengthy period in which little is known of Taushiro history, a period which, given extant and available sources, spans the rest of the eighteenth century, subsumes most the nineteenth century, and includes the early twentieth century.

3 Segmental and prosodic phonology This section is based on my own lexical elicitation unless stated otherwise. I note particular divergences from Alicea Ortiz’s (1975a) analysis at the end of the section. Taushiro exhibits 17 phonemic consonants (Table 22.1). The only bilabial segment is the labiovelar glide /w/. Five fricatives are contrasted at three different places of articulation. Alveolar and velar stops include plain and prenasalized counterparts. Glides exhibit nasal allophones in the context of a nasal vowel. Phonemes with restricted distributions are enclosed in parentheses: /ɾ/ is found only in grammatical morphemes; for discussion of nasals, see below. The velar nasal occurs only in one grammatical morpheme, the recipient suffix -ŋɨ (often realized only as nasality on a preceding vowel, but present in careful speech). Similarly, the prenasalized velar stop is attested only in two grammatical forms, /aŋka/ and /niŋka/, described in Section 7.1. In contrast, the alveolar nasal and prenasalized alveolar stop are widespread. More generally, prenasalized stops are posited due to the existence of a very small number of roots that exhibit them in initial position.5 Currently only two roots are known to exhibit the prenasalized

Tab. 22.1: Taushiro consonants. bilabial stop prenasalized stop nasal fricative affricate flap ikpglide

alveolar

palatal

velar w

t n

t n ɕ cɕ͡

kk (ŋk) (ŋ) x

glottal ʔ

h hj hw

(ɾ) w [w, w̃ ]

j [j, ɲ]

5 An alternative analysis could posit a nasal underspecified for place of articulation, which placeassimilates to a following voiceless stop. While this analysis would be fruitful if [nt] and [ŋk] sequences were restricted to word-internal position – that is, with the placeless nasal restricted to coda position – it would be less expected if these sequences are found word-initially, as is in fact

1004

Zachary O’Hagan

Tab. 22.2: Taushiro vowel qualities.

high mid low

front

central

back

i, ĩ e, ẽ

ɨ, ĩ ̵

u, ũ o, õ

a, ã

alveolar stop word-initially, as in (1), although it is widely attested in word-medial position – see (5) for examples. (1)

a. /ntóɕì/ ‘catfish sp.’ b. /ntíjā/̀ ‘bird sp.’ (Sp. piyapiya)

The distribution of the prenasalized velar stop depends on the analysis of a multipurpose subordinator discussed in Section 7. This form is often realized as [ŋka] (or simply [ga]), but may also be realized as the longer [áŋkà] (or [áŋgà], with voicing), making it seem as if the former is a reduction of the latter. This is indeed the analysis I adopt in this chapter, in which case there are no instances of word-initial /ŋk/ in my data. However, if future work shows there to be two distinct markers, then the subordinator would be an instance of word-initial /ŋk/. For the time being I posit a prenasalized velar stop in parallel with /nt/. Contrasts between fricatives at three places of articulation are illustrated with the triplet in (2). Contrasts between the three glottal fricatives are illustrated in (3). (2)

a. /ɕìːné/ ‘morning’ b. /xèhí/ ‘tapir’ c. /hèʔí/ ‘jaguar’

(3)

a. /àhã/́ ‘our house’ b. /àhjã/́ ‘manioc’ c. /àhwã/́ ‘toucan sp.’

There are six phonemic vowel qualities with contrasts in orality and length (Table 22.2). Nasality seems to spread leftward within a word, targeting the glides /w/ and /j/, with other segments opaque to this process.

the case, and so I do not adopt it here. Finally, avoiding positing coda nasals simultaneously avoids positing complex codas, as would be necessary for sequences of the glottal stop and an alveolar nasal (e.g., [hwáʔntù] ‘corn’).

1005

22 Taushiro

Tab. 22.3: Some attested syllable shapes. pattern

root

gloss

V.CV V.CVC V.CVː Vː Vː.CV CV.CV CV.CV CVC.CV

àwá àwáʔ àhã́í ̵ òó ììná wàjá wàntá tìʔjó

‘tree’ ‘achiote’ ‘palm sp.’ (Sp. irapay) ‘liana’ ‘wasp’ ‘pot’ ‘pineapple’ ‘mud’

Attested syllable shapes are V(ː), CV(ː), and CV(ː)C, that is, with or without length. The only attested coda is /ʔ/.6 Words seem to need to be at least bimoraic, such that V and CV are attested only in words of two syllables or more. Most simplex roots do not appear to be more than four syllables. Some phonotactic patterns are given in Table 22.3. For ease, I represent long vowels orthographically as two vowels, each hosting its own tone mark. Taushiro distinguishes two surface tones: high (H) and low (L). Minimal tone pairs do not exist in my data, though there are near-minimal pairs and groupings of forms of similar phonological shape – as shown in (4) and (5) – for which tones are not predictable. (4)

a. /ùwájù/ Nyctidromus albicollis (bird sp., Sp. tohuayo) b. /ùnàjú/ ‘my knee’

(5)

a. b. c. d.

/ùnúntù/ ‘owl sp.’ /ùnàntá/ Mauritia flexuosa (fruit sp., Sp. aguaje) /àwántù/ ‘Yellow-crowned parrot’ /ùtàntú/ ‘my heart’

LHL LLH LHL LLH LHL LLH

The mora is the tone-bearing unit, as evidenced by combinations of H and L on bimoraic long vowels. Some attested tone patterns are given in Table 22.4 (see Table 22.3 for examples of combinations of H and L on long vowels). Note that roots may be all-L, and contain more than one H, evidencing a tone system as opposed to a stress system. All-H roots are not attested in my data. However, the overwhelming majority of roots exhibit only one H tone.

6 The distribution of the glottal stop in particular warrants attention. In some cases its distribution seems to be prosodically conditioned. For now I represent the /ʔ/ in such positions as phonemic.

1006

Zachary O’Hagan

Tab. 22.4: Some attested tone patterns. pattern

root

gloss

LH HL LLLL LLH LHL HLL LHH LHLH

tàhá ɕánè wàànàhà ànètá àníwì héʔnìwà àwéʔwí aɕíàʔné

‘scorpion’ ‘salt’ ‘woolly monkey’ ‘shotgun’ ‘spear’ ‘tree sp.’ (Sp. leche caspi) ‘man’ Puma concolor (Sp. lluichupuma)

The prosodic properties of morphologically complex stems are not yet well understood, and so the tones of such stems are not represented in this chapter, although the tones of underlying roots are provided in the interlinearization when such roots are attested elsewhere in isolation. It is worth noting that Alicea Ortiz (1975a: 1–13) provides different phonological inventories of consonants and vowels that do not align with my data. Most strikingly, I find no evidence for the complex phone [kpw], an allomorph of /kw/, which she lists as a bilabial, her only bilabial segment besides /w/. For her, prenasalized stops are nasal-stop clusters; she posits a single phoneme /n/ that place-assimilates preceding /k/. She analyzes /ɾ/ as an allophone of /t/, but without a conditioning environment. I find that its distribution is extremely limited (as she does), predominantly occurring in affixes, but it does not follow from that that it is not phonemic. She does not recognize a distinction between the palatal fricative and affricate, analyzing both as an alveopalatal affricate with unpredictable allophony. However, their distribution is unpredictable, and so they must be phonemes, as seen in contrasting with /àntèɕíjò/ ‘anteater’ with /ùntícɕ͡ ì/ ‘agouchi.’ (And /ɕ/ does not occur only adjacent to /i/, e.g., /àɕówà/ ‘tree sp.’ (Sp. ojé), or /hèʔɕá/ ‘barbasco root’.) Relatedly, she recognizes only /h/, but not /hj/ or /hw/, as a phoneme. The latter are treated as combinations of /h/ with /i/ and /u/, respectively. However, these are not full vowels in my data. In general, Alicea Ortiz’s claims regarding allophony are difficult to interpret, for example, /ʔ/ with allophones [n], [t], and [k]. I do not take up those issues further here. In terms of vowels, Alicea Ortiz posits /ɨ/ with (unconditioned) allophones [ɨ] and [e]. I analyze these as separate phonemes. She analyzes nasal vowels as allophones of oral vowels, and long vowels as allophones of short vowels. Although minimal pairs are hard to come by, the distribution of these segments is unpredictable, and so they must be phonemes. (See Table 22.3 for some examples.) She seems to analyze word-level prosody as a stress system consisting of left-aligned iambs with numerous exceptions.

22 Taushiro

1007

4 The noun phrase In this section I describe pronouns, postpositions, possession, quantifiers, and numerals. Possessors, quantifiers, and numerals are frequently not adjacent to the noun that they modify. The Taushiro nominal lexicon distinguishes alienable from inalienable nouns but does not categorize nouns along any other axis (e.g., gender or animacy).

4.1 Pronouns Taushiro utilizes independent pronouns that distinguish three persons in addition to an inclusive (Table 22.5). There is no distinction in number, first-person exclusive being expressed by the simple first-person forms. Long forms occur as core verbal arguments, while short form pronouns are the complements of postpositions (see Section 4.2). Tab. 22.5: Independent pronouns. person

long

short

1 1incl 2 3

úì áì íì nácɕ͡ ò

ú á í á

4.2 Postpositions Non-core arguments are introduced into the clause via a set of postpositional suffixes. (Core arguments do not bear case marking.) The majority of these postpositions express spatial relations, some of which are summarized in Table 22.6 – see (21) for an example. Tab. 22.6: Postpositions (selection). suffix

gloss

-ŋì ̵ -ha -kú -wà -wì ̵ -kúʔkì ̵ -wẽʔ́ wì ̵ -kàné -tàkí ̵

recipient instrumental, comitative locative locative allative illative ablative superessive adessive

1008

Zachary O’Hagan

4.3 Nominal modification 4.3.1 Possession Nouns are possessed via a series of prefixes that distinguish four persons but not number, as summarized in Table 22.7. Note that un- and j- occur before vowel-initial stems. Tab. 22.7: Nominal possessive prefixes. person

prefix

1 1incl 2 3

u(n)aij(a)-

Possessive constructions distinguish alienable from inalienable nouns. When the possessum is inalienable, possessive prefixes attach directly to the root, as in (6a); when the possessum is alienable, the possessor is expressed via the suffixation of cɕ͡ ã́ to a short pronoun, shown in (6b). (6)

a. u-hã́ 1-house ‘my house’

b. u-cɕ͡ ã́ wànántà 1.pro-poss dog ‘my dog’

Nominal possessors similarly combine directly with inalienable nouns and indirectly with alienable nouns via -cɕ͡ ã́. When the possessor is third person, the possessor is always expressed by the short pronoun suffixed with -cɕ͡ ã́, even if the possessum is inalienable (7). (7)

a. i-ɕàwá 2-maternal.aunt ‘your maternal aunt’

ɕàwá b. ja-cɕ͡ ã́ 3.pro-poss maternal.aunt ‘his/her maternal aunt’

4.3.2 Quantifiers The notions expressed by English many and all are expressed by the Taushiro adverb àɕíntù, which need not be contiguous with the noun it modifies (see (13)). A suffix -ɕíà expresses that the quantified referent is unique. This suffix attaches to -cɕ͡ ã́ poss when they co-occur.

22 Taushiro

(8)

1009

ìʔìtú ucɕ͡ aɕia wànántà. wànántà Ø-iʔitu u-cɕ͡ ã-ɕia 3-flee 1.pro-poss-exh dog ‘My only dog ran away.’

There is no morphological expression of number in the language.

4.3.3 Numerals Two morphologically simplex numerals are attested, wàɕìkántù ‘one’ and àɕìʔní ‘two.’ They may be discontinuous with a noun and may also function predicatively, as in (9) and (10), respectively. (9)

wàɕìkántù hóʔɕì úì tàjá ànètáhà. tàjá ànètá-hà wàɕìkántù h-Ø-oʔɕi úi one 1-3-kill 1.pro white-lipped.peccary shotgun-ins.com ‘I killed one white-lipped peccary with a shotgun.’

(10) àɕìʔní ùcɕá ùné. àɕìʔní u-cɕã́ ùné two 1.pro-poss hammock ‘I have two hammocks.’ (Lit. ‘My hammocks are two.’) There seem to be morphologically complex numerals of higher value, at least through ten, but these forms need further investigation, as at least one was said to span two values.

4.3.4 Endocentric denominal derivation Taushiro does not seem to have an elaborated system of endocentric denonimal derivation. Three suffixes are readily mentioned: augmentative -jo (e.g., tàʔá ‘lake’ vs. tàʔjó ‘big lake’), diminutive -ʔka (e.g., hòhwá ‘machete’ vs. hòhwáʔkà ‘knife’), and hydronymic -noʔ, deriving the names of waterways. The latter attaches to nouns, the referents of which are prevalent in the region of the waterway.

5 The verbal word (simple clauses) Taushiro exhibits basic VSO word order, with focused noun phrases (including long-form pronouns), occurring preverbally. Alignment is nominative-accusative.

1010

Zachary O’Hagan

sbj- obj:dir-/refl- ni- ROOT -desid -neg -fut -asp (obl) sbj obj (obl) Fig. 22.1: Verb phrase, partial template.

Core arguments are morphologically unmarked; non-core arguments are introduced via postpositions (Section 4.2). Nominal and pronominal arguments (Table 22.5) commonly occur in addition to these markers, although they are not obligatory, and function to disambiguate instances of syncretism in the subject agreement of transitive verbs. Certain non-core arguments can intervene between the verb and its subject; see (11). Verbal inflection includes future temporal reference, aspect, and negation, all suffixes. Derivational morphology includes causative, reflexive, and incorporated nouns (see (26)), all prefixes (Figure 22.1).

5.1 Person-marking Taushiro verbs mark the person of their subjects and objects via a series of prefixes that resemble the nominal possessive prefixes above. A, S, and O are marked on the verb with – for first and second person – one and the same set of prefixes. Thirdperson S, however, shows additional markers (namely h- and i-) that do not seem to mark A or O. Transitive verbs agree with the subject and object. I first introduce person-marking for intransitive verbs, then transitive verbs.

5.1.1 Intransitive verbs Intransitive subject agreement is summarized in Table 22.8. An example of first-person agreement is shown in (11), and third person-agreement, in (12). (11) àwàkìɕìó ùjùnté íwì̵ úì. úì àwàkìɕìó u-junte í-wɨ alone 1-come 2.pro-all 1.pro ‘I’ve come alone to visit you.’ (Lit. ‘I’ve come to you alone.’) (12) àtìtí nácɕ͡ ò. a-titi nácɕ͡ ò 3-run 3.pro ‘He’s running.’ First- and second-person u- and i- glide to [w] and [j], respectively, before vowelinitial stems. First-person h- occurs specially before stems beginning in a round

22 Taushiro

1011

Tab. 22.8: Intransitive subject agreement. person

prefix

1 1incl 2 3

u-, h(a)i(a)-, h-, i-

vowel (i.e., /u/ or /o/). In contrast, inclusive and third-person a- deletes before vowel-initial stems.7 There is syncretism in the paradigm with third-person h- (cf. footnote 7), which, like its first-person counterpart, occurs before stems beginning in /u/ or /o/. In addition, third-person marking seems to reveal what might be characterized as lexically distinguished verb classes. While some verbs mark third-person subjects with a-, as in (12), others mark it with i-. The latter is the case for junte ‘come’ shown in (11). The third-person form is not *ajunte but rather ìjùnté, making it homophonous with the second-person form of the verb. The distinction does not appear to be one of unergative versus unaccusative intransitive verbs. The personmarking system of Taushiro especially warrants further research.

5.1.2 Transitive verbs A set of prefixes (namely u-, i-, and a- for first, second, and third persons, respectively) similar to those that inflect intransitive verbs for their subjects is used to inflect transitive verbs for both their subjects and objects, in two adjacent morphological positions, as in (13).8 The result is a number of morphophonological alternations based on the glideability of the vowel of the prefixes in question, which are summarized in Table 22.9. The polysemy of a- in denoting first inclusive and thirdperson referents, in combination with its non-glideability (and subsequent deletion), yields several points of syncretism in this paradigm. Table 22.9 also includes the reflexive prefix ini- in those cells in which it is semantically appropriate. An example of a third-person subject and third-person object is given in (13) – compare the form wàxénà, with a first-person subject. Note that Alicea Ortiz (1975b, c) misanalyzes many transitive verb roots as consisting of an initial /a/. This is

7 It is possible that the inclusive differs from third person in exhibiting an initial glottal stop. Evidence for this comes from the realization of “declarative” t(a)- (Section 6.1) when it precedes the inclusive. Whereas with the third person the vowel of t(a)- deletes, as it does with first and second person, with the inclusive the initial sequence is [taʔa], followed by the remainder of the stem. 8 The degree to which h- figures into the transitive person-marking paradigm for first and third persons requires further research, though it can be established with confidence that i- does not enter into the transitive person-marking paradigm for third persons.

1012

Zachary O’Hagan

Tab. 22.9: Transitive verb agreement. subject

object 1

1

1incl

2

3

w-ini-C w-in-V

w-i-C u-j-V

j-u-C i-w-V Ø-u-C a-w-V

j-ini-C j-in-V Ø-i-C a-j-V

w-a-C u-Ø-C Ø-a-C Ø-Ø-V j-a-C j-Ø-V Ø-a-C Ø-Ø-V

1incl 2 3

Ø-a-C Ø-Ø-V

because most of her examples do not include first- or second-person objects that would demonstrate that a different prefix can occur in place of a- (though see the examples in Alicea Ortiz 1975b: 96–97). (13) àɕíntù àxénà ìjṍ wàjùntùnàjá. àɕíntù Ø-a-xena i-ijṍ wàjùntùnàjá many 3–3-make 2-mother clay.pot ‘Your mother makes many clay pots.’ Events notionally involving three participants are syntactically transitive; that is, the language lacks ditransitive verbs. In the case of ‘give,’ for example, the indirect object is marked by the recipient postpositional suffix -ŋì̵, shown in (14); object agreement is with the theme. (14) jàná úŋì̵ kànã́nì. i-a-na ú-ŋɨ kànã́nì. 2-3-give 1.pro-rcp caimito ‘You’re giving me this caimito.’

5.2 Tense, aspect, and modality The tense-aspect-modality system of Taushiro is not well understood. There are no obligatory verbal categories in this domain, and the only marker with a meaning that can currently be identified with confidence is the future marker.9

9 In particular the temporal-aspectual properties of a suffix -kɨ warrant further investigation.

22 Taushiro

1013

5.2.1 Future -ha Taushiro exhibits a two-way tense system, with a distinction between non-future and future tenses, the former being morphologically unmarked. Future tense is expressed by the verbal suffix -ha, as in (15).10 (15) ùʔwàhá úì ecɕ͡ iwake. uʔwa-ha úì écɕ͡ ì-wake go-fut 1.pro garden-all ‘I’m going to go to my garden.’ Alicea Ortiz (1975b: 40) additionally describes uses of the future that she characterizes as expressing probability.

5.2.2 Regressive -ɾo The suffix -ɾo is frequent in my data. Alicea Ortiz describes it as a marker of repetition, given translations as in (16) with ‘again.’ I tentatively analyze it as a regressive aspect. (16) Yova’ro. i-Ø-owa-ɾo 2-3-wash-reg ‘Wash it again.’ (Alicea Ortiz 1975b: 42) However, many translations of verbs with -ɾo lack ‘again’ altogether, as in (17). (17) hòwàɾò úì ùʔkìkùná. h-Ø-owa-ɾo úì u-ʔkikuna 1-3-bathe-reg 1.pro 1-daughter ‘I already bathed my daughter.’ The future and regressive can combine – see (21) – in which case the future precedes the regressive.

10 In addition, there appears to be a periphrastic future tense construction in which uʔwa ‘go’ precedes a lexical verb that does not agree for person. This construction is not well understood.

1014

Zachary O’Hagan

5.3 Modality 5.3.1 Desiderative -tã Desiderative modality is expressed with the verbal suffix -tã, as in (18). (18) hùʔwàtã́ úì ìntùtòwàké. h-uʔwa-tã úì intuto-wake 1-go-desid 1.pro Intuto-all ‘I want to go to Intuto.’ Alicea Ortiz (1975b: 36–37) describes -tã as a separate word, but there is no evidence for this in my data. In addition, she describes a construction in which the desiderative complement is clausal, consisting of what appears to be an initial “dummy” element ni, to which the desiderative attaches, followed by a subject pronoun and then the clause. In three of the five examples, the desiderative is followed by what she glosses as future -ha, whereas other verbal morphology such as negative -hiɾo (Section 5.4) remains on the verb, suggesting that some verbal morphology may be clitic-like in nature.

5.3.2 Positive and negative abilitative Alicea Ortiz (1975b: 39–41) describes a series of constructions expressing the ability to do something well and the inability to do something. The first consists of ‘well,’ optionally in combination with the verb , which the author glosses as ‘know.’ When ‘well’ co-occurs with ‘know’ in her examples, it always takes the form , shown in (19), whereas when it precedes the verb it is , a distinction that is not understood. If a lexical verb occurs (i.e., know how to do X), it follows ‘know’ and precedes ‘well.’ (19) U nu’chi ineha ui. ‘I know well.’ (Alicea Ortiz 1975b: 39) The negative abilitative is described as a marker of futility, realized by the particle . It precedes the clause that cannot be realized, which in all of the author’s examples is headed by a verb that is also marked for negation by -hiɾo (Section 5.4). The negative clause may occur in a biclausal construction.

22 Taushiro

1015

5.4 Negation -hiɾo 5.4.1 Standard negation Main clause negation is expressed with the verbal suffix -hiɾo. Negation does not result in the suppression of any verbal categories, and is thus symmetric (Miestamo 2007). (20) uʔkahiɾo íwì̵ õː. u-ʔka-hiɾo í-wɨ õː 1-come-neg 2.pro-all yesterday ‘I didn’t come to visit you yesterday.’ (Lit. ‘I didn’t come to you yesterday.’) (21) ujuntehiɾohaɾo íwì̵ ɕí. u-junte-hiɾo-ha-ɾo í-wɨ ɕí 1-come-neg-fut-reg 2.pro-all tomorrow ‘I won’t come visit you tomorrow.’

5.4.2 Negation with ‘no longer, not anymore’ Negation of a clause denoting a proposition that was formerly true (cf. English not X anymore) is expressed with one of three morphosyntactic strategies. Two involve the clause-initial elements ɕáʔ ntà or kojṍ, see in (22) and (23). (22) ɕáʔ ntà àʔtúnè ìtéʔcɕ͡ ì jèjú. ɕáʔ ntà Ø-aʔtune itéʔcɕ͡ ì jèjú no.longer 3-walk person old ‘The old man doesn’t walk anymore.’ (23) kòjṍ àtìtí áì. kòjṍ a-titi áì no.longer 1incl-run 1incl.pro ‘We aren’t running anymore.’ The third involves a sparsely attested verbal suffix -ʔi, illustrated in (24). (24) ùtìtíʔì úì. u-titi-ʔi úì 1-run-no.longer 1.pro ‘I’m not running anymore.’

1016

Zachary O’Hagan

The differences among these constructions are not currently understood. However, the positive polarity of the verbs in all three constructions is striking. It may suggest that these constructions do not express negation per se, but perhaps rather an aspectual category in which the culmination of the event proximally precedes topic time.

5.5 Valence-changing operations Taushiro exhibits two main valence-changing operations: causativization and reflexivization. A third, passive, operation may be present. Antipassives and applicatives are not attested.

5.5.1 Reflexive iniReflexives are derived via the prefix ini-, as in (25). This prefix seems to occur immediately to the left of the verb root, except when an incorporated noun intervenes; see (26). (25) wìnìkwí̵ úì hòhwáhà. u-ini-kwɨ úì hòhwá-ha 1-refl-cut 1.pro machete-ins.com ‘I cut myself with a machete.’ (26) wìnìtòk wí̵ úì àxí̵hà. u-ini-tò-kwɨ úì àxí̵-ha 1-refl-foot-cut 1.pro axe-ins.com ‘I cut myself on the foot with an axe.’

5.5.2 Causatives n- and -xɨ Alicea Ortiz (1975b: 23–24, 28–29) describes two causative affixes. One, a prefix n-, seems to be relatively unproductive, occurring with only five verbs (all intransitive), as seen by comparing the simplex verb uʔunti ‘lie’ in (27) with its causativized stem in (28). (27) Tu unti ui u’neva. ta-uʔunti úì ùné-wa decl-lie 1.pro hammock-loc ‘I’m lying in the hammock.’ (Alicea Ortiz 1975b: 24)

22 Taushiro

1017

(28) Tu vanu’unti antu’ u’ntu’ca. ta-u-a-n-uʔunti àntúʔ úʔ ntù-ká decl-1-3-caus-lie thing land-loc ‘I’m laying the thing on the ground.’ (Alicea Ortiz 1975b: 24) The other, a suffix (probably -xɨ), “always seems to go after the verbal nucleus” (Alicea Ortiz 1975b: 29, translation mine). This suffix occurs with what Alicea Ortiz describes as adjectives (ibid.: 23), as well as with what she glosses as intransitive verbs (n.b., examples with transitive verb roots are given). The two causatives also seem to be able to co-occur, although the only example of this involves what appears to be an infixal use of -xɨ (see the verb root aʔtune ‘walk’ in (8) on her page 29).

5.5.3 Apparent passive CThere is tentative evidence for a passive construction, consisting of a single consonant (C-), the realization of which is conditioned by the phonological shape of the following verb root. If the root begins with a vowel, this inner prefix is cɕ͡ -; if it begins with a consonant, the prefix is ʔ-; if it begins with a vowel but the following consonant is a prenasalized stop, then the prefix is ɲ- (see summary chart following Alicea Ortiz 1975b: 22). Verbs marked with this prefix seem to obligatorily take i- as the marker for a third-person intransitive subject, as described in Section 5.1.2. This construction is not well attested in my data, but it is present (29). Alicea Ortiz claims that it “can be analyzed as a passive mood or as participles; but in translating them to Spanish or from Spanish they seem to be more of a completed past tense with a benefactive” (1975b: 20, translation mine). (29) iɕoʔ ntohã núʔwì̵? núʔwì̵ i-ɕ-oʔ nto=hã 3-pass-cook=int meat ‘Has the meat cooked?’ The claim that there are benefactive semantics comes from the fact that an apparent demoted subject is introduced by the benefactive -ŋɨ, as in (30). (30) Ichonto iya’an nu’vu. ìjáʔ-ŋɨ núʔwì̵ i-ɕ-oʔ nto 3-pass-cook father-ben meat ‘The meat has been cooked by my father.’ (Alicea Ortiz 1975b: 20)

1018

Zachary O’Hagan

In this vein, it is noteworthy that in Spanish benefactive complements and the demoted subjects of passives are both introduced by the preposition por. My sense is that the apparent Taushiro passive is not a passive at all but rather an artifact of translation in elicitation, where a request for a translation of a demoted subject with por was interpreted as a benefactive, and so is marked expectedly with the benefactive -ŋɨ. The construction described here, then, may be more like middle voice, one in which the verb is indeed detransitivized (as with a passive) but where the erstwhile subject cannot be reintroduced. This may be more in line with Alicea Ortiz’s participle hypothesis.

6 Other verbal clause types Taushiro shows special marking of (polar and content) interrogatives, prohibitives, and hortatives. Declaratives and positive directives (i.e., imperatives) bear no special marking.

6.1 “Declarative” t(a)A ubiquitous morpheme in my data is a verbal prefix t(a)-, the vowel of which deletes preceding vowel-initial stems. Alicea Ortiz (1975b: 72) analyzes it tentatively as a marker of declarative sentential mood. While it does not seem to occur in interrogative clauses, there are many declarative clauses in both my and Alicea Ortiz’s data that lack t(a)-; thus, it cannot express declarative mood per se. Indeed Alicea Ortiz herself states, “I am not completely sure of the meaning of the morpheme” (ibid., translation mine). This morpheme is especially frequent in the elicitation of verbal meanings based on Spanish infinitives. That is, the translations in Taushiro are fully inflected verbs with an initial t(a)-, as in (31). Furthermore, while the resulting back translations are often with Spanish progressives, they may be periphrastic futures and other types, suggesting that t(a)- does not express a temporal or aspectual category. (31) tùhóʔnẽ ̀ úì. ta-u-hoʔnẽ úì decl-1-blow 1.pro ‘I’m blowing.’ I suspect that this prefix expresses an epistemic-like category having to do with whether the speaker assumes the addressee is familiar with the content of the proposition, hence its ubiquity in elicitation, since the linguist has just asked for a translation of a sentence with content with which they are clearly familiar.

22 Taushiro

1019

6.2 Interrogatives 6.2.1 Polar questions Taushiro polar questions are formed via the interrogative second-position clitic =hã, demonstrated in (32). (32) ìjṍ, ɲòʔ ntóhā̀ núʔwì̵? ìjṍ i-Ø-oʔ nto=hã núʔwì̵ mother.voc 2-3-cook=int food ‘Mother, have you cooked any food?’ Evidence that =hã is a second-position clitic comes from the fact that it attaches to whatever constituent is in first position, as shown in (33), where it attaches to a preverbal noun. (33) Inihiahan ñuntu’? i-nihiã=hã i-untu 2-food=int 2-eat ‘Did you eat food?’ (Lit. ‘Was it your food you ate?’) (Alicea Ortiz 1975b: 74)

6.2.2 Content questions Content questions are formed via a sentence-initial interrogative pronoun that takes =hã. (34) àntúhā̀ júntù íì? àntú=hã i-Ø-untu íì what=int 2-3-eat 2.pro ‘What are you eating?’ SVO and OVS orders are prevalent in responses to content questions that question the subject and object of a clause, respectively, suggesting a preverbal information focus position. Interrogative pronouns are summarized in Table 22.10.

Tab. 22.10: Interrogative pronouns. pronoun n

à tú kú ìná hí ìnté

gloss ‘what’ ‘who’ ‘when’ ‘where, which’ ‘why’

1020

Zachary O’Hagan

6.3 Directives 6.3.1 Imperatives Positive second-person directives are formally indistinct from their declarative counterparts; that is, every directive sentence has a non-directive counterpart. (35) jèʔtú íì àtúà! i-Ø-eʔtu íì àtúà 2–3-shoo 2.pro chicken ‘Shoo the chicken [away]!’ ~ ‘You shooed the chicken [away].’

6.3.2 Prohibitives Negative second-person directives (prohibitives) differ from their positive counterparts in the prefixation of h(a)-. Note that the /a/ of this prefix deletes before a vowel-initial stem. This negation strategy differs from the negation of declarative main clauses with -hiɾo (Section 5.4). (36) hìtìtí íì! ha-i-titi íì proh-2-run 2.pro ‘Don’t run!’

6.3.3 Hortatives First-person directives (hortatives) are expressed with the sentence-initial particle hàʔwí̵ and first-person inclusive subject agreement on the verb. (37) hàʔwí̵ àʔké áì hwáʔ ntù ècɕ͡ ì-wá. hàʔwí̵ Ø-a-ʔke áì hwáʔ ntù écɕ͡ ì-wa hort 1incl-3-plant 1incl.pro maize garden-loc ‘Let’s plant maize in the garden.’

7 Combining clauses Clause combining is striking in Taushiro in that relative, reason,11 and temporal clauses all exhibit an initial particle aŋka, with no other differentiating marking. 11 For an example of a reason clause marked with áŋkà, see (49).

22 Taushiro

1021

This form is often reduced to the prosodically independent [ŋga] or [ga], but it may also cliticize as [ŋga] to the word to its left, which is not part of the same clause. Purpose clauses have distinct marking, as detailed in Section 7.3; conditional clauses are not described.12 I note that, in elicitation, back translations of biclausal sentences with Spanish conjunctions were reworked to be unmarked juxtaposed clauses in Taushiro.

7.1 Relativization Relative clauses are formed via aŋka, which follows the head noun. The relativized noun is gapped, and agreement is preserved. At least subject and and direct object relativizations are formed in this way; information on other types (e.g., of indirect objects) is lacking.13 Example (38) demonstrates subject relativization and (39) shows object relativization. (38) unotakɨhiɾo àwéʔwí áŋkà ìjùnté ùhāk̀ úʔkì̵ . u-notakɨ-hiɾo àwéʔwí áŋkà i-junte u-hā-̀ kuʔkɨ 1-know-neg man rel 3-arrive 1-house-ill ‘I don’t know the man who has arrived at my house.’ (39) ìcɕ͡ úkwàntàké ànòwá áŋkà wàɕó héʔĩ̵.̀ i-cɕ͡ -ukwanta-ke ànòwá áŋkà u-a-ɕo héʔĩ̵ ̀ 3-pass-tear-? clothes rel 1-3-buy a.while.ago ‘The clothes that I bought a while ago have already torn.’

7.2 Temporal The example in (40) illustrates áŋkà in the expression of a temporal clause-linking construction translatable as ‘when,’ appearing in the in the temporally anterior clause. As with relative clauses, there seems to be a preference for the clause marked with áŋkà to be non-initial. (40) uʔwahaɾo úì Ìntùtòwàké áŋkà juʔwaharo ìnì̵wí̵. uʔwa-ha-ɾo úì Intuto-wake áŋkà i-uʔwa-ha-ɾo i-nɨwɨ go-fut-reg 1.pro Intuto-all when 2-go-fut-reg 2-homeland ‘When you go back to where you’re from, I’m going to go back to Intuto.’ 12 It seems that there is considerable formal overlap in the expression of temporal and conditional clause-linking constructions (see Section 7.2; cf. Alicea Ortiz 1975b: 59, 61–62). 13 Alicea Ortiz (1975b: 11) mentions another of what she refers to as relative pronouns, namely . The distinction is not explained, but it is possible that occurs when the relative clause is headless.

1022

Zachary O’Hagan

Alicea Ortiz’s (1975b: 50–51) examples of this kind of clause-linking attest related and different constructions that are difficult to reconcile with each other. In addition to examples parallel to (40), it is common for ina to occur at the beginning of the temporally posterior clause, which she glosses as Spanish entonces ‘then.’ This is not attested in my data, but the form seems to be the same as the interrogative pronoun ìná ‘when’ (Section 6.2.2). In another construction, one of two verbs is marked with a prefix n-. In most cases it is the verb of the temporally anterior clause that is marked in this way, but in at least one example it is the temporally posterior clause. Finally, it is possible that n- is a reduction of a free form na (see Alicea Ortiz 1975b: 59, 84). Further research is needed to determine the differences among these Taushiro constructions and among other kinds of temporal clause-linking (cf. English before, after, while). What is clear at this juncture is that na (and n-) is restricted to the temporal-conditional domain (see footnote 12), whereas áŋkà spans other kinds of clause-linking.

7.3 Purpose -wɨ The verb of a purpose clause is marked with the allative postposition -wɨ. (41) tàʔákèhí àhjã́ jànáʔkì̵wí̵. i-a-na-kɨ-wɨ ta-Ø-a-kèhí àhjã́ decl-1incl-3-plant manioc 2-3-give-?-all ‘We’re going to plant manioc for you to sell (it).’

8 Nonverbal Clauses 8.1 Nominal predication Three nominal predication constructions are currently attested. In the first, only the nominal predicate is overt, with the subject understood to be third person, as in (42).14 (42) a. ìʔkàhã́? i-ʔka=hã 2-son=int ‘Is he your son?’

b. úʔkà. u-ʔka 1-son ‘He’s my son.’

14 This example shows evidence for the clitic status of interrogative =hã, as it attaches to a noun – see (32).

22 Taushiro

1023

In the second construction, the nominal subject is be overt, as in (43), with a second-person subject. (Third-person subjects can also be overt, with the pronoun nácɕ͡ ò.) (43) ùwẽ́ íì. u-wẽ íì 1-same.sex.sibling 2.pro ‘You’re my brother.’ In the third construction, the nominal predicate occurs clause-initially and is marked by the suffix -ne, which receives a high tone. This predicate is followed by a fixed element jànùhúà that I tentatively analyze as a copula, followed by the subject, as in (44). The function of -ne is not well understood, and it is not currently attested elsewhere. (44) ìʔtàwàné jànùhúà àwá. ìʔtàwá-ne jànùhúà àwá kapok.tree-? cop tree ‘The tree is a kapok tree.’ The copula jànùhúà is often translated with Spanish llamarse ‘be named’ in back translations, but it is not restricted to contexts of naming; see (45). (45) prèsìdèntèné jànùhúà úì. presidente-ne jànùhúà úì president-? cop 1.pro ‘I’m [community] president.’

8.2 Locative predication Predication of the location of an object does not require any dedicated marker: a subject noun can be juxtaposed (following) a noun referring to a location, the latter of which is marked with a postposition, as in (46). (46) ìhã́kù íì. i-hã-ku íì 2-house-loc 2.pro ‘You’re in your house.’ This pattern holds of both interrogative and declarative clauses, as can be appreciated in the following question-answer pair. Two forms of the question were volun-

1024

Zachary O’Hagan

teered – with or without the non-inflecting element í̵nì̵ – which happens to appear in the volunteered response. It is not clear what other function(s) this element may have. (47) a. híhā̀ (í̵nì̵) Pepe? hi=hã Pepe where=int Pepe ‘Where is Pepe?’

b. Lìmàʔ ntú í̵nì̵ Pepe. Lìmà-ʔ ntu í̵nì̵ Pepe Lima-loc ? Pepe ‘Pepe is in Lima.’

Responses to the Topological Relations Picture Series (Bowerman & Pederson 1992) typically elicit responses with verbs like ‘sit,’ or other kinds of verbal descriptions of scenes. See Alicea Ortiz (1975b: 55–56) for other examples.

8.3 Existential -tua Existential constructions are formed by suffixing -tua to the noun referring to the entity whose existence is predicated, as in (48). (48) héʔìtùà ùnùʔ ntùká. hèʔí-tua un-úʔ ntù-ká jaguar-exist 1-land-loc ‘There are jaguars where I’m from.’ This construction is also used in the expression of external possession, in combination with the recipient suffix -ŋɨ (i.e., there is X to Y). Stems derived with -tua may be negated with -hiɾo (Section 5.4), although it does not appear to derive a stem with any other verbal properties. (49) ahjãtuahiɾo áŋkà ecɕ͡ ituahiɾo áŋì̵. áŋkà écɕ͡ ì-tua-hiɾo a-ŋɨ àhjã́-tua-hiɾo manioc-exist-neg because garden-exist-neg 1incl-rcp ‘There is no manioc because we don’t have a garden.’ Finally, it is worth noting that Alicea Ortiz (1975b: 90) cites a form of the existential with a final glottal (i.e., -tuaʔ), but there is no evidence for this in my data apart from the clan names mentioned in the introduction. These names may be formed on a different affix.

8.4 Presentationals Presentational constructions are formed with one of two clause-initial particles, with one of two markers following the noun. For a proximal object, the particle is

22 Taushiro

1025

cɕ͡ ṍ, and the noun is marked with the suffix -na, as shown in (50). Note that cɕ͡ ṍ is not an adverb that elsewhere in the language means ‘here.’ (50) cɕ͡ ṍ ùcɕ͡ ã́ ùnénà. u-cɕ͡ ã́ ùné-na cɕ͡ ṍ pres.prox 1.pro-poss hammock-? ‘Here is my hammock.’ For a distal object, the particle is kanṍ, and the noun is followed by the adverb háʔ ntà ‘there,’ as shown in (51). (51) kanṍ ùcɕ͡ ã́ ùné háʔ ntà. ùné háʔ ntà kanṍ u-cɕ͡ ã́ pres.dist 1.pro-poss hammock there ‘There is my hammock.’

8.5 Quotative Alicea Ortiz illustrates in a discussion of “pronominalization” a quotative construction. This construction is not attested in my data, and so I represent the relevant form orthographically the relevant as (i.e., either ne or nɨ). The particle follows the quote and is itself followed by nominal expressions corresponding to the addressee and the speaker of the quote, respectively. The first is marked with -ŋì̵, as in (52). (52) Va’jenaharo in u’ne − ne un iño. u-a-xena-ha-ɾo i-ŋì̵ ùné ne u-ŋì̵ ìjṍ 2-3-make-fut-reg 2.pro-rcp hammock quot 1.pro-rcp mother ‘“I’m going to make a hammock for you,” my mother said to me.’ (Alicea Ortiz 1975b: 81)

9 Conclusion This chapter has described what can currently be known with confidence about the phonology and grammar of Taushiro, based on fieldwork with Amadeo García and the published work of Nectalí Alicea Ortiz. The language still has at least one native speaker, and the potential for learning more about the language history of the Taushiro people is great. In terms of phonology and grammar, I have shown 17 phonemic consonants and 12 phonemic vowels (6 qualities). Attested syllables shapes are V(ː), CV(ː), and

1026

Zachary O’Hagan

CV(ː)C, that is, with or without length; the only attested coda is /ʔ/. Basic word order is VSO, and the language is headmarking, with prefixal slots for both subject and object agreement. The nominal domain exhibits an alienability distinction (with prefixal marking of possessors) and discontinuous modifiers. The verbal domain shows prefixal derivational morphology (including a causative, reflexive, and possibly a passive), suffixal TAM and negation marking, and a relatively underspecified set of clause-linking markers. Nonverbal clauses are relatively simple, with no special marking of locative predicates. I also illustrated nominal and existential predicates and presentational constructions. There is still much to learn about Taushiro grammar, even in the domain of simple clauses, especially person and TAM marking. Modal categories in particular go nearly wholly undescribed in Alicea Ortiz’s work, as in this one, and it is possible that the person-marking system is considerably more complex than described here. Multi-clause constructions are also in need of further differentiation. First and foremost, however, a finer-grained understanding of the phonological system is in order – especially as concerns glottalization and tone in morphologically complex stems – in order to develop more reliable representations that can be used in future work. These representations will also be of immense benefit to the Taushiro people, for whom they can immediately be utilized in the production of lexical materials, for example, for pedagogical purposes. Lexical documentation of Taushiro is of particular importance, given its status as an isolate in a region with many isolates. Patterns of borrowing (and thus contact between ethnolinguistic groups) will only be detectable with a relatively rich lexical dataset. This in turn has implications for the understanding of the precolonial history of the region.

10 Acknowledgments I am thankful to Amadeo García García for teaching me about his language. I am grateful to José Álvarez Alonso for the use of his home in Iquitos and for his dedication to seeing my initial fieldwork through. Wagner Tello Macedo facilitated travel, and Herlinda Nolorbe made sure everyone remained comfortable in Iquitos. I am indebted to Msgr. Miguel Ángel Cadenas Cardo and Fr. Manuel Berjón Martínez for putting us all in contact. I have benefited from discussions with Amalia Skilton about Taushiro phonology.

11 References Alicea Ortiz, Neftalí. 1975a. Análisis fonémico preliminar del idioma taushiro. Lima: SIL. Alicea Ortiz, Neftalí. 1975b. Análisis preliminar de la gramática del idioma taushiro. Lima: SIL. Alicea Ortiz, Neftalí. 1975c. Vocabulario taushiro. Lima: SIL.

22 Taushiro

1027

Alicea Ortiz, Neftalí. 1976. Apuntes sobre la cultura taushiro. Lima: SIL. Ayala, Jaime, Marlene Ballena Dávila, Jorge Chávez, David Coombs, Abe Koop, Debbie Koop, Giuliana López de Hoyos & Carolynn Parker. 2006. Pueblos del Perú. Lima: SIL. Bowerman, Melissa & Eric Pederson. 1992. Topological relations picture series. In Stephen C. Levinson (ed.), Space Stimuli Kit 1.2: November 1992, 51. Nijmegen: Max Planck Institute for Psycholinguistics. Casey, Nicholas. 2017. “Thousands once spoke his language. Now he bears the burden of its survival.” The New York Times, 27 Dec. p. A1. Figueroa, Francisco de. 1986[1661]. Informe de las misiones en el Marañón, Gran Pará o Río de las Amazonas. In Regan, Jaime (ed.), Informes de jesuitas en el Amazonas, 143–309. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía. Grohs, Waltraud. 1974. Los indios del alto Amazonas del siglo XVI al XVIII: Poblaciones y migraciones en la antigua provincia de Maynas. Bonn: U. Oberem. Maroni, Pablo. 1988[1738]. Noticias auténticas del famoso río Marañón. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía. Miestamo, Matti. 2007. Negation – An overview of typological research. Language and Linguistics Compass 1(5). 552–570. Riva Herrera, Martín de la. 2003[1656]. Traslado de los auto fechos por el General Don Martín de la Riva Herrera en razón de la fundación y población de la ciudad de Santander de la Nueva Montaña. In Fernando Santos Granero (ed.), La conquista de los motilones, tabalosos, maynas y jíbaros, 238–251. Iquitos: Centro de Estudios Teológicos de la Amazonía. Santos, Thomas. 1986[1684]. Relación que da el padre Thomas Santos de la conquista y entrada que por el rio del Tigre hizo á cuatro naciones que son los asouinatoas, los pinches, los cenicientos y habitoas. In Regan, Jaime (ed.), Informes de jesuitas en el Amazonas, 341–355. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía. Tessmann, Günter. 1930. Die Indianer Nordost-Perus: Grundlegende Forschungen für eine systematische Kulturkunde. Hamburg: Friederischen, de Gruyter & Co. Veigl, Francisco Xavier. 2006[1798]. Noticias detalladas sobre el estado de la Provincia de Maynas en América meridional hasta el año de 1768. Iquitos: Centro de Estudios Teológicos de la Amazonía. Velasques de Medrano, Salvador. 2003[1657]. Certificación de Don Salvador Velasques de Medrano, Capellán Mayor, Cura y Vicario general de la conquista que a su cargo tiene el General Don Martín de la Riva Herrera, que refiere todo lo hecho en ella. In Fernando Santos Granero (ed.), La conquista de los motilones, tabalosos, maynas y jíbaros, 238–251. Iquitos: Centro de Estudios Teológicos de la Amazonía. Villarejo, Avencio. 1943. Así es la selva: Estudio geográfico y etnográfico de la Provincia de Bajo Amazonas. Lima: Compañía de Impresiones y Publicidad. Zárate, Andrés de. 1988[1735]. Relación de la mision apostólica que tiene á su cargo la Provincia de Quito, de la Compañía de Jesús, en el gran Rio Marañón, en que se refiere lo sucedido desde el año de 1725 hasta el año de 1735. In Pablo Maroni, Noticias auténticas del famoso río Marañón, 293–339. Iquitos: Instituto de Investigaciones de la Amazonía Peruana; Centro de Estudios Teológicos de la Amazonía.

Katherine Bolaños

23 Tinigua 1 2 3 4 5 6 7 8 9 10 11

Introduction Phonology Word classes and morphological structure The noun phrase Verbs and verbal morphology Simple clauses Clause combining Conclusion Ackowledgments References Appendix A. 100-word Swadesh list

1 Introduction Tinigua (tini1245) is a nearly extinct language whose speakers once inhabited the lowlands of the Caquetá Department in the Andean-Amazonian fringe, in the eastern part of Colombia. Today, the only known speaker of the language is Sixto Muñoz (or, in Tinigua, Sɨsɨthio, meaning ‘knife’), who is about 90 years old. Sixto currently lives in the Jiw1 village of Barrancón near the principal mestizo town of the Guaviare Department, near his last remaining child, eight grandchildren, and some 21 greatgrandchildren. All these family members grew up speaking Jiw, the language of their Jiw mother and relatives, in villages of the Guaviare River basin. Before moving to Barrancón in 2019, Sixto lived for many years in the Serranía de la Macarena, in the Meta Department. There are other individuals who identify as ethnically Tinigua in the town of La Macarena, although apparently none speak the language, as they were not sufficiently exposed to it while growing up. Today there are no known Tinigua communities in which the language is spoken or Tinigua cultural practices are maintained. There are also no known semi-speakers, consistent with rapid language shift. At the time of writing, there is no apparent interest in revitalizing Tinigua among Sixto’s relatives or other ethnic Tiniguas. In 2018, the author of this chapter attempted a trip to the Caguán River region in the Caquetá Department to search for possible speakers of Tinigua in the area of Cartagena del Chairá, a small settlement of colonos on the Caguán River. Nineteenth century missionary documents indicate that speakers of Tinigua once lived in this

1 Also known as the Guayabero, speakers of a Guahiban language. Katherine Bolaños, Universidad de los Andes https://doi.org/10.1515/9783110432732-010

1030

Katherine Bolaños

23 Tinigua

1031

region; moreover, a farmer from the area of La Macarena had recently informed Sixto that he had, some seven years previously, spoken with people living around Cartagena del Chairá who spoke “Sixto’s language”, even providing a few Tinigua words that he claimed to have learned in that region. Unfortunately, the author’s trip had to be ended prematurely due to security concerns, and no speakers were located. In 2018 the Colombian National Statistics Bureau (Departamento Administrativo Nacional de Estadística, or DANE), published a census of ethnic groups in which a Tanigua group with 145 members and 28 speakers of the language is included (Tinigua is not mentioned). Whether this represents a misspelling of Tinigua or the name of a wholly different group is not clear; furthermore, the census does not specify the location of this group or how the data were collected. There are no evident references to the Tinigua people or their language from the colonial era. Possibly the earliest mention, while indeterminate, is found in a book of missionary reports by the Jesuit priest Juan de Rivero (1883 [1736]: 35), in which he mentions a “Bamigua nation” who were very numerous and settled on the banks of the Guaviare River; no further information is given. Bamigua bears an obvious similarity to Pamigua, the name associated with a brief wordlist published by Ernst (1891) and evidently a variety of Tinigua (see Section 1.1). The etymology of the name Tinigua itself is unclear, though it could conceivably be based on the root tiní ‘old’ plus an unidentified suffix -gua (cf. Tobar Ortiz 2000: 669; Castellví 1940: 94). The first explicit mention of the Tinigua people and their language was not made until the 20th century, by the Capuchin monk Estanislao Les Corts. Les Corts (1931) authored a four-page manuscript of Tinigua lexical and phrasal material aimed at carrying out confession. Some mention of the Tinigua is also given in his earlier report of a two-month missionary expedition through the Caquetá and Caguán Rivers, during which he found only three Witoto families to evangelize (Les Corts 1926). In describing one of these families, he notes that “the men wear only a piece of tree bark, made flexible by pounding, and tied around the waist with a thread from a palm. In Witoto it is called mesití; in Muinane, moye; and in Tinigua, yafóyasa” (Les Corts 1926: 60). While Les Corts provides no further mention to the Tiniguas in this text, his mention gives the impression that the presence of Tinigua people in the area was common knowledge at the time. (The term yafóyasa also corresponds quite closely to the Tinigua word provided by Sixto for loincloth, or clothing in general: jaɸʷajhatsá.) Les Corts and fellow Capuchin Fructuoso de Manresa (1935–1936) also described a group of Tinigua people who were living at that time in the Yarí plains in the Caquetá Department, one of whom was Sixto’s father, Agapito Muñoz.

1.1 Classification Our current understanding of Tinigua supports its characterization as a language isolate (see Fabre 1998: 104, Adelaar with Muysken 2004: 162). However, as this

1032

Katherine Bolaños

Tab. 23.1: Castellví’s (1940: 95) Tinigua and Pamigua comparison alongside contemporary Tinigua data. gloss

Tinigua

Pamigua

Tinigua (provided by Sixto Muñoz 2019)

eye water fire woman dog jaguar corn manioc let’s go chili pepper good plantain spirit man five

zőti, zɘ̀ti ñikwáiš i ičísa ñíza, ñísä xamno, xámiu číña, ǰíña, xiña t’óka, tióka xaačá manaxǎí ţáxa ayuxáǐ madóxa pan-kianóso psäţeyá xopa-kuáxa

sete, xete nikagé ekísa nixtá xannó xiñaga xukxá xoayoa menáxa saxa ayoxagua [‘good morning’] mandotá kinoxá [‘enemy’] piksiga saksu-kuaxa

eleven

čimatóse-kiésä

čipsé ipa-kiaxi

sıt́ ̵ i ɲikʷájtʃi hikʰítsa ɲísa hanó kʰíɲa ~ tʃíɲa jóʔhá komáha minahá tsák ha hajohási mandót ha hamajiéha tsɨtsía tsátok wahá (tsát ho-kwaʔa ‘left.side-hand’) tapásaɲóha

section explores, previous evaluations have alternatively treated it as a member of its own small family or as a distant relative of the Sáliban or Guahiban languages. Proposals of a “Tiniguan” or “Pamiguan” language family (Loukotka 1968: 151, Kaufman 1990: 40, 1994: 56, 2007: 65; Campbell 2012: 106) make reference to at least two languages, Pamigua and Tinigua. Loukotka (1968) also includes a third sister, Majigua, but no record of this language exists. Sixto reports that his fellow Tiniguas used the name Majigua to refer to a group of Guahiban people with whom they were in contact. The Capuchin priest Fructuoso Manresa describes Agapito Muñoz (Sixto’s father) as the son of a Tinigua father and a Majigua mother, but Sixto reports that his father’s mother was Tinigua. Sixto’s own mother was apparently a mestizo woman from the Caguán area in the Caquetá Department. As for Pamigua, the only record of this language is a word list of 34 entries (Ernst 1891). Ernst reports some 200 speakers of the “Pamigua language” but gives little further detail. Capuchin missionary Marcelino de Castellví (1940: 95) provides a comparative assessment of these Pamigua data with their Tinigua counterparts, suggesting an “evident” relationship between the two languages. Castellví’s comparison is reproduced in Table 23.1, together with corresponding contemporary Tinigua forms provided by Sixto. As can be seen here, the Pamigua and Tinigua forms are in fact very similar, and the observable differences may be as much an artifact of orthographic representation as of linguistic distance. It seems probable that Pa-

23 Tinigua

1033

migua and Tinigua are at most merely dialectal variants of what could be considered a single language. Suggested long-range affiliations have not been supported by closer inspection. The proposal for a Tinigua-Sáliban family (Loukotka 1935: 8, 1942: 10) was discarded by Rosés Labrada (2016) on the basis of lexical data available at that time. Rivet (1924) includes Pamigua within the Guahiban family (see also Voegelin & Voegelin’s 1965 “Guahiboan-Pamiguan” family and other proposals that mention “Majigua” as a Guahiban language); however, Rivet’s (1948) later work on Guahiban languages attributes the lexical similarities between these languages and Tinigua/Pamigua to borrowing (Rivet 1948: 196). Rivet (1948: 237–238) provides a 30-item wordlist comparing Pamigua, Tinigua, and Guahibo, noting apparent resemblances; Ortíz and Queixalós (1981) also provide a comparative list of fauna terms for Guahiban languages and Tinigua with indications of possible loanwords (drawing on Tinigua data from Olivares 1962).

1.2 Origin and migration From the reports of Jesuit and Capuchin missionaries in the late 1800s, it can be supposed that the Tiniguas inhabited the lowlands of the Caquetá Department at the time (where it is possible that some speakers still remain; see above). Palacio Hernández and Bolaños (2019), drawing on their work with Sixto Muñoz, suggest a possible origin of the Tinigua group in the Caquetá River region on the border between the Caquetá and Putumayo Departments in southwestern Colombia. A village in this area, inhabited today by mestizo families, is named La Tagua, plausibly from Tinigua tagʷa ‘town/settlement’. Sixto Muñoz’s accounts of the history of the group and of his own life indicate that his group lived in the Yarí Plains region (also in the Caquetá Department) during the first part of the 20th century and later moved north towards the Serranía de La Macarena, then eastwards (along waterways by canoe) as far as the town of San José del Guaviare. Along this trajectory, the Tinigua people established relationships (some of alliance and others of enmity) with other indigenous groups speaking different languages. According to Sixto’s own narratives and to the missionary reports of the 19th century,2 the areas of the Caquetá and northern Putumayo regions that the Tinigua people passed through were inhabited by groups of Witotos (Witotoan language family), Tamas (West Tukanoan), Guahiban peoples, and Carijonas (Cariban). Chibchan languages also reached as far as the colono settlement of San Juan de Arama (formerly Concepción de Arama), where Ernst (1891) encountered the Pa-

2 During our work with Sixto, we read him some of these missionary reports, but they seemed not to interest him, or perhaps he needed to tell his own version of these migratory movements from his and his family’s own experience.

1034

Katherine Bolaños

miguas. In the Meta Department and Guaviare regions, according to Sixto’s reconstruction of the history of the group and relationships with other groups, the Tinigua people were in contact with speakers of Jiw (Guahiban) and, again, of Witoto (Witotoan); this latter group, according to Sixto, had a hostile relationship with the Tinigua, while the Jiw people were martial allies. The Tinigua appear to have been a relatively semi-nomadic people, at least during the historical period. According to our conversations with Sixto and his accounts of his own group’s history, longer periods of movement were interspersed by shorter periods of settlement (lasting approximately 6–8 months) in inland forest locations. The periods of displacement were more frequent in some years, less so in others, and typically involved revisiting familiar locations; these activities included planting small gardens which they harvested upon their return some 9–12 months later (see Palacio Hernández & Bolaños 2019: 46–70 for further discussion). This history of mobility appears to be consistent with their relatively simple approach to house construction and with the evident lack of Tinigua toponyms in territories they inhabited; except for the previously-mentioned name La Tagua, most toponyms in the Caquetá area are of Tama (West Tukanoan) origin, and in the Guaviare are they are of Jiw origin. The limited ethnographic literature addressing the Tinigua suggests that they moved from the Yarí plains to escape pressures relating to the Colombian guerrilla war and rubber exploiters in the area (e.g., see Valdés Arcila 1996 and Tobar Ortiz 1995); however, our work with Sixto and with the missionary expedition reports from the late 1800s suggest that a pre-established tradition of mobility may have also played a role. By the late 1940s, the Tiniguas had dwindled to a single group of fewer than 50 people who were living near the town of La Macarena, in the Meta Department. It was there that a group of dissidents from the liberal guerrilla, led by Hernando Palma, launched a massacre in the area of La Macarena (then named El Refugio) in which many of the Tiniguas were killed. Others fled downriver toward the Guaviare and inland through the forest to the Sierra de la Macarena. Only a handful of Tiniguas survived, scattered in different areas. At the time of the massacre, Sixto and Severo (Sixto’s cousin, baptized with the Spanish last name Polanía, whose name in Tinigua was Mandókʰi ‘plantain’) were already living in the Guaviare among speakers of Jiw. Sixto’s brother Criterio and aunt Esther, escaping Palma, fled into the forest in the Sierra de la Macarena, were they lived in isolation until approximately 1972 (Esther died soon after they emerged, while Criterio lived until 2005). Sixto’s father, Agapito Muñoz, died in the 1960s, and the few remaining family members died soon afterward; according to Sixto, it was as if a curse had hit them. A younger brother of Sixto’s fled to Venezuela with a group of Guahibos, but his destination and fate are unknown. Today Sixto is the last known speaker of the language; in addition to his own descendants, who are Jiw speakers, two nephews and a niece (children of his cousin Rogelio) live today among colonos and claim not to know the language because they grew up with their mother, a mestizo woman.

23 Tinigua

1035

The death of Sixto’s brother Criterio ended any possibility of documenting conversation and other interactive discourse in Tinigua. While most of the non-missionary anthropological and ethnographic information concerning the Tinigua is focused on the group’s near-extermination in the mid-20th century, information offered in Franco (1989) and in Valdés Arcila (1996), alongside the materials produced by Capuchin priests (e.g., San Martivel 1926; Les Corts 1926 and 1931; Pinell 1929; Castellví 1938; Igualada 1938), help to flesh out our understanding of the Tinigua migration and of their history of contact with both indigenous and non-indigenous groups in the past.

1.3 Linguistic research Linguistic documentation of Tinigua, like the ethnographic record, is very limited. As noted previously, Les Corts and Manresa worked with speakers of Tinigua in the Yarí plains during the 1930s; they gathered lexical material, a few phrases, and a short translation of a text for carrying out a confession, all of which amounts to a four-page manuscript that is archived in Barcelona in a Capuchin repository (Les Corts 1931; see Castellví 1940: 94). Manresa worked from 1935–1936 with Agapito Muñoz and four other consultants (see Castellví 1940: 94), but it is unclear what he was able to gather since the only record is Castellví’s reference to Manresa’s work; the work itself is unavailable. Castellví (1940) published a list of words compiling data gathered by Les Corts (1931) and by Manresa (1935–1936). According to the ICANH (Instituto Colombiano de Antropología e Historia), in 1959 Sixto’s father, Agapito Muñoz, was brought to Bogotá, to the headquarters of the ICANH, to work on audio recordings of the language. These recordings, however, have so far not been found. In the 1960s an ornithologist, Antonio Olivares, gathered a list of bird names with the help of Agapito Muñoz, published in Olivares (1962). Other ornithologists from the Universidad Nacional de Colombia – Pedro Ruiz, Alberto Cadena and Juan M. Rengifo – reportedly collected stories and a few songs in the language in 1976 through work with Sixto’s brother Criterio while on an expedition to record birds’ and frogs’ songs in the Serranía de la Macarena. These recordings are also missing. In the 1990s, linguist Nubia Tobar Ortiz worked with both Criterio and Sixto. A four-page phonological description resulting from this work was published in Tobar Ortiz (1995) and republished in Tobar Ortiz (2000), to which notes on morphology are added. In 2017–2020, this author and Ricardo Palacio Hernández worked with Sixto to produce a documentation and description of Tinigua. An audiovisual documentary corpus of the language, along with a reconstruction of the history of the group, stories of their origin, myths, and a few other genres can be freely accessed in Bolaños and Palacio Hernández (2017). Palacio Hernández and Bolaños (2019) provide a biographical account of Sixto’s life and personal history.

1036

Katherine Bolaños

As this overview makes clear, the bulk of the documentation and description of Tinigua was carried out at a time when very few speakers remained (Tobar Ortiz’s work in the 1990s involved Criterio and Sixto, while by 2017 only Sixto was left). These limitations present many challenges for analysis and for making sense of variation across the available sources, as well as for assessing the degree of language attrition that Sixto himself may have experienced. As the description throughout this chapter indicates, there are many questions concerning the structure of Tinigua that we may never be able to answer.

2 Phonology The phonological description presented here has many gaps and probably various inaccuracies. In light of the severe limitations in the source material, many phonological contrasts are not conclusively clarified. The breakdown of contrasts is a well documented outcome of language attrition (see, e.g., Campbell & Muntzel 1989), and providing phonological oppositions can be a difficult task when working with a last speaker. Some of the presumed contrasts presented in this chapter showed much variation in Sixto’s speech, such that it was not possible to establish a clear phonological distinction among some sounds (e.g., [ts] ~ [tʃ]). In light of these challenges, this section presents our best approximation of Tinigua phonology. The linguistic data presented throughout the chapter are presented as a phonemic transcription, according to our current understanding of the phonology. However, the transcription reflects the still tentative state of analysis. A palatal glide, for example, is represented as a segmental j through the text, but it is not fully clear at this stage whether it should be analyzed in particular contexts as a process of palatalization or as a phonological segment.

2.1 Phonological inventory According to our analysis, Tinigua has 20 contrastive consonants and 6 vowels. Consonant and vowel inventories are presented in Tables 23.2 and 23.3, respectively. In lexical roots, the most frequent vowels seem to be /a/, /i/, /ɨ/, /o/, and /u/, in that order. The front mid vowel /e/ occurs mostly in function morphemes and appears to be very rare in roots. There are no phonologically contrastive long vowels (vowel lengthening is understood to be a phonetic process in word-final open syllables), nor are there phonemic nasal counterparts to the oral vowels. As seen in Table 23.2, aspirated voiceless stops are understood to contrast with non-aspirated stops: /tʰ/ versus /t/ and /kʰ/ versus /k/. See (2) for examples that illustrate the contrast with minimal or near-minimal pairs.

1037

23 Tinigua

Tab. 23.2: Tinigua consonants.

voiceless stops voiced stops aspirated stops voiceless labialized stops voiced labialized stops affricates voiceless fricatives voiced fricatives nasals approximants

bilabial

alveolar

p

ɸ m w

velar

glottal

t

k

ʔ

tʰ

kʰ kʷ ɡw

ts s z n

palatal

ɡ

tʃ h ɲ j

Tab. 23.3: Tinigua vowels.

high mid low

front

central

back

i e

ɨ

u o

a

(1)

a. b. c. d. e. f.

/towána/ /tʰowáha/ /hipótoha/ /hatʰó/ /kíʔtaʔ/ /kʰitʰáh/

‘soil/earth/land’ ‘palm species’ ‘mountain deer’ ‘kill with shotgun’ ‘vein’ (kíʔ ‘tree’ + -taʔ clf:long/thin?) ‘bone’

(2)

a. b. c. d. e. f.

/kamátso/ /kʰámba/ /haká/ /tsákʰa/ /píko/ /jakʰóɲa/

‘caiman’ ‘hen’ ‘stream/creek’ ‘chili pepper’ ‘vomit’ ‘tobacco’

The voiced velar stop /ɡ/ has a limited distribution and has not been found to occur word-initially; furthermore, /ɡ/ has only been attested before the high front vowel /i/ and the low central vowel /a/. Near minimal pairs illustrating the contrast between /ɡ/ and /k/ are shown in (3) in contexts where the voiced counterpart /ɡ/ is licensed to occur.

1038 (3)

Katherine Bolaños

a. b. c. d.

/pajɡíanoso/ /hamúki/ /tíɡa/ /jiaká/

‘devil’ ‘howler monkey’ (Alouatta seniculus) ‘lightning’ ‘wait’

Minimal (or near minimal) pairs illustrating the contrast between the voiced labialized velar stop /ɡʷ/ and the voiceless labialized velar stop /kʷ/ are presented in (4), together with non-velarized counterparts. (4)

a. b. c. d. e. f.

/híkʷa/ /héɡʷa/ /máʔɡa/ /haɡʷá/ /kaɲíɡahe/ /jákʷajkʷánajtʃa/

1sg.pro ‘no’ ‘macaw’ ‘today’ ‘what’ ‘wing’

Stops are always voiced when following a nasal consonant, as in (5), whereas voiced stops cannot in general occur word-initially (see Section 2.2. for the two known exceptions). (5)

a. b. c. d. e. f. g. h. i.

/hasampá/ /hompá/ /kʰámpa/ /pimpí/ /hintá/ /sɨntása/ /jontó-tsɨ/ /hánki/ /pánki/

[hasambá] [hombá] [kʰámba] [pimbí] [hindá] [sɨndása] [jondó=tsɨ] [hánɡi] [pánɡi]

‘be red’ ‘fall’ ‘hen’ ‘toad’ ‘curassow’ ‘skin boil’ ‘flute’ (dance/sing=clf:cylindrical) 3sg ‘forest’

As presented in the consonant inventory in Table 23.2, a phonological contrast is presumed to exist between /ts/, and /tʃ/, evidenced in examples like those in (6). (6)

a. b. c. d. e. f.

/hatʃa=pa/ /hatsa=ha/ /=tʃa/ /=tsa/ /tʃınzatso/ ̵́ /tsıɲa/ ̵́

‘manioc meal’ (manioc=clf:powder) ‘two’ (two=pl) clf:pointed clf:bundle ‘downwards’ ‘young man’

However, many instances of apparent neutralization and/or free variation between these two sounds are found in the same phonological environments, as seen in (7).

23 Tinigua

1039

Nevertheless, in many words the alveolar affricate /ts/ and the post-alveolar affricate /tʃ/ were not accepted as variants of each other when prompted (i.e., [ts] was not accepted in place of [tʃ] and vice versa; neither were these two sounds replaceable by /th/, /s/, /ts/ or /tʃ/ (e.g., (8), and see below for discussion). No prediction could be made to account for the variable occurrence of /ts/ and /tʃ/, as there is no evident phonological context to condition their alternation. (7)

a. [hatʃa=pa] ~ [hasa=pa] ‘manioc meal’ (manioc=clf:powder) b. [hatʃí] ~ [hatsí] ‘lie (down)’

(8)

a. b. c. d.

/tsıwa/ ̵́ /tʃınzatso/ ̵́ /tsietsı/̵́ /tʃipawína/

‘now/at.first’ ‘down’ ‘suck’ ‘sand’

*tʃı̵w ́ a, *tʰı̵w ́ a, *sı̵w ́ a *tsı̵ń zatso, *tʃı̵ń zatʃo, *sı̵ń zatso *tʰı̵ń zatso *tʃietʃɨ, *siesı̵,́ *tʰietʰı̵́ *tsipawína, *sipawína, *tʰipawína

The affricates /ts/ and /tʃ/ also show variations with other sounds in Sixto’s speech; for example, /ts/ shows variation with /th/ (apparently only in intervocalic contexts), as in [patsá] ~ [patʰá] ‘cut’ (although again this variation is not consistent even intervocalically, as seen in (8c)). The relationship between /tʃ/ and /kʰ/ is also not fully clear. The sounds [tʃ] and [kʰ] occur variably when preceding a high front vowel in many words, such as those in (9). However, certain minimal pairs illustrate a contrastive distribution for /tʃ/ and /kʰ/ in this environment, such as /tʃiɲíhi/ ‘lungs’ versus /kʰiɲíhi/ ‘back’. Substitution of /tʃ/ for /kʰ/ and vice versa was also not accepted in some other words, for example, /kʰía/ ‘river’ (never [tʃía]), and /mákʰikʰi/ ‘egg’ (never [mátʃitʃi]). To the extent that the variation exists, it might be explained as a palatalization process affecting /kʰ/; however, /tſina/ in (9) is a likely candidate for a loanword (from Arawakan, cf. Payne’s 1991 reconstruction *tsini) and thus is probably the original form in this case. (9)

a. b. c. d. e.

[sıkʰa] ~ [sıtʃa] ̵́ ̵́ [=kʰi] ~ [=tʃi] [hasuj-kʰi] ~ [hasúj-tʃi] [kʰíɲa] ~ [tʃíɲa] [haʔháj-kʰí] ~ [haháj-tʃí]

‘pineapple’ clf:liquid ‘manioc beer’ ‘jaguar’ ‘blood’

Other sounds that seem to vary sporadically with each other are [kʰ] and [t]. Both of these sounds, however, have ample evidence supporting their phonemic status: for example, /kʰaháj/ ‘hear/know/listen’ and /tatʰiá/ ‘brown woolly monkey’ (Lagothrix lagotricha); /kʰokʰój/ ‘cough’ and /towána/ ‘soil/earth/land’ (see also (1)–(2)). The variations discussed here may be due to idiosyncrasies in Sixto’s speech, perhaps associated with a partial loss of contrasts via language attrition. It is also

1040

Katherine Bolaños

possible that at least some of these variants are (or were historically) allophones of other sounds. The phonological analysis presented here contrasts in some respects with that offered in the only previous analysis of the language. Tobar Ortiz’s (2000: 670) inventory includes a set of voiced and voiceless palatal stops /c/, /ɟ/; I consider /tʃ/ to be the phoneme corresponding to Tobar Ortiz’s /c/, while I have found no evidence for the phonemic status of /ɟ/. Similarly, Tobar Ortiz (2000: 670) includes an aspirated voiceless bilabial stop /pʰ/ within the set of aspirated stops and indicates a contrast between /pʰ/, /p/, and /ɸ/. Nonetheless, in the data gathered with Sixto in 2017–2020, no instances of [ph] were found, and Sixto consistently gave [ɸ] in cases where Tobar Ortiz described /ph/. It is possible that a contrast between /pʰ/ and /ɸ/ did exist previously but that language attrition has led to a neutralization of the distinction. Examples illustrating the phonological contrast between /p/ and /ɸ/, as gathered through work with Sixto, are presented in (10). (10) a. b. c. d. e. f. g.

/ɲíɸo/ /haɸótʰio/ /hipóto/ /hoɸʷı/̵́ /ɸɨʔɸı/̵́ /hapıjtʃa/ ̵́ /haɸı/̵́

‘sun’ ‘manioc starch’ ‘savanna deer’ ‘command’ ‘scratch’ ‘night’ ‘cast spell/make prayer’

A further difference involves the status of glottalization. The present analysis includes a glottal stop as a phonemic consonant with a restricted distribution; it is not robustly attested in word-initial position. Tobar Ortiz’s analysis, on the other hand, considers glottalization to be a property of vowels (which she represents orthographically as a sequence of a vowel plus a glottal stop ), such that a series of glottalized vowels is a counterpart of the plain vowels. The examples in (11) illustrate the occurrence of glottalization in Tinigua words, here understood as a glottal stop functioning both as a (word-internal) syllable onset and as a coda. (11) a. b. c. d. e. f. g. h. i. j.

/jiʔú/ /nɨʔá/ /sɨʔáto/ /máʔga/ /máʔzɨʔ=na/ /ɲoʔkíʔa/ /tiʔkwúj/ /kíʔtaʔ/ /jihéʔ/ /háʔ/

‘eat’ ‘partner’ ‘grater’ ‘macaw’ ‘liver’ (liver=clf:long/round) ‘stream/creek’ ‘dove (n.)’ ‘vein’ ‘heart’ dem.prox

23 Tinigua

1041

2.2 Phonotactics Most syllables are either CV or CVC (with CVV understood to be a phonetic lengthening effect in word-final open syllables, already noted). Free variation appears to exist in word-initial syllables involving the vowel /a/, which may surface without any detectable onset or be preceded by a glottal stop [ʔ] or a glottal fricative [h]. For example, see (12); however, a distinction between #ʔV and #V is difficult to distinguish word-initially when not embedded in a phrase.3 In light of the general phonotactic patterns observed for Tinigua, these words are assumed to have /h/ as their underlying onset. Otherwise, syllables lacking an onset are licensed only in noninitial position, as in kʰí.a ‘river’. (12) a. b. c. d. e. f.

[ajohási] ~ [ʔajohási] ~ [hajohási] [ajúsi] ~ [hajúsi] [atʃáh] ~ [ʔatʃáh] ~ [hatʃáh] [apátsɨ] ~ [ʔapátsɨ] ~ [hapátsɨ] [atihíj] ~ [ʔatihíj] ([hatihíj] is not attested) [andej] ~ [handej] ~ [handaj] ~ [endíj]

‘be good’4 ‘many, group’ ‘manioc bread’ ‘quickly’ ‘be heavy’ element associated with subjects (sbj)

In general, any consonant can function as a syllable onset. However, word-initial position is restricted, as noted previously; the glottal stop and voiced stops are not attested word-initially (see exceptions presented for a set of words in (12)). Two apparent exceptions have been encountered: babatátsɨ ‘rubber’ (Tobar Ortiz 2000: 672), which may be a loanword (cf. balatá, a type of rubber derived from the Manilkara bidentata tree), and the element dása within the lexical construction ɲitsása nakʰóso dása ‘grinding stone’ (ɲitsása ‘stone’, nakʰóso ‘squeeze’, dása ‘?’) Word-internal codas are typically /ʔ/, /j/, /w/, /n/, or /m/, as in (13). (13) a. b. c. d. e. f. g.

[zɨj.tʃa] [kíʔ.wa.sá] [ɸʷáw.kʰí.o.ha] [tʰɨw.tsı]̵́ [ha.men.dó=ha] [ha.tsam.báj] [man.dó.tʰa]

‘aunt’ ‘toucan’ ‘motorboat’ ‘sugarcane’ ‘fly/mosquito [fly=pl]’ ‘be red’ ‘plantain’

Word-final position is also relatively restricted. Coda consonants are rare in this position, and they are limited to the glottal stop /ʔ/, glottal fricative /h/, and ap3 This variation is also encountered in the plural enclitic =ha in combination with pronouns. 4 The variants of hajohási ‘be good’ that lack initial [h] are found most frequently when it functions as an adverb or as a discourse marker relating to emphasis.

1042

Katherine Bolaños

proximant /j/, as in (14). The only known exception is hiáp ‘thin’. Word-final vowels are also phonetically lengthened, as noted above. (14) a. b. c. d.

[háʔ] [tiʔkwúj] [hatʃáh] ~ [atʃáh] ~ [ʔatʃáh] [hakomá]

‘this’ ‘dove (n.)’ ‘manioc bread’ ‘be blue’

Verb roots are most frequently encountered with a final syllable je (or its variant j; the motivations for the allomorphy are unclear), both in discourse and when uttered in isolation (i.e., in elicitation), except when occurring with other bound morphology. This element presumably is (or was historically) a verbal suffix -je; however, no function has yet been identified for -je, and in some cases its synchronic status as a segmentable morpheme is unclear. See Section 4 for further discussion. Interestingly, aspiration tends to occur on all voiceless consonants within a word, possibly suggesting some form of word-level assimilation process, as seen in (15). However, as was encountered in many areas of our work with Sixto, this pattern showed considerable variation. For example, the /k/ in panátʰiokʰo ~ panátʰioko ‘show/teach’ was consistently aspirated in narratives; however Sixto produced both aspirated and non-aspirated variants when uttering the same word in isolation, and he accepted both options as correct. (15) a. b. c. d. e.

atʰatʰokʰía kʰitʰáh tʰatʰátʰi hútʰiojítʰo hajotʰí=kʰi

‘Guayabero River’ ‘bone’ ‘drum’ ‘seat’ ‘beer’ (beer=clf:liquid)

Although it seems that virtually any combination of vowels is possible within the morpheme, as illustrated in the examples presented in this section, morphemes show a general preference for vowels of the same quality, for example: hanzá ‘be hot’, hɨtsı̵= ́ kʰi ‘urine’ (urine=clf:liquid), jondó ‘dance/sing’. In the relatively rare cases where two vowels occur in sequence (i.e., without an intervening consonant), /i/ is usually the first vowel, as in tʰiána ‘all’ or hisía ‘louse’. The sequences /ea/ and /oa/ are also attested, as in sapéahisí ‘four’, tʃáʔtsɨtoa ‘ear’, koáma ‘ayahuasca’, and in the nominalizing suffix -sea. These adjacent vowels are understood here to be distinct syllable nuclei, rather than diphthongs, and they pattern independently with respect to stress. Lexical roots frequently have three to four syllables, while function morphemes tend to be mono- or bisyllabic. Stress normally follows a metrical rule, falling on the penultimate syllable of the word, as seen in (16). However, stress appears on the syllable preceding the

23 Tinigua

1043

verbal suffix -je even when it is realized by its variant -j, creating a stressed final syllable, as in (16). Additional counterexamples to the penultimate stress pattern have also been identified; see (18). Accordingly, stress is represented orthographically throughout this chapter as an acute accent above the stressed vowel. (16) a. b. c. d.

zoɲátʰo jɨsı̵t́ o tʰiána haɸı̵-́ je

‘chest’ ‘canoe’ ‘all/everything/every’ ‘spell/prayer’ (cast a spell/pray-vsuff)

(17) e. f. g. h.

kʰaɸʷá-j ɲɨnzı̵-́ j pandiá-j matsohá-j

(18) a. b. c. d. e. f. g. h.

pimbí nɨtʰiá kʰaɸʷaɲisó janzı̵́ jası̵́ pawá hanó tatʰiá

‘burn’ (burn-vsuff) ‘see’ (see-vsuff) ‘laugh’ (laugh-vsuff) ‘be rotten’ (be.rotten-vsuff) ‘toad’ ‘spouse’ (and variant nɨʔá) ‘underneath/downwards’ ‘be ugly/bad’ (cf. janzı̵-́ j ‘be ugly/bad’) ‘brother-in-law’ ‘spider’ ‘dog’ ‘brown woolly monkey’

3 Word classes and morphological structure Word classes in Tinigua include nouns, pronouns, quantifiers, verbs, interrogative words, demonstratives, and function morphemes. Nouns and verbs are open classes; all other classes are closed. The distinction between verbs, nouns, and other word classes is not very clear-cut, given that the limited morphology present in Sixto’s use of the language seems to be available to multiple word classes; however, the syntactic function of nouns as arguments of clauses is not available to verbs unless additional morphology is present. Words that indicate properties are usually verbs (e.g., hahó ‘be fat/big’, tiní ~ kiɲí ‘be old’, janzı̵́ ‘be ugly/bad’) but can also be nouns. Their use as modifiers may involve zero derivation or other morphological processes; for example, hajú ‘big’ (cf. hahó-j ‘be big-vsuff’), hatsı̵́ ‘far’ (cf. hatsı̵-́ j ‘be far-vsuff’), and hajítʰo ‘full’ (probably from hají ‘be full’ + =tʰo clf:sturdy). There is no evidence that Tinigua has an adjective or adverb class distinct from nouns and verbs. Tinigua morphology seems to be relatively isolating compared to other northwestern Amazonian languages. The language has a handful of suffixes (e.g., verbal

1044

Katherine Bolaños

-je), prefixes (e.g., the verbal negator ka-; see Section 6), and clitics (which occur with a range of hosts; e.g., the plural marker =ha). However, any morphological encoding of tense, aspect, or mood is unattested in the available record of the language. Temporal and other values are instead expressed by adverbial particles or even by verbs (e.g., hatsı̵́ ‘be far’ for past events, hawá ‘today/now’ for present events, and kʰímba kʰamba ‘the day after tomorrow’ for future events; see Section 4.2).

4 The noun phrase A noun phrase in Tinigua may include pronouns, modifiers, quantifiers, demonstratives, and other elements. As Figure 23.1 illustrates, a demonstrative can fill the first slot of the noun phrase template, followed by a numeral, a possessor, a noun root (or two), and a modifier. The only obligatory element is the noun root. The order of elements presented here is typical for Tinigua, but other orders are also attested. Noun phrases function as clausal arguments and as adjuncts (e.g., ‘(in the) forest’), as in (19). The following examples illustrate noun phrases involving possessors (19), demonstratives, numerals and quantifiers (20)–(21), and other modifiers (22). (19) Sixto, kázɨ kiángi ɲihíje pangí, híkʷaʔa ɲɨnzı̵́ natsı̵h́ a hawá-j endí Sixto, 2sg.pro brother be forest, 1pl.pro see trail say-vsuff sbj hatsamí-j be.white-vsuff ‘Sixto, your brother is in the forest, we saw a trail, said the non-indigenous people.’ hiɸʷá ɲisá, kitʃi (20) hatʰatʰíhima híkʷa tsɨwı̵-́ j long.ago 1sg.pro be.together-vsuff (an)other woman, one hajohá-j be.good-vsuff ‘Long ago I married a woman from another group, for one (year) all right.’ (21) ajohási endí hángi kíʔ ɲisá be.good nom this one woman ‘This (one) woman is/was a good (wife).’ (22) híkʷa tsɨwı̵-́ j hatʃandí ɲisá nɨtʰiá 1sg.pro be.together-vsuff indigenous woman spouse ‘My (first) wife was an indigenous woman.’

23 Tinigua

1045

demonstrative | quantifier | possessor / noun | modifier Fig. 23.1: Structure of the noun phrase.

4.1 Nominal morphology Nominal morphology in Tinigua is limited to classifier endoclitics (which encode shape and gender; see (23)), the plural number endoclitic =ha, in (24), and the associative plural suffix, as in (25). The distribution of the plural clitic is not limited to the noun phrase, as noted previously. Classifiers may occur on multiple elements within the noun phrase and can combine with verb roots to derive nominalizations (see below), as in (23) (where the nominalized verb modifies a noun). (23) jɨsɨ=tʰo ɲı̵w ́ =tʰo canoe=clf:sturdy be.small=clf:sturdy ‘little canoe’ (24) kʰakʰá nɨzı̵= ́ ha jitʰátʰako=ha here(this one) boy=pl be.hungry=pl ‘These boys here are hungry.’ (25) kʰahé-j hégʷa, híkʷaha tʰiáj-ajo, tʰiána híkʷaʔa know-vsuff neg2 1pl.pro deceased-assc.pl all/everything 1pl.pro kʰahé-j hégʷa know-vsuff neg2 ‘(We) didn’t know, our deceased (forefathers and those with them), all of us didn’t know.’ Noun phrases functioning as subject arguments of clauses may be marked by the morpheme handá (or its variants endi ~ anda ~ andej). This morpheme typically occurs when the subject follows the verb, and it is not normally present when the subject appears in its canonical position preceding the verb (see Section 6). Its function is not clear, and it is simply identified loosely here as a subject-related element. However, an association with the subject is not evident in a few examples; it is plausible that its function relates more directly to information structure, but this remains to be explored. (26) hawá-je, ɲaǃ endi hatsamí-j ɲísa say-vsuff here sbj be.white-vsuff woman ‘“Here you go, (have this)!” the white woman said.’ The plural number clitic =ha occurs on nouns – which may have either animate referents (e.g., kʰisí=ha ‘lice’, kuzı̵= ́ ha ‘parrots’; see also examples (24), (30)) or inan-

1046

Katherine Bolaños

imate referents, as in (27). It also occurs on predicates, which usually take =ha in agreement with a subject that is marked as plural or associative plural, demonstrated in (28). See Section 5 for further examples. (27) a. hokiatsá=ha ‘hammocks’ b. kiwí=ha ‘leaves’ c. jaɸʷajhatsá=ha ‘loincloths’ (28) kʰahé-j=ha Witoto-ajó know-vsuff=pl Witoto-assc.pl ‘The Witotos and their people, they knew (us).’ Within our corpus of natural speech data and elicited constructions, it appears that the plural number clitic occurs freely with nouns that are treated as mass nouns in other languages. Sixto provided examples such as jamá=ha ‘feces’ and tʰiózɨndá=ha ‘rice’ as the plural forms of these nouns; he accepted the meaning ‘pieces/units of N’ for these constructions (i.e., jamá=ha ‘pieces of feces’, or tʰiózɨndá=ha ‘grains of rice’), but it is not altogether clear what their semantic range involves. The associative plural or human collective suffix -ajo occurs (probably exclusively) with human nouns, principally kin terms and proper names, and indicates a group formed by the nominal referent and associated entities, that is, ‘Noun + those with Noun’, or ‘Noun group’; see (29). The plural and associative plural markers occur on co-referential noun phrases in (30). (29) capitán kikʷá-ajo captain son-assc.pl ‘the captain’s son (and those with him)’ (30) hawá-j endi kʷa hanoso=ha Tinigʷa-ajó say-vsuff sbj 1pl.pro person=pl Tinigua-assc.pl ‘We people, all the Tiniguas, say…’ Two morphological strategies for deriving nominal constructions from verbs are evident in Tinigua. A verb root can take the nominalizing suffix -sea (e.g., ɲihó-sea ‘seed’, from ɲihó ‘(to) plant’ + -sea), or a classifier (as in hatʃanda=pa ‘ground coffee’ formed by hatʃanda ‘be dark’ + =pa clf:powder); see also Section 5.4. Compounding of two or more noun roots, or a verb and noun root (to form a nominalization), is also attested. Such compounds often result in relatively semantically non-compositional lexical constructions, but they do not normally yield a single phonological word, in that the juxtaposed elements tend to maintain independent stress. Examples include ɸʷáwkʰíoha ‘motor’, from poɸʷáj ‘work’ and kʰía ‘river’ (the final ha element may relate to the plural marker =ha); and tʰɨwtsı̵j́ sakʰá ‘panela rallada’, from tʰɨwtsı̵j́ ‘be sweet’ and sakʰá ‘salt’.

23 Tinigua

1047

4.2 Pronouns and demonstratives Tinigua has both free pronouns and bound person markers. Free pronouns are part of the noun phrase, while bound pronouns are procliticized to verbs; Both sets appear in Table 23.4 for comparative purposes. The distribution and function of the bound pronominal elements is not fully clear; while they may be best understood as cliticized proforms, their ability to co-occur with the free pronouns, as in (31), suggests that they may serve an agreement-marking function (see also Section 4.2 below). The bound forms are encountered most frequently with the verbs of movement like ‘go’ (wa=náha ‘I go’, ɲi=náha ‘we go’) and ‘come’ (mi=nakʷú ‘they come’), and with the verb ‘exist’ (wa=ɲihíj ‘I exist/am (at X location)’); however, their occurrence with these and other verbs, and their specification for particular person and number values, are not consistent across the corpus.5 (31) híkʷa kíʔ-je wa=ɲihí-j 1sg.pro be.one-vsuff 1sg=exist-vsuff ‘I am alone (here).’ In addition to the first-, second-, and third-person pronouns, the indefinite pronoun hiɸʷáj ~ hoɸʷáj ‘other’ is used when the referent is not specified and in distinguishing between two third-person participants whose identity has already been established in the discourse. This element can occur both as an independent proform and as a modifier of another pronoun, as seen in (32), where two participants are deciding who should follow some Carijonas who had passed by and killed a large group of Tiniguas. (32) hiɸʷáj hawá-je hiɸʷáj hángi, “híkʷa natʰá matʰóapi-j other say-vsuff other 3sg.pro 1sg.pro go.downriver paddle.off-vsuff haʔá,” hawá-je here say-vsuff ‘One said to the other, “I am paddling off downriver,” (he) said.’

Tab. 23.4: Free and bound pronouns. free pronouns

st

1 2nd 3rd

pronominal proclitics

singular

plural

singular

plural

híkʷa kázɨ hángi

hikʷáʔa ~ hikʷáha kakʷáha ~ kázɨha ~ kakʷáʔa hitʃákʰá

wa= ma= a=

ɲi= n= mi=

5 For example, Sixto sometimes used the 3pl form mi= when first-person singular reference was apparently intended.

1048

Katherine Bolaños

Tab. 23.5: Demonstratives and adverbial expressions. demonstrative pronouns

singular

plural

proximate (‘this’) distant (‘that’) far distant (‘yonder’)

hángi híkʰa/hingí hitʃági

híkʰáha hitʃákha6

Demonstrative adverbials and related expressions ‘here, now’ ‘here (is) this’ ‘here, this’ ‘close’ (this here) ‘far’ (that far one) ‘at first, now, today’ ‘afterwards’

kʰakʰá ɲá háʔ ~ kʰáʔ nɨʔtsı ́ ̵ hatsı ́ ̵ ̵́ a tsıw hatʰatʰíhima

Demonstrative pronouns in Tinigua specify three degrees of distance (spatial or temporal), and a set of deictic adverbial expressions also encodes meanings relating to spatial/temporal distance; see Table 23.5. The element identified as a proximate demonstrative is identical to the third-person singular pronoun (hángi), and both interpretations may be available, as in (33). tsı̵w (33) hángi hanó=ha ɲihí-j ́ a hanó=ha hapátsaɲo hawá 3sg.pro dog=pl exist-vsuff at.first dog=pl three today ɲihí-j hégʷa exist-vsuff neg2 ‘He (this one here) had dogs, three dogs at first, but today there aren’t (he doesn’t have) any.’ The following examples illustrate the use of demonstratives and spatial/temporal adverbial expressions: far distant hitʃági, (34)–(35); distant hingí, (36); and hatʰatʰíhima ‘afterwards’, (34). (34) ajohási mihíma=ha hitʃági kíʔa hatʰatʰíhima hawá-j anda be.good 3pl.exist=pl far.dist river afterwards say-vsuff sbj hatsamá=ha, “Sixto kazı̵́ kiangi ɲihí-je pangí híkʷaʔa ɲɨnzı̵-́ j white=pl Sixto 2sg.pro brother exist-vsuff forest 1pl.pro see-vsuff natsı̵= ande hatsamí-j ́ ha” hawá-j trail=pl say-vsuff sbj be.white-vsuff ‘For a long time [they, my brother and his wife] were fine at that river. Afterwards, the non-indigenous people told me, “Sixto, your brother is in the forest, we saw trails (there),” said the non-indigenous people.’ 6 The final ha element in the plural demonstratives is presumably the plural marker, at least etymologically.

23 Tinigua

1049

(35) híkʷa nahé-j hitʃági jisɨ=to, nahé-j hatsı̵́ 1sg.pro go-vsuff far.dist canoe=clf:sturdy go-vsuff far ‘I left on that canoe, I went far away.’ (36) ajúsi kiháwa ɲihí-j hingí ɲisá many years be/exist-vsuff dist woman ‘For many years (I) was with that woman.’

4.3 Nominal number and quantification The numeral system in Tinigua is very small, as in many Amazonian languages. Although Sixto provided numeral terms for 1–10 and 20 in elicitation (see Table 23.6), in discourse he used Tinigua numerals only for values up to 3; beyond that he used Spanish forms. Numeral terms above 1 appear to be morphologically complex, although no etymology can be determined for ‘three’ and ‘four’, and only partial etymologies are visible in the other terms. Note also that the words for ‘three’ and ‘six’ are identical.7 There are no numeral classifiers in Tinigua, but the terms for ‘two’ and ‘seven’ appear to involve a nominal classifier (=tʃa clf:pointed), probably as a fossilized element. Quantifiers include tʰiána ‘all/every’, tití ‘a lot’, and tʃasí=ha ‘any’, shown in (37). Some quantifiers appear to have a verbal source or be derived from verbs (and can function adverbially), for example: hajúsi (~ ajúsi) ‘(be) many/much’, as in hajúsi panáj ‘steal/rob a lot’; and nɨwsi=ha ‘a few’ (from nɨw ‘be small’ + -si advb + =ha pl).

Tab. 23.6: Numerals. number

form

possible etymology

one (also ‘single, last’) two three four five six seven eight nine ten twenty

kíʔje hátsa ~ hatsajtʃa hapátsaɲo saʔpéa ~ sapéahisí tsáthokwaha hapátsaɲo tsáthokwahátʃa ɲihíkwahá hínzathoɲihí hopáthokwahá hapatsáɲokwahá

verb kíʔje ‘be.alone’ hatsa+=tʃa ‘?+clf:pointed [no transparent elements] [no transparent elements] kwaha possibly related to kwáʔana ‘hand’? [no transparent elements]; same form as 3 tsáthokwahá+=tʃa ‘five+clf:pointed’ ɲihí+kwahá ‘exist+hand?’ hínzatho+ɲihí ‘?+exist’ hopátho+kwahá ‘?+hand?’ hapatsáɲo+kwaha ‘six+hand?’

7 Similar cases of numeral indeterminacy are attested in some other Amazonian languages; see e.g., Pache (2018: 275) for Irohito (Uru, Andean).

1050

Katherine Bolaños

Still others are derived from other parts of speech, such as hitʃangi=ha ‘some/various’ (far distant demonstrative + =ha pl). Non-numeral quantifiers usually precede the noun being quantified, as in hitʃángi=ha hanoso=ha ‘various people (various=pl person=pl)’. Numerals, on the other hand, seem to have a flexible ordering with respect to the noun they modify; compare (33), where the numeral follows the head noun, and (38), where the numeral precedes the head noun. As seen in the examples here, numerals and other quantifiers can also head noun phrases. (37) tʰiána jiʔú=ha andá ɸʷɨw-jiʔú=ha, ɸʷɨw-jiʔú=ha tití all eat=pl sbj scratch-eat=pl scratch-eat=pl a.lot ‘The Carijonas (scratchers and eaters) ate them all, (there were) lots of Carijonas.’ natʰá-j kíʔje hatʰátʰokʰía (38) kíʔje hanoso=ha tsı̵ʔ́ ɲa one person=pl young.adult go.downriver-vsuff one Guayabero.River natokʷú-je come-vsuff ‘One young adult from (our) people went downriver, one came to the Guayabero River.’

4.4 Nominal classification Table 23.7 presents the full set of classifiers that have been identified to date in Tinigua. Classifiers are normally unstressed (as expected for word-final syllables); however, this pattern is not fully consistent, most notably in the case of =tsɨ ‘hollow/ cylindrical’ (whether this variation is grammatically relevant is unknown). Only one classifier occurs with nouns having animate referents, the form =to ~ =tho ~ =tʰjo, which encodes animate feminine; animate masculine/non-feminine is unmarked. A formally near-identical classifier also occurs with inanimates and is associated with heavy or sturdy semantics. While these two forms are understood here as homophonous classifiers (and thus glossed separately), they are likely historically related. Classifiers occur with both nominal and verbal roots to form nominal constructions, but the degree to which these combinations are productive in Tinigua is not clear from the available data. As observed in Section 3.1, verb roots may combine with classifiers to form nominalizations, as seen in lexicalized constructions such as hı̵t́ sɨ=kʰi ‘urine (hı̵t́ sɨ ‘urinate’ + clf:liquid)’, jiʔú=kʰi ~ jiʔú=tʃi ‘poison (jiʔú ‘eat’ + clf:liquid)’, and haɲí=na ‘sky (haɲí ‘be up’ + clf:long/rounded)’. Classifiers relating to plant parts may also combine with terms that refer generically to the plant (or its fruit), as seen in (39) for =sa clf:palm, =tsa cfl:bunch, and =wea clf:leaf. However, even these constructions do not appear fully productive; for example, the

23 Tinigua

1051

Tab. 23.7: Classifiers. semantics

form

example h

Animate feminine

=to ~ =t o ~ =tʰjo

hanó=tʰjo tiní=tʰjo

‘female dog’ ‘old woman (be.old=clf:f)’

Powdery

=pa

towá=na=pa hatʃá=pa hatʃandá=pa

‘dirt (dirt=clf:long=clf:powder)’ ‘manioc meal’ ‘ground coffee (hatʃandá ‘be dark’)’

Liquid (probably etymologically related to kʰía ‘river’)

=kʰi ~ =tʃi

hasúj=tʃi ~ hasúj=kʰi hatokú=tʃi hatʃandá=tʃi haʔháj-kʰí ~ haháj=tʃi heɲí=kʰi ~ heɲí=tʃi

‘manioc beer, chicha’ ‘rain (hatokú ‘to rain’)’ ‘coffee (drink) (hatʃandá ‘be dark’)’ ‘blood’ ‘honey’

hamá=tʰo makó=to zoɲá=tʰo

‘metal roof’ ‘rock/mountain range’ ‘chest’

jɨsɨ=to

‘canoe’

Heavy, sturdy

=to ~ =tʰo

h

Pointed

=tʃa ~ =t a

hatsikʷáj=tʃa hatʰósɨj=tʃa sıj́ ̵ kʰi=tʃa

‘matches (for fire)’ ‘dart (hatʰósɨ ‘shoot dart’)’ ‘leg’

Hollow, cylindrical

=tsɨ

jondó=tsɨ kʰízɨ=tsɨ kiʔhí=tsɨ

‘flute (jondó ‘sing’)’ ‘nose’ ‘beehive’

tʰɨw=tsı ́ ̵ tsɨkʷá=tsı ́ ̵

‘sugar cane’ ‘throat’

Long or rounded

=na

hatʃá=na maʔzıʔ́ ̵ =na jawahé=na kwáʔa=na

‘manioc cooking pot’ ‘liver’ ‘boat paddle’ ‘hand’

Palm(like)

=sa

mandó=sa jahó=sa

‘plantain/banana tree’ ‘peachpalm tree’

Leaf

=wea

jahó=wea

‘peachpalm leaf’

Bundle, bunch

=tsa

mandó=tsa jamá=tsa

‘bunch of plantains’ ‘intestines’

Intangible, invisible

=ti

hasú=ti hakoma=ti hatʃandá=ti

‘wind’ ‘fog’ ‘Tinigua language (hatʃandá ‘be.dark/indigenous’)’

1052

Katherine Bolaños

plant terms mundí ‘lime’ and tʰiomá ‘seje palm (Oenocarpus bataua)’ occur with juxtaposed nouns referring to plant parts, rather than classifiers, as shown in (40). (39) a. b. c. d. e.

mandó=sa mandó=tsa jahó jahó=sa jahó=wea

‘plantain tree/palm’ ‘bunch of plantains (fruit)’ ‘peachpalm (Bactris gasipaes) (fruit)’ ‘peachpalm tree/palm’ ‘peachpalm leaf’

(40) a. b. c. d.

múndi múndi kíhi múndi kiwíha tʰiomá kiwíha

‘lime (fruit)’ ‘lime tree’ (kíhi ‘tree’) ‘lime tree leaves’ (kiwí ‘leaf’, =ha pl) ‘seje palm leaves’ (Oenocarpus bataua)

Tobar Ortiz (2000: 678) indicates that classifiers may occur with demonstratives, apparently in agreement with the noun. Our own data, gathered with Sixto in 2017– 2020, includes a few cases of apparent classifier agreement within the noun phrase (see (41) and (42)), but this pattern is not consistent. The presence of both derivational and agreement functions is a feature of classifier systems in many northwestern Amazonian languages. (41) jɨsɨ=to hasɨsá=to Pájaro jɨsɨ=to canoe=clf:sturdy be.green=clf:sturdy Pájaro canoe=clf:sturdy ‘The green canoe (at the shore), (that one) is Pájaro’s canoe.’ ̵́ to=ha (42) jiʔú=ha-ajó saʔpéa=to=há jɨsı= natʰiá=ha eat=pl-assc.pl four=clf:sturdy=pl canoe=clf:sturdy=pl go.downriver=pl ‘The Carijonas (Lit. ‘the eaters and those with them’) went downriver in four canoes.’

4.5 Nominal possession Possession of nouns is formed by juxtaposition of the possessor (noun or pronoun) and the possessed: (N1)Possessor + (N2)Possessed, as seen in (43)–(46). There seems to be no formal distinction relating to the expression of alienability. (43) hijóki-j wa=nahá híkʷa kíko ɲihí-je, ɲihí=ha hikʷa abandon-vsuff 1sg=leave 1sg.pro house be-vsuff exist=pl 1sg.pro kiɲí-ajo sási old.person-assc.pl too ‘I left and abandoned my house, my mother-in-law and everyone who was there, I abandoned them too.’

23 Tinigua

1053

(44) nahá nihá-j totoðá kíko go go-vsuff doctor.sp house ‘(They had me) going to the doctor’s house.’ (45) híkʷa jazı̵j́ tʃatsá=ha 1sg.pro clothes=pl ‘my clothes’ (46) hángi jíti 3sg.pro head ‘his head’ Reference to the owner of an inanimate object may be conveyed via the noun nigá ‘owner’, which follows the noun to indicate ‘the owner of N’, as in (47) and (48). (47) kikó nigá jawı̵-́ je Rafico Rafael Gordillo jawı̵-́ je house owner be.called-vsuff Rafico Rafael Gordillo be.called-vsuff ‘The house owner’s name was Rafico, Rafael Gordillo was his name.’ (48) hawá-j endí ɸʷawkióhe nigá say-vsuff sbj motorboat owner ‘The motorboat owner said (to me)…’

4.6 Modification of nouns Nouns may be modified by other nouns and by verbs describing qualities, as in (49)–(51); as noted in Section 2 above, there appears to be no distinct adjective class in Tinigua. The modifier element usually precedes the noun it modifies. No dedicated diminutive or augmentative morphology is attested. (49) nɨzɨtʰiáma tsı̵w hégʷa, tiní-j ɲisá wa=tsı̵w híkʷa sanzɨ ́ ɨ ́ ɨ child partner neg2 be.old-vsuff woman 1sg=partner 1sg.pro want ‘I don’t want a child for a wife, I want an old woman to be my partner’ (50) hajú=tʰó jɨsı̵= ́ tʰo big=clf:sturdy canoe=clf:sturdy ‘big canoe’ (51) tʰiána mi=jondó=ha híkʷaʔa Tinigʷa jondó=ti=ha all 1pl=sing=pl 1pl.pro Tinigua sing=clf:intangible=pl ‘We all sang our Tinigua songs.’

1054

Katherine Bolaños

5 Verbs and verbal morphology Tinigua verbs denote processes, actions, states, and properties, as observed in Section 2 (including concepts that are often associated with an adjective class in other languages; e.g., hasɨtsá ‘be green/blue’, hanzá ‘be hot-’, and (52)–(54)). Verb stems are morphologically regular; there are almost no suppletive forms or other types of allomorphy. The highly frequent verbs ‘go’ (nahí, nahá, nahé) and ‘come’ (nakʷú, naká) do exhibit multiple variants; however, as with many other instances of variation in our data, they do not show a consistent pattern of use and, when prompted, Sixto would accept any variation of these forms. (52) janzı̵-́ j hawá-j endi híkʷa tatá be.ugly-vsuff say-vsuff sbj 1sg.pro father ‘My father said that it is very bad (that the Witotos are killing us).’ hatíhi-sea, hawá-j (53) kakʷá naní-j hégʷa hatsı̵,́ hajúsi-j 2pl.pro go-vsuff neg2 be.far be.many-vsuff be.heavy-nmlz say-vsuff sákʰa, tʰɨwtsı̵-́ j, sakʰá, hatʃanda=pa, hatihí-je salt be.sweet-vsuff salt be.dark=clf:powder be.heavy-vsuff ‘You are not going (to make it) far, (it) is too heavy; the salt, panela rallada, ground coffee, is too heavy, (she) said.’ (54) ajohási ɲi=híma wajíma hikʷáha be.good 1pl=exist be.happy 1pl.pro ‘We were happy.’ Unlike many languages of the region, including those the Tinigua have interacted with in the past (e.g., Witoto, Guahibo, and perhaps Carijona, as evident from stories that Sixto learned from older relatives), Tinigua verbal morphology seems even more limited than that of nouns. As explored below, there are a few prefixes and/ or proclitics (marking imperative mood, negation, and person/number) and a few suffixes and enclitics (plural number and derivational elements). A process involving rounding of a final vowel in verb roots (e.g., pakʷa > pakʷú ‘kill’) may also be morphological, but clues concerning its function (if any) were not available in our data. Verbs are frequently encountered with the final element -je (or its allomorph -j) and are almost always given with this ending in elicitation form. While no specific function has been identified for -je ~ -j, it is evidently a verbal suffix that is still at least somewhat synchronically active; it is usually absent when a verb root occurs in a compound or takes other morphology, such as the plural element =ha, in (55), or the nominalizer -sea. For example, consider hatihí-je ‘be heavy’ and hatihí-sea ‘heavy (thing)’, and jiʔú-j ‘eat’ and jiʔú-sea ‘food’. Here we simply call it a “verbal

23 Tinigua

1055

suffix” (vsuff), since no more precise identification can be made at this time. Only a few verbs have been encountered without -je ~ -j in the absence of other morphology; some examples appear in (56). (55) hikʷáʔa hawá=ha 1pl.pro say=pl ‘We were saying.’ (cf. hawá-j ‘say’) (56) a. b. c. d. e.

matsóso mapízɨ jitáʔtsako ~ jitáʔtʰakʰo nazɨtóko hiap

‘blow’ ‘chew’ ‘be hungry’ ‘walk upwards’ ‘be thin’

Derivational morphology associated with verbs is limited to the nominalizing strategies discussed in Section 4.1 and a possible adverbializer -si (hají-si ‘fully’, cf. hají ‘be full’).8 Compounding of verbs roots is for the most part unattested in Tinigua, aside from a few highly lexicalized exceptions (e.g., the nominal construction ɸʷɨw-jiʔú=ha ‘Carijonas (scratch-eat=pl)’). Examples of juxtaposed verb roots otherwise appear to function as coordinated independent predicates or clauses (see Section 6). The placement of modifiers within the verb phrase varies. Temporal adverbs may follow the verb root or precede it, as in (59) and (60), respectively. Other adverbial modifiers tend to follow the verb, as in (53). There is still much to be understood about verbs in Tinigua, and the description presented in this section may not fully reflect the verbal system as it appeared in the past.

5.1 Verbal person and number marking As observed in Section 4.2, verbs may take subject-marking proclitics that encode person and number; the extent to which these involve agreement is unclear. Verbs may also take the plural element =ha, which agrees with a plural subject argument of the clause. For verbal plural marking to occur, the subject must also be overtly marked as plural (either by =ha or by a plural pronoun; see Section 4.1 and examples (57)–(58)); however, the converse does not hold, in that a plural-marked subject does not require plural marking on the verb.

8 As with many other features of Tinigua, this element does not appear to function consistently in our data and is therefore left unglossed in some of the examples in this chapter.

1056

Katherine Bolaños

(57) ka-ajohási handá hatsmá=ha pakʷá tsı̵= híkʷa ́ ha neg1-be.good sbj be.white=pl kill want=pl 1sg.pro ‘These non-indigenous people are not good, they wanted to kill me.’ (58) hikʷáʔa pakʷá=ha nɨaní=ha 1pl.pro kill=pl peccaries=pl ‘We were/are killing peccaries.’

5.2 Tense, aspect, and modality The Tinigua materials indicate a striking lack of any morphological resources associated with tense, aspect, and modality, a fact that potentially reflects language attrition. The expression of temporality is limited to adverbial resources, such as kohá ‘yesterday’ (59), hímba ‘tomorrow’ and hímba kʰámba ‘after tomorrow’ (60), and haʔ semana ‘this week’ (62). These adverbs are not obligatorily expressed, and temporal reference may simply be understood from the discourse context. (59) ma=nihá hawá kaʔgaɲiséa compadre mi=nakú hoɸú-je kohá 2sg=go say why compadre 3pl=come send-vsuff yesterday ‘Compadre said, “Why did you go (away)? Yesterday they came and sent (for you to come with them)”.’ (60) himba kʰámba hajohási tomorrow after be.good ‘The day after tomorrow (it) will be okay.’ (61) hímba kʰámba híkʷaʔa nakú=ha sási hawá-je híkʷa tatá tomorrow after 1pl.pro return=pl also say-vsuff 1sg.pro father ‘We will come back again the day after tomorrow, my father said.’ (62) háʔ semana híkʷa nahí-j hégʷa this week.sp 1sg.pro go.away-vsuff neg2 ‘I am not going away this week.’ Likewise, no aspectual morphology has been identified. The incomplete or progressive interpretation of the events in examples (63)–(65) is conveyed by the discourse context; however, when prompted, Sixto indicated that other aspectual and temporal interpretations were also possible.

23 Tinigua

1057

(63) hawá-j endi híkʷa tatá comadre míwahá-je híkʷa say-vsuff sbj 1sg.pro father comadre menstruate-vsuff 1sg.pro jiʔú-sea jiʔú hégʷa eat-nmlz eat neg2 ‘My father said, “Comadre, you are menstruating, I can’t eat the food (that you cook).’ [also accepted: ‘comadre you have menstruated/you menstruated/you will menstruate’] (64) haɲitone semana nihá paná-j háʔ semana hégʷa next week.sp go work-vsuff this week neg2 ‘Next week I will go to work, not this week.’ [also accepted: ‘Next week I will be working’] (65) híkʷa nɨʔzı̵́ wa=ɲihí híkʷa kitsá-ajó mi=nakú kʰakʰá 1sg.pro boy 1sg=exist 1sg.pro aunt-assc.pl 3pl=come here atʰatʰókʰía Guayabero River ‘I was a boy (when) they and my aunt came/come here to this Guayabero River.’ Modal values relating to epistemic, deontic, or other categories are also not morphologically marked in our data; nor are any evidential categories. Example (66) illustrates a modal interpretation relating to ability that is conveyed only via negation, expressing the meaning ‘not able to verb’. (66) atʃandáj=tʃi ɲaɲíku-j hégʷa jiʔú-sea hégʷa ɲimá-j be.dark=clf:liquid drink-vsuff neg2 eat-nmlz neg2 be.sick-vsuff hajohasi hégʷa janzı̵-́ j be.good neg2 be.ugly-vsuff ‘(I) can’t drink coffee, (can’t eat) food, (I) get sick, it is not good, it is awful.’

5.3 Valency changing strategies The morpheme jaʔná encodes reflexive meanings (and probably reciprocal as well, but this was not clear from Sixto’s responses). (67) hóʔkʰúʲ-je pakʰakʰá-je jaʔná hiɸwa hatsamá-j hóʔkʰúʲ-je ande climb-vsuff teach-vsuff refl other be.white-vsuff climb-vsuff sbj kitiá wife ‘The wife taught herself to climbed up, [and taught] the other white person to climb.’

1058

Katherine Bolaños

In (68), the interpretation and the position of the reflexive element jaʔná (apparently within the noun phrase rather than adjacent to the verb) suggests that it may have an additional function as a reflexive intensifier. (68) kʰambá híkʷa tatá hoɸú-je anda jaʔná compadre after.day 1sg.pro father send-vsuff sbj refl compadre ‘The other day, the compadre himself sent for my father (sent someone to fetch my father).’

6 Simple clauses In this section, we consider the structure of the simple clause, strategies for negation, and sentential mood. These phenomena involve a heterogeneous set of morphological and lexical resources.

6.1 Constituent order and argument marking Constituent order in declarative and other types of clauses is predominantly AVO, as in (69), and (72), and SV, as in (70). Other orders are also encountered, including VS (71), (A)OV (66), OAV (80), and OVA (68). No differences have been observed between constituent ordering involving (free) pronouns and that involving full noun phrases. (69) hángi pomá-je hajotsí=kʰi, hatsı̵w kázɨ jaɲekʷá-je hégʷa ́ a 3sg.pro buy-vsuff beer=clf:liquid far/long.ago 2sg.pro drink-vsuff neg2 ‘He bought beer, it’s been a long time that you haven’t drunk.’ (70) hatsamí-j ɲisá comadre kʰahí-j hégʷa be.white-vsuff woman comadre know-vsuff neg2 ‘The non-indigenous comadre doesn’t know.’ (71) kʰakʰá ɲihí-j hégʷa hikʷáha here exist-vsuff neg2 1pl.pro ‘We weren’t here.’ As noted in Section 3.1, a subject that follows the verb is usually preceded by the morpheme handá ~ endi ~ anda ~ andej, which we simply label here as a subjectrelated element. No pattern governing this apparent allomorphy has been determined, and we do not rule out the possibility that the variants have some functional differences. The handá (etc.) morpheme is normally absent in the context of the

23 Tinigua

1059

more canonical subject-verb order (though see below for exceptions). Its function is not well understood; whether its use is associated with information structure or some other phenomenon remains to be clarified. Examples illustrating the occurrence of handá (or its allomorphs) with a postverbal subject are presented in (72)–(74); other examples are encountered throughout this chapter. Example (72) includes two clauses, in which the second follows the canonical subject-verb order and does not include handá. (72) híkʷa kiɲíso Irene pakʷá=ha anda hatsama=ha. Palma pakʷu-j 1sg.pro sister Irene kill=pl sbj be.white=pl Palma kill-vsuff híkwa kiɲiso Irene 1sg.pro sister Irene ‘The non-indigenous people killed my sister Irene (or: My sister Irene was killed by the non-indigenous people). Palma killed my sister Irene.’ (73) natʰukó-j andéj híkʷa kikʷá go.down.river-vsuff sbj 1sg.pro son ‘My son went downriver.’ (74) mahánika=ha andá jazı̵j́ =tʃa=tsa=ha híkʷaʔa bring=pl sbj cover=clf:pointed=clf:bundle=pl 1pl.pro jazı̵j́ =tʃa=tsa=ha cover=clf:pointed=clf:bundle=pl ‘(They) brought (us) clothes, clothes (for) us.’ While handá (etc.) is preferred with post-verbal subjects, this pattern is far from consistent in our corpus. Examples like (75) are found in which a subject follows the verb but handá is absent. In others, such as (76), handá occurs with a subject that precedes the verb (possibly triggered here by the clause-initial position of the object kazı̵́ 2sg). Example (77) is even more puzzling, in that handá appears to be associated with an object argument rather than a subject, and it intervenes between two elements of a noun phrase (possessor and possessed, which normally cannot be separated by other material). It is not clear to what extent the variants associated with handá might be acceptable in any of these unusual contexts. (75) kíko janı̵-́ j híkʷa tatá house build.roof-vsuff 1sg.pro father ‘My father was roofing the house.’ (76) Sixto, kazı̵́ endí Isabel nahá-j Sixto 2sg.pro sbj Isabel go-vsuff ‘Sixto, Isabel left you.’

1060

Katherine Bolaños

(77) pakʷá=ha híkʷa endi kíso kill=pl 1sg.pro sbj sister ‘(They) killed my sister.’ Morphological marking of arguments is otherwise unattested in Tinigua. Objects are unmarked regardless of their position with respect to the verb or their association with transitive or ditransitive verbs; see (78) and many other examples throughout this chapter. (78) híkʷaʔa naníko ajúsi jiʔú-sea, kázɨ jiʔú-sea, tʰiána naníko ajúsi, 1pl.pro bring a.lot eat-nmlz 2sg.pro eat-nmlz everything bring a.lot kázɨ kíso hitʰánatsı̵,́ hawá-j 2sg.pro daughter dress say-vsuff ‘We will bring you (back) lots of food, your food, a lot of everything we will bring for you, (we will bring) your daughter a dress, (they) said.’ Adjuncts typically appear in either initial or final position within the clause. They too are morphologically unmarked, and the information about the type of adjunct relation is for the most part conveyed only contextually. Examples (79)–(81) show adjuncts with comitative and instrumental interpretations, respectively. (79) híkʷaʔa tatá waha mi=naha 1pl.pro father 1sg.go 3pl=go ‘They and I (we) went with our father.’ [waha < wa=naha 1sg=go] (80) híkʷa hanó híkʷa hatsó-j kíʔ=tʃa 1sg.pro dog 1sg.pro hit-vsuff stick=clf:pointed ‘I hit my dog with a stick.’ Examples (81)–(83) illustrate directional and locative readings. The verb ɲihí-je ‘exist, be at’ is also used for expressing locative relations (84). (81) nahá-j endí Isabel ɸʷanájajó híkʷa napátokó-je hégʷa go-vsuff sbj Isabel San.José.del.Guaviare 1sg.pro find-vsuff neg2 ‘Isabel went to San Jose del Guaviare, (but) I didn’t find (her).’ (82) híkʷa nahé-j Macarena, hapaná-j handá hanóso=ha kʰámba, 1sg.pro go-vsuff Macarena steal-vsuff sbj person=pl hen kaɲíga nahé-j híkʷa, híkʷa nahí-j hégʷa why go-vsuff 1sg 1sg.pro go-vsuff neg2 ‘If I go to Macarena, people will steal my hens, why will I go? I won’t go (to Macarena).’

23 Tinigua

1061

(83) tʰiána pakʷá=ha handá hatsamá=ha haʔá atʰatʰókia sbj be.white=pl here Guayabero.River everything kill=pl ‘The non-indigenous people killed them all, here at the Guayabero River (I’m the only one left).’ (84) kíʔaʔtso ɲihí-je híkʷa tatá Yarí.River exist-vsuff 1sg.pro father ‘(We were) at the Yarí River with my father.’ Some locative relations, such as ‘under’ and ‘above’, may be expressed via positional verbs. In (85), for example, the relation ‘from above’ is expressed via the verb haɲí ‘be above’. (85) kíʔje hanóso nɨw tsı̵ʔ́ ɲa natʰá-j ɲɨnzı̵-́ je hátsama-j one person young man go.downriver-vsuff see-vsuff be.white-vsuff haɲí-j ɸʷowkí-je kʰihí be.above-vsuff go.up-vsuff tree ‘One young man went downriver to look at the non-indigenous people from above in a tree.’

6.2 Negation Tinigua has two strategies for marking standard negation: the verbal prefix ka- and the particle hégʷa. Hégʷa is used to negate both verbal and verbless clauses. In most examples of negation with the ka- prefix, hégʷa also appears following the verb (86) and (87); but hégʷa is not obligatory with ka-, as shown in (88). (86) híkʷa kikʷá ɲihíje ɸʷanájajó, ka-nakú=ha hégʷa 1sg.pro son exist San.José neg1-come=pl neg2 ‘My son lives in San Jose del Guaviare; (they) don’t come (to visit).’ (87) ka-wá-je ka-pakʰákʰa ka-wá-je ka-ɲí-wa=ha neg1-say-vsuff neg1-teach neg1-say-vsuff neg1-exist-int=pl ‘(He) didn’t say (what had happened), didn’t teach/show (where they had gone), didn’t say why (they) weren’t (there).’ [wá > hawá ‘say’] (88) ka-nanikó=ha hégʷa jiʔú-sea neg-send=pl neg2 eat-nmlz ‘They did not bring the food.’

1062

Katherine Bolaños

However, ka- is relatively infrequent, even in verbal clauses. The use of just the negative particle hégʷa generally seems to be the preferred negation strategy across clause types. (89) kázɨ ɲiná-j hégʷa ɲijní-j natóko hégʷa hatsı̵-́ j Yarí 2sg.pro sleep-vsuff neg2 exist-vsuff return neg2 be.far-vsuff Yarí ‘Don’t fall asleep (on the way back); [if you fall asleep] you won’t get back, the Yarí is far.’ (90) kaɲíga nɨzı̵= ́ ha, mijítátako ajúsi híkʷaʔa, jiʔú hégʷa how child=pl be.hungry a.lot 1pl.pro eat neg2 ‘“How are you, boys?” (my father asked); “We are very hungry, we haven’t eaten.”’ In verbless clauses, hégʷa is the only available resource for negation, as in (91) and (92). It directly follows the negated constituent. tsakʰá hégʷa, ɲitáto=ha hégʷa. tsı̵w (91) sakʰá hégʷa, tʰɨwtsı̵́ ́ a salt neg2 be.sweet chili.pepper neg2 fish.hook=pl neg2 now/first kʰahé-j hégʷa tʰiozɨndá know-vsuff neg2 rice ‘No salt, no panela rallada or chili pepper, no fishing hooks. At first we didn’t know about rice.’ (92) kʰákʰa hanóso=ha hajohá-si hégʷa this person=pl be.good-advblz neg2 ‘These (newly arrived) people are not well.’

6.3 Interrogative clauses There seems to be no distinctive prosody or intonation strategy associated with interrogative clauses in Tinigua. Polar interrogatives occur with the verbal suffix -ga ~ -wa, as seen in (93)–(94), but this marker and its function are not yet fully understood. (93) mi=nániko-ga hatʃá=pa? 3sg=bring-int manioc=clf:powder ‘Did you bring manioc meal?’ (94) hímba hujá-wa hatʃá=pa? tomorrow do-int manioc=clf:powder ‘Are you making manioc meal tomorrow?’

23 Tinigua

1063

Tab. 23.8: Question words. interrogative

form

what who where which how many why when

kaɲígahe ~ kaɲí nıʔ́ ̵ a maʔága ~ maʔá máʔná matisíha ~ matisíga kajesíga ~ kaɲisíga mahátsa

Content questions require an interrogative pronoun or word, which occurs at the beginning of the clause. Table 23.8 provides the inventory of attested question words. Most of these question words are evidently etymologically complex. Four of them involve an initial element ma- ~ maʔ,9 and several appear to include the interrogative suffix -ga ~ -wa (see above). Variants of the forms ‘what’ and ‘why’ involve the element kaɲí-, which may itself derive from the negative prefix ka- and/or ɲihí ‘exist’; ‘why’ may also contain the adverbializer (?) -si. The following examples illustrate content interrogatives in Tinigua. Examples (98)–(99) show the two variants of ‘what’ (kaɲígahe and kaɲí). (95) [maʔá nahá-j Berenice?] [nɨʔá nahá nahá-j Berenice]? where go-vsuff Berenice who go go-vsuff Berenice ‘Where did Berenice go? Who went with Berenice?’ (96) panátʰo hawá-j, hawá-j, [kaɲisíga andá hatʃanda=pa teach say-vsuff say-vsuff why sbj be.dark=clf:liquid pakʰáha-j]? know/think-vsuff ‘“(I) will teach (you)” [my father] said; (he) said, “but why (do you want to) know about coffee?”’ (97) [mahátsa tsı̵w híkʷa nahá hégʷa, ́ a nakú hikʷáha atʰatʰókʰía?] when be.far come 1pl.pro Guayabero.River 1sg go neg2 [maʔá híkʷa nahá]? where 1sg.pro go ‘When would we go back to the Guayabero River? I am not going (anywhere); where will I go?’ 9 Tobar Ortiz (2000:678) presents the interrogative words as built on a base form maʔ; while this may be historically accurate for some forms, it does not apply to all of them, and they do not appear to be synchronically compositional. Tobar Ortiz’s forms are also somewhat different from those we

1064

Katherine Bolaños

(98)

jawɨ-je hawá-j andá hatsamá=ha. híkʷaʔa tatá hawá-j be.called-vsuff say-vsuff sbj be.white=pl 1pl.pro father say-vsuff [kaɲígahe jawı̵-́ je]? what be.called-vsuff ‘This is its name [gasoline], the non-indigenous people said. Our father said, “what is this called?”’

(99)

hatsamá-je kakʷá=ha hawá-si hawá-j, [kʰáɲi hégʷa]? be.white-vsuff 2pl=pl say-advr say-vsuff what neg2 ‘The non-indigenous people, in your language/saying, said “what (if)/why not?”’

6.4 Imperative clauses Imperative clauses are formed by adding the prefix ma- to the verb root, as seen in (100). (100) ma-hapátsɨzɨ! imp-wait ‘Waitǃ’ (101) ma-nahúke! hawá-j endí hatsamá=ha, Caqueteñ o imp-climb.up say-vsuff sbj be.white=pl, Caqueteño ‘“Climb up/pull it up!” the non-indigenous people told me, the Caqueteño.’ Negative imperatives also employ the negative particle hégʷa, as in (102). (102) ma-johó hégʷa! hawá-j handá dotodá imp-bath neg2 say-vsuff sbj doctor.sp ‘“Don’t batheǃ” the doctor told (me).’ Imperatives directed toward second-person plural and first-person addressees (exhortatives) are also marked with ma-, as in (103) and (104). (103) hawá-j endi híkʷa kiʔje, ma-náha kohá ma-náha jɨsɨto jawahéna! say-vsuff sbj 1sg.pro one imp-go dig imp-go canoe paddle.canoe ‘I said myself (to my brother), “let’s go dig, let’s go by canoe and paddle!”’

recorded: maʔ +a ‘where?’, maʔ + ga ‘which?’, maʔ + tsa ‘(from) where?’, maʔ + hawa ‘when?’, maʔ + tsi-ha ‘how many?’.

23 Tinigua

1065

(104) ma-ɲɨnzı̵-́ j hímba jɨsı̵t́ o ma-nakú-j! hawá-j imp-see-vsuff other.day canoe imp-return-vsuff say-vsuff ‘“Tomorrow you all look, you all come back to the canoe!” (he) said.’ We note that ma- is formally identical to the second-person singular pronominal proclitic; however, in light of its use with non-second-person singular addressees (but cf. the inconsistencies in person-marking across our corpus more generally; Section 3.2), we treat it here as a distinct imperative marker.

6.5 Existential and equative constructions The existential verb ɲihíj ‘exist, be at’ is used in expressions of existence and location, as in (105) through (108). (105) híkʷa kíʔje wa=ɲihí-j hawá, Tinigʷa, Tinígʷa hanoso 1sg.pro one 1sg=exist-vsuff today Tinigua Tinigua person ɲihí-j hégʷa exist-vsuff neg2 ‘I am the only one left today, Tinigua, there are no more Tinigua people.’ (106) tʰiána jimá=ha, ɲihí-j hégʷa Tinígʷa-ajó haʔá everything be.dead=pl exist-vsuff neg2 Tinigua-assc.pl here ‘Everyone is dead, there are no Tiniguas here.’ (107) kázɨ kʰahé-j, hatsı̵-́ j ɲihí-j kázɨ 2sg.pro know-vsuff be.far-vsuff exist-vsuff 2sg.pro ‘You know, you (have to) stay far away (from the fire).’ (108) Alcaldía kíko ɲihí, nihákʰa mula, caballo ɲihí hégʷa, mula Alcaldia house exist go mule.sp horse exist neg2 mule.sp ‘The Alcaldía house was (there), (they) went off by mule; there weren’t horses, just mules.’ Equative relations are expressed by the juxtaposition of the subject and predicate noun phrases, without a verb, as in (109) and (110). They are negated via the negative particle hégʷa, postposed to the predicate nominal, demonstrated in (110). (109) híkʷa tatá hangí ɡʷeá 1sg.pro father 3sg.pro chief ‘My father, he was the chief.’

1066

Katherine Bolaños

(110) híkʷa kiɸʷajkiso kíkʷa pastor hégʷa 1sg.pro grandchild son pastor.sp neg2 ‘My grandson is not (the) pastor.’

7 Clause combining Coordinated clauses are simply juxtaposed, as in the following examples. Adverbs like tsı̵w ́ a ‘now’ and mitʰatʰíhi ‘afterward’ may indicate the temporal relationship between clauses, illustrated in (113). Coordination may also involve verb phrases without repetition of the subject, as in (113) and (114). (111) híkʷa hawá hawá-je, japatsó=ha tsátokʷahá, hawá wa=pané-j. 1sg.pro today say-vsuff money=pl five now 1sg=steal-vsuff híkʷa hawá-j ajohási, Ricardo kʰahá-je sási 1sg.pro say-vsuff well Ricardo know-vsuff also ‘Today I am going to talk about money, about (how) I have stolen five (thousand pesos). I am going to tell it straight (well), and Ricardo knows it too.’ Ricardo, nakú-je (112) kohá tsawá nakú-je, tsı̵w ́ a nakú-je yesterday be.far come-vsuff now come-vsuff Ricardo come-vsuff Katherine, nakú-je Mona Katherine come-vsuff Mona ‘The other day (they) came, Ricardo came and Katherine came and La Mona came.’ (113) híkʷa jiʔú=ha ɲihá-j, tsı̵w ́ a mitʰatʰíhi ɲikʷáj 1sg.pro eat=pl pull.out.manioc-vsuff now afterward water mitʰatʰíhima sási ɲihá-j kʰía tsı̵w ́ a tsowáje, pull.out.manioc-vsuff river now soak.manioc-vsuff afterward also hútʰioji-j, tsı̵w hatʃá=pa ́ a sási ɸʷowáhá-je sit-vsuff now also toast.manioc-vsuff manioc.meal=clf:powder ‘For my food [I need to go] and pull out the manioc, and then pull it and put it in the water to soften, and then sit and also toast it to make meal.’ (114) “waní nániko, pómi-j hapı̵-́ j waní,” hawá-j endi, “hõ”, pig come pay-vsuff be.later-vsuff pig say-vsuff sbj yes hawá-j endi Varela say-vsuff sbj Varela ‘“(I) come with a pig and pay (you) for the pig later”, said (the nonindigenous woman); “yes,” said Varela.’

23 Tinigua

1067

Clause-internally, the coordination of noun phrases is likewise carried out simply by juxtaposition, as in (115). (115) nanániká=ha jazı̵j́ tʃatsá=ha, nakʷánohá=ha, jiʔú-se=ha bring=pl clothes=pl wash=pl eat-nmlz=pl ‘They brought (us) clothes, and soap, and food.’ Subordination is perhaps the least understood aspect of Tinigua grammar at this point. There appear to be no morphological markers of subordination, and no other elements have been identified that indicate a formal distinction between a main and subordinate clause. Examples of what appear to be complement clause constructions are presented in (116) through (118). From the available examples, the clause understood as the complement follows the main clause, and its subject is elided; however, it is not known whether this is a fixed pattern or merely a tendency.10 (116) híkʷa tatá sánzɨ hégʷa [kíko janı̵-́ je] 1sg.pro father want neg2 house build.roof-vsuff ‘My father did not want [to build a house].’ (117) híkʷaʔa jaje-j hawá-j [nahá-je hégʷa] 1pl.pro think-vsuff say-vsuff go-vsuff neg2 ‘We think [we are not going].’ (118) hángi jaje-j hawá-j Capitan [hono-je], híkʷa hono-je 3sg.pro think-vsuff say-vsuff Captain heal-vsuff, 1sg.pro heal-vsuff hégʷa neg2 ‘He, the Captain thinks [I am a shaman], but I am not.’ Examples (119)–(120) appear to involve relative clauses. In both examples, the relative clause is separated from the noun it modifies, and it may be best understood as a co-referential, headless relative clause. (119) pakwa-je híkʷa kiangi, Severo tiutsa-ajo, [pana-j hángi kill-vsuff 1sg.pro cousin Severo deceased-assc.pl steal-vsuff 3sg.pro nɨtʰiá ɲisá], híkʷa kiangi Severo pakʷá-j hatsami-j pakʷí-j partner woman 1sg.pro cousin Severo kill-vsuff be.white-vsuff kill-vsuff ‘(They) killed my cousin, deceased Severo and those with him, [the one who stole his wife], they killed my cousin Severo, the non-indigenous people killed (him).’ 10 In examples (117) and (118), the verb jaje ‘think’ appears with the verb hawá ‘say’. We note that complementizers in many other languages have grammaticalized from verbs of speaking, although it is unclear whether such a process is relevant here.

1068

Katherine Bolaños

(120) híkʷa hangi ɲisa kʰahe-j hegʷa hawu-je [Yarí handajajo] 1sg.pro 3sg.pro woman know-vsuff neg2 say-vsuff Yarí town Tinigʷa=tí Tinigua=clf:intangible ‘My woman doesn’t know it, the one [who lives] at the town of Yarí, the Tinigua language.’ Dedicated adverbial clauses are also difficult to identify in Tinigua. Clauses may occur in sequence and be preceded by an expression of time or manner, as seen in (112), but there is no evidence that these are subordinated rather than coordinated. Such time/manner expressions include verb roots like tsı̵w ́ a ‘be far’, which can also express meanings like ‘while’, ‘after’, ‘and so’, ‘now’; and hajohási ‘be good’, that can also be used to relate clauses in discourse (i.e., ‘well then’, ‘and so’). Associations between clauses may also be conveyed merely by context, as in (121). (121) zosi=ha hombá-j zosi=ha jíti, pomá-j endí híkʷa jajı̵h́ akóto hair=pl fall-vsuff hair=pl head buy-vsuff sbj 1sg.pro hat ‘(When/after) my hair fell off my head, (and) I had to buy a hat.’ Expressions of condition and consequence are likewise expressed via coordination with their relationship determined contextually. Alternative readings of (122) and (123) include ‘Uriel didn’t return, (and so) I won’t go to Macarena’; and ‘It hasn’t rained this week, (and so) the creek will dry out too’. (122) Uriel nakú hégʷa, híkʷa nahí-j hégʷa Macarena Uriel come neg2 1sg.pro go-vsuff neg2 Macarena ‘(If) Uriel doesn’t come back, I won’t go to Macarena.’ (123) hawá nɨsáj=tʃi ɲihí-j hégʷa, hatukúj=tʃi hégʷa, sási today storm=clf:liquid exist-vsuff neg2 rain=clf:liquid neg2 also tsɨhá-j nɨó-kʰía be.dry-vsuff small-river ‘(If) it doesn’t storm, (if) it doesn’t rain this week, the creek will dry out too.’

8 Conclusion This chapter presents the current state of analysis for Tinigua, a language with – as far as is currently known – only one remaining elderly speaker: Sixto Muñoz. Our work with Sixto has resulted in a far more robust documentation than existed previously, but conversation and other interactional discourse genres remain undocumented due to the moribund state of the language. Further analysis of the existing

23 Tinigua

1069

data will certainly provide more information about the structure of this language, and perhaps more speakers will eventually be located. Our work with Sixto has revealed a number of typologically interesting characteristics of Tinigua, alongside various features that resemble those of other languages of the northwest Amazon. Tinigua’s five-vowel inventory and high central vowel, as well as the presence of a glottal stop and other contrastive consonantal features (like the set of stops, fricatives, and nasals), are fairly typical of the wider region, as is its preference for bisyllabic CVCV lexical roots. In contrast to other languages of the region, Tinigua has a strikingly simple and highly isolating morphological structure, which may reflect some degree of language attrition. Morphological marking of case, tense, aspect, and mood is virtually unattested, with the possible exception of the poorly understood element endi (and its variants) that tends to occur with subjects. On the other hand, Tinigua does have a small set of nominal classifiers which share characteristics with neighboring languages. Among the many other intriguing features of Tinigua that call for further study, we highlight the following: the asymmetry between the plural and singular forms of the bound person markers, and likewise between the free personal pronouns and bound person markers; the existence of a nominalizer (-sea) but no apparent verbalizing morphology (at least none that is synchronically evident); the presence of identical terms for the numerals ‘three’ and ‘six’, which might imply a strategy of counting units in pairs (as seen in some other languages of lowland South America); and the apparent lack of any dedicated morphology for subordination. Interesting characteristics of the lexicon include formal parallels between certain words, suggesting an etymological connection, such as ɲiná ‘sleep’ and ɲimá ‘die’; the same form (kʰaháj) for both ‘hear’ and ‘know’; and associated terms for ‘sun’ (ɲíɸo) and ‘moon’ (hatsı̵́ ɲíɸo, cf. hatsı̵́ ‘far’). Although efforts to classify Tinigua have proposed a small Tiniguan (or Pamiguan) language family, the lack of sufficient information about Pamigua, and the complete absence of any record of Majigua, means that Tinigua is best treated as a language isolate.

9 Acknowledgments My grateful thanks go to Sixto Muñoz for his patience and collaboration in the documentation of Tinigua. I also thank the various institutions that supported the documentation initiative: the Endangered Languages Documentation Programme, the Instituto Caro y Cuervo in Colombia, the Endangered Languages Fund, and the Jacobs Research Funds. In 2015 Jorge Emilio Rosés Labrada provided the initial impetus for the documentation and description of Tinigua and contributed his own resources toward the first effort to establish contact with Sixto. I am grateful to the editors of this handbook and an anonymous reviewer for their detailed suggestions that en-

1070

Katherine Bolaños

riched this contribution. The work with Sixto and his family was also made possible through the collaboration of anthropologist Ricardo Palacio Hernández, who has contributed greatly to understanding the social dynamics and to reconstructing the history of the Tinigua ethnic group; I thank him for his collaboration and ongoing comments, suggestions, and discussion. The Tinigua documentation resulting from this work is housed in the Endangered Languages Archive, https://www.elararchive.​ org.

10 References Adelaar, Willem with Peter Muysken. 2004. The languages of the Andes. Cambridge: Cambridge University Press. Bolaños, Katherine & Ricardo Palacio Hernández. 2017. Documentation of Tinigua. London: SOAS, Endangered Languages Archive, ELAR. https://elar.soas.ac.uk/Collection/ MPI1079925. Campbell, Lyle. 2012. Classification of the indigenous languages of South America. In Lyle Campbell & Verónica M. Grondona (eds.), The indigenous languages of South America: A comprehensive guide, 55–166. Berlin & Boston, MA: de Gruyter Mouton. Campbell, Lyle & Martha Muntzel. 1989. The structural consequences of language death. In Nancy Dorian (ed.), Investigating obsolescence, 181–196. Cambridge: Cambridge University Press. Castellví, F. Marcelino de. 1938. Materiales para estudios glotológicos. Boletín de estudios históricos 7(84). 365–382. Castellví, F. Marcelino de. 1940. La lengua tinigua. Journal de la Société des Americanistes de Paris 32. 93–101. Departamento Administrativo Nacional de Estadística. 2018. Censo 2018. Available online. Ernst, Adolf. 1891. Über einige weniger bekannte Sprachen aus der Gegend des Meta und oberen Orinoco. Zeitschrift für Ethnologie 23. 1–13. Fabre, Alain. 1998. Manual de las lenguas indígenas sudamericanas. Vol. 2. (LINCOM Handbooks in Linguistics 4–5.) München: Lincom Europa. Franco, Roberto. 1989. Los Tiniguas y la colonización de la Macarena. Trianea 3. 129–142 Igualada, Francisco de. 1938. Musicología indígena de la Amazonia colombiana. Boletín latinoamericano de música 4. 675–670. Kaufman, Terrence. 1990. Language history in South America: What we know and how to know more. In Doris L. Payne (ed.), Amazonian linguistics, 13–74. Austin: University of Texas Press. Kaufman, Terrence. 2007. The native languages of South America. In Christopher Moseley & R. E. Asher (eds.), Atlas of the world’s languages, 61–78. 2nd edn. London & New York: Routledge. Les Corts, Estanislau de. 1926. Informe de las misiones católicas de Colombia relativas a los años 1925, 1926. Santa Fé de Bogotá: Imprenta de la Luz. Les Corts, Estanislau de. 1931. Apuntes para vocabulario Tinigua, de los Llanos del Yari (Año de 1931). Manuscript. Unpublished. Loukotka, Čestmír. 1935. Clasificación de las lenguas sudamericanas. Prague: Tipografía Josef Bartl. Loukotka, Čestmír. 1942. Klassifikation des südamerikanischen Sprachen. Zeitschrift für Ethnologie 74. 1–69. Loukotka, Čestmír. 1968. Johannes Wilbert (ed.), Classification of South American Indian languages. Los Angeles: UCLA Latin American Center.

23 Tinigua

1071

Manresa, Fructuoso de. 1935–1936. Encuestas sobre la lengua y costumbres tiniguas. Unpublished. Capuchin missions reports. Olivares, Antonio. 1962. Aves de la región sur de la sierra de la Macarena, Meta, Colombia. Revista de la Academia Colombiana de Ciencias Exactas, Físicas y Naturales XI/44. 305–346. Ortiz, Francisco & Francois Queixalós. 1981. Ornitología Cuiva-Guahibo. Amerindia. Revue d'Ethnolinguistique Amérindienne 6. 125–147. Pache, Matthias. 2018. Lengua X: An Andean puzzle. International Journal of American Linguistics 84(2). 265–285. Palacio Hernández, Ricardo & Katherine Bolaños. 2019. El hombre sin miedo. La historia de Sixto Muñoz, el último Tinigua. Bogotá: Instituto Caro y Cuervo. Payne, David L. 1991. A classification of Maipuran (Arawakan) languages based on shared lexical retentions. In Desmond Derbyshire & Geoffrey Pullum (eds.), Handbook of Amazonian languages. Vol. 3, 355–499. Berlin: de Gruyter Mouton. Pinell, Gaspar de. 1929. Excursión apostólica por los ríos Putumayo, San Miguel de Sucumbíos, Cuyabeno, Caquetá y Caguán. Bogotá: Imprenta Nacional. Rivet, Paul. 1924. Langues de l’Amérique du sud et des Antilles. Famille Guahibo. In Antoine Meillet & Marcel Cohen (eds.), Les langues du monde par un groupe de linguistes sous la direction de Antoine Meillet et Marcel Cohen, 640–707. Paris: Champion. Rivet, Paul. 1948. La famille linguistique Guahibo. Journal de la Société des Americanistes de Paris 37. 191–240. Rivero, Juan de. [1736] 1883. Historia de las misiones de los llanos del Casanare y los ríos Orinoco y Meta. Imprenta de Silvestre. Bogotá [Jesuit missions]. Rosés Labrada, Jorge Emilio. 2016. Relaciones de parentesco de las lenguas sálibas con otras lenguas amazónicas: una evaluación crítica. Special Mini-Colloquium 2 June 2016. Bogotá: Instituto Caro y Cuervo. San Martivel, Justo de. 1926. Excursiones apostólicas atrevidas. Revista de Misiones (Bogotá) 15. Tobar Ortiz, Nubia. 1995. Los tinigua, en el umbral de una muerte inevitable. In Marta Pabón Triana (ed.), Lenguas Aborígenes de Colombia, Serie Memorias. Vol. 3, 62–74. Bogotá: Centro Colombiano de Estudios de Lenguas Aborígenes (CCELA), Universidad de los Andes. Tobar Ortiz, Nubia. 2000. La lengua tinigua: anotaciones fonológicas y morfológicas. In María Stella González de Pérez & María Luisa Rodríguez de Montes (eds.), Lenguas indígenas de Colombia. Una visión descriptiva, 669–679. Bogotá: Instituto Caro y Cuervo. Valdés Arcila, Mónica Liliana. 1996. En el filo de la muerte: etnohistoria de los tinigua durante el siglo XX. Universidad Nacional de Colombia undergraduate thesis. Voegelin, Charles F. & Florence M. Voegelin. 1965. Languages of the world – Native America fascicle two. (Anthropological Linguistics 7(7).) Bloomington, IN: Archives of the Languages of the World, Indiana University.

1072

Katherine Bolaños

11 Appendix A. 100-word Swadesh list 1.

I

híkʷa

2.

you (sg)

kázɨ

3.

we

hikʷáʔa ~ hikʷáha

4.

this

hángi

5.

that

híkʰá

6.

who

nıʔ́ ̵ a

7.

what

kaɲígahe ~ kaɲí

8.

not

hégʷa (neg2)

9.

all (of a set)

tʰiána

10.

many

hajusi ~ ajútsɨ ~ ajusi

11.

one

kíʔ-je (verb: be one, be first)

12.

two

hatsajtʃa ~ hátsa

13.

big

hajjúko

14.

long (not wide)

hatsı ́ ̵ (be long/far/distant)

15.

small

nıó ̵ (be small)

16.

woman

ɲísa

17.

man (adult male human)

tsɨtsía

18.

person (individual human)

hanoso

19.

fish (noun)

zɨsóha (pl) (zɨsó + =ha pl)

20.

bird

tsɨtsáha (pl) (tsɨtsá + =ha pl)

21.

dog

hanú

22.

louse

hisía

23.

tree (not log)

kíhi

24.

seed (noun)

ɲihósea (ɲihó ‘to plant’, -sea nmlz)

25.

leaf (botanics)

kʰiwíha (khiwí + =ha pl; probably etymologically ‘tree+clf:leaf+pl’)

26.

root (botanics)

kiʔtʰáha (kiʔthá + =ha pl)

27.

bark (of tree)

28.

skin

kʷátatsa

29.

flesh (meat, flesh)

handá ɲisá (handá ‘part of’?)

30.

blood

hahájkʰía (haháj + kʰía ‘river’)

31.

bone

kʰitʰáh (related to ‘root’?)

23 Tinigua

(continued) 32.

grease (fat, organic substance)

33.

egg

34.

horn (of bull, etc.)

35.

tail

matsíɸʷajtʃa (may contain -tʃa clf:long/pointed)

36.

feather (large, not down)

hatsówi (may contain part of the word khiwí ‘leaf’)

37.

hair (of human head)

zozíha (zozí + =ha pl)

38.

head (anatomic)

jɨéte

39.

ear

tʃátsɨtoa

40.

eye

sıt́ ̵ i

41.

nose

kʰízɨsɨ

42.

mouth

kíwa

43.

tooth (front, rather than molar)

jióto (may contain -tʰo/-to clf:heavy/sturdy)

44.

tongue (anatomical)

tʰinútsa (may contain -tsa clf:bunch)

45.

claw

46.

foot

sıḱ ̵ ina (may contain -na clf:long/round)

47.

knee

tʰízɨsa

48.

hand

kʷáʔana (may contain -na clf:long/round)

49.

belly (lower part of body, abdomen)

jamako

50.

neck (not nape)

51.

breasts

jáʔzɨná (may contain -na clf:long/round)

52.

heart

jihéʔ

53.

liver

máʔzɨʔná (may contain -na clf:long/round)

54.

drink (v.)

jaɲekʷá

55.

eat

jiʔú

56.

bite

hakʷı ́ ̵

57.

see

ɲɨnzı ́ ̵

58.

hear

kʰahá

59.

know

kʰahá

60.

sleep

ɲiná

61.

die

ɲimá

62.

kill

pakʷá

63.

swim

ɲisı ́ ̵

mákʰikʰí (may contain -kʰí clf:liquid)

1073

1074

Katherine Bolaños

(continued) 64.

fly (v.)

jakʷjní

65.

walk

kiwá

66.

come

nakú

67.

lie (on side, recline)

hatʃí

68.

sit

hútʰjoí

69.

stand

70.

give

nahánika

71.

say

jajé hawá (also ‘think’)

72.

sun

ɲíɸo

73.

moon

hatsı ́ ̵ ɲíɸo (Lit. ‘far sun’)

74.

star

ʋṍsa11

75.

water (n.)

ɲikʷájtʃe

76.

rain (n.)

hatʰokútʃe

77.

stone

ɲitsátsa (may contain -sa clf:bunch)

78.

sand

tʃipawína (may contain -na clf:long/round)

79.

earth (soil)

towána

80.

cloud

haɲíj natʰí (cf. haɲí ‘be up/above’)

81.

smoke (n., of fire)

sɨnátʰi

82.

fire

hikʰítsa (may contain -tsa clf:long/pointed)

83.

ash(es)

84.

burn (v.)

kʰaɸʷá

85.

path

nátsɨ

86.

mountain

87.

red (color)

hatsambá (‘be red’)

88.

green

hasɨtsá (‘be green’)

89.

yellow

hasaná (‘be yellow’)

90.

white

hatsamá

91.

black

hatʃandá (‘be black/dark’)

92.

night

hapıj́ ̵ tʃa (also ‘late’)

93.

hot

hanzá (‘be hot’)

11 The sound /ʋ/ is not otherwise attested in our Tinigua corpus; possibly this word is a borrowing.

23 Tinigua

(continued) 94.

cold

hotsı ́ ̵

95.

full

hajítʰo (may contain -tʰo, clf:heavy/sturdy)

96.

new

97.

good

98.

round

99.

dry

hapawú (‘be dry’)

100.

name

jawú (be called/named)

hajohási ~ ajohási

1075

Raquel Guirardello-Damian

24 Trumai 1 2 3 4 5 6 7 8 9 10

Introduction Phonology Word classes The noun phrase The verb phrase Simple clauses Clause-linking and discourse Conclusion Acknowledgments References

1 Introduction The purpose of the current paper is to provide an overview of Trumai (trum1247), presenting its main characteristics and updating my 1999 overview of the language (Guirardello 1999). Trumai is spoken in Brazil in the indigenous reserve named Parque Indígena do Xingu, where other ethnic groups also live and several languages coexist. The Trumai people have three main villages: Três Lagoas (with families from the former Terra Preta village), Boa Esperança, and Steinen. More recently, a large Trumai family created an extra village (Wani Wani), which is located outside of the Xingu area, in an indigenous reserve called Terra Indígena Capoto-Jarina. In addition, there are families living in other locations in the Xingu area or in nearby cities. The number of inhabitants of the villages varies from year to year, given the mobility of the families; a fairly recent official count recorded 97 individuals (Ipeax 2011). However, the number of speakers is much smaller – there are 46 individuals who can speak Trumai, with different degrees of proficiency; they are all middle-aged or elderly people. For the younger generations, Portuguese is now the main language of daily communication, together with other Xinguan languages, such as Kamayurá, Suyá, and Aweti. Trumai is considered endangered – Ethnologue classifies it as having “very low language vitality” (Lewis 2009) – but there have been some language revitalization initiatives in recent years, especially through school and educational activities. Trumai is considered to be a genealogically isolate language. Greenberg (1987) classified it as a member of his Equatorial stock, but in addition to problems with

Raquel Guirardello-Damian, University of Bristol, United Kingdom https://doi.org/10.1515/9783110432732-011

1078

Raquel Guirardello-Damian

24 Trumai

1079

the Trumai data he employed, this stock (and his broader Amerind stock of which it is a part) is viewed skeptically by specialists (Campbell 1988, Kaufman 1990, Rankin 1992). Examining the Trumai lexicon, it is possible to find a number of words that have some similarities to Macro-Jê languages (Eduardo Ribeiro, personal communication, 2007), but a more careful examination is necessary in order to determine whether these similarities are systematic and would be an indication of genealogical links. The very first documentation of this language was conducted by Von den Steinen (1884), resulting in a list of basic vocabulary (mainly objects, names of animals, body parts, and kinship terms). Murphy and Quain (1955) carried out anthropological studies, but not on the language. Monod-Becquelin produced the first proper descriptions of Trumai, analyzing its phonology and aspects of its morphology (1970, 1975, 1976). Her work was followed by my studies (Guirardello 1992, 1999, 2005, among others), which expand our knowledge about its grammar. In the next sections, we have a description of the main aspects of the language. Section 2 offers an overview of the phonology, while Section 3 explores word classes and morphology. Sections 4 and 5 are dedicated to the noun phrase and verb phrase, respectively. Section 6 outlines the features of simple clauses, while Section 7 describes complex sentences. Finally, Section 8 presents a summary. For the data in this article, the symbols of the International Phonetic Alphabet (IPA) are employed, and the Trumai words are presented in their phonological form.

2 Phonology Trumai has six vowels: /i/, /e/, /ɨ/, /a/, /u/, /o/. The mid vowels /e/ and /o/ have two allophones in free variation: [e] and [ɛ], and [o] and [ɔ], respectively; in other words, the pronunciation varies between more open and more closed. There are twenty-three consonants, presented in Table 24.1.1 There is only one voiced stop. There is a distinction between dental and alveolar stops, and between plain and ejective stops or affricates (n.b., there is no bilabial ejective). Trumai seems to be the only language in the Xingu to have ejective consonants (these sounds are not attested in other languages in the area; cf. Emmerich 1980; Fargetti 1992; Seki 2000; Dourado 2001), and they may be in the process of being lost, since young people trying to learn the language in school are not always able to make the distinction between plain and ejective consonants.

1 A note should be made here: in Dixon and Aikhenvald (1999: 352), the chart of Trumai consonants is presented. However, the chart in that article is slightly distorted due to formatting problems. The sounds /d/, /n/, /l/, /ɬ/ and /ɾ/ appear there as dental, but they are actually alveolar. With regard to /s/ and /ts/, they are indeed dental (/ts/ is phonetically represented as [t ̪s̪]. At the phonological level, the dental diacritic is not relevant since there is no dental-alveolar distinction for affricates).

1080

Raquel Guirardello-Damian

Tab. 24.1: Consonants bilab voiceless stops

plain

labiodent

p

ejective

dent

alv

t̪ t ̪’

voiced stop m

fricatives

liquids

glides

pal

vel

glot

t

k

ʔ

t’

k’

d

nasals

affricates

postalv

n f

s

plain

ts

ejective

ts’

ʃ

plain

l

lateral fricative

ɬ

tap

ɾ w

x

h

j

Stress is fixed, always falling on the final syllable of the word. There are four syllable types: CV, CVC, V, and VC. Most segments are allowed in onset and codas positions, with no restrictions with regard to the position in the word (i.e., as the initial, medial, or final sound of a word). The only segments that present restrictions are: /ɾ/, which is allowed in onset and coda positions (e.g., mu.ɾiɾ ‘basket’) but not as the initial sound of a word; /h/, which is also allowed in onset and coda positions (e.g., ho.tet ̪ ̪ ‘corn’; ʃah.ni.ni ‘early evening’) but not found as the last sound of a word; and /ʔ/, which is allowed in codas word-internally (e.g., kaʔ.ʃɨ ‘walk’ and piʔ.ta̪ ‘step’) but never attested as the initial or last sound of a word. It is possible to find a sequence of two vowels in the same phonological word, but they belong to different syllables. Sequences of three vowels are avoided; when affixation generates such a sequence, one vowel is dropped: pita̪ + =ea > pitaa ̪ ‘her exit’. Geminate consonants are not allowed; when a sequence of two identical consonants is generated, one is dropped, for example, t ̪’ak + =ki > t’aki ‘manioc ̪ bread’ + dative. Like other Amazonian languages, Trumai has reduplication, which can have two forms: partial, as in (1) and (2), or full, as in (3). It occurs mainly with verbs and adjectives, although it can also be observed with some nouns and minor word classes (adverbs and numerals). As with Tupí-Guaraní languages (Rose 2005), partial reduplication in Trumai involves the copying of one or two light (monomoraic) syllables. However, while in Tupí-Guaraní languages the final syllables of the stem are copied, in Trumai the initial syllables are the ones concerned. Another difference is that in Tupí-Guaraní languages, the same verbal stem can have one or two syllables copied, depending on the meaning expressed (one syllable for a sense of

24 Trumai

1081

successive action, and two syllables for the sense of frequentative action). In Trumai it is not the meaning but rather the size of the stem that matters for copying: one light syllable for monosyllabic stems, as in (1), or two light syllables for stems with two syllables or more, as in (2); the reduplicated material is prefixed to the stem, forming a single phonological form. Full reduplication takes place with a large number of lexical items. In contrast, partial reduplication is more limited, and only occurs under particular conditions: (i) monosyllabic stems that have a heavy syllable with a CVC pattern, as in (1); (ii) stems with two syllables or more that have a heavy last syllable preceded by light syllables (L-H or L-L-H), as in (2). Partial reduplication is productive with disyllabic stems, but not with monosyllabic and trisyllabic ones – some stems undergo it, as in (1) and (2b), while others undergo full reduplication instead, as in (3). Reduplication in Trumai is used to convey various kinds of meanings, such as repeated action (iterative), action performed by many individuals, intensity, or a large quantity of something (Guirardello-Damian 2014). (1)

pi~piɾ

‘sweep many times’ [piɾ = CVC]

(2)

a. laxo~laxod b. ɨɾɨ~ɨɾɨɾɨw

‘sniff many times’ ‘slip many times’

[laxod = CV.CVC] [ɨɾɨɾɨw = V.CV.CVC]

(3)

a. kis kis b. midoxos midoxos

‘steal many times’ ‘call many times’

[kis = CVC] [midoxos = CV.CV.CVC]

3 Word classes 3.1 Classes and inflection Trumai word classes are identified on morphosyntactic grounds – particularly on syntactic ones, given that the language does not have much morphology and words usually consist of only a single morpheme. That is, morphologically speaking, Trumai tends to be an isolating language. There are only a few inflectional affixes and a certain number of enclitics.2 Table 24.2 presents the bound morphemes that codify grammatical information (see also Table 24.3). There are four open classes: nouns, verbs, adjectives, and adverbs. The distinction between nouns and verbs is clear in terms of bound morphology, negator (nouns are modified by the negator anuk, while verbs are negated by tak), and syn̪ tactic behavior. Adjectives share some of the characteristics of nouns and verbs (e.g., they can be modified by both negators), but at the same time they present 2 For the classification of a morpheme as an affix or a clitic, I use the criteria proposed by Zwicky (1989).

1082

Raquel Guirardello-Damian

Tab. 24.2: Grammatical bound morphemes. morpheme

function

tsi- ~ t-̪ =ake ~ =(e)a =(V)k =ts =(V)ɬ ~ =ki ~ =(V)s =k(a) te̪ =(V)n tsi=n ~ =e wa-

third-person possessive (kinship) third-person possessive (body parts) ergative ergative (for first-person singular) dative (one of three markers, depending on the head of the NP) genitive locative3 derives predicative adjectives from modifying ones third-person absolutive (verbal marker) middle voice

features that are unique to them. With regard to adverbs, apart from the prefix te-, ̪ an adverbializer that derives manner adverbs from body part nouns, there is no special morphology, thus adverbs are identified mainly by their syntactic behavior. Closed classes consist of pronouns and demonstratives, which are subclasses of nouns; interrogative words; numerals and other quantifiers; pluralizers, that is, dual and plural; postpositions, auxiliaries; particles such as for imperative, negation, causation, relativization, focus, etc; subordinators and discursive connectors; interjections; and ideophones, which constitute a class of their own. There is also a copula ʃɨ, which has its own nature, that is, it is not a regular intransitive verb, and the morpheme (i)ji, which has also a special nature (cf. Section 4). For nouns there are two important divisions. The first is between alienably and inalienably possessed nouns. With alienable possession, the possessor bears a genitive marker (possr-gen possd); with inalienable possession the possessor bears no special morphology (possr possd). There are also third-person anaphoric inalienable possessors for kinship and body part terms; (cf. Table 24.2). The distinction is further observed in the syntax. There are differences in the configuration of possessive predicates and in noun incorporation: terms for body parts are incorporated in transitive clauses and attributive predicates (Guirardello-Damian 2005: 225–226). The second division is between animate and inanimate nouns. The differences are manifested in the use of imperative particles (Guirardello 1999: 111–115), dative markers (Guirardello 1999: 270–282), posture auxiliaries (Guirardello-Damian 2002), demonstrative pronouns, and pluralizers (see Section 4). Verbs can be subdivided into five classes. Type 1 are intransitive, with an absolutive argument. Type 2 are transitive, with arguments marked as ergative and abso-

3 The morpheme =(V)n is the general locative marker. When speakers want to further specify the location, they employ some body parts in combination with =(V)n, e.g., X natu=n ‘on top of X’ (natu ‘back’), X fax=on ‘in X’ (fax ‘belly’), X daʃa=n ‘behind X’ (daʃa ‘spine’).

24 Trumai

1083

lutive. Type 3 are ditransitives, with ergative, absolutive, and dative arguments. Type 4 are intransitive, but with two positions, arguments marked as absolutive and dative. Type 5 are verbs with variable transitivity, their arguments usually marked as ergative-absolutive, but, when the second participant is inanimate and predictable, the case-marking changes to absolutive-dative. The verbs of type 4, which in Dixon’s terminology would also be called “extended intransitives”, exhibit two arguments, but in terms of transitivity they align with simple intransitives. The class includes verbs of perception (huʔtsa ‘see’, faʔtsa ‘listen’), mental activity (faxla ‘think’, falkamu ‘believe’, pudits’ ‘like’), contact (api ‘grab’, piʔta ‘step’), and habitual events (ma ‘eat’, sone ‘drink’, olem ‘cook’, oɬe ‘bake’, etc). At first sight, one might think that clauses with these verbs are instances of antipassive constructions, but they are not. They are the normal and only possible way of codifying the arguments of these verbs.4 Type 5 verbs constitute a rather small class in comparison to the other verb classes. Only a few verbs allow such variation in case marking.

3.2 Derivation and compounding Trumai derivational morphology includes that listed in Table 24.3. Compounding is used for the creation of new lexical items as well. Noun compounds are prevalent and there are various kinds, including N+N compounds (4), N+adj compounds (5), N+special-modifier compounds, as in (6) – which contains (a)nehene, a classifier found in terms for animals –, N+V+nominalizer compounds (7), and N+adj+relativizer compounds (8). Compounds can also be formed with a relativized verb plus its complement (9). The marking of the complement will vary depending on the type of verb. Note that in compounds, the relativizer has a tendency to cliticize (becoming =k) when the previous word ends in a vowel.

Tab. 24.3: Derivational morphemes. morpheme

function

=t’(a) ̪

exocentric nominalizer (nominal, verbal, adjectival bases), with various possible meanings, e.g., ‘former entity’, ‘former attribute’, ‘result of an action’, and ‘place where an action happened’

=kwaʃ ~ =kewʃa

exocentric nominalizer (verbal bases or VPs), with the sense of ‘the tool used for X’

-i

endocentric nominalizer deriving names of trees from their fruits

te̪

denominal adverbializer (body part bases)

4 See Guirardello-Damian (2010: 209–215) for further details.

1084

Raquel Guirardello-Damian

(4)

a. ateɬa iʔan ̪ sun simulacrum ‘clock’

b. axos mut ̪ child packing/wrapping ‘womb’ (Lit. ‘the wrapping of the child’)

(5)

a. huɾa au bird blue ‘name of a bird’

b. feʔde pat jaguar small ‘cat’

(6)

kupiana nehene Green-tailed.Jacamar similar ‘bird sp. (which looks like the Green-tailed Jacamar)’

(7)

a. huksi wen=kwaʃ eyelashes pull=tool.for ‘tweezer’

b. iʃa xoxan=kewʃa tooth wash=tool.for ‘toothbrush’

(8)

a. an mo=k breath smelly=rel ‘jaguar’ (archaic, Lit. ‘the one with smelly breath’)

b. iʃa naʃa=k tooth curved=rel ‘scorpion’ (Lit. ‘the one with a curved tooth’)

(9)

iʃa=s wen ke tooth=dat pull rel ‘dentist’ (Lit. ‘the one who pulls teeth’)

Trumai also has some verb compounds, although in smaller numbers. There are V+adv compounds (10), V+aux compounds (11), and N + V compounds (12). (10) ami makdits speak painfully ‘offend’ (11) alax tsula hunt be.lying ‘ambush’ (Lit. ‘hunt while lying down’) (12) a. kud tet ̪ ̪e head beautify ‘comb’

b. k’asa waman hammock turn ‘get married’ (Lit. ‘turn the hammock’5)

5 That is, in reference to the fact that people who get married often pack their hammocks and move to their in-laws’ house.

24 Trumai

1085

All these patterns are attested in the Trumai lexicon, but the one most used is N+adj+relativizer. Note that the internal word order observed in Trumai compounds correlates with the order observed in other contexts: adjectives follow nouns, auxiliaries follow verbs, and complement nouns precede transitive verbs (see Sections 4 and 5 below).

4 The noun phrase A noun phrase in Trumai has the internal order [determiner possessor NOUN adjective pluralizer ji], as exemplified in (13). Pluralizers are not used with nouns that refer to inanimate entities, as in (14). (13) huʃtahme hi adifle herohen wan ji ̪ three 2.pro sister beautiful pl ji ‘your three beautiful sisters’ (14) huʃtahme hi=kte̪ tahu dat’ ji ̪ ̪ three 2.pro=gen knife black ji ‘your three black knives’ The position of determiner can be occupied by numerals, quantifiers, and demonstratives. So far, there is no evidence to claim that demonstratives and numerals can co-occur in the same noun phrase. In naturally occurring data, there are no examples of noun phrases with two determiners (i.e., demonstrative and numeral). They do not appear in elicitation either – usually the consultant breaks the information into two juxtaposed noun phrases, as in (15). (15) [ kaʔnak wan ji ]NP [ huʃt ̪ahme kiki wan ji ]NP that pl ji three man pl ji ‘those three men’ (Lit. ‘those ones, the three men’) The possessor precedes the possessed noun. With alienable nouns, the possessor receives the genitive marker =k(a) t ̪e, which occurs at the end of the possessor noun phrase, as in (16). With inalienable nouns, the possessor presents no special morphology, for example, kumaru adifle ‘Kumaru’s sister’, but kumaru ato ‘Kumaru’s arm’. Inalienable nouns can also receive third-person anaphoric possessors. Kinship terms receive the prefix tsi- or t-,̪ which attaches to the head of the noun phrase (17).6 Body part terms receive the enclitic =ake or =ea, which occurs at the end (18).7

6 The allomorph tsi- occurs with C-initial roots, t ̪- with V-initial ones. 7 The allomorph =ake occurs with absolutive NPs, while =ea occurs elsewhere.

1086

Raquel Guirardello-Damian

(16) [di wan]NP=kate̪ tahu ̪ woman pl=gen knife ‘knife of the women’ (17) huʃahme t-adifle wan ji ̪ three 3poss-sister pl ji ‘his three sisters’ (18) a. [kuʃ dat’]NP=ake hair black=3poss ‘his black hair (absolutive)’

b. [kuʃ dat’]NP=ea letsi hair black=3poss ins ‘with his black hair’

The head of a noun phrase can be occupied by a noun, a personal pronoun, or a demonstrative pronoun. The personal pronouns for first- and second-person singular are ha and hi.8 When the noun phrase is not absolutive, the first-person singular pronoun is hai. These pronouns can combine with the dual and plural markers to produce other forms: ha a ‘we two’, ha wan ‘we (pl)’, hi a ‘you two’, hi wan ‘you (pl)’. There are also special plural forms to indicate inclusion of the listener: ka a ‘we two (incl)’, ka wan ‘we (pl.incl)’. While ha can occur by itself indicating first-person singular, ka only occurs in combination with the pluralizers. Strictly speaking, there are no third-person personal pronouns in Trumai. The forms that were analyzed as such in previous works (Guirardello 1992: 72–76; Guirardello 1999: 28–34) can actually be considered demonstrative pronouns. They occur as pronouns, but some of them can also be employed as modifiers of a noun, that is, [dem N]. Rather than a two-term demonstrative system as previously proposed, Trumai has a three-term system, expressing distance from the speaker, the addressee, or both.9 The three-way distinction is observed with locative adverbs as well: nina ‘here (close to the speaker)’, ina ‘there (close to the addressee)’, kaina ‘there (far from the speaker and addressee)’. The demonstrative pronouns also distinguish masculine and feminine gender, and animacy (humans, animals, objects). Like the personal pronouns, they can combine with the pluralizers to produce non-singular forms. Table 24.4 presents the set of demonstrative pronouns in Trumai. Note that when a demonstrative is used to refer to an inanimate entity, the forms ni and niʔde (and similarly, in and inde, kaʔin, and kaʔne) can freely substitute for each other. My hypothesis is that the inanimate set originally was simply ni, in, and kaʔin, but became more diverse with the spread of niʔde, inde, and kaʔne. This spread may have occurred because of contact with Portuguese (Guirardello-Damian 2018).

8 These two forms are also the corresponding possessive pronouns. 9 For more details on this new analysis, see Guirardello-Damian (2005: 242–245) and GuirardelloDamian (2018).

24 Trumai

1087

Tab. 24.4: Demonstratives Pronouns. speaker-prox

addr-prox

speaker-addr-dist

niʔde niʔdaɬ niʔdak a niʔdak wan

ine inaɬ inak a inak wan

kaʔne kaʔnaɬ kaʔnak a kaʔnak wan

animals

niʔde

ine

kaʔne

inanimates

ni niʔde

in inde ~ ine10

kaʔin kaʔne

humans

masc sg fem sg masc/fem du masc/fem pl

A head noun can be modified by an adjective, and usually only one adjective comes after it. Speakers do not employ sequences of multiple adjectives, seeming to prefer using relative clauses to further qualify the noun, for example, ‘my yellow hammock, which is big’. When the head noun refers to an animate entity, it can also be modified by the pluralizers wan plural and a dual (or ana dual in non-absolutive NPs). These morphemes have an interesting and unique behavior. When the head noun of an absolutive noun phrase is not present, the pluralizer stands by itself and the thirdperson absolutive enclitic =n/=e appears on the verb, as in (19b). This pattern may be linked to the origins of the pluralizers: it might be that originally they were nouns with the sense of ‘group of two’ and ‘group of more than two’, thus their capacity to stand alone as a noun phrase. This is plausible because Trumai has other nouns that refer to pluralities, such as xol ‘set (of objects)’ and paine ‘group (with all the individuals)’, as in kujpew xol ‘set of spoons’, or axos paine ‘group of all the children (of the village)’. However, xol and paine never appear by themselves and do not cooccur with the third-person absolutive enclitics, as the pluralizers do. Thus, even if the pluralizers were originally nouns, now they have a more grammatical nature. (19) a. [axos wan]NP ain child pl play ‘The children play.’

b. [wan]NP ain=e pl play=3abs ‘They play.’

Finally, in third-person noun phrases, the morpheme (i)ji can appear as its rightmost element. The function of this morpheme is not totally clear, but apparently it conveys pragmatic information, and requires further investigation.11 When there is some element in the noun phrase (i.e., a noun, pronoun, or pluralizer), the reduced form ji is employed; when there is no head or other elements in the NP, we have the

10 This is an alternation that varies between speakers. 11 For more details see Guirardello-Damian (2005: 281–285).

1088

Raquel Guirardello-Damian

full form iji. The reduced form can occur in all types of noun phrases (ergative, absolutive, dative, and genitive), while the full form only occurs in absolutive noun phrases. (20) a. [axos ji]NP watkan child ji cry ‘The child cried.’

b. [>iji]NP watkan=e iji cry=3abs ‘He/she cried.’

A noun phrase can be followed by phrasal enclitics that indicate its grammatical role (see Table 24.2). It can also be followed by postpositions, which are free forms: allative ita; nik ‘without’; t’atske ̪ ablative lots’; instrumental letsi; comitative tam; ̪ ̪ ‘after’; xuik ‘near’. Below, we have some examples of Trumai postpositions. (21) a. Kumaɾu dat ̪ ita̪ Kumaru house/home all ‘towards Kumaru’s house’ c. ha aɬe tam ̪ 1.pro mother com ‘with my mother’

b. Brasília lots’ Brasília abl ‘from Brasília’ d. jakɨɾ nik salt without ‘without salt’

e. pike xuik house near ‘near the house’

5 The verb phrase 5.1 Main configuration The absolutive NP (S or P) occurs immediately before the verb. There are various pieces of syntactic evidence that this NP is part of the VP in Trumai. First, when the absolutive NP is not immediately preverbal – for instance, because it was fronted in order to be highlighted – a special particle (ke) appears after the verb, as in (22b); although similar in shape, this particle is distinct from the relativizer ke, which exhibits different behavior. Second, it is not possible to insert adverbs between the absolutive NP and the verb without triggering this extra morphology. And third, the focus/tense particles ka-in and ʃɨ-in also indicate that [SV] and [OV] form constituents, as in (23). These particles occur in second position, after elements that are clearly constituents, such as NPs or PPs. They cannot occur between words in a phrasal constituent, as in (24b), nor after a sequence of words that do not form a constituent (see below for more information on these particles).

24 Trumai

1089

(22) a. [Alaweɾu=k]A [kodeʃɨʃ ]O [disi]V Alaweru=erg snake beat ‘Alaweru beat the snake.’ b. [kodeʃɨʃ ]O hen [Alaweɾu=k]A [disi]V ke snake then Alaweru=erg beat ke ‘Then Alaweru beat the snake.’ (23) a. [[pet’ew] ̪ S [aʃɨkida]V]VP ka-in frog jump foc/tns ‘The frog is jumping.’ b. [[aɬat ̪ ]O [mapa ]V]VP ka-in [Kumaɾu=k]A clay.pot break foc/tns Kumaru=erg ‘Kumaru broke the pot.’ (24) a. [pike lots’]PP ka-in ha pita house abl foc/tns 1.pro go.out ‘I went out of the house.’ b. *pike ka-in lots’ ha pita. The ergative and dative NPs are outside the verbal phrase. The typical position of the ergative is to precede the VP, while the dative follows it: NPERG [ VP ] NPDAT. Internally, the verbal phrase has the configuration [NPABS V modifier]. The modifier can be the particle of intensity (jumane), the causative particle (ka), or auxiliaries of aspect-mood, body posture, or direction. Auxiliares and the causative particle can co-occur, with two attested orders: [V aux caus] or [V caus aux]. Auxiliares and the intensity particle can also co-occur, with one order only: [V aux intens]. The same can happen to the particles of intensity and causative, with the preferred order [V intens caus]. So far, there is no evidence that the three elements (aux, intens, caus) can co-occur. Adverbs can occur inside the VP as well, but more often they are found outside it, modifying the whole clause. There are six auxiliaries for body posture: la ‘be standing’, katsi ‘be sitting’, chumuchu ‘be lying’, tsula ‘be lying in a high place (not the floor)’, mula ‘be in a closed place’, pila ‘be in a liquid medium’. With regard to directional auxiliaries, they codify information about motion and path, but some also codify motion and path and manner, such as kahmi ‘go uphill, continuously’; lahmi ‘go uphill, with pauses’; iwda ‘go upriver, continuously’; kiwa ‘go upriver, with pauses’. The directionals are based on an absolute reference system, that is, they are not objectcentered or viewpoint-centered, but rather geo-centered, involving externally fixed directions. There is only one form that uses a relative reference: k(e)tsi, exemplified in (25), which codifies motion towards the place where the speaker is. There is an

1090

Raquel Guirardello-Damian

extensive set of directionals, some of which cliticize to the verb (Guirardello-Damian 2012). Internal reconstruction suggests that they were originally motion verbs in serial constructions which were reanalyzed as auxiliaries. (25) wan wal lako=ktsi=n hen pl sing dir(downhill)=dir(towards.speaker)=3abs then ‘They come down (to the village) singing.’ (26) hai=ts ha mut ̪ xoxan katsu 1.pro=erg 1.pro clothing/wrapping wash dir(towards.river) ‘I went to the river to wash my clothes.’ The verb does not receive person affixes. The only person marker is the enclitic =n/=e, which attaches to the last element of the VP and occurs when the absolutive NP lacks a head noun, as in (27b), (28b), and (29b). The allomorph =n is used with words ending in vowels, =e with words ending in consonants. (27) a. [pɨtɨk]NP oɾa monkey scream ‘The monkey screams.’

b. [

]NP oɾa=n scream=3abs ‘It screams.’

(28) a. [di wan]NP sa woman pl dance ‘The women dance.’

b. [

(29) a. hai=ts [hi adif]NP midoxos 1.pro=erg 2.pro brother call ‘I called your brother.’

b. hai=ts [ 1.pro=erg ‘I called him.’

wan]NP sa=n pl dance=3abs ‘They (the women) dance.’ ]NP midoxos=e call=3abs

5.2 Temporal reference, aspect and modality Trumai does not have affixes on the verb that express tense. Temporal reference is expressed via other elements at the sentence level: adverbs (e.g., ka’neʃaj ‘yesterday’, xodakaki ‘tomorrow’, hat’ke ̪ ‘futurely’), or the particles ka-in and ʃɨ-in. These two particles have a very interesting nature. The combinations ka + in or ʃɨ + in do not form a single morpheme, because some adverbs can be inserted between them (e.g., ka + de + in; ka + nuk + in). On the other hand, they always work together, almost as if they were a lexicalized unit; given these facts, I represent them using a hyphen, in order to indicate that they contain internally different elements. Ka-in and ʃɨ-in often occur in answers to questions, indicating new information (in other words, they indicate focus), but at the same time they also convey information

24 Trumai

1091

about the time when the event is taking place: ka-in indicates that the event is happening now or in the recent past, ʃɨ-in that it happened in the past.12 (30) misu=s ka-in ha sone water=dat foc/tns 1.pro drink ‘I am drinking water’. (answer to the question ‘What are you drinking?’) When a clause does not contain adverbs or these particles, temporal reference is inferred from the context (in a text, at least one clause will have some overt temporal information). Aspect and modality are expressed via auxiliaries: desiderative t(a)ke; perfec̪ tive k(u)ma; prospective laketsi; inchoative napta; ̪ delayed inchoative ʃ(ɨ)ketsi; habitual kawala; hup ‘be able’; and hupma ‘get used to’, as in (31). Imperfectivity can be conveyed through the use of huk’an ‘still’, as in (32), which is not an auxiliary, but rather an adverb (unlike auxiliaries, huk’an can occur in various positions in the sentence, which is a typical feature of Trumai adverbs). (31) a. ha katnon ta̪ke ̪ 1.pro work desid ‘I want to work.’ c. ha katnon napta̪ ̪ 1.pro work inch ‘I am beginning to work.’

b. ha kat ̪non kuma 1.pro work pfv ‘I have finished working.’ d. ha kat ̪non ʃɨketsi 1.pro work del.inch ‘I am beginning to work (after waiting for it).’

e. wal kawala=n hai=ɬ sing habit=3abs 1.pro=dat ‘He has the habit of singing to me.’ (32) ha ma huk’an 1.pro eat still ‘I am still eating.’ There are no epistemic modal verbs in Trumai. To express doubts or certainty regarding how likely a proposition is to be true, speakers can use adverbs such as huka ‘perhaps’, hat’ke ̪ ‘in future (certainly)’, ifke ‘in future (less certain)’, etc. There are no deontic modal verbs either. In order to express the notions of obligation (must), permission (may), or recommended behavior (should), Trumai speakers employ ordinary sentences and deontic meanings are inferred from context.

12 See Guirardello (1999: 169–178) for the origin and internal configuration of these particles.

1092

Raquel Guirardello-Damian

(33) ajej maʔtsi ka-in, jenuk tsiets oke=s sone=n alokeloke grandpa sick foc/tmp thus cnsq medicine=dat drink=3abs soon ‘Grandpa is sick, thus he should take medicine soon.’ (Lit. ‘Grandpa is sick, thus he drinks medicine soon’) (34) hi esa tenuk, lapts’a tak t ̪enuk hi ʃɨ ̪ ̪ 2.pro dance in.this.case stand.without.moving neg in.this.case 2.pro cop ‘You must dance, you should not stand without moving.’ (Lit. ‘In this case you dance, you do not stand without moving’)

5.3 Valency Valency-increasing is done through the use of the causative particle ka, which can modify any kind of verb. In the causativization of intransitive verbs, the causee is marked as absolutive and the causer as ergative, as in (35a). In the causativization of transitive verbs, the causer is marked as ergative as well as the causee, and the causer always precede the causee, as in (35b).13 (35) a. Kumaɾu=k Atawaka sa ka Kumaru=erg Atawaka dance caus ‘Kumaru made Atawaka dance.’ b. Kumaɾu=k Atawaka=k aɬat ̪ mapa ka Kumaru=erg Atawaka=erg clay.pot break caus ‘Kumaru made Atawaka break the pot.’ Valency-decreasing is done via argument suppression: the argument is omitted and is not recoverable by context, that is, it is not ‘pro drop’. The suppression of an argument can produce various semantic effects depending on the type of argument, such as a passive reading, as in (36b), an antipassive reading, as in (37b), or even a reflexive meaning, as in (38b). If Trumai speakers want to convey clearly the reflexive interpretation, the word falapetsi ‘do by oneself’ is employed in combination with the argument suppression, as in (38c). (36) a. Kumaɾu=k ha hotaka Kumaru=erg 1.pro deceive ‘Kumaru deceived me.’ b. [

]A ha hotaka de 1.pro deceive already ‘I was deceived (by somebody, not mentioned).’

13 For more details, see Guirardello (1999: 301–320).

24 Trumai

1093

(37) a. axos=ak ka-in kaɾakaɾako husa husa child=erg foc/tns chicken tie tie ‘The boy is tying the chicken.’ b. axos=ak ka-in [ ]O husa husa child=erg foc/tns tie tie ‘The boy is tying (something; it does not matter what, he is just playing).’ (38) a. Kumaɾu=k ha tɨʃɨ Kumaru=erg 1.pro scarify ‘Kumaru scarified me.’ tɨʃɨ ]A ha 1.pro scarify ‘I was scarified (by somebody).’ ~ ‘I scarified myself.’

b. [

c. ha falapetsi letsi ka-in [ ]A ha tɨʃɨ 1pro do.alone ins foc/tns 1pro scarify ‘I scarified myself.’ (Lit. ‘I did it alone, I scarified (myself)’)

6 Simple clauses 6.1 Main features The two main types of simple clauses are those involving nominal predication and those involving verbal predication. For purposes of comparison, we focus on the differences between clauses with nominal predicates and those with intransitive verbs, finding that, in the former case, the default order is [pred S cop], as in (39), while in clauses with verbal predicates the order is [S pred], as in (40). As discussed in Sections 6.2 and 6.3, there are also differences in the negation and imperative particles used in each clause type. (39) ha pine hi ʃɨ 1.pro friend 2.pro cop ‘You are my friend.’ (40) ha katnon de ̪ 1.pro work already ‘I have already worked.’ Looking more closely at verbal predicates, the default orders are [S V] for intransitive clauses, [A O V] for transitives, and [A O V R] for ditransitives. For clauses with

1094

Raquel Guirardello-Damian

verbs of type 4 (extended intransitives), the second participant comes after the verb: S V NPDAT. These are the pragmatically neutral orders, but they can undergo changes, given that in Trumai an element that is highlighted comes in first position in the clause.

6.2 Negative clauses Negative clauses are formed with two particles: anuk for nominal predicates, as in (41), and tak ̪ for verbal predicates, as in (42). (41) ha pine anuk hi ʃɨ 1.pro friend neg 2.pro cop ‘You are not my friend.’ (42) ha katnon tak ̪ ̪ 1.pro work neg ‘I have not worked. Attributive predicate clauses, which contain adjectives, also employ these particles, but in a different way. This predicate allows both word orders: [S pred], when one makes a statement (e.g., ‘The coffee is bitter’), and [pred S cop], when one replies to a question (e.g., ‘Is the coffee sweet? No, the coffee is bitter’). When the order is [pred S cop], the negator used is t ̪ak, as in (43a), and when the order is [S pred], the negator is anuk, as in (43b). In other words, when the clause exhibits word order typical of nominal predicate clauses, a “verbal” negator is used, and vice-versa. (43) a. xeɾeɾe tak ha ʃɨ ̪ wet neg 1.pro cop ‘I am not wet.’

b. ha xeɾeɾe anuk 1.pro wet neg ‘I am not wet.’

The use of anuk is possible with the order [pred S cop], but in this case the meaning of the clause changes, indicating a more permanent attribute. For further information about attributive predicates, see Guirardello-Damian (2005: 290–293). If the verb is accompanied by an auxiliary or the intensity particle, the negator comes after them, as in (44). Apparently, there is some incompatibility between the negation particle and the marker =n/=e 3abs: when the verb is modified by t ̪ak, the third-person absolutive marker is not used, as in (45b). The particle t ̪ak is also employed for negation of adverbs and quantifiers, as in (46). (44) a. ha joɬ take tak ̪ ̪ 1.pro sleep desid neg ‘I do not want to sleep.’

b. ha joɬ jumane tak ̪ 1.pro sleep ints neg ‘I did not sleep much.’

24 Trumai

(45) a. huʔtsa=n hi=ɬ see=3abs 2.pro=dat ‘He saw you.’

1095

b. huʔtsa t ̪ak hi=ɬ see neg 2.pro=dat ‘He did not see you.’

(46) kometani tak ka-in hi ami ̪ ̪ slow neg foc/tns 2.pro speak ‘You are speaking fast.’ (Lit. ‘You are speaking not-slow’) Trumai exhibits no dedicated negative indefinites, such as English nobody, nothing and never, but it does exhibit ways of expressing these notions. The sense of nobody can be achieved by the negation of the verb plus the use of the noun jaw ‘people, human beings’ (i.e., ‘I did not see people’ = ‘I saw nobody’). The idea of nothing can be expressed by the negation of the verb plus the suppression of the patient argument (i.e., ‘I did not drink Ø’ = ‘I drank nothing’). And the sense of never is produced by the use of the privative morpheme =t ̪(a)ke, which is homophonous with the desiderative but is a different morpheme. The privative is usually employed with nominal roots, indicating that somebody is deprived of a particular element, as in (49a–b), but it can also modify verbal roots generating the sense of never, as in (49c). (47) ha huʔtsa tak jaw=as ̪ 1.pro see neg people=dat ‘I saw nobody.’ (48) ha sone tak [ ̪ 1.pro drink neg ‘I drank nothing.’

]PAT

(49) a. aɬe=tke=k ̪ mother=priv=rel ‘orphan’ (Lit. ‘the one deprived of mother’) b. faxdo=tke=k ̪ inner.ear=priv=rel ‘deaf person’ (Lit. ‘the one deprived of inner ear’) c. ʃɨkida=tke ke ha ʃɨ ̪ travel=priv rel 1.pro cop ‘I have never traveled.’ (Lit. ‘I am one deprived of traveling’) As a final note, it is worth mentioning that negative clauses in Trumai can present some variations in word order, depending on whether the clause is a simple statement or an answer to a question. When the negated verb comes in the first position

1096

Raquel Guirardello-Damian

of the sentence, the clause presents a nominal-predicate configuration (Pred S Cop), as in (50b). This is an interesting phenomenon, but further investigation is necessary in order to understand why the changes in word order trigger such a configuration. (50) a. ha sa tak ̪ 1.pro dance neg ‘I did not dance.’ (simple statement) b. sa tak ʃɨ-in ha ʃɨ ̪ dance neg foc/tns 1.pro cop ‘I did not dance.’ (answer to the question: Did you dance?)

6.3 Imperative clauses For nominal predicates, the imperative is formed with the particle wanaʃ, which comes after the predicate (pred imp), as in (51). For verbal predicates, intransitive verbs and verbs of type 4 employ the particle wana, as in (52), while transitive and ditransitive clauses use the particles wa and waki, selected according to the type of NP that occurs in the O role (wa for animates, waki for inanimates), as in (53); these particles precede the verb (imp pred). (51) ha pine wanaʃ ! 1.pro friend imp ‘Be my friend!’ (52) a. wana kat ̪non! imp work ‘Work!’

b. wana faʔtsa! imp listen ‘Listen!’

(53) a. waki husa! imp tie ‘Tie it (thing)!’

b. wa husa! imp tie ‘Tie him (animal, people)!’

Attributive predicates allow both the particles wanaʃ and wana, with the same patterns in the word order: wanaʃ follows the predicate; wana precedes it, as in (54). The semantic difference between these two uses is unclear and needs further investigation. (54) a. taxa wanaʃ ! hard imp ‘Become hard!’

b. wana taxa! imp hard ‘Become hard!’

24 Trumai

1097

6.4 Interrogative clauses Polar questions involve the use of the particle =a, which cliticizes to the last word of the constituent under interrogation, as in (55). Wh-questions require an interrogative word (such as te̪ ‘who’ or hamuna ‘where’), which comes in the first position of the clause and is followed by the focus particle in, as in (56). (55) a. [hi joɬ take] ̪ VP=a? 1.pro sleep desid=q ‘Do you want to sleep?’

b. [ni ]AdvP=a de hi ʃɨ? here=q already 2.pro cop ‘Are you here?’

(56) hamuna in hi joɬ take? ̪ where foc 2.pro sleep desid ‘Where do you want to sleep?’

6.5 Case system Trumai exhibits both morphological and syntactic ergative-absolutive alignment; in other words, this language has deep ergativity. In morphology, the absolutive (S or O argument) is unmarked (-Ø). As previously mentioned, when a third-person S or O argument is not overt due to givenness, the verb receives the third-person enclitic =n/=e. The ergative (A argument) is marked by =k, and the dative (R argument) is marked by =(V)ɬ, =ki, or =(V)s, depending on the characteristics of the head of the NP (Guirardello 1999: 270–282), as summarized in Table 24.5. Ergative-absolutive alignment is manifested syntactically in a number of ways, including constituent order, choice of relativizer in relative clauses, reflexivization strategies, and raising in complement clauses. Beginning with constituent order, S and O typically occur inside the verb phrase, right before the verb; if they are not in their typical position, extra morphology is employed. In contrast, A and R typically occur outside the verb phrase (A precedes the verb phrase, R follows it), and if they change position, no extra morphology appears (see Section 5). With regard to relative clauses, when the NP that is relativized is S or O, the relativizer employed is ke, while when it is A or R, the relativizer is ʃɨk (see Section 7.4 below). In the case of reflexive clauses, as mentioned in Section 5, reflexive events can be expressed in

Tab. 24.5: Argument marking. clause type

marking

intransitive transitive ditransitive

A-k A-k

S-ø O-ø O-ø

V V V

R-ɬ / -ki / -s

1098

Raquel Guirardello-Damian

Trumai by using the word falapetsi ‘do by oneself’ in combination with argument suppression. What is interesting to observe in the reflexive clause is that S or O is always the argument preserved, while A or R can be suppressed (Guirardello-Damian 2010: 221–225). Finally, regarding raising in complement clauses: in one type of subordinate clause, when the S or O argument is not overt, the third-person enclitic =n/=e is employed, but instead of coming at the end of the subordinate VP, it is raised to the main clause (see Section 7.2). In addition to the above manifestations of ergative-absolutive alignment in the syntax, Trumai also exhibits a nominative-accusative alignment in the distribution of posture auxiliaries, which always refer to the posture of the S or A argument.14 And in the selection of verb types, the language is also showing signs of moving in a more nominative-accusative direction: some semantic events can be expressed by two possible verbs, one belonging to class 2 (erg abs V), the other one to class 4 (abs V dat). For instance, in order to express ‘bite’, one can use tako (class 2) or ̪ make (class 4), as in (57). These pairs of verbs were quite numerous in the past according to Monod-Becquelin (personal communication, 2004), who collected data in the 1960s, but nowadays only a few are still observed (Guirardello-Damian 2010: 215–217). Several verbs of class 2 fell out of use and only their equivalent of class 4 remained; in other words, the language favored the construction [abs V dat], which places the agent of the action in an absolutive position, and the second participant in a non-absolutive role – a pattern that is more like an “accusative” alignment. The existence of verb class 5 (described previously in Section 3.1) could be another sign of a move towards accusativity. It might be that these verbs originally belonged to class 2, presenting only the ergative-absolutive marking, but later developed the absolutive-dative marking as well. (57) a. kasoɾo=k axos tako ̪ dog=erg child bite ‘The dog bit the child.’

b. kasoɾo make axos=aɬ dog bite child=dat ‘The dog bit the child.’

With regard to the control of coreference, we do not observe any particular alignment, because in Trumai this control is not syntactic, but pragmatic (GuirardelloDamian 2010: 219–220).

14 For the posture of O or R arguments, another construction is employed. See Guirardello-Damian 2010: 217–218.

24 Trumai

1099

7 Clause-linking and discourse 7.1 Coordinated clauses Clauses can be coordinated by using two main strategies: (i) juxtaposition plus socalled comma intonation, that is, the clauses are under the same intonational contour; (ii) juxtaposition plus use of discursive connectors (inis, in lots’, in t’atske). ̪ The connectors, which can be broadly translated as ‘and then’ or ‘after this’, consist of a combination of the demonstrative pronoun in (which is used anaphorically, referring to the event mentioned in the previous clause) with other morphemes. The connector occurs in the first position of the second clause and is usually followed by the adverb hen ‘then’. (58) kodeʃɨʃ homne hai=ts, inis hen disi=n hai=ts snake find 1.pro=erg disc.con then beat=3abs 1.pro=erg ‘I found a snake, then I beat it.’

7.2 Complement clauses With regard to subordination, there are two main types of complement clauses: (i) dative complement clauses (the complement of verbs of type 4, such as huʔtsa ‘see’); and (ii) absolutive complement clauses (the complement of transitive verbs). They are embedded in the main clause, in the typical position of a dative or absolutive argument, and receive the expected case-marking: (59) S V [subordinate clause]=dat (60) A [subordinate clause] V The verb in a dative complement clause does not bear the enclitic =n/=e 3abs, but rather the third-person possessive marker =(e)a; in other words, it behaves as a possessed noun, as in (61). The verb in an absolutive complement clause does not receive a possessive marker, but it does not receive the third-person absolutive enclitic either, which is raised to the main verb. The raising is clearly observed in absolutive complement clauses that contain an intransitive verb, as in (62b). With subordinate clauses with a transitive verb, it is possible to find the raising in data obtained through elicitation, as in (63), but not in natural speech; in natural uses, speakers seem to prefer to use coordination (e.g., ‘Kumaru was scarifying her and I waited’) or a temporal subordinate clause instead (e.g., ‘When Kumaru was scarifying her, I waited’). It could be that they avoid examples like (63) because the A argument (haits) is too far from its associated verb (padi), which would make the sentence hard to process and understand.

1100

Raquel Guirardello-Damian

(61) a. ha huʔtsa [Koinu sa]CompCl=ki. 1.pro see Koinu dance=dat ‘I saw Koinu dancing.’ b. ha huʔtsa [sa=a]CompCl=ɬ 1.pro see dance=3poss=dat ‘I saw her dancing.’ (62) a. hai=ts [Sula huma]CompCl padi 1.pro=erg Sula take.bath wait ‘I waited for Sula to take a bath.’ b. hai=ts [huma]CompCl padi=n 1.pro=erg take.bath wait=3abs ‘I waited for her to take a bath.’ (63) hai=ts ʃɨ-in [Kumaɾui=k tɨʃɨ ]CompCl padi=nj 1.pro=erg foc/tns Kumaru=erg scarify wait=3abs ‘I waited for Kumarui to scarify herj.’

7.3 Adverbial clauses Adverbial clauses are not embedded in the main clause. Temporal/conditional clauses precede the main clause, as in (64). Other types follow the main clause, such as purpose clauses, as in (65), or consequence clauses, as in (66). In an adverbial clause, the subordinate verb is finite (i.e., it can bear the 3abs enclitic) and a subordinator is used. The subordinator comes at the end of the VP of the subordinate clause, as in (64) and (65), except in the case of tsiets’, which comes at the beginning of the subordinate clause in combination with the adverb jenuk ‘thus’, as in (66). axos watkan (64) a. [ha ma=s]AdvCl ʃɨ-in 1.pro eat=tmp foc/tns child cry ‘When I was eating, the child cried.’

Tab. 24.6: Subordinators used in adverbial clauses. subordinator

function

=(i)s ~ =(e)s (a)hak iets’ ~ =ak tsiets’

temporal or conditional clauses purpose clauses reason clauses consequence clauses

24 Trumai

1101

b. [ma=n=es]AdvCl ʃɨ-in axos watkan eat=3abs=tmp foc/tns child cry ‘When he was eating, the child cried.’ (65) a. hai=ts t’ak wawa=ktsi [hi ma hak]AdvCl ̪ 1.pro=erg manioc.bread carry=dir 2.pro eat purp ‘I brought manioc bread for you to eat.’ b. hai=ts t’ak wawa=ktsi [ma=n ahak]AdvCl ̪ 1.pro=erg manioc.bread carry=dir eat=3abs purp ‘I brought manioc bread for him to eat.’ (66) k’awixu=k ha disi [jenuk tsiets fade=k ha xol]AdvCl rain=erg 1.pro beat thus cnsq flu=erg 1.pro get ‘I got rained on, thus as a consequence I have a flu.’ (Lit. ‘The rain beat me, thus as a consequence flu got me.’) The clauses marked by =(i)s/=(e)s usually express a temporal sense, as in (64). The conditional sense can be produced when the sentence contains the adverbs hat’ke ̪ ‘in future (certainly)’ or ifke ‘in future (less certain)’, as in (67). Since future events are unrealized, they are more a possibility than a real fact, thus the conditional reading; as Thompson and Longacre (1985: 193) point out, the distinction between “when and if clauses is simply one of degree of expectability,” and not all languages codify it. With regard to conditionality in the past, such a sense is produced when the sentence contains the adverb kamʃɨk ‘potentially’, indicating that the event was merely a potential situation which did not materialize, as in (68). (67) fakdits=e=s ha watkan hat’ke ̪ die=3abs=tmp 1.pro cry in.future ‘If he dies, I will cry.’ (68) kaʔʃɨ=ktsi=n=es kamʃɨk ha wakadima walk=dir=3abs=tmp potentially 1.pro be.happy ‘If he came, I would be happy (but he didn’t come).’ There is another kind of temporal clause, but it has a very different shape. This other clause is used to express that an event is happening simultaneously with another one. It has a non-finite verb (i.e., the verb does not bear the third-person absolutive enclitic, but rather the third-person possessive marker =(e)a), which is followed by the comitative postposition tam. Thus to express the idea of ‘The child ̪ cried while taking a bath’, the Trumai sentence would be something like ‘With her bathing, the child cried’.

1102

Raquel Guirardello-Damian

(69) [huma=a tam]AdvCl axos watkan take.bath=3poss com child cried ‘The child cried while taking a bath.’

7.4 Relative clauses Taking into consideration Comrie’s (1989) NP accessibility hierarchy, Trumai does not seem to present restrictions, allowing the relativization of most noun phrase positions: subject (70a); direct object (70b); indirect object (70c); and location (70d). With regard to the relativization of possessors, at present it is not possible to make further claims, given that this issue has not been properly investigated. (70) a. di [Ø huma ke]RelCl woman take.bath rel ‘the woman who is bathing’ b. di [hai=ts Ø midoxos ke]RelCl woman 1.pro=erg call rel ‘the woman whom I am calling’ c. di [hai=ts kafe kɨtɨ ʃɨk Ø]RelCl woman 1.pro=erg coffee give rel ‘the woman to whom I give coffee’ d. esak [axos tsula ʃɨk Ø]RelCl hammock child lie rel ‘the hammock where the child lies’ A gap is left in the original position of the relativized NP, and the verb is followed by a relativizer: if NPREL is absolutive, the relativizer particle ke is employed, as in (71) and (72); if NPREL is non-absolutive, the verb is modified by ʃɨk, which is a combination of the copula plus a cliticized form of the relativizer, as in (73). (71) hai=ts huk’an di padi [Ø esa ke]RelCl 1.pro=erg still woman wait dance rel ‘I am waiting for the woman who is dancing.’ (72) hai=ts huk’an di padi [hai=ts Ø midoxos ke]RelCl 1.pro=erg still woman wait 1.pro=erg call rel ‘I am waiting for the woman whom I am calling.’ (73) hai=ts huk’an di padi [Ø mawmaw kuhmu ʃɨk ]RelCl 1.pro=erg still woman wait papaya throw rel ‘I am waiting for the woman who is throwing the papaya away.’

24 Trumai

1103

The relative clause may or may not be immediately adjacent to its head, and may be postnominal, prenominal, or may come outside the main clause (74). Apparently, the prenominal position is only possible when the head noun is in the beginning of the main clause, either because it is S or A (i.e., it occupies its natural position), or because it is an O which was fronted. (74) a. The man [who arrived] slept. b. [Who arrived] the man slept. c. I am waiting for the woman now [who is dancing]. Note that in (74c), the head is not immediately adjacent to the relative clause, but it is still present in the sentence. However, it is also possible to have relative clauses that are completely headless, as in (75). (75) wal ke ji=ki ʃɨ-in ha huʔtsa sing rel ji=dat foc/tns 1.pro see ‘I saw the one who is singing.’ It seems that the role of the head NP in the main clause influences the position of the relative clause: when the head is S or A, the relative clause tends to follow it, as in (76a–b). In contrast, when the head is O, or the second argument of verbs of type 4, the relative clause tends to come outside the main clause, as in (76c) and (76d). (76) a. b. c. d.

[ S [relative clause] V ] [ A [relative clause] O V ] [ A O V ] [relative clause] [ S V NPDAT ] [relative clause]

These order variations need to be further investigated, but if this analysis is correct, the relative clauses would present an interesting scenario: with regard to the relativizer modifying the verb, we see an absolutive alignment. However, with regard to the position of the relative clause in relation to its head noun, we would have a nominative alignment, although this seems to be just a tendency, not a strict rule. In sum, with regard to subordination, Trumai clauses can have different degrees of integration in relation to the main clause. Complement clauses are very integrated, since they occur inside of the main clause. Adverbial clauses are less integrated, given that they precede or follow the main clause, but are not within it. And the position of the relative clause is flexible, since it can occur in the main clause or not.

1104

Raquel Guirardello-Damian

8 Conclusion Trumai is an Amazonian language that has several interesting features. In terms of sounds, it presents a set of ejective consonants, which are not found in other languages spoken in the Xingu area. Its reduplication processes share some characteristics with Tupi-Guarani languages, but there are differences as well. In terms of morphology, Trumai has very few inflectional morphemes, showing a tendency more towards being isolating (i.e., words with one morpheme only). The distinction between inalienably versus alienably possessed nouns is present in the language, as well as an animacy distinction. Verbs have five subclasses, and the class of auxiliaries could be grouped into two main types: aspect-mood and spatialorientation auxiliaries (with the subdivision of body posture and directionals.) The Trumai directionals are based on an absolute frame of reference. In syntax, the order of nouns and adpositions is harmonic with the verb-complement order found in transitive clauses (i.e., the noun precedes the adposition, and the complement precedes the transitive verb). The verb phrase contains the absolutive argument (S or O), while the ergative and dative arguments are external to it. Trumai is a language with deep ergativity, with ergative-absolutive patterns manifested in various constructions, but the syntax also presents a nominative-accusative pattern in the use of posture auxiliaries. Word order can undergo changes due to pragmatic reasons, and negative clauses present interesting configurations due to such changes. A number of features of Trumai still need further investigation to be fully understood, such as the nature of the morpheme (i)ji and its semantic aspects; the occurrence of the prefix tsi- in attibutive predicates; certain changes in word order in verbal predicates; the use of modifiers in possessive predicates; and lexical choices, among other issues. These are topics to be explored in future studies.

9 Acknowledgments I would like to thank Patience Epps and Lev Michael for their detailed and constructive comments, and to the anonymous reviewer who carefully provided feedback to the article. I would also like to express my gratitude to the Trumai consultants who helped me along various years of research.

10 References Campbell, Lyle. 1988. Review of Language in the Americas by Joseph H. Greenberg. Language 64(3). 591–615. Comrie, Bernard. 1989. Language universals and linguistic typology. Chicago: University of Chicago Press.

24 Trumai

1105

Dixon, R. M. W. & Alexandra Y. Aikhenvald. 1999. The Amazonian languages. Cambridge: Cambridge University Press. Dourado, Luciana. 2001. Aspectos morfossintáticos da língua Panará (Jê). Campinas, Brazil: State University of Campinas (UNICAMP) PhD dissertation. Emmerich, Charlotte. 1980. A fonologia segmental da língua Txikão: Um exercício de análise. Rio de Janeiro: Departamento de Antropologia, Museu Nacional & UFRJ. Fargetti, Cristina M. 1992. Análise fonológica da língua Juruna. Campinas, Brazil: State University of Campinas (UNICAMP) MA thesis. Greenberg, Joseph. 1987. Languages in the Americas. Stanford: Stanford University Press. Guirardello, Raquel. 1992. Aspectos da morfossintaxe da língua Trumai e de seu sistema de marcação de caso. Campinas, Brazil: State University of Campinas (UNICAMP) MA thesis. Guirardello, Raquel. 1999. A Reference Grammar of Trumai. Houston: Rice University PhD dissertation. Guirardello-Damian, Raquel. 2005. Fonologia, classes de palavras e tipos de predicado em Trumai. Boletim do Museu Paraense Emílio Goeldi (Ciências Humanas) 1(2). 193–306. Guirardello-Damian, Raquel. 2012. Um estudo sobre o léxico Trumai: Verbos e auxiliares de movimento. In Cristina Martins Fargetti (ed.), Abordagens sobre o léxico em línguas indígenas, 171–195. Campinas: Editora Curt Nimuendajú. Guirardello-Damian, Raquel. 2014. Reduplication and ideophones in Trumai. In Gale G. Gómez & Hein van der Voort (eds.), Reduplication in Indigenous Languages of South America, 217–246. Leiden: Brill. Guirardello-Damian, Raquel. 2018. Trumai: Non-contrastive exophoric uses of demonstratives. In Stephen C. Levinson, Sarah Cutfield, Michael J. Dunn, N. J. Enfield & Sérgio Meira (eds.), Demonstratives in cross-linguistic perspective, 242–256. Cambridge: Cambridge University Press. Kaufman, Terrence S. 1990. Language history in South America: What we know and how to know more. In Doris L. Payne (ed.), Amazonian linguistics: Studies in lowland South American languages, 13–73. Austin: University of Texas Press. Lewis, M. Paul (ed.). 2009. Ethnologue: Languages of the world, edn. 16. Dallas: SIL International. Monod-Becquelin, Aurore. 1970. Multilinguisme des Indiens Trumai du Haut-Xingu (Brésil central). Languages. 78–94. Monod-Becquelin, Aurore. 1975. La practique linguistique des Indiens Trumai. Paris: Selaf. Monod-Becquelin, Aurore. 1976. Classes verbales et construction ergative in Trumai. Amerindia 1. 117–143. Murphy, Robert F. & Buell Quain. 1955. The Trumai Indians of central Brazil. Seattle: University of Washington Press. Rankin, Robert L. 1992. Review of Language in the Americas by Joseph H. Greenberg. International Journal of American Linguistics 58(3). 324–351. Rose, Françoise. 2005. Reduplication in Tupi-Guarani languages: Going into opposite directions. In Bernard Hurch (ed.), Studies in Reduplication, 351–368. Berlin & New York: Mouton de Gruyter. Thompson, Sandra & Robert E. Longacre. 1985. Adverbial clauses. In Timothy Shopen (ed.), Language typology and syntactic description 2: Complex constructions, 171–234. Cambridge: Cambridge University Press. Von den Steinen, Karl. 1940. Entre os aborígenes do Brasil central. Separata da Revista do Arquivo Municipal, vols. XXXIV a LVIII. São Paulo: Departamento de Cultura. Zwicky, Arnold M. 1989. Clitics and Particles. Language 61(2). 283–305.

Knut J. Olawsky

25 Urarina 1 2 3 4 5 6 7 8 9

Classification and background Phonology Word classes Noun phrase structure Verb phrase structure Simple clauses Multi-clause constructions Summary References

1 Classification and background Urarina (urar1246) is an isolate spoken by fewer than 3,000 people living around the Río Chambira and its tributaries in the Urarinas District in northwestern Peru’s Loreto department. The language exhibits a range of unusual grammatical characteristics that are rare or absent in other languages, such as OVA constituent order. Urarina was originally also spoken between the Urituyacu and Corrientes Rivers, but language shift has contributed to Spanish dominance on the Urituyacu River. While Urarina is still spoken on the Corrientes River, it is subject to increasing pressure from Spanish. Generally, in the more remote areas, Urarina is the first language of all speakers, including children. At the time of the author’s research and fieldwork (2000–2005), language contexts and attitudes varied, ranging from monolingual Urarina speakers and positive attitudes towards indigenous language and culture in more remote and secluded areas to Urarina-Spanish bilingualism and ideologies favoring Spanish-dominant culture in areas frequently visited by merchants and other external parties. A training program for native Urarina teachers who operate in most villages of the Chambira basin (central Urarina country) has proved immensely useful, as it focuses on Urarina as the language of education, contributing to hopes that the language may be maintained efficiently into the future. There are minor differences between the various dialects of Urarina which are mostly relevant to phonology. The speaker area can be subdivided into four main dialect zones, roughly characterised as Zone A (Western area, including Tigrillo and Espejo – the dialect which served as a basis for this outline), Zone B (Lower Chambi-

Knut J. Olawsky, Mirima Dawang Woorlab-gerring Language and Culture Centre & Language and Culture Research Centre, James Cook University (Cairns) https://doi.org/10.1515/9783110432732-012

1108

Knut J. Olawsky

25 Urarina

1109

ra dialects, including Asna and Airico rivers), Zone C (Upper Chambira with all its tributaries), and Zone D (Corrientes). Other names by which the Urarina have been known include Shimacu and Itucale, which are no longer in use. Urarina speakers also refer to themselves as Cacha ([katɕa]; Lit. ‘man’). Other nearby languages (some of which have few or no remaining speakers) include Záparo and Iquito to the north, Pinche and Omurana in the west, Cahuarano and Yameo in the east, and Cocama in the south. However, travel distances through the dense rainforest are immense, and the Urarina tend to stay isolated from other groups. Urarina’s genetic affiliation has been subject to numerous assumptions and hypotheses, ranging from Andean (Greenberg 1960: 794) to Tupian (Figueroa & Acuña 1986: 253). Based on lexical and grammatical comparison to the small amount of data on languages supposedly related to Urarina, it is difficult to find similarities between Urarina and any of the language groups so far mentioned in the literature. One argument in favour of assigning an isolate status to Urarina is the lack of lexical overlap with any of the languages spoken in the wider vicinity. This overview of Urarina is largely based on research published in Olawsky (2006) – a comprehensive grammar of Urarina which spells out the details for each area of the grammar. These details are only summarised here in highly condensed form. The grammar was based on a database of over a hundred texts of different genres from 24 different speakers, recorded over a span of five years, and archived with the Endangered Languages Archive (ELAR). It should also be noted that Olawsky (2006: 887–895) distinguishes between the traditional and innovative language, a distinction based on the ages of speaker groups. The innovative language tends to simplify certain aspects of lexicon and grammar. However, the main grammatical features remain identical, and the analysis provided in this article is based on the traditional language unless otherwise indicated. Other literature on and in Urarina is not extensive and includes articles by Manus (1992), as well as a few bilingual and Urarina publications by the educational NGOs Su Camino and Red Educativa de la Etnia Urarina; the latter is fully controlled by Urarina speakers and has started to replace mestizo teachers with bilingual teachers in most Urarina communities.

2 Phonology 2.1 Phonological inventory and alternations The sound system of Urarina contains 16 phonemic consonants and a number of allophones (Table 25.1). Most consonant phonemes occur relatively freely in word-initial and word-internal position, before any vowel. The few restrictions include the absence of the following combinations: /b/+/u/; /fw/+/u/; /hj/+/e, i/; /ɲ/+/i, ʉ/; /ʃ/+/ʉ/. Exceptions

1110

Knut J. Olawsky

Tab. 25.1: Urarina consonants (allophones in square brackets).1 labial

alveolar

b

t d

stop

labialized stop

kw

nasal

m

voiceless fricative labialized fricative

palatal

velar

glottal

k

n

ɲ

s

ʃ

[ŋ] h [hw]

fw

palatalized fricative

hj

affricate

tɕ [dʒ]

lateral

l

flap glide

retroflex

ɽ [ɾ] [w]

[j]

[ɰ]

to many of the generalizations presented below may occur in ideophones and in speakers who are highly proficient in Spanish. Phonological structures occurring in loans from Spanish include /p/, /g/, and a range of consonant clusters, such as stops followed by /l/ or /r/. The Urarina consonant inventory contains four stops (/b, d, t, k/) and, in terms of phonemic distinctions, the feature [Voice] is distinctive only in one case (/d/ versus /t/). A voicing contrast otherwise only exists as a non-phonemic alternation for the affricates [tɕ/dʒ], which largely occur in complementary distribution. The voiceless alveolopalatal affricate /tɕ/ mainly occurs word-internally as a syllable onset (e.g., [ka.tɕá] ‘man’). It is only attested word-initially, before /a/ and /ʉ/, where it is usually realized with the voiced affricate [dʒ], and it is therefore not regarded as a phoneme (e.g., [dʒa.ka.rí] ‘crocodile’). In word-initial position before other vowels, the affricate is realised as [tɕ]. The phonemic nasal consonants use three different places of articulation (/m, n, ɲ/), plus a non-phonemic velar nasal [ŋ] whose occurrence is restricted to a predictable environment. Labialization of /k/ and /h/ and palatalization of /h/ are widespread features; while their occurrence is largely predictable based on vowel copying, the consonants /kw/ and /hj/ do have phonemic status. The process of vowel copying can be

1 The transcription used in this article follows previous publications and it largely aligned with the IPA (but differs substantially from the official orthography which is based on the Spanish spelling system). Exceptions are /ɽ/, which is transcribed as /r/, used as a placeholder from the rhotic and its allophones, and labialization, which is transcribed as plain rather than a superscript.

25 Urarina

1111

characterized as homorganic vowel insertion after certain consonants, which include /h/, /k/, and /r/. For example, /h/ is affected by the presence of /i/, /u/, and /ʉ/. As a result of an optional rule, these vowels spread rightward through the consonant and are realized as palatalization, labialization, and “retroflexization”, respectively. For instance, /laauhiri/ ‘small’ is realised as [laa.u.hwi.rí], /kiha/ ‘paddle’ surfaces as [ki.hjá], and the vowel /ʉ/ carries through /h/, as in /kʉhana/ [kʉ.hʉa.ná] ‘wind’.2 The two fricatives [fw] and [hw] phonetically differ minimally from each other and are in fact not distinguished by most younger speakers. They function as allophones in the innovative language but are distinct in the traditional language. Examples (1a) and (1b) show the traditional language, and (1c) and (1d) the innovative language. (1)

a. /ufwa/ [o.fwá] ‘father’

b. /uhwa/ [o.hwá] ‘huicungu tree’

c. /fwʉitua/ [fwʉ.i.to.á / hwʉ.i.to.á] ‘cover:3(a)’

d. /fwafwakaa/ [fwa.fwa.kaá / hwa.hwa.kaá] ‘lift up:3(a)’

The rhotic /r/ has several possible realizations (retroflex flap [ɽ] alveolar tap [ɾ]) conditioned by word-internal versus word-initial position. The labialized velar ([kw]) can occur as a result of vowel copying from preceding vowels and as a phoneme, in root-initial position only. The voiceless palatal fricative /ʃ/ is infrequent and mainly occurs in loans. The glottal fricative /h/ can occur as a syllable onset, in any position, with most vowels. However, it is extremely rare before /ʉ/, except in cases where this results from vowel copying. This consonant also occurs as postvocalic aspiration, as in /ate/ [ah.té] ‘fish’ and /bʉkʉ/ [bʉh.kʉ́ ] ‘bone’. This phenomenon is optional and non-distinctive and can vary between speakers. As a tendency, postvocalic aspiration occurs after short vowels, on non-final, light syllables, and before voiceless or lateral consonants. The sound /hj/ occurs mainly in word-initial position, with only a few exceptions. Its distribution is restricted, as it occurs before the vowels /a/, /u/, and /ʉ/ only, as in [hja.na.ri.hí] ‘sugarcane’, [ka.hju.né] ‘clothing’, and [hjʉ.si.á] ‘scratch’. The palatal glide [j] is not a phoneme, but is realized in hiatus between two vowels (e.g., /here-i/ [he.ri.jí] ‘want’-neg:3(a)). Similarly, the labial glide [w] has a transitional function between /u/ and a following vowel (e.g., /kalaui/ [ka.la.wí] ‘son’). Phonetically, [w] takes the place of an onset for the subsequent syllable.

2 The accent marked on vowels indicates high pitch, which is discussed further in Section 2.3.

1112

Knut J. Olawsky

There are five vowel phonemes /i, e, u, a, ʉ/ and five diphthongs (/aj, ae, ej, au, aʉ/), again involving some alternations. Nasalization is a distinctive feature throughout the phonology; all vowel phonemes also have a nasalized counterpart, though there are no minimal pairs attested. Examples for some nasalized vowels include /arãala/ ‘tapir’ for /ã/, /rũa/ ‘side’ for /ũ/, and /hãʉ/ ‘because’ for /ãʉ/. There is a high degree of variation between [u] and [o] (/u/ is usually realized as a close vowel but sometimes surfaces as [o]). Examples for this are /uku/ [u.kú] ‘needle’ versus /umari/ [o.ma.rí] ‘basket’. The occurrence of the two allophones is not entirely predictable from the phonological environment, which suggests that these vowels are in free variation. The question to what extent vowel length is distinctive is controversial. It is significant on the grammatical level, since the combination of roots with suffixes results in vowel lengthening under certain conditions (for instance, when the imperative -ʉ is suffixed to a verb whose root ends in /ʉ/, the form results in a long vowel). On the lexical level, vowel length is lexically distinctive in isolated examples only, involving just a handful of minimal pairs that differ from each other by vowel length. Diphthongs contrast to vowel sequences in that they occupy only one syllable. The second component involved in a diphthong is realized as an offglide. In Urarina, a variant of /ʉ/ can function as an offglide, in addition to the common [j] (for /i/) and [w] (for /u/). Urarina phonology features a wealth of phonological alternations, both automatic and of a morpho-phonological nature (see Olawsky 2006: 30–119). Among the main alternations are the following: insertion of transitional glides [j, w, ɰ] (e.g., /rũa/ → [rũ.wá] ‘side’); vowel copying, that is, homorganic vowel insertion after certain consonants, including /h/, /k/, and /r/ (e.g., /enanihá/ → [e.na.ni.hʲá] ‘canoe’; /kukajtɕa/ → [ko.kʷaj.tɕá] ‘inhabitant’); dissimilatory vowel raising of /e/ → [i] before /a/ (e.g., /te.á/ → [ti.á] ‘give’-3(a)); diphthong raising of /aj/ → [ej] word-finally (e.g., /komasáj/ → [ko.ma.séj] ‘wife’); palatalization of /n/ after /i/ (e.g., /urarína/ → [u.rʷa.rí.ɲa] ‘Urarina’); rightward nasal spreading through adjacent vowels and /h/ (e.g., /sẽeuha/ → [sẽẽ.õ.hʷã]́ ‘big’). Urarina has a range of morpho-phonological alternations, with the main ones as follows. The plural marker -ʉrʉ on nouns and verbs has the allomorphs -kʉrʉ and -tɕʉrʉ. Various further alternations involving [k] and [tɕ] also appear on verbs (e.g., with first- and third-person forms and imperatives). When a verbal root ends in a long vowel or in a vowel sequence, the syllable /he/ is inserted as a stem extension before some suffixes (e.g., the participle form -ĩ is realised as -heĩ in combination with the respective verb roots). In both nouns and verbs the vowels /a/ and /e/ are lengthened if a subsequent suffix begins with a [+high] vowel, (i.e., /u, ʉ, i/). There are some suffixes that consistently merge with the final vowel of the root they are attached to (e.g., the combination of the suffix -e 3(e) with preceding /a/ resulting in [e]). Various rules of shortening and simplification specific to certain morphemes

25 Urarina

1113

also occur. An example for vowel shortening is the realization of the demonstratives before nouns: kaa ‘this’ and nii ‘that’ may surface as ka= and ni= respectively. The nasalized diphthong /ãʉ/ can be simplified by deleting its component /ʉ/ in either stressed or unstressed syllables, as in /hãʉ/: [hãʉ] ~ [hãá] ~ [hã]́ ‘because’ (cf. Section 2.3).

2.2 Phonotactics The prototypical syllable type in Urarina is CV. However, syllables do not require an onset, and other syllable types exist, summarized by the following canonical syllable inventory: V, VV, CV, CVV (where VV stands for a long vowel). Coda consonants are absent; CVC syllables do not occur. Diphthongs are analyzed as tautosyllabic, occupying a single V position. Syllable weight plays an important role for certain tone assignment rules (see Section 2.3); however, the relevance of moraic structure still requires further analysis. There is a tendency in Urarina to prefer words that are longer than a CV syllable. However, this cannot be formulated as a word length minimality constraint as there exist a number of verb forms and nouns which have a CV structure. For instance, the word for ‘skirt’ is hu, and ‘rat’ is sa. Most underived simple verbs in their citation form contain two to three syllables (e.g., [ʉ.á] come-3(a), [a.mʉ.á] walk-3(a)); the average noun involves three or more syllables (e.g., [a.ka.nó] ‘snake’, [e.na.ni.hjá] ‘canoe’).

2.3 Prosody Following Hyman (2009), I will refrain from categorizing Urarina in terms of assigning labels such as “tone language”, “stress language”, or “pitch accent language” and instead briefly summarize the main prosodic features in relation to stress and tone. As a general observation, heavy, (C)VV, syllables inherently attract stress, which is manifested by higher intensity/amplitude. However, analysis of stress patterns is complicated by the fact that tone and stress interact to some degree, and no detailed analysis is available at this stage. The Urarina tonal system contrasts H(igh) tone, which is lexically specified, and toneless syllables, which are L(ow) by default. The tone system can be characterized as “restricted” on the lexical level in that there is maximally one tone per morpheme. Despite the apparent simplicity of the tonal contrast (“tone versus no tone”), a complex system determines the position of the tone on the units to which it is assigned. While the vast majority of nouns (and verbs) carry a final H tone on the surface in isolation, their tonal structure is lexically predetermined insofar as they are responsible for tone assignment on adjacent words. That is, even though

1114

Knut J. Olawsky

Tab. 25.2: Tone groups for trisyllabic nouns: typical members. type A

type B

type C

type D

/ku.ku.ri/ (‘armadillo’) /bi.to.a/ (‘trap’)

/le.no.ne/ (‘food’) /ta.haa.e/ (‘bird’)

/ko.ma.saj/ (‘wife’) /kwi.tɕa.na/ (‘blood’)

/e.o.rí/ (‘termite nest’) /aj.tɕu.né/ (‘Espejo River’)

words have a final H tone in isolation, their underlying tonal structure may differ significantly from that pattern. Nouns can be divided into four tonal classes (called Types A, B, C, and D here) and, depending on the tonal class they belong to, H tone is assigned to a particular syllable of the word to the right. Type A surfaces all-L and assigns an initial H tone to a word that follows. Type B surfaces all-L and assigns an H tone to the third syllable of a word that follows. Type C surfaces all-L and assigns a final H tone to a word that follows. Type D contains an H tone but does not assign an H tone to a word that follows. These patterns are illustrated in Table 25.2. In addition to the one-tone-per-morpheme principle, tone assignment rules in Urarina are related to syntactic structure: as a tendency, each constituent (rather than each word) receives only one H tone. Another important feature is the existence of construction-types based tonal patterns that override any underlying tonal structures, such as in possessive or associative constructions.

3 Word classes Urarina has a highly polysynthetic morphology, especially in relation to verbs. The language features an abundance of suffixes and enclitics and only isolated cases of prefixing or proclitics. Proclitics exist for the singular forms of first, second, and third person objects and possessives (see Table 25.5), plus a proclitic form of the numeral ‘one’, le= (as in le=laʉriraana ‘one group of peccaries’). Prefixes include the reflexive ne- (also used as an intransitivizer), the associative ku- (as in aheri kuteru: stone+assc-axe = ‘stone axe’), and the rare associative modifier b- (e.g., lomaj b-eene: ‘Lomai’ (name)+assc-woman = ‘the woman Lomaj’). Urarina is head marking in the NP, seen through possessive constructions only and further illustrated in Section 4.4. Due to the absence of case marking, evidence regarding head/dependent marking cannot be gathered from this area. Nouns can be marked for number, possession, and location (in some cases). Classification of nouns into different subclasses is possible through a number of parameters, including whether nouns of a certain class can co-occur with a demonstrative, an adjectival modifier, or whether they function as a modifier themselves (see Section 4.3). As discussed in Section 4.4 on possession, Urarina nouns can be further classified relating the semantic features of the possessum. Most nouns can take a plural form, but not all (discussed in Section 4.2).

25 Urarina

1115

Tab. 25.3: Word classes and functional slots. function

verb

noun

adverb

head of intr. clause head of trans. clause head of np modifier in np copula complement argument of postposition modifier of verb

Yes Yes Derived (nomSbj) Derived (nomSbj) Derived (nomSbj) Derived (nomSbj) Derived (ptcp)

No No Yes Some Yes Yes Rare/derived (ptcp)

Derived (ptcp) Derived (ptcp) Some Some No Yes Yes

Verbs can be subject to a large number of inflection types. These include person and number, polarity, tense, aspect, mood, and various kinds of modality or Aktionsart. Valency-changing suffixes, such as those marking passive and causative, can attach to some verbs. Generally, verbs have three different forms of inflection for person/ number/polarity, depending on clause type and/or constituent order (Section 5.1). Another set of suffixes is used to cross-reference objects and intransitive subjects under certain conditions. All grammatical categories are realized as suffixes on the verb; the only productive prefix is the intransitiviser ne-. Urarina verbs fall into two main classes – transitive versus intransitive – with several subclasses. In summary, there are transitive, active intransitive, stative intransitive, and reflexive or intransitivised verbs. However, the differences become clear only through the derivational morphology. For instance, in nominalizations different suffixes are assigned to transitive and intransitive verbs. Ambitransitivity is a marginal feature attested for one verb only. Ideophones form an open word class along with nouns and verbs. They include onomatopoeia for various kinds of sounds and noises, with some variability between speakers. While nouns and verbs form the major part of the Urarina lexicon, adverbs are underrepresented in that there are few underived members, and they form a closed class. However, the creation of adverbs is possible through derivation with the participle form. Almost all underived adverbs are time-related. Adverbs can be categorized into five types related to time, place, manner, degree, and epistemics. Nouns, verbs, and adverbs are contrasted in Table 25.3 based on the functional slots they can occupy. Property concept-denoting words do not constitute a homogeneous, independent word class in Urarina, as they differ significantly from each other by means of function, position, and morphology. While there is no genuine, underived class of adjectives in Urarina, there is a range of “adjective-like” words which resemble nouns in that they can all be copula complements and they can function as nominal heads (except for one, which is derived). For example, enamanaa ‘young man’ is a noun which can also be used with modifying function, as in raana enamanaa

1116

Knut J. Olawsky

‘young peccary’. There also is a group of derived verbs which resemble adjectives whose behavior is similar to that of stative verbs. For instance, the word lanahaj ‘red’ is in fact a derived form of the verb lana-ha-j (be red-drv-nmlzSbj; Lit. ‘the red one’). The main difference between adjective-like words and other word classes lies in their function as a modifier. Pronouns and demonstratives can be analyzed as subtypes of nouns. There is one set of free pronouns and a corresponding set of proclitics in the singular. Demonstratives distinguish three levels of spatial reference; pronouns and demonstratives are discussed in more detail in Section 4.1. Postpositions are noun-like in that they can take a plural marker to optionally mark number agreement with the noun. This is exemplified in (2), where kʉanajʉrʉ ‘inside-pl’ refers to the noun which is absent from the clause. (2)

kʉanaj-ʉrʉ nianatiĩ hja-ʉr-e inside-pl even urinate-pl-3(e) ‘They even urinated in them [the clothes].’

Quantifiers do not represent a homogeneous word class as some can be analyzed as fossilised participle forms of verbs. For example, sati-ĩ ‘all’ can be analyzed as a form of the verb root ‘end’ combined with an intensifier and the participle form. Others resemble nouns with regard to their syntactic and morphological properties, such as taba-ʉrʉ ‘some’, which is composed of the verb root ‘be big’ and the plural marker. Numerals have been analyzed as special types of nouns and verbs. The native numbers include lexemes from ‘one’ to ‘five’ (treated as verbs), while the other numerals are loans from Quechua (analyzed as nouns). What remains are small classes of highly functional categories such as conjunctions and interjections. While Urarina has a comparatively large number of clause conjunctions, there are no phrase conjunctions. Instead, two nouns or noun phrases are associated by juxtaposition, as in kanʉ ii (1sg + 2sg) ‘you and I’. Urarina has a small class of words (listed in Table 25.4) that function as clauseinitial elements which indicate clause type and other categories, such as negation or mood. These “introducers” are obligatory, in addition to verbal morphology. The properties these words have in common involve their occurrence in clause-initial position and their co-occurrence with specific clause-type inflections. The fact that their syntactic position is specified and that most require a particular type of inflection on the verb, or elsewhere in the clause, suggests that these introducers form a separate word class, even though this may be typologically unusual. Urarina’s complex morphological system is restricted to nominal and, especially, verbal markers. One particularly atypical feature for a morphologically complex language is that there is hardly any productive inflection specifically assigned to word classes other than verbs.

25 Urarina

1117

Tab. 25.4: Clause introducers. introducer

use

co-occurs with

nihjauria, hjaui ɲe kwa kwatia kwane nabana baana ta dʒatera

“Strong” prohibitives “Weak” prohibitive on verb Negative (optional: emphasis) Hortative Apprehensive (optional) Irrealis/Conditional (optional) Negative question Suggestive clause

Negative clause-final enclitic Neutral inflection on verb Negative inflection Hortative inflection kwataa ‘so that not’ Conditional subordinate clause Final negative marker Special inflection

4 Noun Phrase structure 4.1 Pronouns and demonstratives Urarina marks person with bound pronouns on the verb. In addition to these, there is one set of free/cardinal pronouns which can be added for emphasis. Apart from the common distinction between first, second, and third person for both singular and plural, Urarina differentiates between inclusive and exclusive forms in the firstperson plural and dual. As Urarina does not have morphological case, personal pronouns do not differ with respect to their function as subject or object of a verb. Animacy also does not play a role. Table 25.5 shows the paradigm of free pronouns and the corresponding proclitics, in the singular. The Urarina deictic system distinguishes three degrees of spatial reference, each of them involving the position of the participants: kaa ‘this’ (close to the speaker), nii ‘that’ (close to the listener), and taa ‘that over there’ (distant from both participants). Each of the demonstratives can be combined with a number of elements that form locative or directional expressions, such as ka-hia ‘over here’ (near the speaker), nii-tɕae ‘over there’ (near the listener), and ta-hia ‘over there’ (away from speaker and listener). The parameters involved in their distinction involve indicating or pointing, which further distinguishs items that are in sight, or not. Further parameters refer to punctuality, distance, and location versus direction.

Tab. 25.5: Free pronouns and proclitics. singular free pronoun

plural proclitic

1 kanʉ

ka=

2 ii 3 aka

i= n= (ne=, ni=)

1 incl 1 excl 2 3

kana kanakaana ɲaara akaʉrʉ

1118

Knut J. Olawsky

4.2 Nominal number and quantification Plural can be marked on most nouns, including mass nouns or abstract nouns, where plurality is employed to express “different kinds of”. Plural is also marked on verbs and some postpositions. A plural marker may cross-reference a constituent, typically the subject or the object of a transitive clause, under certain conditions. As detailed in Table 25.6, Urarina has two predominant plural suffixes. One is -tɕe, which occurs with verbs only; the other suffix is -ʉrʉ, functioning as a nominal plural marker, as well as a pluralizer for third-person verbal marking. Other suffixes referring to number are the plural object marker -naha for transitive verbs, the distributive suffix -akwa, and the plural marker -ana for stative verbs. One factor that all plural markers have in common is their tendency to be optional. While plural marking on nouns or as morphological agreement on verbs may be the more frequent option (3a), it is not uncommon to use the singular form if plurality is implied by the context. In (3b), for instance, plural is not marked on either verb or noun, as the context provides this information. (The sentence is from a story where two men escape from two jaguars). (3)

a. itɕa-kʉrʉ-a leotɕa-ʉrʉ do-pl-3(a) other-pl ‘The others did it.’ b. nete akaʉru ɲasi-lanaala amʉ-e nii hanulari but 3pl.pro harm-without walk-3(e) that jaguar ‘But without doing them any harm, those jaguar(s) went (away).’ [About two jaguars mentioned earlier]

The most frequent case of number disagreement occurs between a noun in the plural form and a verb that remains unmarked for plural. This is illustrated in (4), with absent plural agreement. (4)

iɲoaelʉ ne-nakaʉrʉ katɕa-ʉrʉ ketau=te erʉ-rehete earlier be-those.who man-pl hrs=foc find-hab1:3(e) ‘It is said that the ancient people found them.’

Tab. 25.6: Overview of plural markers. marker

person and grammatical role

verb type

-tɕe -ʉrʉ -ana -naha -akwa -akwa

2 sbj 3 sbj 3 sbj 3 obj 3 sbj.distr 3 sbj.distr

All Transitive, and active intransitive Stative intransitive Transitive Intransitive Transitive

25 Urarina

1119

The verbal suffix -tɕe occurs as a plural marker for second person, as both subject marking on verbs in the declarative mood and for imperative marking. It also functions as a politeness marker in conversation between members of the opposite sex and in specific in-law relationships. The plural marker -ʉrʉ has a wider distribution than -tɕe, since it can occur on nouns, postpositions, and verbs, where it functions as a cross-referencing plural marker for third-person subjects. Urarina has three forms for first-person plural based on a distinction between inclusive versus exclusive, and dual versus plural, but this three-way distinction is realized only in verbal marking. There are only two free pronouns for 1pl, which are kana incl and kanakaanʉ excl, regardless of the number of speakers involved. The verbal suffix -akatɕe 1pl.incl includes the speaker and the listener, under the condition that the listener or the speaker group consists of more than two. In general, Urarina has number agreement between verb and subject (S/A). Since personal pronouns and other subject referents are optional and usually omitted, number concordance is not relevant. There are limitations to plural agreement between the head of a noun phrase and its modifiers. One context which demonstrates that plural marking on nouns is omissible is with numerals and quantifiers, as in (5). (5)

setʉ-akwa-a arahiĩ enʉa rot-distr-3(a) many tree ‘Many trees rotted.’

Plural marking on a noun preceded by a numeral or quantifier is not obligatory. However, an important distinction can be made between nouns with regard to animacy; in the majority of examples in the corpus, nouns referring to humans are more likely than not to be marked for plural in combination with numerals, shown in (6a). As exemplified with quantifiers in (6b), the same regularities apply: nouns with human referents are typically marked with the plural, while other nouns are not. (6)

a. le=tɕuŋka-j katɕa=ʉrʉ=te kaa ita kuruatahane-ĩ amiane work:3(e) one=ten-nmlzSbj man=pl=foc this recp help-ptcp ‘Ten people are working [here], helping each other.’ b. u-akwa-a arahiĩ katɕa-ʉrʉ die-dstr-3(a) several man-pl ‘Several people died.’

4.3 NP phrasal syntax The order of elements within the Urarina Noun Phrase is variable to a certain degree. However, certain tendencies can be summarized as follows. Demonstratives,

1120

Knut J. Olawsky

Tab. 25.7: Summary of noun modifier combinations. position

n

dem

num

possr

qnt

a

mod

rel

n

Precedes Follows

– –

X ✓

X ✓

X ✓

some some

some some

some some

some ✓

dem

Precedes Follows

✓ X

– –

✓ X

✓ X

✓ ✓

✓ X

✓ X

✓ X

num

Precedes Follows

✓ X

X ✓

– –

X ✓

✓ ✓

✓ X

✓ X

✓ X

possr

Precedes Follows

✓ X

X ✓

✓ X

– –

✓ ✓

✓ X

✓ X

✓ X

qnt

Precedes Follows

some some

✓ ✓

✓ ✓

✓ ✓

– –

✓ X

✓ X

✓ X

a

Precedes Follows

some some

X ✓

X ✓

X ✓

X ✓

– –

✓ X

✓ ✓

mod

Precedes Follows

some some

X ✓

X ✓

X ✓

X ✓

X ✓

– –

? ?

rel

Precedes Follows

✓ some

X ✓

X ✓

X ✓

X ✓

✓ ✓

? ?

– –

numerals, and possessors obligatorily occur in prenominal position. Demonstratives typically occur on the left edge of a noun phrase; their position is relatively fixed. Only quantifiers can precede demonstratives. Of all other modifier types, at least some members can occur in prenominal position. Numerals occur before the head noun, but they follow demonstratives and possessives. Thus, for those modifiers that obligatorily precede the noun, there is a preferred order: dem+possr+num+n. Adjectives, nominalizations (with relative clause function), and other noun-like modifiers typically occur close to the head of the noun phrase. At least some members of these three modifier types can occur in either prenominal or postnominal position, but most adjectives occur after the noun while most relative clauses occur before it. Adjectives precede other noun-like modifiers and relative clauses. However, a preferred order between the three noun-like modifiers is difficult to draw, as these groups are heterogeneous. Table 25.7 summarizes potential combinations and orders between the head noun and its various modifiers. It is observed that the variability with regard to the position to each other mainly concerns word types with a high lexical load, that is, adjectives, modifying nouns, and nominalized verbs. These types, which happen to relate to noun-like categories, typically occur adjacent to the head noun but have a comparatively free order amongst each other. In contrast, demonstratives, which have a more functional use, have a relatively fixed position in the noun phrase and need not occur adjacent to the noun. Numerals, which also are less noun-like than other modifiers, are fixed in their position in that they precede the head noun; how-

25 Urarina

1121

ever, their position with regard to other modifiers is comparatively free. Quantifiers, (which are not a separate underived word class) exhibit the most diverse order pattern, as they do not constitute a homogeneous group.

4.4 Possession In the noun phrase, possession is expressed by the use of possessive pronouns or proclitics, with an optional possessive marker between possessor and the head. The Possessor always precedes the Possessum. The different strategies to express possession in the noun phrase differ slightly from each other with respect to their function or occurrence only with certain nouns. While inalienability is a contrastive feature in the traditional language, younger speakers tend not to draw a clear line between inalienably and alienably possessed nouns. Several features suggest a certain degree of remaining inalienability for body parts and kinship terms but not for other nouns. A possessor can be human and is expressed through a full noun, a proper noun, or a pronoun. However, nouns lower on the animacy hierarchy can also function as possessors. Table 25.8 shows different kinds of nouns found in possessor function, categorized with respect to animacy. Note that examples for abstract nouns as possessors are not attested. Urarina noun phrases involving possessives occur in a variety of ways. In all cases, the position of the possessor is fixed, as it precedes the possessum. Possessive marking is optional and can be zero (by juxtaposition, as in (7a)); alternatively it can be marked by raj which otherwise means ‘for’ (when functioning as a postposition, as in (7b)), or by a proclitic, shown in (7c). (7)

a. alau bihi beeʉite nii ʉnee spider.monkey hand cut.off:3(e) that kinkajou ‘The kinkajou cut off spider monkey’s finger.’ b. rʉaka-ĩ ama-ʉre raj loanari asae carry-part take-3pl(e) poss shed under ‘They carried him and took him under his shed.’ c. raaheniane ni=hitɕana te-ĩ kʉane nesari-tɕʉrʉ-a nerajɲe self 3=blowgun give-part inside hunt-pl-3(d) must ‘They had to make their own blowguns and hunt with them.’

Multiple marking of possessive forms is not unusual in Urarina. Most commonly, the possessive marker raj is followed by a possessum that is cliticised with an additional possessive proclitic. Example (8) shows double marking with raj + n=.

1122

Knut J. Olawsky

Tab. 25.8: Types of possessors (possessor is underlined). possessor type

example

translation

pronoun human proper noun spirit animal body part plant nature-related inanimate

kanʉ lureri dʒ=ukwala komasaj Oriasi komasaj kana+kwaaʉnera letono alau bihi bihi kutʉri hiɲori kari nʉna kokwajtɕakʉrʉ lureri alo

‘my house’ ‘your brother’s wife’ ‘Uria’s wife’ ‘the envoy of God’ ‘spider monkey’s hand’ ‘hand’s head’ (= ‘finger’) ‘the bark of a hinyori tree’ ‘the inhabitants of the jungle’ ‘the roof of the house’

(8)

[raj n=itɕafwa] tɕoae kodoonete poss 3=rifle on heap.up:3sg(e) ‘It [the tree bark] heaped up on his gun.’

The high degree of optionality and the fact that double marking of possession is possible might indicate that the differences are due to diachronic changes in the system, even with speakers of traditional Urarina. One might suspect that Urarina had inalienably possessed nouns at an earlier stage, but this feature has been largely lost over time. Possession can also be marked by two verb-based strategies. One is with siiria, a transitive verb meaning ‘have’; the other involves the copula in combination with a free or bound pronoun followed by the possessive marker raj.

4.5 Word formation Urarina has a complex system of derivation, including highly productive processes for nominalization and less productive cases of verbalization. A verb or clause can be relativized through nominalization, of which the different types are summarized here. Intransitive subjects and objects attached to intransitive or transitive verbs are nominalized with -i. This strategy is shown in (9) with subject function on the intransitive verb root laʉhʉ ‘sit’. (9)

kʉ [nii ajrinia laʉ-hʉ] raʉsi there that outside sit-nmlz:sbj- bite:3(e) ‘It bit the one that was sitting outside.’

With transitive verbs only, nominalization of the agent is done with -era, as illustrated on the transitive verb ku ‘drink’ in (10).

25 Urarina

1123

(10) hajti=te ne iɲono ku-ur-era still=foc be:3(e) ayahuasca drink-pl-ag ‘There are still those who drink ayahuasca.’ The suffix -naa is used for nominalization of agents or abstract nouns, mainly attached to intransitive verbs. An example for this is the noun amiane-naa ‘work’ which is derived from the verb amiane ‘(to) work’. For plural intransitive subjects, nominalization can occur through constructions with -nakaʉrʉ. In (11), the suffix occurs on the intransitivized verb for ‘buy’ (resulting in the meaning ‘sell’), where ‘traders’ literally are ‘those who sell things’. (11) heriane=te ʉʉ-re=ĩ rʉkʉele ke ne-kurete-nakaʉrʉ maybe=foc come-irr:3(e)=ass things appl intr-buy-those.who ‘Maybe the traders will come.’ Nominalizations that realize both locative and purposive functions are formed with -naha, ‘a place to …’. For instance, in combination with the verb root enoala ‘cook’, the form enoala-naha is used to mean a ‘place to cook’. Verbalization with -oka (or allomorph -koa, after a vowel sequence) typically refers to long-term possession, such as in lana-oka ‘she has a husband’. The suffix only applies to certain nouns, including terms for kinship and body parts but also other important possessions, such as the words for ‘house’ and ‘axe’.3 There is a wide range of verbal derivations which do not involve a change of word class. Reduplication is another instance of word formation that applies to verbs. Urarina has no less than five different types of verbal reduplication (plus subtypes), each of which have a slightly different grammatical function and exhibit a high degree of overlap between pluractional interpretation (‘repeatedly’), spatial distribution (‘in various areas’), and associated motion (‘move/run’). All types are instances of total reduplication; differences between the various types are indicated by additional morphemes such as suffixes. Table 25.9 overviews the attested types. It should be noted that most types of reduplication are very productive with PSC (position/shape/color) and affect verbs but hardly, or do not, occur with other verb classes (except Type 3). Thus, it may not be surprising that some reduplication types occur as “pairs”, each distinguished for big and small items, as is observed with suffixal derivation. In fact, each type of reduplication is enriched by the addition of other segments that are also partly found in suffixal derivation. Nine more or less productive types of compounding can be identified. All types involve nouns in some way, and each type has its own semantic structure. In most compound types, the rightmost element (typically a noun) functions as the head. Table 25.10 gives an overview of common types of compounding. 3 While this may indicate a potential distinction between alienable and inalienable nouns, there are other, largely morphological, factors related to possession which do not allow this conclusion.

1124

Knut J. Olawsky

Tab. 25.9: Types of reduplication (r = Root, underlined; red = Reduplicant; [vv] = long vowel).4 no.

type

example

functions

1a

red~r-[vv]ka

teme~teme-eka ‘get/move to stand’

Position/state and motion; ‘move in a certain state or position’; big items

red-a~r-akoa

baʉ-a+baʉ-akoa ‘be moving in lying position’

1b

red-e~r-ekoa

laʉ-e~laʉ-ekoa ‘get/move to sit’

Position/state and motion; ‘move in a certain state or position’; small items

2a

red~r-[vv]oka

lana+lanaa-oka ‘be moving quickly as something white’

Position/state and quick motion; ‘run in a certain state or position’; big items

2b

red~r-eka

laʉ~laʉ-eka ‘quickly get to stand’

Position/state and quick motion; ‘run in a certain state or position’; small items

3

red-e~r-eka

saʉ-e~saʉ-eka ‘cut again and again’

Aspectual: repeatedly, as a habit

4

red~r-a

saʉ~saʉ-a ‘cut various times’

Aspectual meaning (repeatedly); On PSC verbs: Distributive ‘in various areas’; big items

5

red~r-eria

teme~teme-eria ‘lean against various places’

Distributive ‘in various areas’; small items

Tab. 25.10: Types of compounding. type

head

structure

example

N+N

N2

Possessive NP (N1 = possr)

tihja+sʉ́ ri ‘lower leg’ + ‘belly’ (= ‘calf’)

N + ku-N

N2

Associative NP (N1 = source or goal)

aheri+ku-téru ‘stone’ + ASC-‘axe’ (= ‘stone axe’)

N+N

N2

N1 = modifier

laano+dʒakári ‘cassava’ + ‘crocodile’ (= ‘white crocodile’)

N+N

exocentric / N2

Possessive NP

baka+isítɕo ‘cow’ + ‘breast’ (= type of plant called teta de vaca)

V+N

exocentric / N

N modified by V

maʉe+bíhi ‘be short’ + ‘hand’ (=‘person missing a finger’)

N+N

N1

N1 modified by N2; N2 = nominalized verb

fwanara + lanáhaj ‘banana’ + ‘red’ (=‘red banana’, guineo type)

N+V

N

N modified by V (V can be interpreted as nominalized)

nʉkʉe+súru ‘river’ + ‘run’ (=‘tributary river’)

V+N

N

N modified by V

ere+enʉ́ a ‘talk’ + ‘tree’ (=‘talking tree’)

N + V; N+N

V N2

N is the object of V N1 is the object of N2 (nominalized verb)

sʉʉhʉ+tía ‘heart’ + ‘give’ (= advise’); kunu+muka-éra ‘light’ + ‘burner’ (= ‘lighter’)

4 The roots are glossed as follows: teme ‘stand’; laʉ ‘sit’; saʉ ‘cut’.

25 Urarina

1125

5 The verb phrase Urarina verbs can be formally divided into transitive and intransitive verbs, with a number of subclasses. Stative intransitive verbs are distinguished from other verbs by a special plural suffix that is only used with this subclass. Intransitive verbs that refer to position, shape, or color (PSC verbs) form another special subclass of intransitives which is manifested through derivational regularities. Reflexivity may be expressed by verbs with the prefix ne-. Within the class of transitive verbs, there is a subgroup of “affect” verbs which differ from others with respect to their derivational behavior. Urarina has no ditransitive verbs. Verbs such as ‘give’ are regular transitive verbs that cannot be formally distinguished from others. Urarina also has a copula which has some properties of both transitive and intransitive verbs; otherwise, ambitransitivity is not a predominant feature in Urarina. Furthermore, there are some verbs that refer to a question, and a few more verbs have special properties (cf. Section 5.4). These do not represent a homogeneous class.

5.1 Person inflection classes Each Urarina verb can occur with up to three different verb forms dependent on the grammatical categories of person and polarity marking (and, partly, irrealis), here labeled as D-form (in finite dependent clauses only, (12c)), A-form (whose 3sg form is -a, see (12b)), and E-form (whose 3sg form is -e, see (12a)). The examples in (12) illustrate the different forms. (12) a. iɲoaelʉ hetau=te enejtɕʉ-kʉrʉ kʉraanaa ne-ĩ ɲe=lʉ earlier hrs=foc monkey-pl chief be-ptcp be:3(e)=rem.tns alau spider.monkey ‘In ancient days, the spider monkey was chief of the monkeys.’ b. iani-a ʉsi burn-3(a) fire ‘A fire is burning.’ c. ʉnee bʉa basihjaʉ-a alau=ne kinkajou bag steal-3(d) spider.monkey=cond ‘When the spider monkey stole the kinkajou’s bag’ (Title of a narrative) For negative polarity, a separate paradigm exists for each inflection class (see Section 6.3). Note that I do not classify these classes as “conjugation”, as they do not refer to subclasses of verbs, as in Romance languages. Instead, the distinction depends on the grammatical context in which a verb occurs, meaning any verb can

1126

Knut J. Olawsky

Tab. 25.11: Obligatory contexts that determine the choice of person inflection type. context

person inflection type

Citation form Greetings Polar questions With introducer kwatia Content questions foc marker preceding verb Dependent clause

A-form A-form A-form A-form E-form E-form D-form

Tab. 25.12: Optional contexts that influence the choice of person inflection type. context

A-form

E-form

Short utterances (independent clauses) Negated verb Narrative style After a dependent clause Before a dependent clause

typical typical with third person possible possible possible

uncommon possible typical typical typical

be inflected for any of the three types if it occurs in the correct context. From a functional point of view, the D-form is easiest to characterize, as it occurs only in dependent clauses, as in (12c). The E-form is the preferred person inflection type used in narratives; see (12a). One typical feature of narratives is also the elaborate use of complex syntactic structures, which in turn coincides with the use of the E-form. The A-form is preferred with verbs that are inflected for third person negative; in short statements like (12b); in everyday conversations between speakers; and in characterizations of simple events, such as picture descriptions. The various contexts for each inflection type are summarized in Table 25.11. The distinction between the A-form and the E-form in non-obligatory contexts involves factors such as polarity and syntactic complexity, as well as style. Table 25.12 gives an overview of the factors that play a role when obligatory conditions do not apply.

5.2 Verbal morphology Urarina has 17 suffix and 7 enclitic positions (including 5 clause enclitic positions). However, person is the only feature that is obligatorily marked on every finite verb. Other categories only surface when necessary. For instance, polarity is only marked for the negative, number is only marked in the plural, etc. In contrast, the potential for the morphological realization of grammatical categories is considerable and, de-

25 Urarina

1127

Tab. 25.13: Summary of postverbal slots: Suffixes. slot

suffix

function

order variability

 1  2  3  4  5  6

-a -erate -naka -he -era -reheto -nahauka -akwa -naha -uri -hee -si -ana -ʉrʉ -kaj -ri (Various) (Various) -tɕe

causative-1 causative-2 Possibility Aspect: continuous Impersonal passive Aspect: habitual-1 Aspect: habitual-2 distributive plural object Velocity diminutive Aspect: completive In-law talk plural-1 Probability irrealis negation Person plural-2

– – – May vary – May vary (unclear) May vary May vary May vary May vary – – May vary – – According to paradigm – –

 7  8  9 10 11 12 13 14 15 16 17

pending on the circumstances, a large number of features can be marked in one verb. Though person is the only obligatory marking on verbs, the large number of suffixes that can appear (with a few theoretical exclusions) makes Urarina a highly polysynthetic language. To give an overview, the order of suffixes is summarized in Table 25.13. The table also indicates whether the order of the respective slots may vary or not. Urarina does not have any formal marking explicitly used for tense. For instance, the remoteness marker =lʉ is usually used with past reference, but it can also refer to remote future when used with some words (as in ɲoae=lʉ netohweĩ ‘forever’). There also is no suffix to mark future tense on its own. Instead, future tense is a construct of the irrealis marker -ri (and allomorph -re) and the assertive enclitic =ni. Without the clitic, the form is understood as irrealis. While -ri can be translated as ‘would’, the addition of the assertive expresses the certainty of an action or event happening in the future. In terms of position changes of suffixes within the verb, there is a “variable area” between slots 4 and 9, with actively floating morphemes in the positional slots 7, 8, and 9. For instance, the position of the plural suffix -ʉrʉ may change in combination with the habitual suffix (slot 6). Optional changes in order are also observed with slots 6, 8, 9, and 10, though it should be mentioned that at this stage not all possible combinations are attested, and further data is required.

1128

Knut J. Olawsky

Tab. 25.14: Summary of postverbal slots: Enclitics. slot

clitic

function

order variability

18 19 20 21

=tɕe =ni =ni =he =lʉ =tau =na =ne =ra =ta =naare =naate =te

Politeness assertive (future) Evidentiality: witness Evidentiality: reportative remoteness (past) Reassurance Mood: interrogative Mood: negative question; prohibitive Attitude/Emotion: (positive) emphasis Attitude/Emotion: negative affect Attitude/Emotion: Warning Attitude/Emotion: Fear Rhetorical question

According to paradigm According to paradigm – May vary May vary – – – – – – – –

22 23 24

As a general rule, any enclitics follow the suffixes discussed above. There are seven clitic slots, some of which are best characterized as “zones”, as they include several enclitics, as shown in Table 25.14. The order of enclitics is less subject to variation than the suffixes. Most exclusions occurring with enclitics are based on pragmatic incompatibilities. As can be seen in the Table 25.14, evidentiality can be marked for two values: witness and reportative. With the latter, the addition of Slot 22 =tau) results in a hearsay evidential hetau which is frequently used in narratives (see Section 7.5).

5.3 Politeness One peculiar feature of Urarina verbal morphology is the various strategies it offers to express politeness. Unlike many other languages, where politeness is largely based on situational variation, most polite forms used in Urarina follow strict patterns that depend on gender and family relations. The enclitic =tɕe is used in conversations between members of the opposite sex and between male in-laws. The function of =tɕe may be described as honorific, as the addressee is to be treated with respect. However, the group of people with whom this enclitic is used is predetermined by gender and kinship relations. The enclitic is usually attached to the verb. In verbless clauses such as (13), which illustrates woman to man speech, the politeness marker is attached to the first constituent, which here is the pronoun. (13) hẽ, kanakaanʉ=tɕe kaa=ra itɕene na-a kʉ-ʉrʉ-a hey 1pl.excl.pro=polit this=aff.emph uncle say-ntr go-pl-3(a) ‘“Hey, uncle, this is us” they went to say.’

25 Urarina

1129

An additional, obligatory form of expressing politeness again regards the relationship between in-laws. It is expressed by a verbal suffix that is used when referring to – not talking to – persons to whom the speaker is linked through certain in-law relationships. The presence of the suffix -ana (or allomorphs) in a verb automatically implies that this condition applies to the subject of the clause; see (14) for an example with -ana. (14) kwajteĩ kohwanoo ari-a kʉ-akaanʉ hãʉ leijhiĩ tabiitɕa again the.next.day seek-ntr go-1pl.excl because one finally rʉ-hi-ana-e find-dim-ilt-3(e) ‘As we went to seek [turtles] again on the next day, he (my son-in-law) finally found one.’ In addition to these two “fixed” parameters, Urarina also has strategies for expressing politeness that are based on the situation. In these cases, morphemes whose general function is to mark other grammatical categories, such as diminutive -hee and velocity -uri, are employed to convey politeness, usually associated with a request.

5.4 “Special” verbs Urarina has a number of verbs that do not form a homogeneous class but which exhibit unusual features different from those found with other verb types. Example (15) overviews these verbs. (15) a. b. c. d. e. f. g. h.

naaohwa nukuja ɲaaohwaa ɲatahaa ɲasia dʒatoania tonaa ãaka

enumerative ‘be in vain’ ‘be how much’ ‘be how much’ ‘do how’ ‘be how’ ‘sound how’ ‘be where’

The numeral verbs for ‘one’ through ‘five’ may also be added to this list. However, note that not all numerals behave in the same way.

5.5 Valency-changing mechanisms Most valency changes in Urarina are marked by verbal suffixes. There is a verbal derivation with passive function which is best characterised as a case of nominaliza-

1130

Knut J. Olawsky

tion. In (16), the passive marker -noi (here as allomorph -nohwi) is attached to the verb su ‘kill’, effectively turning the verb into a nominal, here used as a modifier “the killed fish”. (16) neleere nii nii-tɕa su-nohwi ate become.big:3(e) that that-only kill-pass fish ‘The fish that he had killed became big.’ From a morphological and syntactic point of view, the “passivized” verb becomes a noun, which is reflected in that: it is compatible with the nominal plural -ʉrʉ; it is not inflected like a verb, but requires an auxiliary; it can modify another noun; and it can be possessed. The general meaning of the passive in discourse is not entirely clear, but, typically, it may be chosen when the agent of a clause is unknown. Example (17) illustrates this with the nominalizer -noi, here followed by a plural suffix. (17) [amia-rate-noi-tɕʉrʉ katɕa-ʉrʉ=te]A [ena]Adv [ita kuruatahane-ʉre]V work-caus2-pass-pl man-pl=foc now recp help-3pl(e) ‘The people who had been forced to work now help each other.’ Another form with valency-decreasing properties is ne-, which functions as a general intransitivizer. For instance, the transitive verb relaa ‘teach’ becomes intransitive by prefixation to result in nerelaa ‘learn’. Urarina has two suffixes that mark causative, -erate and -a, which are used with verbs of different transitivity classes. The semantic interpretation of the causative -erate is context-dependent. It always focuses on the fact that the causer does not conduct the action, but delegates others to do it. The examples in (18) illustrate the occurrence of -erate with different verb types. When an additional argument occurs as a result of increased valency, it is followed by the postposition ke as shown in (18b). (18) a. [kaa katɕa-ʉrʉ]O [amia-rate-reheto-a=lʉ]V this man-pl work-caus2-hab1-3pl.(a)=rem.tns ‘He used to make these people work.’ [akaʉrʉ ke] … b. [arusu]O [temʉ-erate-ĩ]V rice plant-caus2-part 3pl.pro posp ‘making them plant rice…’ c. ne-raʉta-erati-a ʉ-ʉ-tɕe=ra! intr-heal-caus2-ntr come-imp-pl=aff.emph ‘Come to be healed!’ [NT: Luke 13:14] The causative suffix -a is less productive and attaches to intransitive verbs only. The argument structure of verbs causativized with -a corresponds to those that carry

25 Urarina

1131

-erate. The functional difference between the two causative suffixes is manifested in that -a implies “direct personal involvement” of the causer, whereas -erate points at “indirect” cause, for example, the causer sends someone else to give the order for the action. See (19) for an example of this alternation. (19) a. [surutiĩ=te]Adv [atero-e]V [katɕa]S quickly=foc get.tired-3pl(e) man ‘The man became tired quickly.’ b. [kanʉ]O [atero-a-a]V [amiane-naa]A [hataĩ]Adv 1sg.pro get.tired-caus1-3(a) work-nmlz much ‘The work has tired me a lot.’ Both suffixes can occur in combination on the same verb, as in (20), which typically refers to a delegated task and involves an additional participant. (20) katɕa eno-a-erati-a man enter-caus1-caus2-3(a) ‘He had the man enter.’ The transformation from an intransitive verb to a higher valency is unproblematic, since the resulting verb behaves in the same way as a transitive verb. Given that Urarina does not have ditransitive verbs, the new argument (causer) becomes A, the old O argument remains O, and the old A argument becomes oblique, which is accounted for by marking it with the postposition ke, see (18b).

5.6 Serial verb constructions Apart from a range of other multi-verb constructions (Section 7), Urarina makes use of serial verbs. These constructions involve a closed class of verbs of direction and movement. While the first verb (V1) can be transitive or intransitive, the nature of the second verb (V2) is restricted. In the majority of serial verb constructions, V2 is represented by the intransitive verbs ʉa ‘come’ or kʉa ‘go’. In (21), the two verbs appear between square brackets. (21) obaa-heĩ hetau, alau [kwara-a kʉ-e] nii ʉnee get.angry-part hrs spider.monkey see-ntr go-3(e) that kinkajou ‘The kinkajou got angry and went to see the spider monkey.’ The main feature of Urarina serial verb constructions is that all verbs involved share the same subject, automatically excluding switch-subject serialization. The first

1132

Knut J. Olawsky

verb bears the neutral marker -a (or allomorphs), which is also the suffix used in the citation form of verbs. In general, serial verbs refer to one event (or parts of it), which is reflected by the fact that all morphology marked on V2 scopes over the entire construction. V1 is usually less morphologically marked, but it can be characterized as the semantic head of the construction, while ‘go’ and ‘come’ mainly describe the direction or result of the event.

6 Simple clauses 6.1 Basic constituent order The basic constituent order of Urarina is OVA/VS (extending to free pronouns), making the language a typological rarity. Any deviations from this order are rather specific and predictable. The typical clause structure can be characterised by the following features. Subject and object noun phrases are frequently omitted and understood by the context. Note that pronouns are not required to refer to a thirdperson object. The order for affirmative and negative sentences is identical, with possible variation in the negative. Adjuncts and postpositional phrases usually occur in the periphery of the core clause. Typically, they precede the object. In many sentences, adverbs or other items such as ‘therefore’ are focused (using a focus marker). This does not affect the order of other constituents. Filler words such as ɲae ‘already’ and kʉ ‘there’ can occur in almost any position in the sentence. However, they do not occur within constituents. Evidence for the basic constituent order is provided by a statistical sample of order types in seven selected texts of different genres by different speakers (a total of 799 clauses). The most common clause type is verb-only without overt arguments/noun phrases (58 %), followed by OV (21 %) and VS (10 %). Where arguments are present, the OVA/VS order is five times more frequent than deviating order types (35 % vs. 7 %). A minority of clauses exhibit a different order. These can be explained by factors such as focus and a few other special forms, which include negation and the presence of the ‘must’ construction; see (26). Constituent order within the dependent clause is largely identical to the independent clause. Naturally, full noun phrases are even rarer in dependent clauses, as the referent of subjects or objects is usually mentioned in the main clause (if mentioned at all). Unmarked OVA order is demonstrated in independent clauses, as in (22a), and dependent clauses, as in (22b). (22) a. [nitoaneĩ hetau=te]Adv [katɕa]O [lemʉ-e=lʉ]V [Lomaj]A [edara like.that hrs=foc man sink-3(e)=rem.tns Lomai water.people ne-ĩ kʉ-ʉrʉ-a=ne kujɲa]DEP be-part go-pl-3(d)=sub so.that ‘Lomai sank the people like that so that they would become water people.’

25 Urarina

1133

b. [ʉnee bʉa]O [basihjaʉ-a]V [alau=ne]A kinkajou bag steal-3(d) spider.monkey=cond ‘When the spider monkey stole kinkajou’s bag…’ Further examples illustrate the absence of overt arguments, such as omission of the subject in (23a) and object in (23b), showing OV and VA orders, respectively. (23) a. [hitariĩ kaʉte-na ii kuriki]O [ʉ-a-re-ʉ̃ =ni]V all remain-nmlz 2sg.pro money come-caus1-irr-1sg(e)=ass ‘I will bring all of your money that remains.’ [itɕa-kʉrʉ-a]V b. [niriu kokwajtɕa-kʉrʉ ajɲa ne-rela-ĩ]DEP Marañon inhabitant-pl with intr-teach-part do-pl-3(a) [leotɕa-ʉrʉ]A other-pl ‘Learning with the people from the Marañon River, the others did it [too].’ In intransitive clauses, the unmarked order VS prevails, as in (24). (24) [nii hãʉ hetau=te]CNJ [ubaa-he]V [nii ʉnee]S that because hrs=foc get.angry-3(e) that kinkajou ‘Therefore, the kinkajou got angry.’ The main condition which can affect the default word order is focus. Any constituent of a clause can be focused; in this case, it is followed by a focus marker and shifted to the front. Most commonly, non-core constituents such as adverbs or discourse-related expressions (e.g., nii hãʉ ‘therefore’) are marked for focus. A change in the basic constituent order only applies when the subject of a transitive, in (25a), or intransitive, in (25b), clause is focused. Typical contexts for focus involving constituent order deviations are content questions and fronting of subjects. (25) a. [raj kalaui=te]A [fwei]O [bajhja-ĩ ama-e]V poss son=foc firewood carry.on.shoulder-part take-3(e) ‘Her son carries firewood on his shoulder and takes it along.’ b. [lejhiĩ katɕa=te]S [hataĩ]Adv [nesoonete]V one man=foc very be.scared:3(e) ‘One man is very scared.’ [from picture description] Negation and focus markers cannot co-occur, and, therefore, fronting of the subject (if required to indicate focus) occurs without a focus marker.

1134

Knut J. Olawsky

(26) a. [hataĩ ɲoaelʉ ne-nakaʉrʉ]S [niji]Adv [enanihja kʉane]PP very earlier be-those.who not.at.all canoe inside [amʉ-ʉrʉ-i=lʉ]V walk-pl-neg:3(a)=rem.tns ‘Those who lived very much earlier did not go at all by canoe.’ letono]CC [ni-a-ʉ̃ =ta]V! b. [kwaita kanʉ]CS [itɕene not 1sg 2sg=place envoy be-neg-1sg(e)=neg.aff ‘I am not the envoy for your place!’ The tendency to place syntactic material other than objects into preverbal position is further illustrated by dependent clauses. While it is possible for a dependent clause to follow the main clause, the more typical position is before the main clause. Complement clauses generally precede the main clause, with very few exceptions.

6.2 Sentential mood Sentential moods other than declarative clauses comprise interrogatives and imperatives which are marked lexically and/or morphologically. Interrogatives as well as imperatives can be divided into several types, each with specific features. Content questions contain an interrogative pronoun, which is obligatorily fronted to clause-initial position and can be marked for focus.5 The basic interrogative pronoun from which most other forms are derived is dʒa. Content questions require the use of the E-form on the verb. In contrast, polar questions always occur with the A-form (Section 5.1). These are marked with the interrogative enclitic =na, which is attached to the verb in non-focus clauses but to the first constituent in focus constructions. Negative questions feature double marking by the negative question introducer ta in clause-initial position plus the final negative marker =ne in final position. The inflection on the verb is affirmative, that is, negation is not marked on the verb itself but only through the accompanying introducer and final negative enclitic. Urarina has a complex system for marking imperative mood which features a separate clause type. It is marked by a suffix for positive imperative, as listed in Table 25.15, which can be followed by the plural suffix -tɕe in first-person plural inclusive and second-person plural. There are special forms for first and third person imperatives (hortative, jussive). The hortative requires the presence of a clause

5 The presence or absence of a focus marker tends to influence the interpretation of the interrogative pronoun with respect to animacy: its absence results in the interpretation of dʒa as inanimate, which in turn implies its function as an O argument. The insertion of the focus marker in an otherwise identical question results in a preferred reading in which dʒa is interpreted as a subject with human reference.

1135

25 Urarina

Tab. 25.15: Comparison between verbal person marking with imperative and declarative mood. person

positive imp

positive decl

negative decl

strong proh

weak proh

1sg

-e

–

-akaanʉ -aka -aka-tɕe

-a-ʉ̃ -ene-ʉ̃ -a-akaanʉ -a-aka -a-aka-tɕe

–

1pl.excl 1pl.du 1pl.incl

-anʉ -ʉ̃ -akaanʉ -aka -aka-tɕe

– – –

– – –

2sg

-ʉ -riʉ -ʉ-tɕe

-i

-e-i

-i=ɲe

-a

-i-tɕe

-e-i-tɕe

-i-tɕe=ɲe

-a-tɕe

-miĩ -iɲe -ʉrʉ-miĩ -ʉrʉ-iɲe

-a -e -ʉrʉ-a -ʉre

-i -ene -ʉrʉ-i -ʉrʉ-ene

-a=ne

–

?

–

2pl 3sg 3pl

introducer under certain conditions. Clause introducers such as nihjauria and hjauiɲe, which require special types of inflection on the verb (also see example (27)), are also used with the prohibitive form. As an additional feature, there is a distal form -ni, which is only attested with positive imperatives. Other factors related to commands are: marking of imperatives for non-distant future with the enclitic =ni, and the suggestive forms -riʉ 1sg, -miĩ 3sg, and -iɲe 3pl, as included in Table 25.15, which provides a summary.

6.3 Negation Negation can be expressed through a wide range of strategies in Urarina. It can be marked in any clause type, that is, declarative, imperative, and interrogative. The scope of negation is the clause, and single constituents cannot be negated. There are several allomorphs for the negative suffix based on the three different person inflection classes (Section 5.1). Regarding the occurrence of negation in different moods, a few minor differences between negative forms are observed. There are several clause-initial prohibitive markers – nihjauria, hjauiɲe (or variant ɲaauiɲe), and kwa – all of which can be glossed as ‘don’t’ but which differ from each other with respect to urgency and strength, as in (27) and (28). The introducers nihjauria or hjauiɲe require the same person inflection as in declarative clauses, but they are only attested with forms of the A-conjugation class. Constructions with nihjauria or hjauiɲe involve the clause-final enclitic =ne, again seen in (27). (27) nihjauria nekajritɕa-i=ɲe kuriki siiri-i=ɲe! don’t suffer-2=neg.fnl money have-2=cond ‘Do not suffer if you do not have money!’ [Implied: ‘but borrow some from me’]

1136

Knut J. Olawsky

With the prohibitive introducer kwa all person marking is neutralized and the neutral suffix -a is attached to the verb, illustrated in (28). (28) kwa ka=su-a i=tɕej biri bi-anʉ=ne kujɲa=ra! don’t 1sg=kill-ntr 2sg=for piripiri tell-1sg(d)=sub so.that=aff.emph ‘Don’t kill me and I [can] show you piripiri!’ Urarina also has a number of inherently negative lexemes. Some of these are interjections for ‘no’, such as hjauiɲera (strong) and aji which is the negative form of an auxiliary. In addition, there is the conjunction kwataa ‘(that) not’ which functions as the negative counterpart to the purposive conjunction kujɲa ‘so that’ and can optionally be accompanied by the negative introducer nabana at the beginning of the clause. However, rather than requiring negative marking on the verb, this conjunction occurs with affirmative forms only.

6.4 Clause-level particles Urarina has a range of clause-level particles, some of which share characteristics with clitics. However, speakers tend to easily identify all particles as separate words which justifies listing these separately. The particle hjã shown in (29) has an encouraging or mitigating function, translatable as ‘simply/just’ in English as in “simply (do it), it’s all good”. Most occurrences are combined with a form of command, which are “softened” by the presence of hjã. (29) eresi nii ora ke tɕʉ hjã ajto-re-ʉ̃ =ni tomorrow that hour posp crtn just say-irr-1sg(e)=ass ‘I will just tell [you] tomorrow at that time.’ [from conversation] The emphatic particle laẽ (which may also be glossed as ‘just’) marks a contrast to a previous situation. Usually, this refers to a temporal change, as most occurrences of laẽ follow ɲaẽ ‘already, now’ or other time-related adverbs, as it does in (30). Therefore, laẽ could be translated as ‘now yes’ or ‘but now (as opposed to earlier)’, indicating a change of situation. (30) ɲãe laẽ ena kwaaʉne-ĩ kwajte-ri-a-akatɕe=ĩ na-akaanʉ already just now create-part repeat-irr-neg-1pl.incl=ass say-1pl.excl ɲaẽ nitoani-akaanʉ hãʉ, ena=ra kanakaana already be.like.that-1pl.excl because now=aff.emph 1pl.excl.pro nesoone kãʉ fear because.of ‘“From now on we won’t do [this] again”, we said already, because of our fear, because we were in this situation.’

25 Urarina

1137

The particle noane (lexically related to the verb ‘lie’), marks self-pity and compassion, and it can be translated as ‘unfortunately’. The emphasis, however, is not to show compassion for another person, but to express one’s own pitiful situation, illustrated in (31). (31) netitɕara noane kukwarue-hee-ka katɕa=ne … maybe pity remain-dim-3(d) man=cond ‘Poor me, maybe if someone has remained [of all those who have died] …’

7 Multi-clause constructions 7.1 Dependent clauses Urarina has three main types of dependent clauses, each of which has different syntactic properties – subordinate clauses with or without conjunctions, participle clauses, and complement clauses. D-type person inflection (Section 5.1) is used only in subordinate clauses since the other two types constitute non-finite forms. Subordinate clauses with temporal or conditional function are marked by the subordinate marker =ne and do not require a conjunction, as illustrated in (32a), where =ne occurs without conjunction and has a conditional function. When a conjunction follows =ne, the meaning of the clause is determined by the semantics of the conjunction. Conjunction always appear at the end of the dependent clause. The use of the subordinate marker =ne is optional before conjunctions. This is shown in (32b), where the conjunction hana ‘when’ is not preceded by =ne. ajane,]DEP2 [raa-ni-ʉ=te]Main (32) a. [esiɲae ajto-i=ɲe,]DEP1 [dʒaha-ĩ truth say-2=cond come.on-part yes receive-dist-imp=foc [kwara-ka]DEP3! see-1pl.incl ‘If you tell the truth, come on and get her, so that we can see her!’ b. [kʉ-he-ʉrʉ-a hana=te]DEP [muku-e akaʉrʉ raj lintereno go-ipfv-pl-3(d) when=foc burn-3(e) 3pl.pro poss flashlight fwoko]Main lamp ‘When they were going, the lamp of their flashlight went out.’ Participle clauses involve a non-finite verb marked with -ĩ, illustrated in (33). Subsequently, person inflection is marked on the verb in the main clause. Participles have a variety of functions which include sequence, overlap, or adverbial modification of events. There also are two further non-finite verbal suffixes which exhibit a behavior similar to the participle – the suffix -laĩ ‘while’ and the privative -elanaala ‘without’.

1138

Knut J. Olawsky

(33) [kauato-a hãʉ,]DEP1 [here-ĩ]DEP2 [mʉkʉ-ĩ]DEP3 [ajɲa]PP [rãasa-ʉre]V be.nice-3(d) because want-part catch-part with dance-3pl(e) ‘As she was beautiful, they wanted her, caught her, and danced with her.’ There are two basic types of complement clauses, both of which occupy the slot of an O argument. One involves -na, which is used for same subject constructions and the infinitive (see (34a)); the other one is with the subordinate marker =ne. In (34b), =ne in combination with D-type inflection in the dependent clause indicates that the clause refers to a different subject. (34) a. [ʉ-na]DEP [heri-to-anʉ]V come-inf want-ints-1sg.a ‘I (really) want to come.’ b. [tʉrʉ-a=ne]DEP [heri-ji]V arrive-3=sub want-neg:3 ‘He (A) does not want him (B) to come.’ (Lit. ‘arrive’) Both complementation types require a complement-taking verb (control verb) out of a limited set of verbs. The two major types of control verbs of a complement clause are verbs that can take an infinitive clause, which is marked by the suffix -na (e.g., heria ‘want’, kwaaʉka ‘think’, or najnia ‘be able’) and those that control a different subject complement clause, marked by =ne (e.g., esenetaa ‘believe’, kwaasia ‘be afraid’, or aunaa ‘hear’). Some control verbs can take both functions, including heria ‘want’ and kwaasia ‘be afraid’.

7.2 Clause coordination In contrast to the combination of participles or other dependent clauses with a main clause, the conjunctions tɕãe ‘also’ and nete ‘but’ are clear instances of coordination, as both clauses are independent from a grammatical and semantic point of view. Example (35a) demonstrates tɕãe and (35b) illustrates the use of nete. (35) a. [Preceding: ‘Okay, I will buy [the watches] tomorrow.’] [kurete-ʉ]V [kuruata Soru ne-ĩ [tɕãe ʉhʉee+kwi]O also diarrhea+medicine buy-imp two Sol be-part tableta]DEP tablet ‘Also, buy medicine against diarrhea, [the one for] two Soles.’ [Sol is the national currency]

25 Urarina

1139

b. [tabaʉrʉ=te]A [niki]PART1 [kwitʉkʉ-ʉre]V1 [nete]CNJ1 [ɲãe]ADV2 some=foc frus know-3pl(e) but already [itɕa-kʉr-ene]V2 do-pl-neg:3(e) ‘Some [people] know [how to make the mixture], but they do not make it any more.’ However, coordination in Urarina can be regarded as a relatively minor feature of the grammar, based on the following considerations. Firstly, both words described here as conjunctions are not genuine conjunctions, as tɕãe also has adverbial function and nete is a lexicalized form composed of clitics with different functions. The use of a Spanish loan for ‘but’ (pero) by younger people could also reflect the need for an alternative to express coordination. Secondly, these conjunctions are comparatively rare in the database, which implies that they might not represent popular strategies to conjoin clauses. In contrast, clause dependency as expressed by the various means discussed in this chapter is an extremely common way to join clauses. Thirdly, the use of the conjunctions described above has no impact on the syntactic or morphological structure, which is different from some dependent clauses, which may involve special inflection (in subordinate clauses), or specific syntactic positions, such as in complement clauses.

7.3 Reference tracking Reference tracking in Urarina discourse can be problematic, as it is not signaled very clearly, which is especially noticeable in narratives. One reason for this is that Urarina has no syntactic pivot and no formal switch-reference marking. Thus, in combination with omitted arguments, the interpretation of a clause can be ambiguous and much depends on the context and the general knowledge of the listener. There are several ways to indicate a switch of subject reference in discourse: insertion of the respective noun phrase, use of the demonstrative nii ‘that’ in place of a noun phrase, subject marking through person inflection on the verb, and through context. The last case involves no formal marking at all and may result in ambiguities. A change of speakers in quotations can be marked by ideophonic discourse particles.

7.4 Information structure: Focus The most prominent clause level particle in Urarina is the focus marker (Section 6.1). Its use involves attaching an enclitic directly to the focused constituent (with fronting where necessary). The marking of constituents with the focus enclitics =te, =ne, and =na is a pervasive feature of Urarina syntax and occurs with all constituent types in positive declarative clauses and in content questions. These markers are

1140

Knut J. Olawsky

cross-referenced to the subject as marked on the verb of a clause and require the use of the E-form person inflection. An example involving =te for third person focus marking can be found in example (25); example (36) illustrates the use of =ne in a 1sg context. (36) eresi=ɲe ʉ-re-ʉ̃ =ni tomorrow=foc:1sg come-irr-1sg(e)=ass ‘I will come tomorrow.’ The function of the focus markers ranges between indicating focus and topic, and their interpretation depends considerably on the context. For instance, (36) appears to mark contrastive focus by highlighting ‘tomorrow’, whereas the examples in (25) appear to topicalize a fronted subject. The insertion of a focus marker is obligatory only with a fronted subject, except in negative clauses. The marker may occur after a constituent that is in its original position, such as an object or an adverb. In this case, it tends to take focus function.

7.5 Discourse strategies and expressive elements The organization of discourse in Urarina texts varies according to genre. The main types of discourse on which current research is based are narratives and conversations. Narratives may involve the frequent use of the hearsay evidential hetau. Conversations exhibit much more frequent use of particles than narratives. This is reasonable, as most particles express information about the speaker’s emotional state – which may be less relevant in a third-person narrative. This is similar to the use of the attitude markers =ra (expressing urgency as well as truthfulness) and =ta (a negative affect marker that can convey a wide range of negative emotions, including dislike, anger, shock, and disappointment), which exclusively occur in direct discourse in the traditional language. See (37a) for =ra and (37b) for =ta. (37) a. hja-rito-anʉ=ra! urinate-ints-1(a)=aff.emph ‘I (really) have to urinate!’ b. dʒɑnʉna kanʉ bʉa basihjaʉ-i=ta? why 1sg.pro bag steal-2=neg.aff ‘Why have you stolen my bag?’ Features that are not normally part of the phonological system are commonly used to emphasize dramatic events, or to “spice up” the description of actions or events, especially in narratives. Most typically, such exceptional sound patterns occur in

25 Urarina

1141

ideophones. This may include the gemination of consonants and extreme vowel lengthening, which is clearly beyond the normal length of long vowels. The most significant feature of morphological marking related to the organization of discourse relates to the different person inflection types discussed in Section 5.1. As pointed out there, the E-form is not only the obligatory inflection type for clauses that contain a focus marker, but it is also the typical inflection type for a particular text genre – narratives. In contrast, the A-form is the preferred type used in conversation. Another characteristic of narratives, which may be described as reflecting phonological and morphological factors, is reduplication. This tends to create a more lively effect in the narration of stories, adding intensity, authenticity, or iterative action. Sometimes this can be expressed by triplication of the verbal root. There are a number of syntactic properties that are typically observed in Urarina narratives. While omission of constituents may occur, the insertion of syntactic material is much more predominant and may include iconic repetition, tail-head style clause chaining, and concatenation of several dependent clauses into very long sentences. The introduction of a quotation in traditional narratives can be accompanied optionally by certain ideophonic discourse markers. Occurring only in traditional narratives, hẽ is a frequent discourse introducer which indicates that someone is starting to say something. This can also mark a change of speaker in quotations when no quotative verb is present. The adversative particle niki is relevant for discourse structure, as it does not directly refer to the clause in which it occurs. Instead, it marks in anticipation that something negative is going to happen as a result of the action described, like adding ‘but…’ at the end of a clause. In (38), niki gives the listener an idea that the described hunting trip will involve an unexpected (in this case negative) event. (38) ɲoaelʉ hetau=te niki hano-a kʉ-e raj komasaj ajɲa earlier hrs=foc frus hunt.overnight-ntr go-3(e) poss wife with lejhiĩi katɕa one man ‘Once a man went hunting overnight with his wife (but …).’ Another typical feature of narratives is common expressions that begin and end a story. Though there are no fixed literal formulae in this regard, the expressions used in this context very much resemble each other. Traditional narratives typically begin with ɲoaelʉ hetau=te (‘in early days reportedly’), which can be translated as ‘once upon a time’ and end with ɲae saa (‘already finished’) to mean ‘this is the end’.

1142

Knut J. Olawsky

8 Summary In conclusion, Urarina can be characterized as a language with typologically interesting features, most prominently exhibited by its OVA/VS constituent order – an order that was previously assumed to be non-existent or extremely rare. The regularity with which it occurs in Urarina and the fact that deviations can be predicted or explained adds further weight to this structural peculiarity. A polysynthetic verb system, person inflection classes, and emotional expressions such as through the frustrative form make Urarina an interesting language for further study.

9 References Figueroa, Francisco de & Cristóbal de Acuña. 1986. Informes de Jesuitas en el Amazonas 1660– 1684. (Monumenta Amazónica). Iquitos: IIAP-CETA. Greenberg, Joseph H. 1960. The General Classification of Central and South American Languages. Men and Cultures: International Congress of Anthropological and Ethnological Sciences 5. 791–794. Hyman, Larry M. 2009. How (not) to do phonological typology: the case of pitch accent. Language Sciences 31. 213–238. Manus, Phyllis. 1992. Proposiciones de desarrollo y de apoyo y su expresión en las oraciones del urarina. Revista Latinoamericana de Estudios Etnolingüísticos 7. 15–45. Olawsky, Knut J. 2006. A Grammar of Urarina. (Mouton Grammar Library 37). Berlin: Mouton de Gruyter.

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

26 Wã́nsöjöt (Puinave) 1 2 3 4 5 6 7 8 9 10

Introduction Phonology Word classes and morphological structure The noun phrase The verb complex Simple clauses Complex sentences Topic and focus Conclusion References

1 Introduction 1.1 Demographics and sociolinguistic background ́ ́ ̀ ɤ̃t̀ ] (puin1248) is spoken by approximately 6800 people in 32 comh Wãnsöjöt [ˈw̃ ãnsɤ̃ munities along the shores of the Inírida River in Guainía, Colombia and about 470 people in ten communities along the Orinoco River on the border between Colombia and Venezuela. The language is also known as Puinave, Camaku del Guaviare, or Camaku del Inírida; see Girón (2008: 1–2) for discussion, as well as de Wavrin ́ and Girón (1998). Up until the 1950s, the Wãnsöjöt communities had at most 80 people each and were quite mobile, moving every seven to ten years in order to find more arable land and as a means for reducing social tensions. However, in recent years, as a result of evangelization and the introduction of modern institutions such as schools and health posts, these settlements have become fixed and their populations have increased, sometimes to upwards of 250 people. In addition, many families have migrated to the city of Inírida or Puerto Ayacucho. There are no known ́ dialects of Wãnsöjöt, but speakers in Colombia report that settlements further east along the Orinoco, such as Patacame and Caño Minisia in Venezuela, exhibit lexical ́ and intonational differences compared to the other areas where Wãnsöjöt is spoken. ́ The Wãnsöjöt communities are largely riverine settlements, with clearings within forested areas for the cultivation of manioc (Manihot esculenta). In addition to fishing and hunting, they also participate in the extraction and sale of forest prod-

Adam J. R. Tallman, Friedrich Schiller Universität Cynthia Hansen, Grinnell College Jesús Mario Girón, Universidad Nacional Abierta y a Distancia (UNAD), Colombia https://doi.org/10.1515/9783110432732-013

1144

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Wã´ nsöjöt

26 Wã́nsöjöt (Puinave)

1145

ucts such as piassava palm fiber (Leopoldinia piassaba), which is typically used for ́ making brooms and rope. The traditional culture of the Wãnsöjöt started to change with the introduction of Evangelical Christianity and incorporation into local economies. Shamanism and traditional beliefs have been largely replaced by Christian ́ mythology since the middle of the twentieth century.1 All speakers of Wãnsöjöt are bilingual in Spanish, except some very old speakers (Girón 2008: 441). According ́ to a linguistic auto-diagnostic carried out in Colombia in 2009, the ethnic Wãnsöjöt population reached 6390 individuals, of which 5400 were older than two years old ́ and fluent native speakers of Wãnsöjöt (87 %); approximately 476 (7.7 %) do not ́ speak Wãnsöjöt or have only a basic understanding of the language.2 A reported 84 % of children and teenagers use the language daily. However, despite the apparent vitality of the language, some community leaders do not think the language will last more than two generations, due to the serious threat imposed by Spanish. Spań ish has entered Wãnsöjöt communities through economically-driven contact, the education system, and the media, even in the most distant communities. As a consé quence, in recent years there has been an effort to introduce Wãnsöjöt in the ́ schools, with Wãnsöjöt courses being offered alongside language revitalization programs, while communities and recent regulations seek to reassess ancestral culture and language.

1.2 Classification and distant relationships ́ While earlier research proposed that Wãnsöjöt was genetically related to “Makú” ́ languages (see below), Wãnsöjöt is usually classified as an isolate, although recent research suggests it may be distantly related to Kakua-Nukakan. Epps and Bolaños ́ (2017) observe similarities between Wãnsöjöt and Kakua-Nukakan in their respective pronoun inventories and basic vocabulary. They also present a number of regular sound correspondences between the two languages. Using a likelihood model of potential cognancy applied to the Swadesh list, Brown (2017) also argues that there ́ is evidence for a relationship between Wãnsöjöt and Kakua-Nukakan. Epps and Bolaños (2017) argue that the relationship cannot yet be confirmed, however, because ́ of the lack of morphological similarity between Kakua-Nukanan and Wãnsöjöt and the possibility that apparent cognates emerged from contact. The relationship is

́ 1 Girón (2008: 2–3) notes that the Wãnsöjöt share, along with other groups in the region, the ritual ́ tradition and myths of the Yuruparí (referred to as bon among the Wãnsöjöt), and that despite evangelization efforts from Christian missionaries, a single local group resisted these efforts and preserved a few traditional songs and dances up until the late 1990s/early 2000s. 2 These data were obtained by Jesús Mario Girón. Figures can be accessed at https://www.mincul​ tura.gov.co/SiteAssets/documentos/Despacho/Lenguas/​infor​me_​preliminar_autodiagnostico_so​cio​ lin​guis​iptico.pdf.

1146

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

also difficult to assess at this point because of the relative paucity of information on Nukak (see Volume 3). Other distant relationships are not as well supported. Based on some similaŕ ities in lexical forms, Rivet and Tastevin (1920) proposed that Wãnsöjöt was related to the Makú languages (see Epps and Bolaños (2017) for the languages typically assumed to be part of this group, and why Naduhupan and Kakua-Nukak are better designators today). Other authors assumed this relationship but did not explicitly defend it (Mason 1950; Huber & Reed 1992; Martins 2005; see also Girón 2008: 418– 433). However, Epps and Bolaños (2017) argue that a genetic relationship between ́ Wãnsöjöt and Naduhupan is unlikely; sound correspondences are not robustly supported, there are no similarities in the pronoun inventory, and the few similarities in basic vocabulary are likely due to chance or borrowing, not common ancestry.

1.3 Prior research ́ Until recently there has been little information on Wãnsöjöt. Jean Caudmont (1954) provided a preliminary phonemic analysis of the language. Ellen Peirsen (1980) discussed the pragmatic function of a few markers of tense, aspect, and modality. Ethnographic studies were done in the second half of the twentieth century (see Pabón 1979; Camacho 1982; Triana 1985), but they do not contain linguistic data. The first ́ substantial text of Wãnsöjöt was a translation of the Bible initially written by Sophia Müller and more recently revised by the missionary Larry (Lorenzo) Richardson, who has prepared a new version of the New Testament. Girón (2008: 14–17) provides ́ a list of other documents that contain information concerning the Wãnsöjöt language, including word lists produced by travelers, ethnographers, and missionaries. Most of the data for this sketch are from the descriptive grammar by Jesús Mario Girón (2008), supplemented by the study on segmental phonology by Girón (2006) and tone by Girón and Wetzels (2007). In scope and coverage, Girón’s grammar greatly surpasses all other previous materials. By and large the analysis presented below follows Girón (2008), but in certain cases, the analysis is presented in a simplified form for reasons of space and to make the description correspond with the conventions used in the rest of this volume. In what follows, we diverge from Girón’s phonological representations in that we use instead of and instead of . Examples pulled from Girón (2008) have also been edited to reflect the most recent morphological and tonal analyses.

2 Phonology ́ Wãnsöjöt has a relatively small segmental phonemic inventory and a tonal system that requires more research. There is some degree of complexity relating to the high number of allophonic variants that each phoneme exhibits.

26 Wã́nsöjöt (Puinave)

1147

2.1 Segmental phonology 2.1.1 Segmental categories ́ Wãnsöjöt has ten consonant phonemes. The voiced stops surface as obstruent oral voiced sounds [b, d, ɾ] before oral syllable nuclei and as nasals [m, n] before nasal syllable nuclei.3 The alveolar voiced stop is realized as a flap in intervocalic position. This flap has an allophone [ɽ] that is sociolinguistically conditioned (see Girón 2008: 54). Glides [j, w] are true approximants at the onset of a syllable and, like the voiced ̂ stops, are nasal before nasal vowels, for example, /jĩk-ot/ > [ɲĩgot] ‘bead-pl’, ́ ɯ̀ k] ‘bag-shaped fishing net’. The consonant phonemes and their ́ ɯk/ > [w̃ ɤ̃b /wɤ̃b allophones are displayed in Table 26.1. ́ Wãnsöjöt has 11 vowel phonemes. Seven of these vowels are oral, and four are nasal; see Table 26.2.

Tab. 26.1: Consonant phonemes. labial

alveolar

vcless stop

/p/

/t/

voiced stop

/m/ → [mṼ] → [bV]

/n/ → [nṼ] → [dV] → [VɾV]~[ɽV]

fricative glide

palatal

velar

glottal

/k/ → [k, g]

/ʔ/

/s/

/h/

/w/ [w, ß, w̃ , u]

/j/ [j, ɟ, ɲ, i]

Tab. 26.2: Vowel phonemes. front oral high

oral

nasal

unrounded

oral

nasal

u i

ĩ

ɯ

rounded unrounded

low

nasal

back

rounded unrounded

mid

central

e a

o

õ

ɤ

ɤ̃

ã

3 We have chosen to orthographically represent labial and alveolar voiced stops /m, n/ as m or b, and n or d, depending on whether the nucleus that follows the consonant is nasal or oral.

1148

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

́ Wãnsöjöt exhibits a rounded and unrounded distinction in its back-vowel phonemes, as can be seen in (1). (1)

a. [hɯ́ ʔ] hɯ́’ ‘axe, iron’

b. [hùʔ] hu’ ‘hawk’

́ Wãnsöjöt makes a phonemic distinction between nasal and oral vowels, (2), but this contrast is only evident for a subset of the vowel inventory. There is no nasal contrast in the high back vowels [*ũ], [*ɯ̃ ], nor is there a mid front nasal vowel [*ẽ]. (2)

a. [hahɤ̂j] ha-hɤ̂ j 3sg-clitoris ‘her clitoris’

b.

[hahɤ̃ĵ ] ha-hɤ̃̂ j 3sg-whistle ‘S/he whistles.’

́ Wãnsöjöt’s syllable structure is (C)V(C). Complex onsets can surface in certain circumstances. Glides in coda position become obstruent onsets of the next syllable; contrast (3a) where the glide is in onset position with (3b), where the glide is initially in the coda of the penultimate syllable. The labio-velar glide [w] becomes the bilabial fricative [β] in this context, and the palatal glide [j] becomes the palatal stop [ɟ], as in (4). Word-finally, glides participate in diphthongization, (5). (3)

a. [ndig.wóʔ] dikwó’ worm ‘worm’

b. [ha.mã.hé.βì] ha-mã-hêw-i? 3sg-2sg-look.for-int ‘Do you want it?’

(4)

a. [ijík] i-jík attr-large ‘large’

b. [nɟùk] juk ‘hollowing stick’

(5)

a. [dáp˺ɟuj] dap-jû j hand-clothing ‘ring’

b. [hàwǔj] ha-wǔ -j? 3sg-happen-int ‘Did this happen?’

VV combinations can occur as long as they are word-internal and each vowel is parsed as the nucleus of the syllable, (6).

26 Wã́nsöjöt (Puinave)

(6)

a. [sùk˺.sù.ót] suksu-ot flower-pl ‘flowers’

1149

b. [ha.hûn.sì.àk] ha-hû nsiak 3sg-fin ‘(its) fin’

Complex onsets can emerge on the surface from underlying CGlideV(C) sequences in the following circumstances: (i) the onset consonant is voiceless; (ii) the nuclear vowel is [–high]; (iii) the total number of sounds in the syllable does not exceed four. Examples of complex onsets are provided in (7). Elsewhere, the vowels are realized as distinct nuclei, as in (6) (see Girón 2008: 30–42 for a precise formalization of vowel syllabification rules). (7)

́ a. [mãj.hwa] mãjhwa ‘medicinal plant sp.’

b. [swâm.kà] swámka ‘achiote (Bixa orellana)’

Although all consonant phonemes can occur in onset position, voiced stops /n, m/ will surface as prenasalized oral stops word-initially or when the preceding syllable ends in a nasal segment, but only if the voiced stop precedes an oral vowel. The glide /j/ is realized as a prenasalized palatal plosive in these environments. (8)

a. b. c. d.

/móp/ /nɯ́ t-ot/ /mõ-mon/ /jăn/

> > > >

[mbóp] [ndɯ́ ɾə̀t] [mõ-̀ mbòn] [nɟăn]

‘arrow’ ‘ropes’ ‘indf.pro-man; someone’s husband’ ‘kingfisher’

The distribution of sounds is more restricted in coda position. For instance, the voiced stops /n, m/ do not surface as voiced oral stops in codas; for example, [n] in (8). This is true whether the preceding vowel is oral or nasal, for example, pupdɯ́m ‘belly’ and kãn̂ ‘corn’. Voiceless oral stops are unreleased in all coda positions, (9). The velar stop /k/ can optionally be realized as voiced [g] in coda position, (9c). The fricatives [s] and [h] do not occur in codas. (9)

̌ / > [çɤ̃p ̌ ˺] a. /sɤ̃p ‘serpent’ ‘Cuniculus paca’ b. /nét/ > [ndét˺] c. /kǒk/ > [kŏg˺] ~ [kŏk˺] ‘pepper’

2.1.2 Segmental morphophonology There are a number of morphophonological processes that influence the realization of phonological segments; that is, changes in form from underlying phoneme to

1150

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

surface representation require reference to structural categories such as roots, affixes or words. A velar deletion rule eliminates morpheme-final /k/ at root-root and root-suffix junctures (Girón 2008: 82–84). An example is provided in (10), where /k/ deletes at the boundary between the adjective root pîk ‘black’ and the nominalizer -ju. (10) [kapíju] ka-pîk-ju 3pl-black-nmlz ‘type of paste/paint to make something black’ According to Girón (2008: 82–84), this deletion rule is not “a systematic phenomenon” in the sense that it sometimes does not apply. It is restricted to certain rootroot or root-suffix junctures and not others. A voicing and/or flapping process occurs intervocalically. Labial and alveolar plosives /p, t/ are realized as [b, ɾ] intervocalically within a single morpheme or a combination of bound morphemes. Intervocalic voicing and flapping is illustrated in (11) through (13). Voicing and flapping does not occur at the boundary between free forms, nor at the boundary between certain affixes and roots, as illustrated in (13). In (13), neither /t/ nor /k/ are voiced or flapped despite occurring intervocalically. Note that there is no phonemic distinction between voiced and voiceless velar stops. (11) a. [mbóp˺] bóp arrow ‘arrow’

b. [mbóbòt˺] bóp-ot arrow-pl ‘arrows’

(12) a. [kɤ́t˺] kɤ́ t star ‘star’

b. [kɤɾot˺] kɤt-ot star-pl ‘stars’

(13) [ɲãmdá] [kàtɯkɯ́ t˺] jãm-da ka-tɯ̌k-ɯ́-t 2sg-ass 3pl-tickle-ag.nmlz-pl ‘You tickle them.’ A number of morphophonological processes are specific to individual morphemes rather than more abstract structural categories such as root or suffix. For instance, the general process of voicing and flapping described above is blocked at a juncture with allative -a (Girón 2008: 87), as illustrated in (14).

26 Wã́nsöjöt (Puinave)

́ (14) a. [kamãpa] ̂ ka-mãp-á 3pl-fish-all ‘to their fishing’

1151

b. [pitatá] pitat-á hospital-all ‘to the hospital’

There is an agentive suffix that also displays morpheme-specific morphophonology. It is realized by copying the same vowel that precedes it or as /m/. The conditions for the application of this rule are currently not well understood but seem to be lexically conditioned. The precise inventory that conditions the /m/ form requires future research. The vowel realizations are illustrated in (15a) and (15b) and the /m/ realization in (15c). ́ (15) a. [itĩmĩ] ̂ i-tĩm-i attr-lick-ag.nmlz ‘he who licks’

́ ʔ] ̀ b. [kàhãnã ̂ ka-hãn-ã’ 3pl-hide-ag.nmlz ‘one who hides’

c. [hàʔɤ̀ɟúm] ha-’ɤjú(k)-m 3sg-bring-ag.nmlz ‘the one who brings’

The plural morpheme -ot displays a somewhat similar type of allomorphy in that its vowel assimilates to the preceding vowel, but it only occurs in specific lexical items ending in the low vowel [a], and does not appear to be morphologically or phonologically motivated. The contexts conditioning this allomorphy require future research. An example is provided in (16) (see Girón 2008: 93–95 for more details). (16) a. [képaʔat˺] képa-’ot ancient-pl ‘ancient ones’

b. [káwaʔat˺] káwa-’ot chicken-pl ‘chickens’

̀ ʔót˺] ~ [wõgɽɤ̀ ̀ ót˺] c. [wõgɽɤ̀ wõkdɤ́ -’ot sweet.manioc-pl ‘sweet manioc (more than one)’

Causative -’ɤ́ k displays morpheme-specific allomorphy. Its initial glottal stop always deletes unless it is right-adjacent to the phoneme /k/ (Girón 2008: 85). In these cases, the glottal stop is retained but the [k] is deleted, as in (17b). (17) a. [hàmòtàɾɤ́k˺] ha-mõ-tat-’ɤ́ k 3sg-indf.pro-point-caus ‘to point (a finger at something)’

̀ ʔɤ́k˺] b. [hàmõʝɤ̀ ha-mõ-jɤ̀ (k)-’ɤ́ k 3sg-indf.pro-leave-caus ‘throw out’

1152

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

2.2 Suprasegmental phonology 2.2.1 Suprasegmental categories ́ Wãnsöjöt tones are associated with syllables. Underlyingly, a syllable can contain H, HL, or LH tone melodies. The tonal features of the underlying contour tones can associate to adjacent syllables on the surface (syllables they are not associated with underlyingly) through a variety of tone sandhi processes. According to Girón and Wetzels (2007), syllables do not contain a unitary L tone feature underlyingly. They are assigned an L tone either through tone sandhi rules that break up contour tones across more than one syllable, or they are assigned to toneless syllables by default. The evidence for this is discussed below. L tones are marked as V,̀ H tones as V,́ HL ́ melodies as V,̂ and LH melodies as V.̆ The tonal phenomena of Wãnsöjöt are highly complex and the interested reader should consult Girón (2008) and Girón and Wetzels (2007) for more details. Monosyllabic syllables exhibit four surface tone melodies: H, L, HL, and LH. These tonal melodies are illustrated with minimal pairs in (18) through (23). Example (18) contrasts H and L tones; (19) contrasts H and HL; (20) demonstrates a minimal triplet contrasting H, LH, and L; and (21) to (23) distinguish L and HL and LH, and HL and LH. (18) a. dén ‘woman’

b. dèn ‘tick’

(19) a. -kák ‘count’

b. -kâ k ‘cook’

(20) a. há’ ‘pendare (Couma macrocarpa)’

b. hǎ’ ‘canoe’

c. hà’ ‘stone’ (21) a. sɯ̀ k ‘suck’

b. sɯ̂ k ‘clean’

(22) a. kõǹ ‘owl’

b. kŏn ‘sassafras (Ocotea cimbarun)’

(23) a. kãn̂ ‘corn’

b. kãn̆ ‘hammock’

However, the four-way contrast is not paradigmatic in forms with more syllables. Only four tonal melodies are found for bi- and trisyllabic forms, as illustrated in Table 26.3.

26 Wã́nsöjöt (Puinave)

1153

Tab. 26.3: Contrastive tonal melodies in multisyllabic forms. tonal melody

example

trans.

example

trans.

bisyllabic

L.L L.H H.L HL.L

[kò.tèʔ] ̀ .síʔ] [ɲãm [í.ɾàk] ̂ ]̀ [ɲã.ɲĩ

‘bird sp.’ ‘fish sp.’ ‘fish sp.’ ‘here’

[pì.sàw] [wàk.dúk] [dú.mãt̀ ] [sôj.kàk]

‘heron sp.’ ‘spider’ ‘now’ ‘monkey sp.’ (Lagothrix lagotricha)

trisyllabic

L.L.L

[dà.mã.̀ nãʔ̀ ]

‘iguana’

[à.kà.ɟàʔ]

L.L.H L.H.L H.L.L

[jù.tà.ɾák] [ɟɯ̀ k.sí.kɯ̀ j] ̀ ] [ɲṍ.pã.̀ wãn

‘cockroach’ ‘turtle sp.’ ‘cloud’

[ɟè.sì.pík] [mõ.̀ ɾé.nõ]̀ [mã́.kà.pàʔ]

‘tree sp.’ (Anacardium occidentale) ‘lip’ ‘white heron’ ‘catfish sp.’

The following constraints on the position of underlying tones can account for these tonal patterns. Patterns that contradict these constraints are not permitted. (24) a. Initial position: LH, HL, H, Ø (monosyllabic, bisyllabic, trisyllabic) b. Internal position: H, Ø (trisyllabic) c. Final position: H, Ø (bisyllabic/trisyllabic) (25) a. b. c. d.

One H-tone per form (culminativity) L-tones assigned to toneless syllables H of LH associates to the syllable to its right HL.L.L is banned

In general, these conditions apply within complex stems as well. Notice that an important distinction between HL and LH tonal melodies emerges from the data above. An underlying H feature of an LH contour breaks off and associates to the right (stated in (25c)), but the same process does not apply to the L feature of HL contours, which is why HLL patterns are attested. This process can further be illustrated with the minimal pair in (26). An underlying LH.Ø surfaces as L.H, and underlying HL.Ø surfaces as HL.L. (26) a. /jǔm/ → [ɟùmót] jù m-ót jungle.mother-pl ‘jungle mother (type of spirit)’

b. /jûm/ → [ɟûmòt] jû m-ot mosquito-pl ‘mosquitos’

2.2.2 Suprasegmental morphophonology Suprasegmental morphophonology refers to the way that tonal melodies surface across morpheme or constituent junctures. Some of the tone sandhi processes men-

1154

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

tioned above also apply in complex forms. However, some of the constraints put forth in the previous section need to be relaxed to account for the tonal melodies of multimorphemic combinations. Finally, as with segmental morphophonology, some morphemes display morpheme-specific tone sandhi rules.

2.2.2.1 LH > L.H across root-suffix / suffix-suffix juncture An L floating tone association rule applies at the juncture between roots and affixes. This rule takes an underlying LH contour on a root and reassigns the underlying H tone to a syllable to its right on the surface (see Giron & Wetzels 2007: 138 for details). (27) [àsàátdà kàdĭʔ] ̌ a-sã-at-da ka-dǐ(k)’ 1sg-left-abl-ass 3pl-be.there ‘They are there to my left.’

2.2.2.2 High tone insertion in complex forms A complex form composed of one or more roots, affixes, or clitics is assigned an H tone two syllables to the right of the left edge of the root (the root must have an underlying L tone; see Girón 2008: 139–140 for details). The rule is illustrated in (28) and (29). The root in (28) is -dok ‘tongue’. Two syllables from the left-edge of the root is the morpheme -da assertive, and this is where H tone is assigned. The root in (29) is hupɤ́ t ‘onward’. Two syllables from the left-edge of the root is -a allative, and H tone is assigned there. This rule does not apply to simplex forms (compare to Table 26.3). In the examples below the first line represents the surface tones and the second line represents which tones are present underlylingly. (28) [hà-dòk-hèt-dá] hadòkhetda 3s-tongue-com-ass ‘(The cat washes himself) with his tongue.’ (29) [kàhùpɤ̀t-áwìndà] ka-hupɤ́ t-a-win-da 3pl-onward-all-gen-ass ‘the tallest of them’

2.2.2.3 Tonal affixes A small number of affixes have underlying tones. Some of these tonal affixes condition morpheme-specific tonal allomorphy. For instance, the topic marker -hɤ́ j has

26 Wã́nsöjöt (Puinave)

1155

an H tone that is never deleted (30), (31). The tonal behavior of this morpheme may be connected to its pragmatic function. ̀ (30) ojàt-hɤ́ j hɤ́ t epìn-òt-hɤ́ j kà-wɤ̃ ̀ p-tó-dik-mã-dì so-top these people-pl-top 3pl-shoot.arrows-quant-nfin-rpt-pst ‘So, these people, used to shoot many arrows they say.’ (31) hɤ́ ’ɤ̀ hǎ-hɤ́ j Hwàn-pín-dà ’ô’o dem canoe-top Juan belong-ass 3sg.pro ‘As for that canoe, it belongs to Juan.’ The restrictive -wí morpheme interferes with the tonal melody of the previous element. The morpheme can delete an H tone of a prior LH contour. (32) [óìdìkwí] o-i-dǐk-wí 3sg-attr-be-rest ‘always’ (33) [ʔótdi hàkàwúʔàk bɤ́át hàdìgwí-tèp] ’ót-di ha-ka-wu’ak bɤ-at ha-dĭk-u-wí-tèp 3pl-pst 3sg-3pl-to.call indf-obl 3sg-be-ade-rest-emph ‘They called someone who was around.’ The nominalizer -ju has an underlying H tone, but it is deleted when there is an H tone on the root. We see this deletion in (34b), where the nominalizer is then subject to the default assignment of low tone, and in (34c), where the LH combination on the root is realized as L.H across the root and nominalizer suffix. (34) a. kɯk ‘to sweep’

>

kɯ̀ (k)-jú sweep-nmlz ‘broom’

b. bék ‘to finish’

>

hà-bék-jù 3sg-finish-nmlz ‘end of it’

c. jǒj ‘to laugh’

>

mõ-jòj-jú indf.pro-laugh-nmlz ‘laughter’

Some morphemes display a different tonal melody depending on their position. This is illustrated for the mirative -kõm in (35), which has a low tone as a prefix and a HL tone as a suffix.

1156

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

̂ ̀ ̀ ̀ m (35) hehsit! hà-kõm-hèj-mõ k-mà dén-pèk-mã-kõ intrj 3sg-mir-well-say-rpt woman-adult-rpt-mir ̀ hà-kõm-sɤ̀ k-mã ̀ hɤ́ ’ɤ̀ jòdán-pèk 3sg-mir-female.in.heat-rpt dem jaguar-adult ‘Uh! Immediately it came because of the calling. It was an adult female jaguar in heat.’ ́ Wãnsöjöt has a tone erasure rule that deletes all H tones in a nominal compound except the leftmost one. This is illustrated in (36), where the first component of the compound retains its high tone, but the other two components do not. (36) [ʔám pòk pì] ’ám pók pí bee juice be.black ‘black honey’ ́ Wãnsöjöt also has a class of verbs whose root forms have extended exponence such that they can be separated by certain functional morphemes (see Section 5). The two parts of the separable verb can each have their own high tone, as shown in (37b). (37) a. bìdút-dà kà-hétɤ̂ ̛j-ɤ-t 1pl-ass 3pl-wait-ag.nmlz-pl ‘We (are) who are waiting for them.’

̀ ̀ -tɤ́ j b. hà-hé-mã-ũ 3sg-wait-2sg-imp-wait ‘Wait for him.’

3 Word classes and morphological structure A broad distinction between lexical and grammatical categories can be made in ́ Wãnsöjöt. Lexical roots are an open class and canonically tone-bearing. Grammatical morphemes form a closed class, almost all atonal (with a few exceptions, see Section 2.2.2), and they select for a lexical morpheme such that they cannot surface without one. That said, there is a closed class of tone-bearing, free-standing, monomorphemic particles that do not neatly fit into either of these categories: they express both lexical and grammatical categories, namely adverbial concepts of time, manner, and place, as well as negation, and they do not take any inflectional morphology (see Girón 2008: 176–177). These characteristics are summarized in Table 26.4. Lexical categories include nouns and verbs, which can be robustly distinguished. These properties are described below and summarized in Table 26.5. Among the properties that distinguish verbs and nouns is that verb roots require some index of an argument, either as a full NP, a pronoun, or a pronominal prefix, illustrated in (38) and (39), whereas nouns do not require any index as in bon ‘man’ (39). Although

26 Wã́nsöjöt (Puinave)

1157

Tab. 26.4: Wã́nsöjöt morphological types. open or closed?

tone-bearing? semantic characteristics

additional characteristics

lexical roots

Open

Yes

Primarily verbs and nouns; a few bare adjective roots

Verb roots require an argument; nouns do not

grammatical morphemes

Closed

Atonal with some exceptions

Modify nouns and verbs

Always bound to a lexical root

particles

Closed

Yes

Adverbial concepts Free and monomorphemic; of time, manner, and do not take any inflectional place; negation morphology

there are inalienable nouns that exhibit obligatory possession with these same pronominal prefixes, this bound pronominal marking does not encode an argument (see Section 4.4). (38) bɤ́ n a-péwãj-at ha-’ɤnṍk rem.pst 1sg-brother-erg 3sg.o-bring ‘My brother already brought him/it.’ (39) bon ka-ha-dû k-nĩ ka-jɤ́ k-dik ka-’ot-té-at man 3pl-3sg-see-rec.pst 3pl-come.out-nfin 3pl-house-iness-abl ‘The man saw them leave their house.’ Nouns, but not verbs, can occur in non-verbal predicate constructions, (40). (40) hɤ́ ’ɤ hɤ́ j a-tóda dem top 1sg-female.cousin ‘She is my cousin.’ Verbs, but not nouns, can be derived into participial-like forms with the prefix i-, and in so doing, express adjectival meanings, (41). This is the major distributional difference between verbs and nouns. (41) a. júk c. pû n

‘come’ ‘stick’

b. i- júk d. *i-pû n

‘one who comes, coming’

Verbs, but not nouns, can be modified by the nonfinite marker -dik, as in the form ̂ mĩnjuk-dik ‘arriving’, (42).

1158

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Tab. 26.5: Characteristics of nouns and verbs. characteristic

nouns

verbs

Take an obligatory argument

No. But inalienable nouns must be marked with a pronominal prefix indicating the possessor.

Yes. The argument may be a full NP, pronoun, or pronominal prefix.

Occur in non-verbal predicates

Yes

No

Used to derive adjectival meanings No with the prefix i-

Yes

Marked with nonfinite marker -dik

No

Yes

Marked with gender and number

Yes

No

̂ (42) ka-bok ka-mĩnjuk-dik 3pl-walk 3pl-arrive-nfin ‘They arrived walking.’ There are also a large set of categories that can modify only verbs and not nouns, and vice versa (see Girón 2008: 145–208). For instance, gender and number are only relevant for nouns (see Sections 4 and 5). ́ There appear to be few bare adjective roots in Wãnsöjöt. However, adjectival concepts are productively derived with the prefix i-. A class of adverbs cannot be defined in a consistent way. As with many languages, the class of morphemes that encode adverbial meanings are too heterogeneous for them to be designated a single part of speech class. In general, closed class morphemes are highly heterogeneous in terms of their distributional properties. We will refer to closed class elements as “grammatical categories” (as opposed to the lexical categories of nouns and verbs); however, they vary widely with respect to how much lexical content they express. Grammatical morphemes will be discussed in the sections dedicated to nouns and verbs, because most can be seen as modifiers of these categories. ́ Wãnsöjöt can be considered agglutinating in the sense that there is typically a one-to-one correspondence between meaning and (segmental) form. An important exception to this generalization is the fact that certain types of verb bases in the language can be separated by certain morphemes, but each of the separated pieces cannot be understood as a unitary morpheme.4 See example (37), where the verb

4 In other words, verb bases display “extended exponence” (see Anderson (2015) and Blevins (2016: Ch. 2) for a review of the different types of deviations from agglutinative patterns).

26 Wã́nsöjöt (Puinave)

1159

base hétɤ̂ j ‘wait’ is contiguous in (37a) and split in (37b). Separable verbs are also discussed in Section 5.2. ́ With respect to the analytic-synthetic status of Wãnsöjöt, the language can be considered polysynthetic based on the following properties: (i) pronominal arguments can be bound and obligatorily left-adjacent; (ii) sentences can be composed of long strings of bound elements (holophrasis); (iii) this long string of bound elements usually corresponds to the stress domain and is the domain for tone-sandhi rules; (iv) many lexically heavy elements occur in such strings of bound morphemes; (v) compounding (noun-incorporation) processes are evident in the verbal complex. The convergence of these properties is illustrated in (43), except for nounincorporation which is discussed in Section 6.2 and tone sandhi rules which are treated in Section 2. ́ (43) ja-je-si-pathéj-wã́j-tenõk-ja 2pl-eat-prosp-sprl-eat-dim-fut ‘You all will still be eating.’ A few caveats need to be mentioned regarding the polsysynthetic designation. First, pronominal arguments such as the prefix ja- in (43) are not agreement markers that can coindex an argument with an overt NP or pronoun. In most constructions, these prefixes and full NP arguments are mutually exclusive; see Sections 5.1 and 6.1. Second, there are a few reasons why the form provided in (43) may not necessarily be a word with respect to morphosyntactic and phonological principles. The verbal domain occasionally corresponds to two stress domains, if certain adverbial elements intervene between the pieces of the verb base. There are also some forms that can interrupt holophrastic structures, such as the aspectual morpheme pathéj ‘still, for a long time’ that splits the verb ‘eat’ in (43) (discussed in Section 5.3). The affixal status of many of the elements that can occur in such large verb complexes can also be called into question because of their distributional freedom and potential to display variable orderings. The categorization of the language as “polysynthetic” hinges on a more detailed analysis of wordhood domains, which requires future research (cf. Bickel & Zúñiga 2017; Haspelmath 2018 for recent discussion). It is beyond the scope of this sketch to provide a detailed discussion of wordhood ́ domains in Wãnsöjöt. A final comment should be made concerning the distinction between affixes ́ and clitics in Wãnsöjöt. While morphemes are classified variously as affixes or clitics based on their positions in the noun and verb templates, a general definition of how the affix-clitic distinction should be drawn does not appear in Girón (2008). Given our preliminary understanding of the language, we do not feel that such a sweeping distinction would be helpful, even for expositional purposes. Thus, we will not make an affix-clitic distinction here. Instead, we focus on providing a general characterization of the categories that modify nouns, verbs, and clauses with some preliminary comments on how they seem to be distributionally distinct.

1160

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

4 The noun phrase 4.1 Nominal morphology Noun roots can be either free or bound. Bound nominal roots are almost always inalienably possessed nouns: the possessor may be marked by a preposed noun or a possessive pronominal prefix (Section 4.4). There is also a small set of bound nominal roots that function like classifiers (Section 4.2). Free nominal roots are frequently arguments of the verb and can combine with a variety of inflectional and derivational morphemes. They can also occur in a sentence without additional marking, such as tɤ̃ jwãǰ ‘boy’ and du’ ‘firewood’ in (44). (44) héde’ tɤ̃ jwãǰ du’ ha-’ɤjúk-nĩ dem boy firewood 3sg-bring-rec.pst ‘That boy brought the firewood.’ Case is marked on nouns in certain contexts. There is an ergative marker for agent NPs of transitive verbs and a set of case markers for expressing spatial, directional, and other oblique relationships. See Section 4.6 for more detail.

4.1.1 Number and gender Number is obligatorily expressed on nominal roots: the singular is unmarked, and the plural is marked by the suffix -ot, for example, a-pem ‘1sg-shoulder; my shoulder’, a-pem-ót ‘1sg-shoulder-pl; my shoulders’. The plural morpheme is atonal, but it adds an H tone to nouns with a L or LH tone, resulting in a L-H melody, for example, bu.an-ót ‘curare (poison)’, which has a bisyllabic L root, and wo.wa-ót ‘bird sp. (Caprimulgidae)’, which has a bisyllabic L-LH root. There are a few exceptions to this pattern, for example, ’ɯ̌t ‘water/river’ becomes ’ɯ́t-ot ‘rivers’, instead of the expected *’ɯt-ót. There is not a robust gender system, but there are a handful of suffixes that are used with some nouns and nominalized attributive constructions to express masculine or feminine gender, for example, -di masculine and -da feminine (human referents only), and -bon ‘male’ and -dén ‘female’ (used for non-human animate referents but only in cases where the sex of the individual is relevant to the discourse). On nominal roots, only the feminine is marked (e.g., képa ‘old man, elder’ vs. képa-da ‘elder-f, old woman’). In a few nominalized attributive constructions, both masculine and feminine markers are used (e.g., tesáw-di ‘liar-m’ vs. tesáw-da ‘liar-f’, both derived from the verb tesáw ‘lie’), but this marking is not applied productively. Nominal roots may also be marked with TAME morphology when they are in the first position of the clause, but the scope of this marking extends to the clause

26 Wã́nsöjöt (Puinave)

1161

and not to the specific constituent being marked. For instance, the reportative evidential and past tense marking both occur on the noun for peach palm fruit in (45). This past tense marker is the only indicator of tense for the entire clause. (See Section 5.3 for more on TAME morphology.) ̂ (45) mõn-ot-mã-di ha-hɤ̃́ mka-at ’ó’o nã́’ã bon peach.palm.fruit-pl-rpt-pst 3sg-planting-pl 3sg.pro dem man ‘It is said the peach palm fruits were the plantings of this man.’ There are two morphemes that encode nominal tense when combined with nouns. The nominalizing past marker -hĩn can affix to nouns or verbs (nominalizing the latter). It is used to indicate that an event, state, or characteristic described by the root has become a habit or fixed state (e.g., dén-hĩn-pek ‘woman-nmlz.pst-very; womanizer’) or to express a sense of expiration (e.g., a-‘u-‘i-hĩn ‘1sg-spouse-fathernmlz.pst; my ex-father-in-law’). The resultative -pɤn indicates the result of a finished process, as in sîm-pɤn ‘foot-res; footprint, tread’. These two morphemes can be used in conjunction with one another, as in tim-pɤ́ n-hĩn ‘yuca/plot-res-nmlz.pst; abandoned plot’ (compare to tǐm-pɤn ‘yuca/plot-res; harvest stubble’). Finally, nominal roots may be optionally marked with diminutive (-tɯ̌j) or augmentative (-swaj) suffixes, for example, jot-tɯ́j ‘dog-dim; puppy’ and jǔ m-swaj ‘mosquito-aug; big mosquito’.5

4.2 Noun compounds and nominal classifiers ́ Wãnsöjöt makes extensive use of the juxtaposition of roots to create new lexical items. Noun compounds can be formed from two nouns, a noun and a verb (in either order), and from nouns plus additional derivational morphology, (46). (46) a. dáp-juj hand-dress ‘ring’ (N-N) c. tek-tɤ̃́ m-õt sit-one/unity-pl ‘pile, heap’ (V-N)

b. jak-dúk moss-enter ‘burrow’ (N-V) d. ’ut-’ɯ̂ j-ot bone-neg-pl ‘insect’ (Lit. ‘one without bones’)

In addition to noun compounding, nouns can also be derived via the addition of a nominal classifier. Like compounding, these classifiers provide semantic content

5 Size can also be expressed via a stative verb construction; see Girón (2008: 187–188) for both diminutive and augmentative options.

1162

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Tab. 26.6: Nominal Classifiers. classifier

semantic domain

example

-ka, -kam

fruit

mõ-dáp-kâm indf.pro-hand-clf:fruit ‘finger’

-dɤ’

tree/bush

wôk-dɤ́’ food-clf:bush ‘sweet yuca’

-pok

juice; also used with liquids more generally

dɤ́-pok fire-clf:liquid ‘gasoline’

-ta

dough; meat; also used with large things

hudé-ta forest-clf:dough ‘dense forest’

-su’

resin

pû n-su’ tree-clf:resin ‘resin’

-je’

oil

debom-jé pig-clf:oil ‘porkfat’

-hudé

group or crowding of something (for nature or environmental things)

wo-húde seje-clf:crowding ‘seje (Jessenia polycarpa) field’ sawan-húde cotton-clf:crowding ‘cotton field’

about the noun, for example, dɤ́ -pok ‘fire-clf:liquid; gasoline’. However, these classifiers are bound nominals that do not have underlying tone, and they only attach to noun roots (juxtaposed to the right of the head noun). They are only used derivationally; they do not function anaphorically or as a means of agreement marking. A list of the attested classifiers is given in Table 26.6; see also Girón (2008: 182– 185).

4.3 Pronouns The pronoun class is further subcategorized into personal pronouns (which have both free and bound forms), demonstratives, dative pronouns, interrogative pronouns, and indefinites. Like nouns, the pronouns have singular and plural forms.

26 Wã́nsöjöt (Puinave)

1163

Tab. 26.7: Personal pronouns (free and bound forms). function

1 2 3

singular

plural

free

bound

free

bound

’ãm mãm ’ó’o

amãha-

bidú ~ bidút jãm ’ót

bijaka-

demonstratives discourse proximal distal indefinite non-specific

hɤ́’ɤ ~ dú’u nã’́ ã héde bɤ́j

hɤ́t ~ dút nãt́ hétet bɤ́t mõ-

The personal pronouns are given in Table 26.7. There are distinct forms for each person, in both the singular and plural, but there is no clusivity distinction, gender marking, or animacy marking. There are morphologically free forms that occur as subjects and direct objects, although these arguments can also be expressed via bound forms (see Section 5.1 for verbs and their arguments). The bound forms also function as possessors when bound to a noun (see Section 4.4.). There is a distinct set of free forms for indirect objects; these dative pronouns are given in Table 26.8. Table 26.7 also includes the demonstrative pronouns and indefinite pronouns. The demonstrative forms hɤ́ ’ɤ ~ dú’u and their plural counterparts are used anaphorically to indicate a third person or object already active in the discourse. They can also combine with the third-person general-purpose pronoun ó’o for emphasis. The demonstrative pronouns nã́’ã/ nã́t and héde/ hétet encode a speaker proximal/ distal distinction. Girón (2008: 174) labels the distal demonstrative héde/hétet as ‘distal (visible).’ If the referent is not visible, the speaker will use one of the other pronoun paradigms: namely ’ó’o / dú’u (sg), ót / dút (pl), or the discourse pronoun hɤ́ ’ɤ, hɤ́ t. The indefinite pronouns bɤ́ j, sg and bɤ́ t, pl are used to introduce a new or indefinite topic to the discourse or to refer to a known argument that is considered unfamiliar or distant; these pronouns are glossed variously as ‘a’, ‘other’, ‘someone/ something/some’. The non-specific person marker mõ- only appears as a dative argument and in possessive constructions when a possessor is required but not specified. Table 26.8 presents the set of dative pronouns that are used for indirect objects. These pronouns are free morphemes that are independent of the verb. There are two paradigms; the first paradigm is formed via the corresponding personal pronoun plus the ablative suffix -at, and the second paradigm is a phonological reduction of the first paradigm forms (see Girón 2008: 172–173 for the specific processes). These paradigms are interchangeable with one another, but the second paradigm is more frequent in spoken discourse. It is not clear what conditions this distribution.

1164

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Tab. 26.8: Dative pronouns. singular

1 2 3 demonstrative non-specific

plural

paradigm 1

paradigm 2

paradigm 1

paradigm 2

’ãm-at mãm-at ’ó-at hɤ́-at ~ dú-at

’ãn mãn hãn

bidúd-at jãm-at [ɲãmãt] ’ót-at hɤ́t-at ~ dút-at

bídat jãn [ɲãn] kádat

mõ-dat

Interrogative pronouns are derived from the indefinite pronoun root bɤ́ -. See (47) for a sample of the possible forms and Girón (2008: 209–214) for the full list along with examples. These question words occur at the beginning of interrogative sentences, but they can also be used in subordinate clauses that are not interrogative constructions. In these latter cases, the pronoun fills the position of an argument. Interrogative pronouns can also take modifiers for number, modality, directionality, tense, and aspect (see Section 5.3 for clause-typing morphemes, including details on question formation). (47) a. bɤ-i-dík indf-attr-exist ‘what’

e. bɤ̂ n/bɤm indf(dist)/indf(prox) ‘where’

b. bɤ́’ɤ ‘who (sg)’

f. bɤ́t ‘who (pl)’

c. bɤ́-at indf-abl ‘to/for whom’

g. bɤ́t-at indf.pl-abl ‘to/for whom’

d. bɤ́-dów indf-date ‘when’

4.4 Possession Possession is marked in two ways: by marking the possessed noun with the bound form of the personal pronouns (e.g., bi-hǎ’ ‘our canoe’), or by juxtaposing the two nouns together, following the order possessor-possessum (e.g., hom-hṍ’ ‘bird’s wing (bird-wing)’). These prefixes, given above in Table 26.7, are also used as prefixes to verbs to express arguments (see Section 5.1). Inalienable nouns – namely body parts, kinship terms, and part-whole relationships – are obligatorily possessed, but

26 Wã́nsöjöt (Puinave)

1165

no special marking is used to differentiate inalienable from alienable possession. For instance, in (48), the same possessive prefix is used for kǒk ‘chili pepper’ (not obligatorily possessed) and -pem ‘shoulder’ (obligatorily possessed). (48) a. kǒk ‘chili pepper’ b. mõ-pem ‘(one’s) shoulder’

c. a-kǒk ‘my chili pepper’ d. a-pem ‘my shoulder’

There are two markers that occur with obligatorily possessed nouns when the possessor is unknown: the non-specific person marker mõ- ‘one, someone, something’ (e.g., mõ-sîm ‘one’s foot’) and the indefinite marker ébu- ‘thing’ which is specifically used with mass nouns (e.g., ébu-je’ ‘thing-grease; oil of something’). In these examples, the indefinite pronoun mõ- functions as a true possessor, whereas the indefinite marker ébu- serves as an indefinite participant of a genitive relationship.

4.5 Noun modification As mentioned in Section 3, adjectival concepts are expressed in a variety of ways, but the most productive form for designating qualities is by prefixing attributive ito an intransitive stative verb.6 Attributive i- has a wide variety of functions (see also Sections 5.4, 6.1, and 7.1). Depending on the semantic structure of the intransitive verb, then, the attributive prefix plus a verb produces either a verbal adjective (49) or an agentive, (50). (49) i-bák-ot attr-abundant-pl ‘many’ ́ (50) i-nõk-ot attr-go-pl ‘those who go’ All terms referring to color and other physical characteristics and states are formed with this pattern, see for instance ‘black’ in (51). (51) nã́t jót-ot i-pík-ot dem dog-pl attr-black-pl ‘these black dogs’

6 Stative verbs function as finite predicates when they are prefixed by person markers. (See contrasting examples in Girón 2008: 296–297)

1166

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Tab. 26.9: Numeral terms. 1 2 3 4 5 6

7 8

tɤ̃̂m kâ w pɤ́j kaw-nṍk two-go dáp-tɤ̃m hand-one tɤ̃̂m-dáp-bat-a-bat one-hand-cross/opposed-all-cross Lit. ‘one crossing to the opposed hand’ kâ wdábattabat pɤ́jdábattabat

 9 10

kawnṍkdábattabat kâ w-mõ-dáp two-indf.pro-hand Lit. ‘someone’s second hand’ or dáp-at-kaw hand-abl/compl-two Lit. ‘from the second hand’ 11– Numbers between 11 and 19 add the 19 word for ‘foot’ and are formed similarly to 6–9 20 tɤ̃̂m épin (Lit. ‘one person’) 40 kâw épin (Lit. ‘two person’)

The attributive prefix is also used with numerals. Numerals precede the noun they modify and also take the plural suffix -ot when relevant, (52). (52) ójat ka-mõk-ma i-pɤ́ j-ot mõt, dukjín-ot then 3pl-say-rpt attr-three-pl man.pl orphan-pl ‘Then the three men, orphans, said …’ Numeral terms are provided in Table 26.9. The terms for one through three are etymologically opaque. Four is analyzed as ‘two’ plus the suffix -nõk, which is a grammaticalized form of ‘go’. The terms for five through twenty are based on counting the fingers of the hands, using ‘hand’ as a base. There are two forms for ‘ten’, both based on combining ‘hand’ with ‘two’. The term for ‘twenty’ introduces a new base (‘person’), which can be used to produce numbers up to 40 (‘two persons’). However, Spanish is replacing numeral terms above 5, and most speakers stop using ́ Wãnsöjöt terms around 20.

4.6 Case markers Core arguments are marked for grammatical relation in certain contexts. For instance, agent NPs of transitive verbs are marked with ergative -at. This suffix is also used for marking dative arguments. However, intransitive subject NPs and object NPs are unmarked (see Section 6.1 for more detail). There is also the set of dative pronouns given above in Table 26.8. ́ Wãnsöjöt also has a series of case markers that attach to noun roots to express spatial, directional, and other oblique relationships. In most cases, these markers are suffixes that attach to a noun root, but a few can also be attached to verb or adverb roots. These morphemes are considered to be case markers rather than post-

26 Wã́nsöjöt (Puinave)

1167

Tab. 26.10: Spatial cases. suffix

gloss

example

-at

ablative (abl) ‘from’ Also marks ergative case and dative objects

ha-júk bɯ’-at 3sg-come torrent-abl ‘S/he is coming from the torrent.’

-a

allative (all) ‘to, towards’

mõ-jãn-a-dá ha-dí’ indf.pro-village-all-ass 3sg-cop ‘S/he has come to town.’

-te/-te’é

inessive (iness) ‘within, inside’

wãm-te’é pot-iness ‘in a/the pot’

-u’

adessive (ade) ‘in contact with’ (also a general locative)

bɯ’-ú’ torrent-ade ‘in the torrent’

-ha’

translative (trnsl) ‘across, towards’

bɤ́dem-da a-si-nõḱ -ja hudéta-ha tomorrow-ass 1sg-prosp-go-fut mountain-trnsl ‘Tomorrow I will go to the mountain.’

-je

perlative (per) ‘through’

ka-ha-túhuk wodíju-ot bentana-je ó’-at 3pl-3sg-throw garbage-pl window-per 3sg-erg ‘S/he threw the trash (pl) through the window.’

positions because tense and modality markers that occur with nouns can also be attached after these case endings. Table 26.10 presents the spatial case markers that encompass concepts such as trajectory, interior/exterior distinctions, and contact between entities. These morphemes can also combine in pairs, in which case they have a distinct tone pattern of H-L or L-H. These morphemes attach in order of proximity: markers expressing interior localization attach closer to the noun; exterior locations and notions of trajectory or directionality are further from the noun root. See Girón (2008: 200–203) for examples of these combinations. Not all possible combinations are attested. Data from naturalistic speech and elicitation confirm that more than two markers are not allowed. Table 26.11 presents a series of other case markers that combine with noun roots. Unlike the spatial case markers, these markers do not combine with one another.

1168

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Tab. 26.11: Other case markers. suffix

gloss

example

-het

comitative/ associative/ instrumental ‘with’ (with nouns, verbs, and adverbs)

ha-dok-hét 3sg-tongue-com ‘with its tongue’

-sɤm

motive/cause/ reason

bɤj-dík-sɤm indf-exist-motive ‘why’

-nĩ

comparative ‘like’

’ó’o ha-kík-nĩ ó-he i-tɤ̂m tɤ̃jwãj́ -nĩ 3sg.pro 3sg-cry-rec.pst 3sg-distr attr-one child-cmpr ‘He cried like a child.’

sɤ́t

opposition/ antagonism (antg); also distal (dist) ‘more X than …, as far as’

ka-nɤ̃̂n-da nãt́ jót-ot hɤ́t sɤ́t 3pl-fierce-ass dem dog-pl 3pl antg ‘These dogs are more ferocious than those.’

jãwãnã’́

purpose/utility (purp) ‘for’

pída-ot mã-sãn-’ɤjút-te bi-dintérna-ot jãwãnã’́ battery-pl 2sg-neg-bring.neg-int 1pl-flashlight-pl purp ‘You didn’t bring batteries for our flashlights?’

5 The verb complex This section is concerned with a description of the verb complex. The verb complex is defined as a verbal base and its modifiers. A clause is defined as the verb complex including NP arguments and obliques. Tense, aspect, mood, and evidentiality marking are discussed in this section, even though they occur in various positions within the verb complex and contribute to the interpretation of the clause. Clauses are described in Section 6.

5.1 The verb complex and finite clauses ́ An important distinction in Wãnsöjöt is the one between finite and non-finite verb complexes and/or clauses. Non-finite and nominalized clauses are described in Section 5.4. A finite verb complex minimally contains a verb base and one or more overt arguments, depending on the transitivity of the verb base. If the overt arguments are not expressed through separate NPs, they must be realized as bound pronominal ́ prefixes that encode person and number. Thus, the clause in Wãnsöjöt can be a minimal free form containing only a verb root and pronominal prefixes, as in (53).

26 Wã́nsöjöt (Puinave)

1169

(53) ka-ha-péw-di 3pl-3sg-carry-pst ‘He carried them [on his shoulder].’ The order of the pronominal prefixes determines their grammatical function. In transitive clauses with multiple preverbal prefixes, the left-most prefix must express O (e.g., ka- 3pl in (53)) and the pronominal prefix closest to the verb must express A (e.g., ha- 3sg in (53)). The bound pronominal prefixes do not vary in form according to grammatical function (whether S, O, A). All bound argument prefixes have the same form as the bound pronominal prefixes listed in Table 26.7 and described in Section 4.4. The non-specified pronominal prefix mõ- has a variety of functions (see Section 4.4), one of which is to be an indefinite pronoun (e.g., ‘one,’ ‘a person,’ ‘something,’ etc.) as illustrated in (54). In this example, the indefinite pronoun is the only argument that is marked on the verb, and it expresses A; the other argument is expressed by an NP. (54) mõ-’ok-hɤ́ j mõ-sutá-hɤ́ j … hám’õ sutá indf.pro-take.out-top indf.pro-fiber-top then fiber mõ-hetwâw-hɤ́ j indf.pro-prepare-top ‘One takes out one’s fiber … and then one prepares the fiber.’ Pronominal arguments do not have to occur on the verb when that argument is expressed by an overt NP elsewhere in the clause, but they may occur with a coindexed NP argument depending on the construction type. More details concerning the distribution of pronominal arguments and their relationship to full NPs are provided in Section 6.1 in the context of a description of alignment. ́ A rough schema of the Wãnsöjöt clause which summarizes the relative order of elements discussed in this section is provided in Table 26.12. In this schematic template, a slot refers to a morphosyntactic position that can only be occupied by one element at a time, and a group refers to a position which can contain more than one element (not necessarily in a fixed order). The details of the internal structure of groups requires more research. ́ The verbal word in Wãnsöjöt can be defined as a span from the {O,S} pronominal markers to the tense-aspect-modality markers. A number of other modifiers can occur outside this span, but since they can be interrupted by full noun phrases and postpositional phrases, they are considered to be outside the verbal word. That said, ́ identifying this span as the verbal word is somewhat problematic in Wãnsöjöt. First, negation occurs immediately before the verbal base (and after the pronominal markers), making it inside the span of the verbal word, even though it is clausal in scope. Additionally, there are numerous TAME markers (mostly bound, but some free) that

1170

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Tab. 26.12: Schematic template for the Wã́nsöjöt verb complex and clause. type of position

elements

group group

NP Argument, PP Adjunct tame markers / interrogative

slot slot slot group slot group group group

Pronominal {O,S} Extended base (for separable verbs) Pronominal {A} Mood/modal markers Negation Root(s) Aspect/Adverb group tam markers

group group group

NP Argument, PP adjunct tame markers / interrogative Argument, adjunct

clause

verbal word

can interrupt the verbal word (see Section 5.3). We exclude from the verbal word TAME markers that can occur on either side of the verbal base and/or can be interrupted from the verb base by a full noun phrase or postpositional phrase.

5.2 The verb base ́ The verb base in Wãnsöjöt contains one or more verb roots formed through compounding and/or serialization. It is subject to processes of lexicalization in the sense that, when it consists of multiple formatives, the resulting combination is often not semantically compositional. ́ There are five types of verb bases in Wãnsöjöt that can be classified according to four parameters: (i) Simple/Complex; (ii) Separable/Contiguous; (iii) Asymmetric/ Symmetric; and (iv) Non-analyzable/Transparent. The classes of verb bases according to these three parameters are depicted in the terminal nodes of Figure 26.1. Verb bases can be simple or complex; complex verb bases contain more than one part. Often the parts can each be understood as roots, but often not. Examples of simple verb bases are provided in (55). (55) a. a-dɯ́k 1sg-jump ‘I jump.’

b. ha-bi-tǎj 3sg-1pl-burn ‘We burn it.’

Polymorphemic verb bases can be divided into those in which their parts must be contiguous and those that are separable. A separable verb base refers to a verb base composed of parts that can be interrupted by a pronominal argument prefix or one

26 Wã́nsöjöt (Puinave)

1171

Verb base Simple

Complex Separable

Non-analyzable

Transparent

Contiguous

Asymmetric Symmetric

Fig. 26.1: Types of verb bases in Wã́nsöjöt.

or more TAM morphemes. Recall from Section 3 that it is possible for separable verb bases to occur contiguously. We can divide separable bases into two types: non-analyzable and transparent. Those that are non-analyzable cannot be understood as morphemes on their own. This is true of písi in pisí…wá’ ‘know’ as in (56). (56) jãm-da-wí ha-písi-ja-si-wá’ nã́ã pápeda 2pl-ass-rest 3sg-know-2pl-prosp-know dem.prox paper/book You (pl) need to know this book.’ The separable verb base can be transparent in the sense that its form is associated ́ with some other morpheme in Wãnsöjöt and it could be understood as contributing a meaning associated with this morpheme. An example is provided in (57b), where bík ‘eye’ and sãk̂ ‘imagine’ combine to mean ‘think’. ̂ (57) a. mõ-i-sãk indf.pro-attr-imagine ‘imagine’

̂ b. mõ-bík-mõ-sãk indf.pro-eye-indf.pro-imagine ‘think’

Polymorphemic contiguous verb bases can be divided into symmetric and asymmetric types. The latter is one which contains two or more elements that are in an asymmetric relation with one another such that one element can be regarded as a head and the other its dependent or modifier. There are a number of criteria by which asymmetric relations can be established (Bauer 1990). In (58), the head of ̂ the verb base ‘kill’ is ɯj ‘die’. In (59), the elements of the verb base mĩn-jɤk ‘leave from below’ are in a type-of relation such that the meaning of the combination can be viewed as a hyponym of the head jɤk ‘leave’.7 (58) ha-mõ-ɯ̂j-’ɤk 3sg-indf.pro-die-caus ‘Someone killed him/her/it.’ 7 The distinction between symmetric and asymmetric forms follows Girón (2008). We are unsure whether the criteria can always be applied consistently to complex verb bases.

1172

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

(59) ha-mîn-jɤk 3sg-grow-leave ‘S/he leaves from below / leaves lower level to go to higher one.’ ́ Wãnsöjöt does not appear to have a rich system of valency-adjusting morphology. The verb ‘ɤk ‘do’ functions as a causative in complex verb bases. Girón (2008: 223) describes a cline from lexicalized instances of verb bases with ‘ɤk to those that seem more productive and semantically transparent. Some illustrative examples are provided in (60). (60) a. ha-mõ-‘ɯ̂ j-‘ɤk 3sg-indf.pro-die-caus ‘kill’

b. ha-mõ-jɤ‘ɤ́ k 3sg-indf.pro-leave-caus ‘make a mistake, take out’

c. ha-i-wõ-‘ɤ́ k 3sg-attr-eat-caus ‘invite/let someone eat’ ́ Wãnsöjöt does not clearly display a system of grammaticalized associated motion in the sense of containing a paradigm of closed class elements dedicated to associating motion events with main events (see Rose 2015; Guillaume 2016). However, verb roots that express motion events can occur as one of the elements in a complex verbal base such that motion is co-expressed with some other event in one verb complex. An illustrative example is provided in (61). (61) bi-wà(k)-nõk-sɯ́j tãm kadej-ha’ 1pl-dance-go-rep far.away Caney-all ‘We went there to dance again, far away in Caney.’

5.3 Tense, aspect, mood, modality, and evidentiality ́ Wãnsöjöt has a rich inventory of tense, aspect, mood, modality and evidentiality markers (TAME). TAME morphemes vary in terms of their distributional possibilities across these positions. A typical pattern is displayed by past tense -di, which can occur before or after the verbal word as in (62) and (63), respectively. In (62), this marker attaches to the first constituent of the clause, and in (63) it is part of the verbal complex. (62) bɤ̂ n-di ha-ma-dú’ mã-bon? where-pst 3sg-2sg-see 2sg-man ‘Where did you meet (Lit. ‘see’) your husband?’

26 Wã́nsöjöt (Puinave)

1173

héde dén-at (63) ha-tɯ̌j ha-’ɤjúk-súk-di 3sg-son 3sg-bring-desid-pst dem woman-erg ‘That woman wanted to bring her son.’ The prospective morpheme -si-, in contrast, is fixed between the two pieces of a separable verb base (je…wã́j ‘eat’) as in (64). The morpheme -si- will occur prior to the root in cases where the verb base is not complex as in (69) (see also Table 26.12). (64) ja-je-si-pathéj-wã́j-tenók-ja 2pl-eat-prosp-sprl-eat-dim-fut ‘You all will still be eating (little by little?).’ The aspectual morpheme pathéj ‘still, for a long time (also superlative)’ has more distributional freedom than these two morphemes in that it can occupy a position internal to the verbal word and a position outside of it, as illustrated in (65) and (66).8 (65) ha-bu-pathéj-sík-da-tep ikóm 3sg-hunger-sprl-hunger-ass-emph simul ‘…But he was still hungry.’ ̂ (66) … nã́’ã sõm ha-hika-ha-kow-mã pathéj … dem worm 3sg-bear-3sg-bear-rpt sprl ‘…It is said that this worm holds much (for a long time).’ ́ Wãnsöjöt encodes four tense morphemes: -ja future; -nĩ recent past/present perfect; -di past; bɤn remote past. Clauses that are unmarked for any of these tense markers are present tense by default (i.e., they receive present tense interpretations in elicitation), but semantically tenseless such that they can receive their temporal interpretation from discourse context. Sentences with a past time reference do not require the past tense marker -di. These properties of tense marking in ́ Wãnsöjöt are illustrated in (67) which contains a sequence of sentences from naturally-occurring discourse. The first two sentences in (67a) and (67b) contain past tense -di; the rest of the sentences in the narrative do not contain the past tense marker and the temporal reference comes from discourse context, as shown in (67c).

8 This fact, along with many others, undermines the wordhood status of what we have referred to as the “verbal word”. A better approach might be to simply describe the constituency facts of the language and not stipulate an a priori notion of word (e.g., Bickel and Zúñiga 2017; Tallman 2020). Here, we follow the dominant approach to these issues in the field and stipulate a word constituent without motivation or argumentation.

1174

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

(67) a. o-he-di’ a-sík-bát-di ’ãm, saj-dej 3sg-distr-pst 1sg-pain-deic-pst 1sg.pro night-during ‘This is (what it was like) when I was in pain at night.’ b. hɤ̂ n-di-wɤn i-wɤju(k)-nã́k-at-a there-pst-pot attr-day-close-abl-all ‘It would have been as the day was beginning.’ c. hé-a-kuk-tɤ̂ j o-dów jáw ójem pɤj-wɤjú jâ w wait-1sg-frus-wait 3sg-date nothing conj three-day nothing ‘…I waited unsuccessfully, this day, and nothing (after) three days, nothing.’ The recent past marker -nĩ expresses that an event occurred earlier in the day and that the event still has current relevance, (68). It can also express events that encode some subjective sense of recently. (68) ha-t-pew-nĩ písikup 3sg-rdc.ag-carry-rec.pst morning ‘He carried (something) this morning.’ Modal morphemes that also relate to temporal reference are -si- prospective and -wɤn potential. Examples of the prospective are provided in (64) and (69). Examples of the potential are provided in (69) and (70). (69) a-si-hɤ̃̂ j-wɤn 1sg-prosp-whistle-pot ‘I would whistle.’ (70) mã-’ṍm-di-wɤn-da mãm 2sg-shake.in.fear-pst-pot-ass 2sg.pro ‘You would have been scared.’ ́ Wãnsöjöt has a counterexpectational mirative morpheme kṍm. An example is provided in (71). (71) nã́’ã-da kṍm a-’in-’ã́n dem.prox-ass mir 1sg-mother-cassava ‘This (I am just realizing now as I see it) is the cassava of my mother.’ ́ Wãnsöjöt also contains aspect markers -at inchoative, completive and -hin ‘again, repeatedly’ and one inflectional aspect marker -hu imperfective. One of the primary functions of the imperfective is to set up a background event to some story as in (72).

26 Wã́nsöjöt (Puinave)

1175

(72) ha-dik-dí-hu Jadòdu-’u Huan 3sg-be-pst-ipfv Matraca-ade Juan ‘Juan was in Matraca for some time.’ ́ Wãnsöjöt expresses sentential mood through clause-level clitics. Sentences without a sentential mood marker are declarative. There is an assertive marker -da which encodes the speaker’s commitment to the assertion, see for instance (70). The interrogative is marked with -te after the first element of the sentence or with -i after the verbal root. These morphemes are only used in polar interrogatives and ostensibly occur outside of the verbal word as it was defined in Table 26.12. Examples are provided in (73) and (74). (73) mãm-te ha-pew-é ’ó’o? 2sg-int 3sg-carry-ag.nmlz 3sg.pro ‘Did you carry it?’ (74) i-mã-bêp-i-nĩ-hu? attr-2sg-work-int-rec.pst-ipfv ‘Have you been working?’ Information questions are formed with an interrogative word (wh-word) derived from the root bɤ́ -, such as bɤ́ ’ɤ/bɤ́ t ‘who (sg/pl)’, bɤjdík ‘what’ (see (75)), bɤ̂ n ‘where’ (see (62)), and bɤ́ dóu ‘when’ (for a complete list, see Girón 2008: 209–210). These interrogative words occur at the beginning of a question or in a subordinate clause in the place of an argument, and they can be inflected with plural morphology, TAM, or a directional affix. (75) bɤjdík i-mã-wõk? what attr-2sg-eat ‘What are you eating?’ Commands are expressed with the hortative -kú, which follows the verbal root, and the imperative -u, which precedes it, illustrated below. The prohibitive kupɤi- also precedes the verbal root. (76) mã-huj-kú-mĩk! 2sg-get.up-hort-get.up ‘Get up please!’ (77) ja-u-kút-ja 2pl-imp-escape-fut ‘You have to go/escape.’

1176

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

(78) mã-kupɤj-’ṍm 2sg-proh-worry ‘Don’t worry.’ ́ There is one evidential marker in Wãnsöjöt: reportative -mã. It can combine with all sentential moods such as the assertive in (79) and the interrogative in (80), and it can occur as an affix to the verb. (79) mã-’ṍm-mã-da mãm 2sg-shake.with.fear-rpt-ass 2sg.pro ‘It is said that you were shaking with fear.’ (80) dút-te-mã ka-si-pew-é-t-ja? 3pl-int-rpt 3pl-prosp-carry-ag.nmlz-pl-fut ‘Is it said that they will carry them?’ See Girón (2008: 261, 271, 281, 292) for some general comments concerning the relá tive ordering of TAME morphemes. Whether TAME morphemes in Wãnsöjöt can be variably ordered with scopal effects requires future research.

5.4 Nominalizations and non-finite verb forms ́ Wãnsöjöt has a rich set of nominalizing morphemes. For the purposes of this description, we define a nominalized verb complex as one which can occur as an argument of a verb, but in practice, they are most commonly found as predicates in non-verbal predicate constructions. These constructions involve pragmatically focusing one of the arguments and, as such, are marked with respect to information structure. Clauses built from nominalized verb complexes also display special alignment properties distinct from that of main clauses (Section 6). Nominalizations and non-finite verb forms are also used in clause-linkage (Section 7). The nominalizers ́ of Wãnsöjöt are listed in (81). (81) a. b. c. d. e. f. g. h. i.

-ju -hét -dipɤn -nõba -V/-m9 i-sãn-…-i -pɤn -hin

‘Instrument/abstract nominalizer’ ‘Instrument nominalizer’ ‘Locative nominalizer’ ‘Agentive habitual nominalizer’ ‘Agentive nominalizer’ ‘Attributive, agentive nominalizer’ ‘Negative nominalizer’10 ‘Result nominalizer’ ‘Past nominalizer’

9 This alternation with –m happens in the singular only. 10 This strategy employs the clausal negator sãn-; see Section 6.3.

26 Wã́nsöjöt (Puinave)

1177

A few illustrative examples of simple lexical nominalizations with the instrumental -ju and the resultative -pɤn are provided in (82) and (83). (82) a. sîn b. kɯ̀ k

‘filter, sift’ ‘sweep’

→ →

sîn-ju kɯ(k)-jú

‘sieve’ ‘broom’

(83) a. sɤ́ t ̂ b. mãp

‘smoke, be smoked (v.)’ ‘fish (v.)’

→ →

i-sɤ́ t-pɤn ̂ i-mõ-mãp-pɤn

‘smoked thing’ ‘a catch’

Lexical nominalizations can be elaborated by many of the categories that full nouns can (see Sections 3 and 4), for instance, gender marking – as in (84) with the agentive nominalizer -nõbá – and possessive marking and case marking, as in (85) with the locative nominalizer -dipɤn. Lexical nominalizations also exhibit number distinctions in some cases: singular -m alternates with plural -t (Girón 2008: 249–250). (84) a. ját-nõbá-di talk-hab.nmlz-m ‘talker, gossiper’

b. pɯ́tok-nõbá-da fight-hab.nmlz-f ‘female fighter’

́ (85) a-’ép-da dúmãt a-nõk-at a-’ôu-dipɤn-a ja-sɤ́ t 1sg-bad-ass now 1sg-go-inch 1sg-sleep-loc.nmlz-all 2pl-dist ‘Now I am sick, I am going to my room (Lit. ‘where I sleep’) far from you (pl).’ ́ Wãnsöjöt also has a number of morphemes that derive non-finite verb complexes that can also be used to encode subordinate relations (see Section 7.1), (86). (86) a. -dik eventive b. -a ‘locative (allative) non-finite clause’ c. -u ‘protasis in conditional’ Forms derived with these morphemes are not regarded as nominalizations because they cannot serve as core arguments of a verb. Many nominalizations and nonfinite verb complexes can also take full arguments and have important functions in relativization, complementation, and clause-linkage more generally (see Sections 6 and 7). For instance, the eventive affix creates a kind of verbal nominalized word that can take one argument, namely a verbal complex by itself.

6 Simple clauses This section provides a description of three important aspects of simple clauses in ́ Wãnsöjöt: clause-level alignment is described in Section 6.1, noun-incorporation

1178

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

and its relationship to alignment is discussed in Section 6.2, and clausal negation is presented in Section 6.3.

6.1 Alignment and constituent order ́ This section is concerned with describing two aspects of alignment in Wãnsöjöt: 11 constituent order and case marking. Alignment will be described in two domains: pronominal prefixes and full NPs. As noted in Section 5.1, pronominal prefixes display a neutral alignment in that they have the same form across all grammatical relations (see also Girón 2008: 292). As was demonstrated in Section 5.1, for transitive clauses, O pronominal prefixes and A pronominal prefixes occur in distinct positions within the verb complex. The distinct positions can most clearly be seen with complex verb bases as in (56) with O-base-A-base form (see Table 26.12 and also Girón 2008: 292 for a more detailed template). In all clauses which are intransitive, the S argument occurs to the left of the extended base, as in S-base-base in (61) and (65). Thus, if one judges by the positions of the bound pronouns in relation to the elements of the extended base, there is an ergative alignment in the bound pronouns (see Table 26.12 for clarification).12 Pronominal prefixes are only obligatory in cases where an overt NP is not present; otherwise they do not occur. For instance, in (87) the O argument is not expressed as a pronominal prefix because it is an overt NP. In (88), the A argument is not expressed as a pronominal prefix because it is expressed as a full NP. Pronominal prefixes can co-occur with full NP arguments when the latter occur after the verb (see below for details). (87) káwa-bíkdik ka-sek-dí ka-wõk-á chicken-egg 3pl-rob-pst 3pl-eat-all ‘They stole a chicken egg to eat.’ (88) ha-t-dû k-mã doktor-ot-at 3sg-?-see-rpt doctor.sp-pl-erg ‘The doctors saw him.’ ́ With respect to case marking on full NPs, Wãnsöjöt alignment depends on the construction type. The relevant constructions are defined as follows. In finite clauses,

11 That is, it is focused on coding properties and not behavioral properties, which require more research. 12 Although this seems to follow from the available data and Girón‘s (2008) template of the verbal complex (Girón 2008: 292), the question requires future research because in certain cases it seems as if the S argument can occupy both positions.

26 Wã́nsöjöt (Puinave)

1179

neither argument is privileged in terms of focus.13 A marker of tense, aspect, or modality is obligatory in these clauses. Where no full NPs are present, the finite verb does not require any other constituents except pronominal prefixes. In nonfinite or event-focused clauses, neither argument is privileged in terms of focus. The verb is marked with the non-finite marker -dik. A non-finite verb cannot be the only constituent in the sentence (discussed in Section 7). Object-focused focus on the object; the verb complex is marked with the prefix i- attributive. Finally, agentfocused clauses focus on the A subject; the verb complex is marked with the suffix -V agentive nominalizer. Finite and non-finite (event-focused) constructions display an ergative alignment such that the NP in A function is marked with -at. NPs in {S,O} function are unmarked. (89) [dúii]O [hai-sɯk-da]V [héde sikju-at]A Luis 3sg-suck.up-ass dem.dist whirlpool-erg ‘The whirlpool sucked up Luis.’ Object-focused clauses display a semantic alignment. S of a non-active intransitive (SS) and O of a transitive verb are unmarked, as in (90) and (91). ́ V (90) bɤ́ j wɤjudów [i-’ôw-du-kunõk] [nã́’ã dómã’]​Ss​ indf one.day attr-sleep-enter-quickly dem.prox grandfather ‘One day, this old man stayed and fell asleep quickly.’ (91) [pin-át]A [i-t-bénduk-pɤn]V [héde jǒt]O anaconda-erg attr-rcd.ag-swallow-res dem dog ‘The anaconda swallowed that dog.’ In object-focused clauses A of a transitive verb and S of an active verb (SA) are marked with -at as in (91) and (92). This pattern is distinct from the alignment of finite verbs where only A is marked with -at. (92) [i-mã-bêp-da]V [mãm-at]​SA​ attr-2sg-work-ass 2sg.pro-a/sA ‘You are the one who worked.’ (Lit. ‘work happened by you’)

13 This description follows Girón (2008) closely. The concept of focus here seems to be useful in categorizing the construction types according to which alignments they display. However, it is not clear how it relates to the concept of focus as it is understood more broadly in the literature on information structure. Future research is required to flesh out what the relationship is between these constructions and information structure.

1180

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Since marking with attributive i- is a subordinating strategy used for relativization, one might speculate that object-focused clauses are really non-verbal predicate constructions (see Section 7.1). However, since their peculiarities in alignment would not follow directly from this structural analysis, they must still be recognized as a distinct construction. Agent-focused clauses display a neutral alignment in that {A,S,O} are all unmarked for case as can be seen from (93) and (94). (93) [pin-dá]A [ha-bénduk-u]V [héde jǒt]O anaconda-ass 3sg-swallow-ag.nmlz dem dog ‘The anaconda swallowed this dog.’ (Lit. ‘the anaconda is the one who swallowed this dog’) ̂ ́ (94) [’ãm-da] S [ha-hɤ̃ j-ɤ̃ -hu]V 1sg-ass 3sg-whistle-ag.nmlz-ipfv ‘I am whistling at him.’ (Lit. ‘I am the one who is whistling at him’) Agent-focused clauses could be regarded as bi-clausal non-verbal predicate constructions. According to this analysis, the A argument of agent-focused constructions is structurally an S of a non-verbal predicate construction (see the literal translations of agent-focused sentences above). The predicate is a verb complex nominalized with -V~-m that takes an O argument. Example (95) provides a basic schema. (95) s

[ [v / (o)]ag.nmlz ]PRED

This analysis explains both why no case is assigned to the notional A argument in agent-focused clauses (they are structurally S arguments) and also the fact that the verb complex is found with additional plural pronominal marking when the notional A argument is plural. See Girón (2008: 332–335) for more details and a parallel analysis of object-focused clauses. ́ Constituent order is fairly flexible in Wãnsöjöt. For finite clauses, all the orders are permissible except OAV. AOV is permitted only if we include O-incorporated verbs (see Section 6.2).14 Some examples that illustrate the free ordering are provided in (96) through (99).

14 Some readers might regard the sentence in (99) as actually instantiating an SV pattern. However, word-internal structure need not necessarily be encapsulated from argument structure (cf. ́ Sadock 1980). In fact, adopting this assumption seems problematic in Wãnsöjöt, because the ‘S’ argument is assigned ergative case in (100).

26 Wã́nsöjöt (Puinave)

1181

Tab. 26.13: Constituent orders in Wã́nsöjöt across three constructions. clause type

VAO

VOA

AVO

OVA

AOV

AO-V

OAV

Finite verb Object-focused Agent-focused

✓ ✘ ✘

✓ ✘ ✘

✓ ✓ ✓

✓ ✓ ✘

✘ ✓ ✘

✓ ✘ ✓

✘ ✓ ✓

(96) [ha-t-sák-da]V [a-jót-at]A [héde hɤdúk-ta]A 3sg-rdc.ag- bite-ass 1sg-dog-erg dem animal-clf:mass ‘My dog ate that meat.’ (Lit. ‘animal-dough’) (97)

[ha-ka-bɯ́k-di-da]V [padatá’]O [ót-at]A 3sg-3pl-give-pst-ass money 3sg-erg ‘They gave him money.’

[héde síkju-at]A (98) [Dúji]O [hai-sɯ́k-da]V Luis 3sg-suck.up-ass dem whirlpool-erg ‘The whirlpool sucked up Luis.’ (99) [Huan-át]A [mesa-pǎj]O-V Juan-erg table-hit.with.palm ‘Juan hits the table with a palm.’ Agent-focused and object-focused constructions are marginally more restricted than finite clauses with respect to constituent order; they do not allow any verb initial constituent orders (*VAO and *VOA). Interestingly, they both allow OAV, (100), which is banned in finite verb constructions. Table 26.13 provides a summary of the (im)permissible constituent orders discussed in this section. (100) [Patrisia-da]O [sɤ̃́ p-at]A [i-sak]V Patricia-ass snake-erg attr-bite ‘The snake bit Patricia.’ (Lit. ‘the snake is the one that bit Patricia’)

6.2 Noun incorporation ́ As noted previously, Wãnsöjöt displays noun incorporation. An O argument can be expressed by a single noun root which is morphophonologically incorporated into and left-adjacent to the verb base of which it is a dependent, as in (99) and (101), which provides an example of object incorporation in an agent-focused construction. Incorporated nouns do not have phrasal structure and are not their own pho-

1182

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

nological word (they are subject to the tone sandhi processes described in Section 2.2). (101) bon-tɯt-ót-te pû n-pew-é-t-di? man-young-pl-int stick-carry-ag.nmlz-pl-pst ‘Did the young men carry sticks?’ Only the O argument can incorporate, and all the attested examples incorporate inanimate Os. From the available data, it appears that the incorporated O is not compatible with an overt phrasal NP O. O incorporation is incompatible with objectfocused constructions (see Table 26.13 where O-V represents object incorporation). Verb complexes with incorporated objects still assign ergative case to A as in (102), demonstrating that they are not structurally intransitive. (102) Madí-at watjé ka-sɯk María-erg guava-suck.on ‘María sucks guava fruit.’

6.3 Negation Clausal negation is expressed via the morpheme sãn- immediately before the verbal root. The root itself participates in a negation-induced phonological process; syllable final dorsal consonants are realized as [t]. For separable verb roots, negation occurs between the constituents, and the morphophonological process only affects the last part of the verb root. (103) ka-sãn-nó(k→)t ka-hûj-dipɤn-a 3sg-neg-go.neg 3pl-learn-loc.nmlz-all ‘They didn’t go to school.’ ́ In addition to the clausal negation marker, Wãnsöjöt also has three other negative markers. There is a constituent negator ‘ɯ̂ j, which can be used to respond negatively to polar interrogatives, for negating nominal constituents, and for negating nonfinite verbs in order to express strong prohibition. It can also combine with the comitative morpheme -het to mean ‘without’. The negative existential bɤidût functions as a verbal root in that it is separable and can take the interrogative marker found with verbal roots as well as modal morphemes. However, it does not take person markers. Finally, the morpheme jáu is used in discourse as an exclamation to indicate that something is not there, or was not successful, contrary to expectations. See Girón (2008: 259–260 and 293–296) for an overview of all of the negation ́ strategies in Wãnsöjöt.

26 Wã́nsöjöt (Puinave)

1183

7 Complex sentences This section is concerned with the description of sentences which are composed of ́ more than one clause in Wãnsöjöt.

7.1 Subordination ́ Wãnsöjöt does not seem to have any morphemes that uniquely encode subordinate functions. Nevertheless, five strategies can be identified for marking subordinate ́ relations between clauses in Wãnsöjöt: (i) juxtaposition of finite verbs with one TAME marker that modifies the whole sentence, (ii), a clause that occurs with its head verb with attributive i-, (iii) a non-finite construction marked with the eventive -dik, (iv) a clause that occurs with an adverbializer (or nominalizer that can have an adverbializing function) on its head verb, (v) a clause that occurs with a clause linker formed from the third person singular independent pronoun ó’o. The strategies listed and described above are used for different types of subordination, namely relativization, complementization, and adverbialization. They are summarized in Table 26.14. It is important to point out that the last three strategies listed in Table 26.13 do not uniquely encode subordinate functions. There are no morphemes that can be said to only encode clausal subordination. For instance, complexes marked with iTab. 26.14: Subordinate relations and their strategies in Wã́nsöjöt. strategy

relative

complement

adverbial

PRO-V in O position i- ‘attributive’ marked complex PRO-V-dik Marked with adverbializing or nominalizing morpheme

✘ ✓ ✘ ✘

✓ ✓ ✓ ✓15

✘ ✓ ✓ ✓

also form adjectives, numerals, and nominalized verbs, among other functions. Non-finite clauses marked with -dik are not necessarily subordinate. Finally, the nominalizing and adverbializing morphemes have a number of functions apart from clause-linkage. The first strategy is illustrated in (104) and (105) with adverbial and complementation strategies respectively. A typical finite sentence contains at least one TAME

15 Not all adverbializing or nominalizing suffixes can be used for complementation (see below).

1184

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

marker. Semantic dependence between two finite clauses typically emerges when there is only one TAME marker between the two sentences. (104) ha-mã́n-bat-da ha-dɤ’ 3sg-run-deic-ass 3sg-fall ‘As he ran, he fell.’ (105) ha-’á-sák ha-ɯ̂j-di huan-jot-hin 3sg-1sg-imagine 3sg-die-pass Juan-dog-nmlz.pst ‘I imagine that Juan’s missing/disappeared dog died.’ The second strategy is illustrated in (106) with a relativizing function. The attributive i- is a productive strategy for marking relative clauses. (106) nã́t papeda-ót i-ha-’ɤjúk-ot Juan-at jáu-da firmando dem paper.sp-pl attr-3sg-bring-pl Juan-erg none-ass signing.sp ka-kôm-di-dípɤn 3pl-mir-have-inc ‘These papers that Juan brought had not been signed (no one had signed them).’ Recall from Section 6.1 that matrix clauses marked with -dik are a type of non-finite clause that focuses on the event (rather than one of the referents of the arguments of the verb). Non-finite clauses marked with -dik constitute a productive strategy for complementation, as can be seen from (107). (107) ha-písi-ha-wák ha-jedók-ha-wâ w-‘ɤk-dik 3sg-know-3sg-know 3sg-respond-3sg-respond-caus-nfin i-hû j-‘ɤ́ k attr-teach-caus ‘[The student] knows how to respond to what she studied/what was taught.’ ́ Finally, Wãnsöjöt has a number of markers that nominalize and/or adverbialize verb complexes. Combining a matrix clause with constructs derived in such a manner is a productive strategy for clause linkage in the language. Some of these class-changing derivations, listed in (108), also encode aspectual and adverbial distinctions themselves. They suffix directly to the right-most verb root, before any TAME markers. The full list of elements that fall into the category of class-changing derivation is not clear. (108) a. -bat ‘distal (prior) event’ (deic) b. -hé’e ‘purpose’ (purp) c. -hĩn ‘past tense nominalizer’ (nmlz.pst)

26 Wã́nsöjöt (Puinave)

1185

Examples of these morphemes marking subordinate clauses are provided in (109) and (110). (109) ka-dúk ka-hû j-hé’e 3pl-enter 3pl-learn-purp ‘They entered to learn.’ ̂ (110) a-mĩnhok-bat bɤjdû t pǎj 1sg-arrive.from.jungle-deic nothing chicha ‘When I arrived from the jungle (over there), there was no chicha.’ The markers in (108) can also attach to the pronoun ó’o (3sg, non-specific) to form clause linkers; for example, see óbat ‘when’ in example (111). ó-bat bi-mînhu-dik hãm’õ (111) a-hu-sɤ̃́ m-di 1sg-ipfv-small-pst 3sg-deic 1pl-arrive-nfin there ‘I was small when we arrived there.’

7.2 Coordination ́ Wãnsöjöt has a few overt markers of sentential coordination; ójem ‘and’; pɤj ‘as well as’; ikóm~ojkóm~sikóm ‘while’; oś ɤm ‘for this reason, therefore’; -hin ‘in contrast, nmlz.pst’. The conjunctions ojem ‘and’, pɤi ‘as well’, and ikóm~ojkóm~sikóm ‘while’ occur at the end of the clauses which they coordinate, as illustrated in (112) and (113). The conjunction pɤj ‘as well’ is distinct from ojem in that it involves gapping or eliding a part of the verb phrase (‘do not grow’ in (113)). The conjunction ojkóm encodes that the events denoted by the sentences occur simultaneously, as in (114). (112) ó’o-da wágot-ot ’ɤjû -m ka-hu-ha-jik ojem 3sg-ass bags-pl bring-ag.nmlz 3pl-open-3sg-open conj ‘She brought bags and opened them.’ ́ (113) ósɤm-da bi-tim-sãn-mĩn-i bi-tɯ́j-ot pɤj therefore-ass 1pl-conuco-neg-grow-neg 1pl-son-pl as.well ‘Therefore, our conuco (sown field) does not grow, and our children as well (do not grow).’ (114) Huan a-hé-mã-u-tɤ́j-denṍ nãn a-si-wok-tó-nõk ojkóm Juan 1sg-wait-2sg-imp-wait-some.time here 1sg-prosp-take-frq-go simul ‘Juan, wait for me here while I go drinking.’

1186

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

Purpose and causation between adjacent sentences can be marked by ósɤm which falls between the two clauses, (115). ́ (115) Mãdia ha-nõk-súk Bogotá ósɤm bɤ̌ n i-t-bêp ́ Maria 3sg-go-desid Bogotá therefore rem.pst attr-rcd.ag-work padatá’ ha-hêu money 3sg-look.for ‘Mariá wanted to go to Bogotá, therefore, (before this), she worked looking for money.’ The past tense nominalizer has an adversative function, illustrated in (116). (116) a-sãn-júk-sút-hin-nĩ-da nãn ’ãm 1sg-neg-come-desid.neg-nmlz.pst-rec.pst-ass here 1sg.pro ́ a-mã́p-a-da a-nõk-súk-hin 1sg-fish-all-ass 1sg-go-desid-nmlz.pst/advrs ‘I didn’t want to come here, but rather, I wanted to go fishing.’

8 Topic and focus ́ This section provides a brief description of topic and focus in Wãnsöjöt beyond the distinction between finite, agent-focused, and object-focused constructions described in Section 6. We understand topic as the phrase or expression which a given ́ utterance is about (van Valin & La Polla 1997: 203). Wãnsöjöt has a dedicated topic marker hɤ̂ j which marks the constituent to its left as a topic. We do not mean to imply that all clauses with a topic-comment structure must have the topic marker. Most likely there are other constructions or conditions which can mark topic in ́ Wãnsöjöt, as there are in all languages. An illustrative example from naturalistic speech is provided in (117). (117)

Speaker 1: bɤ́ i-há-jek-bɯ ja-hé’e? int.attr-trnsl-be.extended-distr 2pl-purp ‘What is there still left to do for you (to finish)?’ Speaker 2: jaj-at-wí-at bɤjdík-át-wɤn-nĩ ́ dû n pɤ́ j pút-ot-hɤ́ j, small-abl-rest-compl what-compl-pot-cmpr there three tree-pl-top, ́ ót-da i-bi-ket’ɤk-nõk-ot dúmãt 3sg-ass attr-1pl-can-prsc-pl now ‘It‘s just a little that is left, with respect to the three logs there we are finishing (them) now.’

26 Wã́nsöjöt (Puinave)

1187

Sentential focus of a constituent, understood as the expression in a phrase which is not presupposed in the utterance (Lambrecht 1994) is marked by: (i) appearing in the first position in the clause, (ii) occurring before an encliticized TAME morpheme, or (iii) both these distributional properties simultaneously. It is not clear which of these properties is more important in cases where the two contradict one another. A cursory illustration of sentential focus is provided in (118) below, with the focused expression appearing in small caps in the translation. ́ (118) újem-mã i-t-nõ’ɤk Mãdia-at mirror-rpt attr-rcd.ag-lose María-erg ‘It is said María lost the mirror.’

9 Conclusion ́ Current evidence suggests that Wãnsöjöt is either an isolate or related to the KakuaNukakan family, although more information on Nukak is needed to assess this hypothesis. The language has a relatively complex phonological inventory compared ́ to typical Amazonian patterns (Aikhenvald & Dixon 1999: 8) with 11 vowels. Wãnsöjöt appears to be a typical Amazonian language in contrasting two tone heights (L and H) (Hyman 2016). The language displays paradigmatic contrasts of up to four tonal patterns on monosyllabic forms (H, L, HL, LH). However, certain constraints on underlying forms (such as the fact that the full range of tonal specifications are only permitted on the first syllable) result in a smaller number of possibilities than would be expected for longer forms. The language displays a complex interaction between morphosyntax and tone structure. As in other languages of the region, ́ Wãnsöjöt has a nominal classifier system with eight members that relate to the cuĺ turally salient semantic domains. Like many Amazonian languages, Wãnsöjöt has an evidential system, although it only consists of one overt category (the reportative). ́ In Wãnsöjöt grammatical utterances can appear unmarked by TAME morphemes. Many TAME markers can occur in multiple positions in the clause, some in wordinternal and some in word-external positions. Tense morphemes do not occur in a fixed position in relation to the verb, and they do not modify nonverbal part of speech classes, although they can phonologically integrate with them. The morphotactics and syntax of bound TAME morphemes overlap with free forms to such a degree that we found it challenging to draw a helpful division between affixes and clitics, but this issue requires future research. ́ With respect to morphological typology, Wãnsöjöt seems to display a fairly loose distinction between morphology and syntax, a phenomenon that Tallman and Epps (2020) suggest is typical of Amazonia. Overall the language is agglutinating, with the important exception of the extended exponence displayed by its verb bases.

1188

Adam J. R. Tallman, Cynthia Hansen, and Jesús Mario Girón

́ Wãnsöjöt has a set of bound pronominal markers integrated into its verb complexes. These bound pronominals are not agreement markers per se, because they do not necessarily co-occur with a full NP with which they are co-indexed. ́ Wãnsöjöt displays an ergative alignment in its case marking system, with important exceptions related to constituent order and construction type. It is hard to pin down precisely the alignment system of all constructions because they can be alternatively analyzed as nonverbal predicate constructions with nominalized ́ clauses. Wãnsöjöt complements its complex alignment system with object incorporation, but it restricts this to inanimate nouns. ́ Wãnsöjöt displays a number of strategies for marking subordination, but it is typical of Amazonian languages in using clausal nominalization as one of its most prolific strategies.

10 References Anderson, Stephen R. 2015. Dimensions of morphological complexity. In Matthew Baerman, Dunstan Brown & Greville Corbett (eds.), Understanding and measuring morphological complexity, 11–28. Oxford: Oxford University Press. Bauer, Laurie. 1990. Be-heading the word. Journal of Linguistics 26(1). 1–31. DOI: 10.1017/ S0022226700014407. Bickel, Balthasar & Fernando Zúñiga. 2017. The ‘word’ in polysynthetic languages: Phonological and syntactic challenges. In Michael D. Fortescue, Marianne Mithun, &Nicholas Evans (eds.), The Oxford Handbook of Polysynthesis, 158–185. Oxford: Oxford University Press. Blevins, James P. 2016. Word and paradigm morphology. Oxford: Oxford University Press. Brown, Cecil. 2017. Evaluating proposals of language genealogical relationship: The BeckWichmann-Brown (BWB) system. Language Dynamics and Change 7. 252–285. Camacho, Alfonso. 1982. Etnografía Puinave. Bogotá: Departamento de Antropología, UN. Caudmont, Jean. 1954. Fonología Puinave. Revista Colombiana de Antropología 2. 265–276. Bogotá: Instituto Colombiano de Antropología. Dixon, R. M. W. & Alexandra Aikhenvald. 1999. Introduction. In R. M. W. Dixon & Alexandra Aikhenvald (eds.), The Amazonian Languages, 1–22. Cambridge: Cambridge University Press. Epps, Patience & Katherine Bolaños. 2017. Reconsidering the “Makú” language family of Northwest Amazonia. International Journal of American Linguistics 83(3). 467–507. Girón, Jesús Mario. 2006. Contraste de oclusivas y nasalidad. Amerindia 29–30. 81–96. Girón, Jesús Mario. 2008. Una gramática del Wãnsõjõt (Puinave). Utrecht: LOT. Girón, Jesús Mario, & Leo Wetzels. 2007. Tone in Wãnsöhöt (Puinave). In Leo Wetzels, Language Endangerment and Endangered Languages, vol. 1, 129–156. Leiden: CNWS Publications. Guillaume, Antoine. 2016. Associated motion in South America: Typological and areal perspectives. Linguistic Typology 20(1). 81–177. Haspelmath, Martin. 2018. The last word on polysynthesis: A review article. Linguistic Typology 22(2). 307–326. DOI:10.1515/lingty-2018–0011. Huber, Randal Q., & Robert B. Reed. 1992. Vocabulario comparativo: Palabras selectas de lenguas indígenas de Colombia. Bogotá: Asociación Lingüístico de Verano. Hyman, Larry M. 2016. Morphological tonal assignments in conflict: Who wins? In Enrique Palancar & Jean Léo Léonard (eds.), Tone and inflection: New facts and new perspectives, 15–39. Berlin & Boston: De Gruyter Mouton.

26 Wã́nsöjöt (Puinave)

1189

Lambrecht, Knud. 1994. Information structure and sentence form: Topic, focus and the mental representations of discourse referents. Cambridge: Cambridge University Press. DOI: 10.1017/ CBO9780511620607. Martins, Valteir. 2005. Reconstrução fonológica do Protomaku Oriental. Amsterdam: Vrije Universiteit, LOT PhD dissertation. Mason, John A. 1950. The languages of South American Indians. In Julian H. Steward (ed.), Handbook of South American Indians, 281–331. Washington, D.C.: Smithsonian Institution. Pabón, Magdalena. 1979. Relaciones interétnicas de contacto y sus efectos en la comunidad indígena puinave del Río Inírida. Degree monograph. Bogotá: Departamento de Antropología UN. Peirsen, Ellen Jean. 1980. Formas de destacar información en las narrativas del Puinave. Artículos en lingüística y campos afines 10. 89–103. Bogotá: Instituto Lingüístico de Verano. Rivet, Paul, & Constant Tastevin. 1920. Affinités du makú et du puinave. Journal de la Société des Américanistes de Paris 12. 60–82. Rose, Francoise. 2015. Associated motion in Mojeño Trinitario: Some typological considerations. Folia Linguistica 49(1). 117–158. Sadock, Jerrold. 1980. Noun incorporation in Greenlandic: A case of syntactic word formation. Language 56. 300–319. Tallman, Adam J.R. 2020. Beyond grammatical and phonological words. Language and Linguistics Compass 14:e12364. https://doi.org/10.1111/lnc3.12364. Tallman, Adam J.R., & Patience Epps. 2020. Morphological complexity, autonomy, and areality in Amazonia. In Gardani Franceso & Peter Arkadiev (eds.), The complexities of morphology, 230–264. Oxford: Oxford University Press. Triana, Gloria. 1985. Los Puinaves del Inírida – Formas de subsistencia y mecanismos de adaptación. Bogotá: Instituto de Ciencias Naturales UN. Van Valin, Robert D. & Randy J. LaPolla. 1997. Syntax-Structure. Meaning and Function. Cambridge: Cambridge University Press. Wavrin, Robert de, & Girón, J.M. 1998. Lengua Puinave. In Jon Landaburu (ed.) Documentos sobre lenguas aborígenes de Colombia del Archivo de Paul Rivet, vol. II, 187–230. Bogotá: Universidad de los Andes.

Alexia Z. Fawcett

27 Wao Terero 1 2 3 4 5 6 7 8 9 10 11

Introduction Phonology Word classes and morphological structure The noun phrase The verb phrase Simple clauses Clause combining Discourse Conclusions Acknowledgements References

1 Introduction Wao Terero (waor1240) is a language isolate spoken in the Amazon region of Ecuador. In previous work, the language has been referred to as Waorani, Huaorani, Huarani, Wao, Huao, Waotededo, Wao Tiriro, Auca (pej.), Auishiri, Awishiri, and Ssabela/Sabela; however, the endonym for the language of the Waorani is Wao Terero meaning ‘the people’s language’. According to the most recent Ecuadorian census, there are 2,416 people who identify as Waorani, and 1,766 Wao Terero speakers (INEC 2010). This chapter provides a summary of what is known about the structure and use of the language and thus describes some basic and salient aspects of: phonology (Section 2), word classes and morphological structure (Section 3), the noun phrase (Section 4), the verb phrase (Section 5), simple clauses (Section 6), clause combining (Section 7), and discourse (Section 8). Wao Terero is spoken mainly on the Waorani Ethnic Reserve, but it is also spoken in other areas of Pastaza and Napo provinces (such as the cities of Puyo and Coca), the Yasuní National Park, and the Taromenani Tagaeri Intangible Zone. The Reserve is the largest state-recognized Indigenous territory in the country, yet it covers only a fraction of the ancestral Waorani territory stretching roughly between the Napo and Curaray rivers. Today, Waorani typically live in permanent to semipermanent communities – of which there are over 30 and as many as 50 (Alianza Ceibo 2018) – where they participate in mainly subsistence hunting, fishing, gathering, and farming. Trekking in the forest is central to traditional Waorani lifeways (Rival 2002) and continues to contribute, along with movement between communities and migration for employment, to a mobile way of life (High 2006).

Alexia Z. Fawcett, University of California, Santa Barbara https://doi.org/10.1515/9783110432732-014

1192

Alexia Z. Fawcett

Wao Terero

27 Wao Terero

1193

Prior to the creation of relatively permanent communities by missionaries starting in the late 1950s, the Waorani lived in self-isolation. Thorough Evangelical missionization, the associated sedentism, and the ongoing impact of extractive industries in the territory have led to rapid cultural and linguistic change (e.g., Yost 1981, Rival 1992, Rival 1996, Lu 1999, Ziegler-Otero 2006, Lu 2007). While elder speakers are more likely to be monolingual, multilingualism is common amongst younger generations. Due to the existence of state-run schools, inter-ethnic marriages, and economic pressure to gain employment outside one’s community, the number of Spanish-Wao Terero bilinguals and Spanish monolinguals is ever increasing. Bilingualism with Lowland Kichwa or Shuar is also on the rise due to the increasing frequency of Waorani-Kichwa and Waorani-Shuar intermarriage and employment of bilingual education teachers who speak Kichwa or Shuar in community schools (Rival 1996, Gondecki 2015, Korak 2015). Although Wao Terero is now recognized as an isolate (Peeke 1973, 1994; Adelaar & Muysken 2004; van Gijn 2014) or as unclassified (Dickinson et al. 2013), previous accounts attempted to group it with other language families. Ruhlen (1987) argues that Wao Terero belongs to the Urarina family of the Andean substock while Greenberg (1987) classifies it as part of the Itucale-Sabela group. Kaufman (2007) proposes that it belongs to the Yaguan family of the Záparo-Yaguan stock. Campbell (2012) classifies it as the only member of the family Sabela but notes that Loukotka (1968) also included Tiwituey as part of the same family (as does Greenberg 1987). There is no demonstrable relationship between Wao Terero and any language that is spoken in the area today; however, there are two groups that live in voluntary isolation near Wao communities, the Tageiri and Taromenani (see Cabodevilla et al. 2004, Proaño García & Colleoni 2008, Gondecki 2011). The Tageiri are known to speak Wao Terero, as they are followers of Tage who split from those who today identify as Waorani in the face of missionization and oil exploitation. However, the Taromenani, ‘people living at the end of the path’, are viewed as culturally and linguistically “similar but different” from the Waorani (Rival 2004). In the 1950s, Summer Institute of Linguistics missionaries began investigating Wao Terero, which has resulted in materials including: a preliminary tagmemic grammar (Peeke 1973), a set of workpapers addressing discourse with an accompanying collection of texts (E. Pike & Saint 1988), a discussion of phonemes (Saint & K. Pike 1962), Master’s theses about word class and developing literacy materials (Peeke 1962, Kelley 1988), pedagogical materials (Peeke 1979, SIL 1995), word and phrase lists (SIL 1978, Orr & Pearson 1979, Peeke 1991, Borman 1991, Wamöñe & Peeke 2009), and a dictionary (Peeke 2015). While Wao Terero remains under-described, Connie Dickinson, Uboye Gaba, Casey High, Nemonte Nenquimo, Oswando Nenquimo, and Mery Nenquihui have built a video corpus documenting the language in a variety of discourse contexts including narratives, traditional stories, conversation, and political speeches (Dickinson et al. 2013). This video corpus in addition to recordings made during the author’s fieldwork in 2011, 2016, 2017, and

1194

Alexia Z. Fawcett

2018 provide the majority of the data analyzed in this chapter. Because these recordings were created in different communities across Waorani territory, multiple dialects are presumably represented in the corpus; therefore, examples cited in this chapter may also reflect some of this variation. To date, no study of dialectal variation in Wao Terero has been conducted; however, speakers note differences across communities, for example, in the realization of nasal spreading. The first line of each example is transcribed in a practical orthography used by speakers, and examples from published sources that are reproduced in this chapter have been re-transcribed to match this community orthography for continuity.1 While the orthography distinguishes nasal and oral consonant allophones, vowel nasality is not reliably represented. Therefore, the second line of each example and in-text references to morphemes represent nasal vowels with a tilde and do not represent excrescent nasals at morpheme boundaries, but they maintain all other conventions (see inventories).

2 Phonology The consonant inventory for Wao Terero is presented in Table 27.1. The nasal consonants are considered semi-phonemic and thus presented in parentheses (further discussed in the following section), and [ɾ] is an allophone of /d/ realized intervocalically. Fricatives and affricates – only appearing in ideophones – are outside the basic phonology and not represented in the table below. Orthographic conventions, when different from the IPA representation, appear in angle brackets. Despite differing in the representation of vowel phonemes, previous accounts agree that there are five oral and five nasal vowels. The inventory in Table 27.2

Tab. 27.1: Consonant inventory.

voiceless stops voiced stops nasals approximants

labial

alveolar

p b (m) (w̃ ) w

t d (n) (ɾ)

palatal

(ɲ) j

velar k g (ŋ)

1 Other practical orthographies exist in which, for example, nasal vowels are marked with dieresis, or is used for both the stop and flap (represented as and here, respectively). Choice of orthography can index ideology and stance of the speaker regarding missionaries and their work and/or Spanish language dominance. I do not aim to take such a stance in this work other than to respect the orthographic choices made by my collaborators and to maintain clarity and important contrasts for the reader.

27 Wao Terero

1195

aligns with that of Saint and Pike (1962) and differs from Peeke’s (1994) proposed inventory of /i, ɪ, æ, ɨ/, and /a/.

Tab. 27.2: Vowel inventory. front high mid low

central

i, ĩ e, ẽ æ , æ̃

back

o, õ a, ã

While vowel duration appears to be contrastive on the surface (okẽmẽ ‘armadillo’ and okẽẽmẽ ‘return’), long vowels are best understood as sequences of two vowels of the same quality – some cases possibly occurring from consonant elision VCV → VV as has been noted in neighboring languages (e.g., Chicham, Kohlberger 2020: 118). There are clear cases of creak found at word and phrase boundaries; however, whether laryngealized vowels should be considered phonemic remains unclear at this time.

2.1 Nasality Nasality is not always contrastive at the level of the segment as progressive nasalization results in a clear allophonic origin of nasal stops in most environments. In (1), voiced oral stops and approximates alternate with nasals depending on the nasality of the preceding vowel. (1)

a. b. c. d.

/go-bo/ /go-bi/ /go-dãdi/ /go-dõ/

[gobo] [gobi] [goɾãni] [goɾõ]

‘go-1sg’ ‘go-2sg’ ‘go-3pl’ ‘go-caus’

e. f. g. h.

/põ-bo/ /põ-bi/ /põ-dãdi/ /põ-dõ/

[põmo] [põmi] [põnãni] [põnõ]

‘come-1sg’ ‘come-2sg’ ‘come-3pl’ ‘come-caus’

The approximant /w/ does not have a nasal allophone in Peeke’s (1994) account, but Saint and Pike (1962) argue that it is nasalized following nasal vowels and that the effect crosses word boundaries, as shown in (2). (2)

a. /õnõwa/ [õnõw̃ a] ‘foot’ b. /tõmẽkã wa kækã/ [tomeŋã w̃ a kjækã] ‘He does well.’ (Saint & K. Pike 1962: 13)

Voiced oral stops can have pre-nasalized variants following a nasal vowel in two contexts: enclitics, in (3a), and word-initial position. Saint and Pike (1962: 8) charac-

1196

Alexia Z. Fawcett

terize the pre-nasalization in word-initial position as optional in (3b) but obligatory between a nasal vowel and a stop, as shown in (3c). (3)

a. /biwĩ=dia/ [biwĩ=ndia] ‘younger brother (emphatic)’ b. /bekã/ [bɪkã] ~ [mbɪkã] ‘He drinks.’ c. /wĩ bekã/ [wɪ ̃ mbɪkã] ‘He does not drink.’ (Saint & K. Pike 1962: 8–9)

The varied realization of this pre-nasalization could be dependent on individual speaker, carefulness of speech, or another parameter. However, in word-initial position as in /baõ/ [mbãn] ‘meat/flesh’, it can be attributed to venting, the phonetic enhancement of voiced stops in positions that are difficult to produce and/or perceive through nasal leakage (Wetzels & Nevins 2018). In contrast, voiceless oral consonants surface with a preceding excrescent nasal when preceded by a nasal vowel. This necessarily occurs across syllable boundaries given Wao Terero phonotactics (no coda consonants) (4a)–(4c) but also across morpheme boundaries, as in (4d) through (4f). These nasals do not show the same duration as a syllable-initial nasal segment. (4)

a. b. c. d. e. f.

/õta/ /õdõpo/ /ẽkebo/ /ã-te/ /ĩ-pa/ /be tẽ-ke/

[õnta] [õnõmpo] [ẽŋkebo] [ãnte] [ĩmpa] [be tẽŋke]

‘nail’ ‘hand’ ‘egg’ ‘say-cvb’ ‘cop-decl’ ‘meet-intn’

However, voiceless stops do alternate with their nasal counterparts when both preceded and followed by a nasal vowel, as in (5a)–(5e) but not when only preceding a nasal vowel, as in (5f) through (5j). (5)

a. b. c. d. e.

/ã/ /ã-te/ /ã-pa/ /ã-kĩ/ /ã-kã/

[ã] [ãnte] [ãmpa] [ãŋĩ] [ãŋã]

‘say’ ‘say-cvb’ ‘say-decl’ ‘to say’ ‘he says’

f. g. h. i. j.

/a/ /a-te/ /a-pa/ /a-kĩ/ /a-kã/

[a] [ate] [apa] [akĩ] [akã]

‘see’ ‘see-cvb’ ‘see-decl’ ‘to see’ ‘he sees’

Word-initially, we only find nasal consonants preceding nasal vowels where they are not contrastive with oral consonants in the same environment, illustrated in (6).2

2 With one exception – bãmo ‘many’ contrasts with mãmõ ‘bring’.

27 Wao Terero

(6)

a. b. c. d. e. f. g. h. i.

[mãnõmãĩ] [mẽɲete] [mẽmpoga] [mĩ:] [mĩŋkaje] [nãmẽnta] [ɲõwõ] [ɲẽ:nẽ] [ŋõŋa]

1197

‘this manner’ ‘knife’ ‘two times’ ‘new’ ‘wasp’ ‘white, light in color’ ‘now’ ‘large’ ‘fire’

Additionally, there are a limited number of morphemes that begin with a nasal consonant followed by a nasal vowel (NṼ) and always surface as such regardless of the preceding vowel. For example, the first-person plural suffix -mõni does not alternate with a form -bõni in oral environments (instead we find gomõni ‘we go’ and põmõni ‘we come’). Suggesting that nasal consonants have become marginal phonemes in Wao Terero, this phonologization is likely due to these segments often occurring in nasalizing environments. Finally, in these cases we also see that while nasalization does not itself act as a morphological marker of plurality, it plays a role in distinguishing singular and plural person reference forms (see Sections 4.2 and 5.1).

2.2 Other processes Another common process is the raising of /ẽ/ to [ĩ] before palatals – exemplified by instances of demonstrative tõmẽ in Table 27.3 (Section 4.1). Additionally, vowel harmony appears to be limited to /e/ and /ẽ/ when preceding a syllable featuring a consonant and [i] or [ĩ], and it is therefore common in second person forms, as in (7). The contexts of, and possible dialectal differences in, the realization of these processes require further research. (7)

a. /ẽñẽ-bi/ [ĩñĩmi] ‘know-2sg’ b. /be-kĩ-bi/ [bikĩmi] ‘drink-fut-2sg’ c. /kẽwẽ-bi/ [kẽwĩmi]/[kĩwĩmi] ‘live-2sg’

2.3 Phonotactics There are no restrictions on the number of syllables in any word class, with options for CV, CVV, V, VV syllable shapes. The phonotactics of Wao Terero allow for two contiguous vowel segments in sequences of oral-oral, nasal-nasal, oral-nasal, and nasal-oral 3 both within and across syllable boundaries; however, there are no observed instances of more than two contiguous vowels. 3 This nasal-oral sequence could synchronically be limited to ẽi ‘go up’, wẽi ‘go down’, and mẽi ‘bring upward’ (Section 5.5).

1198

Alexia Z. Fawcett

Phonotactic restrictions are highlighted in loan words such as dotoro from Spanish ‘doctor’ where we see only CV syllables and no coda consonants nor consonant clusters. Ideophones have different phonotactic restrictions in that they employ closed syllables, as in teik ‘hit’, when describing unchanging states, results of some completed event, or punctuated actions. This word class also exhibits affricates and fricatives, as in chao chao ‘walk through water’, which are not otherwise attested.

2.4 Stress In isolation, primary stress predictably falls on the final syllable if it is heavy (CVV) but on the penultimate syllable if the final syllable is light (CV), as seen in (8).4 (8)

a. b. c. d. e. f.

[oˈkãã] [ˈõnta] [ˈnõĩŋa] [weˈpẽĩ] [õˈnõne] [ẽnẽˈŋare]

‘short’ ‘fingernail’ ‘correct’ ‘red’ ‘mouth’ ‘wide’

The system is a structural one in which certain syllables are relatively more prominent than others. This is clear in (9), where the addition of a morpheme shifts the placement of stress and does not change word class, meaning that this prominence is not tied to the underlying representation but rather realized at the lexical level with one primary stress per word assigned metrically. (9)

a. b. c. d.

[boˈgĩmã] [bogĩmãnˈkoo] [ˈgi:tã] [gi:ˈtãme]

‘cloud’ ‘clouds’ (cloud-qnt) ‘small’ ‘thin’ (small-side)

Yeti Caiga (2012: 47) notes that there are uncommon instances of antepenultimate stress patterns that are used for emphasis. While primary lexical stress is typically clear in isolation, there is not yet a complete understanding of secondary stress and interaction with phrasal prominence. Considering the limited number of publications about Wao Terero overall, the stress system has received a significant amount of attention. K. Pike (1964) proposes the existence of “stress trains” which assign stress rightward from the beginning of

4 As noted by a reviewer, this may indicate that the VV sequences were historically VCV and the stress remained on the same mora after re-syllabification.

27 Wao Terero

1199

the word and leftward from the end. The train moving in the rightward direction stresses odd syllables of the stem and the train moving leftward from the end of the word stresses even syllables of the suffixes. Halle and Kenstowicz (1991) also provide a bidirectional analysis, but with proposed cyclic and non-cyclic rules applying to the stem and suffix respectively. Keeping with a bidirectional analysis, Hayes (1995) discusses Wao Terero as an example of a stress system with syllabic trochees and claims that it has a weak prohibition on degenerate feet where they are only allowed on the surface in strong position. Fitzgerald (1999: 117) argues against a bidirectional analysis using OT by proposing that “the stress pattern is actually a combination of rightward directional effects combined with a restrictive prohibition on what type of feet occur in suffix trains.” More specifically, she posits “a domain-specific constraint that is violated by the presence of degenerate, nonbinary feet in the suffix train”. Monodirectional analyses are also favored by Lester (1994; using theories of metrical, lexical, and harmonic phonology) and Peeke (1994). K. Pike also writes that “the most frequent [intonational pattern] has a high pitch on stressed syllables, with stressed syllables in the suffix train cascading a bit lower in a contour ‘fade’” (1964: 430). Hayes (1995: 184) interprets K. Pike’s “fade” across syllables in the suffix to mean that “the H* pitch accent must be associated to the rightmost stress assigned within the stem domain.” However, Lester (1994: 27) finds that the acoustic correlates of Wao Terero stress are intensity and duration rather than fundamental frequency.

3 Word classes and morphological structure Wao Terero is a synthetic, agglutinative, and predominately suffixing language. Verbs can take derivational suffixes that change valency (Section 5.4) or denote spatial orientation/direction (Section 5.5), as well as inflectional suffixes for person, tense, aspect, and mood (Section 5.3). As such, there is more complex suffixing on verbs than on nouns; however, nouns can take suffixes for oblique case marking (Section 4.6) and some spatial relations (Section 4.7). Other components of the noun phrase such as pronouns and demonstratives are closed classes. Word class is not always simple to determine morphologically since morphemes such as classifiers (Section 4.5) and person suffixes can co-occur across classes. For example, morphology that is typically found with verbs can also associate with predicate nouns or adjectives, such as wao-bo ‘person-1sg; I am a (Wao) person’. However, word classes can be distinguished using a combination of functional and distributional properties. Adjectives denoting size and other descriptive properties, such as giitã ‘small’, akĩĩ ‘soft’, teẽme ‘heavy’, and mĩĩ ‘raw’, may be derived from verbal material. Addi-

1200

Alexia Z. Fawcett

tionally, there appears to be a marginal strategy deriving adjectives using the suffix -i (wepẽi ‘red’ from wepẽ ‘blood’) (see Section 4.4). Postpositions are used to express spatial relations (e.g., kẽnkarẽ ‘inside’, yeka ‘above’ (Section 4.7)), but there are other suffixing strategies for expressing similar relations (õnko-ro ‘house-edge; outside the house’). Some of these postpositions could have grammaticalized from verbs, and many feature classifiers. There are postpositions for non-spatial relations such as comitative tõnõ or the similative enclitic =baĩ, which can have scope over an adjacent noun (dorãni baĩ ‘like the ancestors’) or the whole phrase. Similar to postpositions, there is a closed class of clause combining elements such as beyẽ ‘because’ (Section 7.4), which occur at the end of the adjoining, non-matrix clause. There is extensive use of ideophones which cover cross-linguistically common semantic domains and follow Dingemanse’s (2012: 663) implicational hierarchy [sounds > movement > visual patterns > sensory perceptions > inner feelings and cognitive states] where sensory perceptions and all preceding categories are attested. Respective examples: yee ‘cry’, yak ‘grab’, ñãn ‘light’, and tĩn ‘salty’ (Gaba 2011). Ideophones can have general meanings when unaccompanied but gain more specific senses in conjunction with a verb or from context. For example, the ideophone tẽn can be used independently to mean ‘stuck’, with the verb ‘sit’ for sitting down (tẽn kontate), or with ‘die’ to mean ‘fall dead’ (tẽn wẽngã). Some ideophones are inherently reduplicated such as ka ka for laughter, but other instances of reduplication denote event structure, duration, or aspect. Ideophones can occur in a variety of morphosyntactic constructions, constituting their own intonation unit as well as modifying and co-occurring with verbs (see Fawcett 2018 for discussion of morphosyntactic integration, and Section 5.7). While the language on the whole is agglutinative, there are some instances of fusional or seemingly non-compositional morphemes or words which display some fossilized morphological complexity, such as the verbal negation marker -rãmaĩ (Section 6.4).5

4 The noun phrase The Wao Terero noun phrase can contain simply a noun, a demonstrative (Section 4.1), a pronoun (Section 4.2), or a numeral (Section 4.3). Numerals, other quantifiers, adjectives, and adjectivized nouns can modify nouns in the phrase (Section 4.4). This section also describes classifiers and their role in derivation (Section

5 Due to nasal spreading from the preceding vowel as discussed in (Section 2.1), all consonantinitial morphemes have two allomorphs beginning with an oral or nasal consonant (i.e., -rãmaĩ/nãmaĩ); however, these morphemes are represented in-text using the oral consonant allomorph.

27 Wao Terero

1201

4.5), case marking (Section 4.6), spatial suffixes and postpositions (Section 4.7), and other nominal morphology (Section 4.8).

4.1 Demonstratives Demonstratives or deictic roots can appear in their bare forms as referring expressions or take classifiers, spatial and location suffixes, or a suffix for time as shown in Table 27.3 and exemplified in (10). While the distinction between ĩ and ĩnĩ appears to be proximal versus distal, the distinction between mãnĩ dem1 and tõmẽ dem2 could be related to information structure, as seen in Section 4.2.

Tab. 27.3: Demonstratives.

dem.prox dem.dist dem1 dem2

bare

location -yõmo

time -yerẽ

path/side -rõ

clf:tree -wẽ

ĩ ĩnĩ mãnĩ tõmẽ

ĩ-ñõmo ĩnĩ-ñõmo mãnĩ-ñõmo tõmĩ-ñõmo

ĩ-ñerẽ ĩnĩ-ñerẽ mãnĩ-ñerẽ tõmĩ-ñerẽ

ĩ-nõ ĩnĩ-nõ mãnĩ-nõ tõmẽ-nõ

ĩ-wẽ ĩnĩ-wẽ mãnĩ-wẽ tõmẽ-wẽ

(10) Maniñomo waponi menonte owokimini epene. mãnĩ-ñõmo wa=põni mẽnõ-te õ-wo-kĩ-mĩni epẽ-nẽ dem1-place well=very construct-cvb be-suspend-fut-2pl water-loc ‘Here you could build a nice home and live at the river (suspended in hammocks).’ (07Feb1105-Oswando-25:34) Demonstratives without suffixes can precede a head noun, as in tõmẽ obe ‘this boa’ or ‘this river’ in (11). (11) Igikii mani epe igi pere ta apa. igikii mãnĩ epẽ igi pere ta aa-pa near dem1 water go.downstream pass.through come.out exist-decl ‘It’s close, the river that comes out (from the mouth of the river).’ (07Feb1105-Ramón-55:39)

4.2 Person reference Wao Terero pronouns, in Table 27.4, show a distinction between singular, dual, and plural for all persons. We also see what has been previously discussed as a clusivity distinction in the first-person plural (Peeke 1973, Crevels & Muysken 2005). While

1202

Alexia Z. Fawcett

Tab. 27.4: Person pronouns. singular 1

plural

incl

botõ

tõmẽmo

mõna(tõ)

tõmẽmõna

bitõ mĩnã ĩngã ĩnã

tõmĩmi tõmĩnã tõmẽngã tõmẽnã

mĩnatõ

f.hon m f

ĩna

2 3

dual

tõmẽmĩna

mõno mõni(tõ) mĩnito

tõmẽmõ tõmẽmõni tõmĩni

tõmẽna

ĩnãni

tõmẽnãni

mõno ‘we’ can be used as an inclusive form, mõni(tõ) ‘we’ is not necessarily exclusive. Furthermore, mõno additionally connotes ‘we as opposed to others’ or ‘you and I in a relationship (however defined)’. There is also a gender distinction in the third-person singular, where the feminine is used for women and the masculine for all others: men, children, unknown persons of all genders, and some non-human animates. There is more than one pronoun per person and number reference; beyond the basic set represented in the left columns of each number category, there are pronominal forms based on deictics (Heine & Song 2010). Example (12) shows an instance of the first set, and (13) shows the second set with demonstrative base tõmẽ that is often translated as ‘I, myself’/‘you, yourself’/‘we, ourselves’, suggesting contrastive focus or emphasis. While not possible in the first-person singular nor any second-person forms, demonstrative mãnĩ can also be used for person reference, as shown in (14). (12) Manomai minito amini boto eñenamai imopa. mãnõ-maĩ mĩnitõ ã-mĩni botõ ẽñẽ-nãmaĩ ĩ-mo-pa dem.manner-sim 2pl.pro say-2pl 1sg know-neg cop-1sg-decl ‘You speak like this, but I don’t understand.’ (29Jun1303-Pablito-53:19) (13) Tomenani apote mamo anoyomo kononkerani. tõmẽ-nãni a-po-te mãmõ ano-yõmõ kõnõ-ke-rãni dem2-3pl ?-clf:hand-cvb bring one-place leave-intn-3pl ‘Those people, preparing the chicha (mixing with their hands), have to come and leave it in the same place.’ (07Jul1601-Mintape-10:48) (14) Kengi ge ente webopa aen maninga ingante. kẽŋi ge ẽ-te we-bo-pa ã-ẽ mãnĩ-ŋã ĩ-ŋã-te food hunger have-cvb worry-1sg-decl say-irr dem1-3sg cop-3sg-cvb ‘Tell her “I’m hungry because there’s no food”.’ (07Jul1602-Mintape-11:41)

27 Wao Terero

1203

Tab. 27.5: Relative possessive pronouns. singular 1incl 1 2 3

mĩno nãno

dual

plural

mõna

mõno mõni mĩni nãni

mĩna nãna

The pronominal forms listed in Table 27.4 can all function as possessors without any additional marking. There is no alienable/inalienable possession distinction: The possessum is unmarked and preceded by the pronoun or nominal possessor, which is also unmarked, as seen in (15). (19) a. b. c. d. e.

mõno wẽmẽ-iri mõni girĩ-ŋã bitõ wẽ-na barã tonĩña-ra ĩŋã õnko

1pl.incl grandparent-coll 1pl.excl relative-m 2sg child-du mother brother-du 3m house

‘our grandparents’ ‘our relative’ ‘your two children’ ‘the mother’s brothers’ ‘his house’

Body parts can be referred to with possessive constructions such as bito onowa ‘your foot’; however, many translation equivalents of body part possession look like (16) in discourse, where classifiers are incorporated into verbs (Section 5.2). (16) Nantakawabopa. nãnta-ka-wa-bo-pa hurt-clf:head-clf:foot-1sg-decl ‘My ankle hurts.’ Peeke (1973: 130) identifies the forms in Table 27.5 as relative possessive pronouns, as they are central to relativization (Section 7.1) and can be used as possessive pronouns. However, these pronouns also appear to denote an affinity, closeness, or reverence. For example, in (17) the speaker uses the relative second singular form mino to address an elder. (17) E, werenke terekimi, Kope e mino ebano ponemi? e wẽrẽnke tere-kĩ-mi Kope e mĩno ebano põnẽ-mi intrj little speak-fut-2sg prop.n intrj rel.2sg how think-2sg ‘Will you speak a little, Kope. What do you think?’ (13Jun1105-Cesar-18:27) These same forms can take suffixes that further define the relationship between the indicated persons. For example, in (18) -kabo ‘group/bunch’ (same as clf:head) is

1204

Alexia Z. Fawcett

suffixed to nani rel.3pl to refer to extended families/clans or other groups, and -kaya ‘between/amongst’ can co-occur with a subset of the above relative possessive pronouns as in (19) and (20). (18) Maniñomo boto weme Waawe ingante tenonani wenganta nanikabo anobain. mãnĩ-ñomõ botõ wẽmẽ Waawe ĩ-ŋã-te tẽnõ-nãni wẽ-ŋã-ta dem1-place 1sg grandfather prop.n cop-3m-cvb spear-3pl die-3m-pst nãni-kabo ãno=baĩ rel.3pl-group one=sim ‘In this place, the group of them speared my grandfather, Waawe, dead.’ (07Ju1601-Mintape-34:29) (19) Monato monakaya Kawiya tono terekemonapa. mõnatõ mõna-kaya Kawiya tõnõ tere-ke-mõna-pa 1du rel.1du-as.pair.siblings Kawiya com speak-intn-1du-decl ‘Kawiya and I are going to talk between us two siblings.’ (04Dec0501-Yolanda-00:15) (20) Nanabarankaya nonkonatapa ante ake kera apa. nãna-barã-kaya nõ-kõ-na-ta-pa ã-te a-ke ke-ra rel.3du-mother-as.pair spin-prog-3du-pst-decl say-cvb see-intn do-3du a-pa see-decl ‘She is about to see and say that the mother and daughter are spinning together.’ (07Jul1601-Mintape-01:30) Finally, the Dravidian-style kinship system of the Waorani (Rival 2002) and traditional cross-cousin marriage practices (preference for marriage with a child of your parent’s opposite-sex sibling) are integral to the understanding of the Wao Terero system of person reference. In this system, one uses the honorific form for mothers and maternal aunts (who are both referred to as ‘mother’ and all their children as siblings) and the dual form not just for reference to two people, as in (21), but for all in-laws and potential in-laws (parent’s opposite-sex siblings and their spouses, etc.), as in (22). (21) Ñonekoo iñomo ñowo mina gomina ñonekoo mani taromengare magi ire oko ire. ñõne-koo ĩ-ñõmo ñõwo mĩna go-mĩna ñõne-koo mãnĩ bamboo-qnt dem.prox-place now rel.2du go-2du bamboo-qnt dem1 tarõmẽgare mãgi ire õnko ire far.inside bring.inward bend house bend ‘You two (can) go to this area where there is a lot of bamboo. (He) brought (them) far inside (the forest) and made a house, bending (the bamboo).’ (24Dec1002-Omare-54:12)

27 Wao Terero

1205

(22) Mina anonke maniyere boto imo mamo teik takamina. mĩna ã-nõ-ke mãnĩ-yerẽ botõ ĩ-mo mãmõ teik rel.2du say-cond-sub dem1-time 1sg cop-1sg bring ideo.pierce taka-mĩna kill-2du ‘In that time, you were saying that you would have come to kill me.’ (20Abr1102-Toka-20:32)

4.3 Number and quantification Using numerals is infrequent; however, there is a small set of basic cardinal numerals for 1, 2, 5, and 10, while the remainder are compositional, as seen in (23). (23) a. b. c. d. e. f. g. h. i.

1 2 3 4 5 8 10 15 20

aro-ke mẽa mẽa go aroke mẽa go mẽa ẽmẽ-po-ke ẽmẽmpoke go mẽa go aroke tipẽ-po-ga tipẽmpoga go ẽmẽmpoke tipẽ-wa6

one-rest two two-plus-one two-plus-two one.side-clf:hand-rest five-plus-two-plus-one span.sides-clf:hand-surface ten plus five span.sides-clf:foot

In the event that they are used, numerals tend to precede the head noun, as shown in (24), or are used as head of the noun phrase, illustrated by (25). (24) Oweta kenkare ongompa aroke geye. oweta kẽnkarẽ õ-ŋõ-pa aroke geye bowl inside be-upright-decl one fish ‘One fish is in the bowl.’ (Topological Relations Picture Series, elicited) (25) Mea go aroke tenomona wepa anani. mẽa go aroke tẽnõ-mõna wẽ-pa ã-nãni two plus one spear-1du die-decl say-3pl ‘They said, “we two speared three (peccary) dead.”’ (07Ene1101-Awa-36:21) Strategies for reference to plural animates differ by word class. The plural forms in (26) are based on adjectival or adverbial roots and use the same person and number marking as appears on verbs (Section 5.1). 6 Some speakers use tipempoga go tipewa [span.sides-clf:hand-surface plus span.sides-clf:foot].

1206 (26) a. b. c. d. e.

Alexia Z. Fawcett

wĩñẽ-nãni pikẽ-nãni õnkiyẽ-nãni bake-rãni do-rãni

young-3pl old-3pl female-3pl unmarried.female-3pl before-3pl

‘children’ ‘elders’ ‘women’ ‘unmarried women’ ‘ancestors’

In (27), pluralizing nominal roots to refer to animates is achieved by the suffix -iri. While this suffix is used with both human and non-human animates, it seems to have a collective reading, which is further evidenced by the fact that the same -iri morpheme can also be suffixed to a proper name with an associative sense, to refer to the group of people associated with or led by that individual. One can refer to Daboiri ‘the Dabo clan/family/group’ after their leader’s given name, Dabo.

(27) a. b. c. d. e.

mẽñẽ-iri oba-iri mẽmẽ-iri nagapa-iri kowore-iri

jaguar-coll armored.catfish-coll grandfather-coll Shuar-coll outsider-coll

‘jaguars’ ‘armored catfish’ ‘grandparents/ancestors’ ‘Shuar people’ ‘outsiders’

The suffix -koo can be used with most nominals including those listed in (28); however, -koo also has a distributive sense or perhaps that of a more general quantifier ‘many’, as in (29).

(28) a. b. c. d. e. f. g.

õnko-koo mẽñeka-koo ore-koo awẽ-koo tubo-koo wẽ-ĩ-koo õnkiyẽ-koo

house-qnt fruit-qnt peccary-qnt tree-qnt tube.sp-qnt die-ptcp-qnt women-qnt

‘houses’ ‘fruits’ ‘white-lipped peccaries’ ‘trees’ ‘tubes/pipelines’ ‘dead ones’ ‘women’

(29) Manino orekoo pon gotapa amo. mãnĩ-nõ ore-koo põ go-ta-pa ã-mo dem1-path peccary-qnt come go-pst-decl say-1sg ‘Many peccaries passed through here (came and went).’ (07Feb1105-Okata-2:16)

This suffix is related to bakoo ‘many’, demonstrated in (30), and in this form it can take person marking bakorãni ‘many-3pl; many people’.

27 Wao Terero

1207

(30) Bakoo onani wen. bakoo õ-nãni wẽ many hunt-3pl die ‘They kill many (animals).’ (14Aug0802-Coba-2:41) Quantification is not obligatory but depends on definiteness. For example, iwa ‘howler monkey’ can refer to a single monkey or howler monkeys generally as a species, but one can refer to multiple howler monkeys more specifically with collective plural marking iwairi or quantification iwakoo. While more of a quantifier semantically, tõmaa ‘all’ behaves like demonstratives (Section 4.1) – appearing in its bare form, as in (31), or with person marking (tõmamĩni ‘all-2pl; all of you’). (31) Tomaa awenii aera amo. tõmaa awẽnẽ-ĩ a-e-ra ã-mo all record-ptcp see-irr-du say-1sg ‘I want you two to see the whole recording.’ (07Feb1105-Oswando-53:02)

4.4 Modification Many adjectival modifiers appear to be morphologically complex given similar sequences across the class. While the morphological analysis of these modifiers is not synchronically transparent, the existence of pairs such as waẽmẽ ‘sweet’ ~ waẽmõ ‘nice/beautiful’7 or teẽmẽ ‘heavy’ ~ teẽmõ ‘hard’, might be signs of fossilized morphological complexity. Adjectival modifiers can appear post-nominally, as in (32)– (33), and pre-nominally, like in (34) and (35). (32) Awemparo menkayonta wepei kon kein. ãwẽ-pa-ro mẽnkayõnta wepẽ-i kõ ke-ĩ tree-flat-edge paper blood-adjr leave do-ptcp ‘On the table red paper is placed.’ (Positionals-Marcia-02:14) (33) Boto awenkata waemo mempo ente ponga. botõ ãwẽnkata waẽmo mẽmpo ẽ-te põ-ŋã 1sg shoe nice father have-cvb come-3m ‘Dad brings my nice shoe.’ (Yeti Caiga 2012: 101) 7 These examples might be derived from verbs/verb phrases, perhaps: wa ‘well’ and ẽ ‘have’.

1208

Alexia Z. Fawcett

(34) Monito ñeene onko menonte… mõnitõ ñẽẽne õnko mẽnõ-te 1pl.excl large house build-cvb ‘We, constructing a big house…’ (13Jun1105-Amoa-52:34) (35) Ñene amo teeme awe maniganka awenpa ganka i. ñeẽne ã-mo teẽmẽ awẽ mãnĩ-gãnka ãwẽ-pa gãnka ĩ big say-1sg heavy tree dem-up.to tree-flat up.to cop ‘It was really big and heavy, up to here like the size of this board.’ (07Feb1103-Okata-53:37) However, in natural discourse, qualities are often expressed as predicates, such as wĩñẽ ĩ-te ‘young cop-cvb; being young’, or with person and classifier suffixes, shown in (36) and (37). (36) a. b. c. d. e.

wĩñẽ wĩñẽ-ŋã wĩñẽ-na wĩñẽ-nãni wĩñẽ-wẽ

young young-3m young-3du young-3pl young-clf:tree

‘child/young’ ‘son’ ‘two children’ ‘children’ ‘secondary forest’

(37) Oon pikewe ii, pikewe, ino ganka awoto taaro ganka goro winewe. õõ pikẽ-wẽ ĩ pikẽ-wẽ ĩ-nõ gã-ka awoto yes old-clf:tree be old-clf:tree dem.prox-path spat-clf:head car taarõ gã-ka go-rõ wĩñẽ-wẽ path spat-clf:head go-path young-clf:tree ‘Yes, it is primary forest (old-growth forest). Through there up to the road is secondary forest.’ (07Feb1101-Okata-01:36) Additional morphemes, including the augmentative -bo, can modify the nominals to which they are suffixed, as in tẽnte-bo ‘snake-aug; big snake’. Basic color terms in Wao Terero have been noted as exceptions to proposed universals of color terminology (Berlin & Kay 1969). Kay et al. (2009) find Wao Terero to be the only language in their World Color Survey (WCS) to have one term for white and yellow (nãmẽnta ‘white, light-colored’) and a separate term for red (obatawẽ ‘red’). Wao Terero conforms to other cross-linguistic tendencies in that there is a term wẽntãmo ‘black, dark-colored’ and wĩñarẽ ‘grue’, but based on the data in the WCS, the color-naming system overall was found to not “fit the structure of perceptual color space” (Regier, Kay, & Khetarpal 2009: 891). As evidenced by the transparently derived term wepẽi ‘red’ and references to color drawing on likeness to other objects, as in (38), color reference is not limited to the basic terms.

27 Wao Terero

1209

(38) Peneme peneme mono emokaro. peẽnẽ-mẽ peẽnẽ-mẽ mõno ẽ-moka-ro plantain-clf:liquid plantain-clf:liquid rel have-clf:neck-nmlz? ‘The one (type of fish) with a very yellow neck (like the color of plantain drink).’ (20Feb0701-Εnqueri-18:51)

4.5 Classification and derivation The classifier system in Wao Terero resembles a multifunctional (Krasnoukhova 2012) or multiple classifier system like those described for other Amazonian languages (e.g., Tariana, Aikhenvald 2000), where multiple word classes take classifier suffixes. For Wao Terero, these have also been analyzed as bound nominal roots (Fiddler 2011). A large portion of those identified to date are related to parts of the body, but there are also classifiers such as -bo for spherical objects and -pẽ referring to various liquids, as in (39). If there is a recoverable meaning for the portion of each word preceding -pẽ in (39a)–(39c) it is not yet known. Fully lexicalized, these independent nouns either contain the classifier historically, or epẽ ‘water’ may have served as the source of the classifier. The diachronic development of these classifiers requires further investigation.

(39) a. b. c. d. e.

epẽ wepẽ gãwapẽ dai-pẽ clay-clf:liquid peẽnẽ-mẽ plantain-clf:liquid

‘water’ ‘blood’ ‘lake’ ‘liquid clay’ ‘ripe plantain drink’

Classifiers also, in what might be considered compounding, play a role in noun formation. Some Wao Terero nouns are transparently built from a root plus a series of classifier suffixes. For example, Peeke (1973: 125) notes that the word for ‘headdress’ okabogata is composed of o- an unidentified root, followed by -ka ‘stone/ head’, -bo ‘egg/sphere’, -ga ‘place/surface’, and -ta ‘shell/covering’. While these single syllable classifiers have significant individual meaning, she also finds that classifier compounding can result in suffixes such as -repo ‘season/year’, composed of -re ‘mouth/rim’ and -po ‘hand/times’. Another example, -karẽ, from -ka ‘head/ top’ and -rẽ ‘abdomen/inside’, is used for objects capable of containment. It is found in lexical items for traditional vessels such as (40a)–(40b). For reference to nontraditional items, -karẽ suffixes to verb forms (40c), loans (40d)–(40e), and demonstratives (40f).

1210 (40) a. b. c. d. e. f.

Alexia Z. Fawcett

tikarẽ wãmõ-karẽ be-kã-ĩ-karẽ boteya-karẽ mareta-karẽ tõmẽ-karẽ

necklace-clf:vessel drink.v-3sg-ptcp-clf:vessel bottle.sp-clf:vessel suitcase.sp-clf:vessel dem2-clf:vessel

‘pot for cooking’ ‘container for accessories’ ‘cup’ ‘bottle’ ‘purse’ ‘this (lampshade)’

As mentioned in Section 4.1, classifiers can appear with deictic roots to create proforms that operate as full noun phrases, such as in (40f) and (41). They also have reference tracking functions; for example, in (42) the copula takes a classifier (see Section 6.1 for more on this object marking construction using copulas) to refer to the same piece of bark mentioned in (41). (41) Inta gante ñowo pina gawae tein inkete manita. ĩ-ta gã-te ñõwo pĩ-na gã-wa-e tẽĩn dem.prox-clf:shell press-cvb now angry-3f press-clf:foot-irr ideo ĩnkete mãnĩ-ta but dem1-clf:shell ‘Now step hard on this piece (of bark). This piece here.’ (12Feb2003-Marcia-01:24) (42) Maninke inta pues aki ininke tarongi impa anata. mãnĩ-ke ĩ-ta pues aki ĩ-nĩ-ke ta-rõ-ŋĩ dem1-rest cop-clf:shell well.sp soft cop-pfv-sub come.out-caus-fut ĩ-pa ã-nã-ta cop-decl say-3f-pst ‘She said to take out just this, the softer piece (of the bark).’ (12Feb2003-Marcia-01:32) Classifiers can function as nominalizers, as in (43), and play a role in spatial descriptions (Section 4.7). They also appear with adjectives (Section 4.4), verbs (Section 5.2), and interrogative words (Section 6.2). (43) Tawariya enkebo rojobo ino kenamai ananta dotoro, tawariya ekebo kenkaremo. tawariya ẽkebo rojo-bo ĩno kẽ-nãmaĩ ã-nã-ta dotoro chicken egg red.sp-clf:round cop eat-neg say-3f-pst doctor.sp tawariya ẽkebo kẽnkarẽ-mo chicken egg inside-clf:round ‘The doctor said that you shouldn’t eat the yolks of chicken eggs.’ (07Feb1104-Cesar-47:20)

27 Wao Terero

1211

4.6 Case and postpositions There are two oblique case markers that are realized as suffixes on the head noun – instrumental -ka and locative -rẽ in (44) and (45), respectively. See Section 6.1 for a discussion of object marking. (44) Ore arobain tapaka tenonani. ore aro=baĩ tapa-ka tẽnõ-nãni white.lipped.peccary one=sim spear-ins spear-3pl ‘They also kill white-lipped peccary with spears.’ (14Aug0801-Coba-1:08) (45) Koyotowekarene ongo weke. koyotowe-karẽ-nẽ õ-ŋõ weke bottle-clf:vessel-loc be-upright frog ‘The frog is inside the bottle.’ (19Jun1601S1-Linda-2:05) The restrictive enclitic =ke and similative =baĩ are often realized on head nouns but have clausal scope (Section 7.4). There is also a comitative, tõnõ, that occurs postnominally, demonstrated in (46) and (47). (46) Wita tono inate awene. Wita tõnõ ĩ-na-te awẽnẽ-ẽ prop.n com cop-2du-cvb take.photo-irr ‘Take the photo with Wita.’ (07Feb1104-Ana-28:20) (47) Bakeka iñomo waana tononke. bakekã ĩ-ñõmo wããnã tõnõ=ke unmarried.female dem.prox-place mothers com=rest ‘The young women were only with the mothers.’ (07Jul1601-Mintape-14:46)

4.7 Spatial suffixes & postpositions Classifiers can suffix to nominal elements to indicate a part of said nominal, such as ãwẽ-ka-bo ‘tree-clf:head-clf:round; tree tops’ or pentibo-karo ‘pambil.palmclf:nose; tip of the pambil palm.’ Select examples of classifiers, their spatial extensions, and related adpositions are listed in Table 27.6 and exemplified in (48) and (49).

1212

Alexia Z. Fawcett

Tab. 27.6: Body parts, classifiers, and spatial descriptors. body parts8

CLF

spatial extensions

related adpositions

okabo ‘head’ õnõne ‘mouth’ õnõnẽ ‘abdomen’ õnõkaro ‘nose’ õnõwa ‘foot’

-ka -re/-ne -rẽ/-nẽ -karo -wa

‘head, top’ ‘at mouth/rim’ ‘inside’, ‘inner part’ ‘tip/end/edge’ ‘base of’

yeka ‘above’, gãnka ‘up to’ yewere ‘at mouth/opening/rim’ kẽnkarẽ ‘inside’ yekaro ‘ahead’, ‘at tip/end/edge’ yewa ‘below’

(48) Awe kenkare pawenka ogogampa. ãwẽ kẽnkarẽ pawẽnka õ-ŋõ-ŋã-pa tree inside owl be-upright-3sg-decl ‘There is an owl in the tree.’ (15Ago1503-Milton-07:35) (49) Maniñabo ganka eirinke. mãnĩ-ñabo gã-ka ẽi-rĩ-ke dem1-clf:leaf spat-clf:head go.up-pfv-sub ‘They had gone up to that leaf.’ (07Ene1101-Awa-51:54) These classifiers are also used for spatial descriptions beyond adpositions, suffixing to verbal predicates, as in õ-ŋõ-karo ‘be-upright-clf:nose; be positioned at tip of the branch(es)’ (Section 5.2). Spatial adpositions unsurprisingly extend to expressions of time, as in ñõwõ-ne gã-ka ‘now-loc spat-clf:head; up to/until today’.

4.8 Other nominal morphology There is extensive use of a suffix on proper names and kinship terms indicating they are deceased, as in (50). (50) a. b. c. d.

Boya-wori Boya-wori-ri mẽmẽ-wori botõ barã-wori

prop.n-dec prop.n-dec-pl grandfather-dec 1sg mother-dec

‘the late Boya’ ‘the late Boya’s clan’ ‘(my) late grandfather’ ‘my late mother’

8 Related body part words are listed here, and some classifier glosses rely on body part terms; however, as noted by a reviewer, these may not necessarily be the source of the classifiers.

27 Wao Terero

1213

5 The verb phrase Verbs can appear as bare stems with no suffixing morphology with inanimate subjects, when the subject is understood from discourse context, and in serial verb constructions or complex predicates (Section 5.7). However, verbs can take multiple suffixes, namely: inflectional ones for person, tense, aspect, and mood; and derivational suffixes that change valency. As such, the verbal template is roughly as shown in Figure 27.1.

stem

valency

spatial

classifier(s)

aspect

tense

person

mood

Fig. 27.1: Verbal template.

Variability exists in the ordering of morphemes. For example, past tense -ta follows the second- and third-person suffixes, as in (51); otherwise, tense suffixes precede person suffixes. Possible explanations for this pattern have not yet been addressed. (51) a. go-ta-bo-pa go-pst-1sg-decl ‘I went.’

b. go-bi-ta-pa go-2sg-pst-decl ‘You went.’

c. go-ta-mõni-pa go-pst-1pl.excl-decl ‘We went.’

d. go-rãni-ta-pa go-3pl-pst-decl ‘They went.’

5.1 Person suffixes Just as with pronouns (Section 4.2), Wao Terero verbal person suffixes distinguish number (both dual and plural), gender in the third-person singular, as well as a female honorific, as seen in Table 27.7. Finite verb forms take person suffixes, cross-referencing the human subject of the clause. Inanimate subjects are not cross-referenced on the verb, but higher order

Tab. 27.7: Person suffixes.

1

incl excl

2 3

f.hon m f

singular

dual

plural

-bo/-mo

-mõna

-bi/-mi -mĩ -kã/-ŋã -rã/-nã

-mĩna

-mõ -mõni -mĩni

-ra/-na

-rãni/-nãni

1214

Alexia Z. Fawcett

animates can be, depending on context. For example, the snake in (52) is not crossreferenced, while the tapir in (53) is. (52) Amo eiyomi tente pokemiñe webi. ã-mo ẽi-yõ-mi tẽnte pokẽ-mĩñẽ we-bi say-1sg go.up-simul-2sg snake bite-clf:buttocks worry-2sg ‘As I said, when you go up there a snake can bite you in the buttocks.’ (07Feb1105-Ana-51:02) (53) Arobai tapaka tenonani wenga tite. aro-baĩ tapa-ka tẽnõ-nãni wẽ-ŋã tite one-sim spear-ins spear-3pl die-3sg tapir ‘They also kill tapir with spears.’ (Lit. ‘They spear, tapir dies.’) (14Aug0802-Coba-00:39)

5.2 Classifiers Classifiers do not behave like the person suffixes outlined above; rather, they are incorporated to the verb (Type IV classificatory noun incorporation, Mithun 1984). There can be more than one classifier expressed on the verb (54), they appear directly following the verb stem and therefore precede suffixes for person when expressed (55), and they can also co-occur with a co-referential noun (56). (54) Ae toma yoge gikape pere tae. a-e tomaa yoge gi-ka-pẽ pere ta-e look-irr all enter enter-clf:head-clf:liquid go.through come.out-irr ‘Look, put your whole head in the water and come out.’ (07Feb1105-Moipa-31:52) (55) Do, do, do wemiñenga. do do do weẽ-mĩñẽ-ŋã ideo.shiver ideo ideo shake-clf:buttocks-3m ‘He’s shaking shaking shaking in the buttocks.’ (07Feb1105-Uboye-09:33) (56) Nani onowa tate ewayomo gote, werenke fotografia kekimoni. nãni õnõwa ta-te ẽ-wa-yõmõ go-te werẽnke fotografia rel foot come.out-cvb have-clf:foot-place go-cvb little photo.sp ke-kĩ-mõni do-fut-1pl ‘We should take some pictures of the spot where they went leaving footprints.’ (25Mar1301-Moipa-14:36)

27 Wao Terero

1215

5.3 Tense, aspect, mood, and evidentiality The following section illustrates select morphological realizations of tense, aspect, mood, and evidentiality including the past, remote past, future, perfective, imperfective, inceptive, declarative, irrealis, and inferential suffixes.

5.3.1 Tense There are three verbal tense suffixes. Remote past -ga- is frequently used when talking about ancestors and how they lived; however, since these cases are not evidenced by the speaker’s own experience, they co-occur with the inferential -ĩ, as in (57). This same remote past marker can be used when talking about an event within one’s own lifetime, such as during their childhood, but other past contexts show the suffix -ta, shown in (58b). There is also the future tense -kĩ, illustrated by (58c), while verbs in the present tense are zero-marked, in (58a). (57) Eeme kete wenke anamain kewegaranin. ẽẽme ke-te wẽ-ke ãã-nãmaĩ kẽwẽ-ga-rãni-ĩ celebration do-cvb die-sub exist-neg live-rem.pst-3pl-infr ‘They (evidently) used to celebrate (a lot) because there were not many killings.’ (07Jul1601-Mintape-30:23) (58) Tomamini imini amonipa, ñowene baronani baminitapa ñowo kekiminipa monito girinani nani tomañomo nani keweñomo. a. tomaa-mĩni ĩ-mĩni ã-mõni-pa all-2pl cop-2pl say-1pl.excl-decl b. ñõwõ-nẽ ba-rõ-nãni ba-mĩni-ta-pa now-loc become-caus-3pl become-2pl-pst-decl c. ñõwõ ke-kĩ-mĩni-pa mõnitõ girĩ-nãni now do-fut-2pl-decl 1pl.excl relation-3pl d. nãni toma-ñomo nãni kẽwẽ-ñõmo rel.3pl all-place rel.3pl live-place ‘We say to all of you (new leaders), today they elected you and now you should work for our people who live all over (Waorani territory).’ (29Jun1303-Pablito-54:55)

1216

Alexia Z. Fawcett

5.3.2 Aspect There is a basic perfective-imperfective contrast realized as verbal suffixes the verb, as seen in (59) through (62). (59) Bakeka inga monte deyeko eigati ente eigatimpa. bakekã ĩ-gã mõ-te deye-koo unmarried.woman cop-3sg marry-cvb spider.monkey-qnt ẽi-ga-tĩ ẽ-te ẽi-ga-tĩ-pa go.up-rem.pst-pfv have-cvb go.up-rem.pst-pfv-decl ‘Marrying the young woman, the spider monkeys had brought her up (into the tree).’ (07Jull1601-Mintape-17:53) (60) Iñomo nantate gotinpa anani. i-ñomo nãnta-te go-tĩ-pa ã-nãni dem-place hurt-cvb go-pfv-decl say-3pl ‘Injured here, they say (he) had gone.’ (14Aug0801-Dayuma- 01:07) (61) De tomaa, no ate tokomonike, peene kente tomonike. de tõmaa no a-te to-kõ-mõni-ke peẽnẽ kẽ-te neg all ? see-cvb laugh-ipfv-1pl.excl-sub ripe.plantain eat-cvb to-mõni-ke laugh-1pl.excl-sub ‘There wasn’t anybody. Looking, we were laughing. Eating ripe plantain, we laughed.’ (07Ene1101-Awa-19:27) (62) Pikemo eñekomo… pikẽ-mo ẽñẽ-kõ-mo old-1sg listen-ipfv-1sg ‘I, an elder, am listening…’ (29Jun1303-Gabamo-09:56)

5.3.3 Reality Status The realis verb forms are either a zero-marked indicative or a declarative – which has been referred to in previous work as assertive (Peeke 1973) – explicitly marked with the suffix -pa on the matrix verb, demonstrated in (63). The -pa suffix can cooccur with verbs in all tenses (present, past, remote past, and future) and persons.

27 Wao Terero

1217

(63) Kengi tano kentamonipa. kẽŋi tãno kẽ-ta-mõni-pa food first eat-pst-1pl.excl-decl ‘First we ate food.’ (07Feb1105-Oswando-40:43) Irrealis forms, marked with -e, appear in purposives, possible or uncertain futures, situations over which the speaker does not have control, and imperative contexts (Section 6.3). Unlike declarative -pa which is always the last morpheme suffixed to the verb, the irrealis marker can either immediately follow the verb root or the person suffixes, as in (64) and (65), respectively. (64) Baane onkone ongoena. baãne õnkõ-nẽ õ-ŋõ-ẽ-na tomorrow house-loc be-upright-irr-3du ‘Tomorrow they will be in the house.’ (07Feb1104-Ana-51:34) (65) Kento inte ami kiñe ente pongae. Kẽnto ĩ-te ã-mi kĩñe ẽ-te põ-ŋã-ẽ prop.n cop-cvb say-2sg quickly have-cvb come-3m-irr ‘Tell Kento to come quickly, bringing (food).’ (14Aug0803-Dayuma-3:24)

5.3.4 Modality and Evidentiality Peeke (1973) notes many modal suffixes including: admonitive -kẽ, urgent -bã, remonstrative -wẽ, speculative -bẽ, dubitative -wo, subjective -rõ, and certain (although unobserved) -keĩ. There is an intentional suffix -ke, which Peeke (1973) analyzes as an inceptive, that is used in desideratives, as in (66), and more generally for future planned action. (66) Oworani monito tenonketante keyomoni. õ-wo-rãni mõnitõ tẽnõ-ke ã-te ke-yõ-mõni be-suspended-3pl 1pl.excl spear-intn say-cvb do-simul-1pl.excl ‘While they were there suspended (in hammocks) we wanted to spear (them).’ (07Ene1101-Awa-15:25) The deontic modal -bawe does not co-occur with any other verbal suffixes, illustrated in (67) and (68).

1218

Alexia Z. Fawcett

(67) Ome diki minte kebawe wi iñomo kewimi. õmẽ diki mĩ-te ke-bawe wĩĩ ĩ-ñõmo kẽwẽ-mi work ideo search-cvb do-deont neg dem.prox-place live-2sg ‘You have to look for work, not stay here.’ (26Ene1302-Amoa-11:11) (68) Tabi ponte kemawe. ta-bi põ-te kẽ-mawe come.out-2sg come-cvb eat-deont ‘Come out. Come and eat.’ (07Jul1602-Mintape-10:54) The only clearly evidential marker identified thus far, the inferential -ĩ, “marks lack of authentication because the speaker did not (or does not yet) observe the action” (Peeke 1973: 51), as seen in (69). (69) … manomai waremiñegaranimpa amo. mãnõ-maĩ ware-mĩñẽ-ga-rãni-ĩ-pa ã-mo dem.manner-sim clean-clf:buttocks-rem.pst-3pl-infr-decl say-1sg ‘… I said they used to wipe like this (after a baby defecated).’ (07Feb1105-Ana-34:17) Beyond verbal morphology, confirming perspective and clarifying information can be achieved through forms of ‘say’ and ‘see’.

5.4 Voice, valency, and transitivity There are both valency increasing and decreasing morphemes suffixed to the verb. Causatives are one valency increasing operation, of which there are both morphological and analytic types: the suffix -rõ,9 as in (70)–(71), and the light verb ke ‘do/ make’, as in (72). (70) Gaan wenoe aye. gãã wẽ-nõ-ẽ aye spat come.down-caus-irr more ‘Lower it a little still.’ (07Feb1105-Moipa-23:30)

9 The causative also appears in lexicalized verbs like põ-nõ ‘come-caus; buy’, go-rõ ‘go-caus; sell’, tẽ-nõ ‘pierce?-caus; spear’, and mẽ-nõ ‘bring.up?-caus; construct’.

27 Wao Terero

1219

(71) Kenenke ton taromi emoe. kẽnẽ=ke tõ ta-rõ-mi ẽ-mo-e manioc=rest ideo come.out-caus-2sg have-1sg-irr ‘Take out just the manioc for me.’ (06Jul1601-Daila-17:08) (72) Ñon kebo wiñenga inga. ñõ ke-bo wĩñẽ-ŋã ĩ-ŋã lie do-1sg young-3sg cop-3sg ‘I lay the girl down.’ (28Jul1802-Uboye-08:09, elicited) Valency decreasing operations such as reflexive -gẽ are also realized as suffixes on the verb, as in (73). (73) Wente ñonge doobe… wẽ-te ñõ-ŋẽ doobe die-cvb lie-refl already ‘Lying down already dead…’ (10Dec1204-Nene-51:38)

5.5 Directional Verbs Using Table 27.8, Peeke (1973: 131) explains that motion verbs reflect an intersection of direction in terms of a deictic center established by the discourse context and topographical direction. The upward and downward forms refer to movement along a vertical axis (such as climbing a tree, but also up- or downhill); however, there appears to be only one directional verb that reflects a riverine orientation. The forms ‘bring’ and ‘take’ are included to show how this paradigm is extended to transitive verbs of motion. An example from discourse is given in (74). (74) Baane taokemo. baãne tao-ke-mo tomorrow go.out-intn-1pl.incl ‘We are going to leave tomorrow.’ (07Feb1104-Ramón-00:59) These motion verbs also serve as the source for spatial adverbs, where they can take complex strings of suffixes such as -mõŋa ‘farther’ to direct the verbal root to a location, illustrated in (75) and (76).

1220

Alexia Z. Fawcett

Tab. 27.8: Motion verbs (adapted from Peeke 1973: 131). deictic

general

upward

downward

downstream

outward

inward

‘come’ ‘go’ ‘bring’ ‘take’

põ go mãmõ mão (mãgo)

ẽ ẽi mẽ mẽi

weẽ wẽi mãwẽ mãwẽi

i igi mãgi mãgi

ta tao (tago) mãta mãta

gi gii mãgi mãgi

(75) Aye emonga ei. aye ẽ-mõ-ŋa ẽi more come.up-dir-place go.up ‘Go farther up.’ (07Feb1105-Ramón-11:20) (76) Wemonga epe pantake weimoni, mina wenamai iina wii eñente kerinke, wemonga mono ore entamo maniñomo epene weimoni. wẽ-mõ-ŋa epẽ panta-ke wẽi-mõni mĩnã come.down-dir-place river swim-intn go.down-1pl.excl 2du wẽ-nãmaĩ ĩ-na wĩĩ ẽñẽ-te ke-rĩ-ke wẽ-mõ-ŋa come.down-neg cop-2du neg listen-cvb do-pfv-sub come.down-dir-place mõnõ ore ẽ-ta-mõ mãnĩ-ñõmo epẽ-nẽ 1pl.incl white.lipped.peccary have-pst-1pl.incl dem1-place river-loc wẽi-mõni go.down-1pl.excl ‘We’re going down to the river to bathe, you two won’t go because you don’t listen, down where we caught the peccary, that’s where we’re going down to the river.’ (07Feb1104-Ana-15:50)

5.6 Adverbs Adverbs have a greater degree of syntactic freedom than most word classes, although they tend to occur verb phrase-initially. In addition to spatial adverbs (Section 5.5), Wao Terero temporal adverbs such as ĩĩmo ‘yesterday’, doobe ‘already’, and weẽne ‘before’ as in (77); manner adverbs such as wa ‘well’ and wĩwa ‘negwell; badly’ as in (78); and degree adverbs such as woro ‘almost’ and nãngi ‘a lot’ as in (79). (77) Awe yiramai, weene waponi intapa. ãwẽ yi-rãmaĩ wẽẽne wa=põni ĩ-ta-pa tree fell-neg before well=very cop-pst-decl ‘It was better before, when they didn’t cut down trees.’ (07Feb1105-Oswando-35:47)

27 Wao Terero

1221

(78) Maniñere winwa baronanita iroin beyenke. mãnĩ-ñerẽ wĩĩ-wa ba-rõ-nãni-ta iroĩn=beyẽ=ke dem1-time neg-well become-caus-3pl-pst shamanism=because=rest ‘In this period, everything was going wrong due to the work of shamans.’ (10Jul1602-Omene-28:39) (79) O, nangi poni koone ketapa. õ nãngi=põni koõnẽ ke-ta-pa yes a.lot=very rain do-pst-decl ‘Yes, it rained really hard.’ (07Ene1101-Awa-33:04) There is an intensifier used with adverbs, =põni ‘very’, shown in (79)–(80), and example (81) shows how the degree adverb nãngi can interact with negation (Section 6.4). (80) Kinante manomain pomini, okemente ñowo poni kiñe poni mani. kĩnãnte mãno-maĩ po-mĩni okẽmẽ-te ñõwõ=põni kĩñe=põni why dem.manner-sim come-2pl return-cvb now=very quickly=very mãnĩ dem ‘And he said, “Why did you come like this? Leave right now and quick”.’ (10Dec1204-Nene-21:15) (81) Tepe gitame tao beka wii tepe nangi poni bikinga. tepẽ gitã-mẽ tao be-kã wĩĩ fermented.yucca.drink.clf:liquid small-clf:liquid go.out drink-3sg neg tepẽ nãngi=põni bi-kĩ-gã fermented.yucca.drink.clf:liquid a.lot=very drink-fut-3sg ‘He left drinking only a little bit of yucca drink, he didn’t drink very much.’ (10Dec1204-Nene-03:15)

5.7 Complex predicates Peeke (1994: 271) notes that compound verbs can be close-knit constructions “expressing motion-motion, motion-action, cause-action, action-reaction, action-observation, or a combination involving three or more of these categories (motion, cause, action, reaction, observation) or a sequence of two reactions.” Wao Terero exhibits serial verb constructions that would be considered prototypical (Aikhenvald 2006) of the label – mono-clausal sequences of verbs that share tense and aspect values as well as arguments, do not take markers of syntactic dependency or coordination,

1222

Alexia Z. Fawcett

and constitute a single predicate expressing a single event. In serial verb constructions where a bare root precedes a fully inflected verb, V1 tends to be one of motion, as in (82) and (83). (82) Yite pon keyonani. yi-te põ ke-yõ-nãni fell-cvb come do-simul-3pl ‘They kept coming to fell (the trees).’ (07Ene1101-Awa-16:49) (83) Mona tano go awenegimona. mõnã tãno go awẽnẽ-ŋĩ-mõna 1du first go record-fut-1du ‘We two are first going to go record.’ (07Feb1105-Oswando-08:02) If present, TAM and person morphology occurs on the final verb of the construction; however, it can go unmarked, as in (84). (84) Omite mamo wo kewa ei. õmite mãmo wo ke-wa ẽi vine.for.climbing bring suspend do-clf:foot go.up ‘Bring a vine so (you) can climb up.’ (20Feb1001-Enqueri-05:45) While ideophones often appear periphrastically, as their own intonation unit and at clause boundaries, there is evidence of complex predicates consisting of an ideophone and a verb (Fawcett 2018). In (85), the uninflected ideophone precedes the light verb which can be inflected for TAM and person, and the ideophone carries the semantic content of the verbal predicate. However, ideophones can also cooccur with verbs that are nearly synonymous, as in (86). (85) Minkaye onko ginta, wanke wanke kete. mĩnkaye õnkõ gĩnta wãnke wãnke ke-te wasp house dog ideo.move ideo.move do-cvb ‘The dog was shaking the wasp nest.’ (19Jun1601S1-Linda-05:00) (86) Ire do do do wemenga, amo. ire do do do weẽ-mẽ-ŋã ã-mo already ideo.shiver ideo. shiver ideo. shiver shake-clf:arm-3m say-1sg ‘But he is already shivering in his arms, I say.’ (07Feb1105-Ana-44:48)

27 Wao Terero

1223

Finally, the third type of collocation, demonstrated in (87), suggests a more adverbial function since the ideophone narrows the manner by which one does the action of the verb with which it co-occurs. (87) Ekano inaa ante wa ayona. Tome obe, dao dao pon. ekãno ĩ-nã ã-te wa a-yõ-nã tõmẽ obe dao who cop-3f say-cvb ideo.look see-simul-3f dem2 boa ideo.step dao põ ideo.step come ‘She saw who it was as the boa came walking up to her.’ (15Aug0801-Ñay-00:47)

6 Simple clauses It appears that SOV is the default constituent order, but in natural discourse, word order is variable, with SOV, OSV, OVS, and SVO all attested. Wilkendorf (1988) addresses word order in her discussion of “clause formulas”; however, further investigation is needed to fully understand constituent order and possible discourse-level motivations for variation.

6.1 Alignment Wao Terero exhibits nominative-accusative alignment, and person marking of the A/S is found on the verb. There is no overt grammatical relation marking on nominals, but objects are differentially marked. Human objects are always followed by an inflected copula as in (88)–(89), which appear to have grammaticalized to function as case markers. (88) Nanogena Omatoke inante atabopa. nãnõ-ŋẽ-nã Omatoke ĩ-nã-te a-ta-bo-pa rel.3sg-spouse-3f prop.n cop-3f-cvb see-pst-1sg-decl ‘I saw his wife Omatoke.’ (14Aug0801-Dayuma-10:00) (89) Wiñenga inga ente go kowañe wirok ke. wĩñẽ-ŋã ĩ-ŋã ẽ-te go kowãñe wirok ke young-3m cop-3m have-cvb go deer ideo.slip do ‘The deer was carrying the boy and slipped.’ (19Jun1601-Linda-06:51)

1224

Alexia Z. Fawcett

Non-human, animate objects of the matrix verb are variably marked: Peeke (1994: 269) notes that “reference to domestic animals and, rarely, reference to other large animals, is marked for object, with or without indication of person and number” as seen in (90) and (91) respectively. Inanimate objects do not receive any marking. (90) Amo arobain ompaka onani. ãmõ aro-baĩ õpa-ka õ-nãni collared.peccary one-sim arrow-ins hunt-3pl ‘They also hunt peccary with arrows.’ (14Aug0801-Coba- 01:33) (91) Piinte tamo taka weete minkaye doobe ginta inga teempote. pĩĩ-te tãmõ ta-ka weete mĩnkaye doobe gĩnta ĩ-ŋã te yell-cvb owl come.out-3sg in.turn wasp already dog cop-3sg step ẽ-po-te have-clf:hand-cvb ‘After yelling, the owl came out (of the house), but the wasp was already pursuing the dog.’ (19Jun1601-Linda-05:38) When objects are marked, the constituent order is relatively free, but in cases with no marking, order is more rigid, as the object tends to immediately precede the verb, which can help disambiguate if both S and O are inanimate (and thus the finite verb lacks person marking).

6.2 Interrogatives Polar interrogatives are achieved with rising intonation and shifting stress to the final syllable. Content questions use interrogative words such as those listed in (92), which tend to appear phrase-initially and feature morphemes and classifiers that suffix to other classes (such as demonstratives Section 2.1). (92) a. b. c. d. e. f. g. h. i. j.

e-kã-no e-bã-no e-ro e-rõ-no e-yõmõ-no e-po-ro e-yerẽ-no kĩ-no kĩ-nante kĩ-wẽ-mẽ-no

int-3sg-?10 int-manner-? int-? int-path-? int-place-? int-clf:hand-? int-time-? int-? int-? int-clf:tree-clf:arm-?

‘who?’ ‘how?’ ‘where?’ ‘which way/direction?’ ‘where/which place?’ ‘how many/much?’ ‘when?’ ‘what?’ ‘why?’ ‘what branch?’

10 The final -ro element is not yet understood well enough to provide an accurate gloss but could be a nominalizer (tere-ro ‘speak-nmlz; language’).

27 Wao Terero

1225

Although not required for interrogative utterances, the matrix verb can take a dubitative or inferential modal suffix as in (93) and (94) respectively. (93) Kino keranitawo ñowo ganka? kĩno ke-rãni-ta-wo ñõwõ gãnka what do-3pl-pst-dub now up.to ‘What have they done up to now?’ (29Jun1303-Nancy-14:33) (94) Nanikabo tenonte de wengaranin? nãni-kabo tẽnõ-te dee wẽ-ŋa-rãni-ĩ rel.3pl-group spear-cvb neg die-rem.pst-3pl-infr ‘Spearing among themselves, everyone in the extended family died?’ (10Dec1204-Oswando-28:55)

6.3 Imperatives As discussed in Section 5.3.3, irrealis -e/-ẽ is used for imperatives. There is structurally no difference between a verb in the irrealis mood and an imperative except that imperatives with a second-person singular addressee take no person marking, as in (95). If an imperative is directed at two or more individuals, the dual or plural marking is used, just as with a verb in the irrealis mood with the same subjects. (95) Nangi gaan gobi aboe, nangi gaan goe. nãŋi gãã go-bi a-bo-e nãŋi gãã go-e a.lot spat go-2sg see-1sg-irr a.lot spat go-irr ‘Move over a lot so I can see, move over a lot.’ (07Feb1104-Moipa-30:57) Negative imperatives are achieved via the verbal suffix -rãmaĩ, as in (96), the same as the negation strategy for indicative and declarative clauses further discussed in Section 6.4. (96) Giporeramai arani. gi-po-re-rãmaĩ a-rani enter-clf:hand-clf:mouth-neg see-3pl ‘Don’t put your hands in your mouth, they (can/will) see.’ (07Feb1105-Ana-45:17)

6.4 Negation The basic strategy for standard negation is addition of negative -rãmaĩ to a verb. The negated verb is not inflected for person or TAM, which is typically expressed

1226

Alexia Z. Fawcett

on the following copula, as in (97), although it is not required; see (98) and (99). Verbs in this form often have a translation of ‘without X-ing’ and are the main strategy for negative imperatives (Section 6.3). (97)

Ñeñeiri keramain inanitaa. ñẽñẽ-iri ke-rãmaĩ ĩ-nãni-ta grandmother-coll do-neg cop-3pl-pst ‘Grandmothers didn’t do (this).’ (26Ene1302-Amoa-08:23)

(98)

Keramain minito doobe yewemonintapa ante pinanitapa. ke-rãmaĩ mĩnitõ doobe yewe-mõni-ta-pa ãnte do-neg 2pl already write-1pl-pst-decl therefore pĩ-nãni-ta-pa be.angry-3pl-pst-decl ‘(You) don’t need to do (it), we have already written, for this reason they were mad.’ (25Mar1301-Ana-35:34)

(99)

Waa kewenamain pegabopa antaa boto bara. wa kẽwẽ-nãmaĩ pe-ga-bo-pa ã-taa botõ barã well live-neg grow-rem.pst-1sg-decl say-pst 1sg.poss mother ‘(We) didn’t live well as I grew up, my mother said.’ (24Dec1002-Omare-52:50)

The constituent-level negator wĩĩ can give “…special emphasis to the negated action, such as in the rebuttal of a positive statement” and when used for negative habituals can be interpreted as ‘never’ (Peeke 1994: 277). In these cases, it negates both finite verb forms, shown in (100), and verbs in subordinate clauses. Wĩĩ can also negate non-verbal elements such as noun phrases and adverbials, where it precedes the negated constituent, as seen in (101). The scope of negation does not extend beyond the clause or phrase it precedes. (100) Wii eñenanipa amo. wĩĩ ẽñẽ-nãni-pa ã-mo neg listen-3pl-decl say-1sg ‘I say they don’t listen.’ (13Jun1101-Wiña-02:20) (101) Bokayakoo wii otore. bokaya-koo wĩĩ otore spider-qnt neg tarantula ‘Spiders, not tarantulas.’ (20Feb0701-Enqueri-58:24)

27 Wao Terero

1227

The negative existential construction using dee is typically followed by the existential verb ãã, which can be inflected for person (102) and TAM (103) or left uninflected (104). (102) Weiyomoni dee anga. wẽi-yõ-mõni dee ãã-ŋa go.down-simul-1pl neg exist-rem.pst ‘We went down, but nobody was there.’ (07Ene1101-Awa-44:52) (103) Kengi ente poginga kengi, ñowo dee ampa. kẽŋi ẽ-te põ-ŋĩ-ŋã kẽŋi ñõwõ dee ãã-pa food have-cvb come-fut-3m food now neg exist-decl ‘He has to bring food, now there is none.’ (14Aug0801-Dayuma-1:51) (104) Dee a kewentamonipa. dee ãã kẽwẽ-ta-mõni-pa neg exist live-pst-1pl-decl Since there weren’t any (colonists), we lived (in peace).’ (07Ene1101-Awa-1:52) There is also an inherently negative verb root õma ‘not have/lack’ in (105)–(106) which can take person marking. (105) Bito nanonge eñabi imi omabi imi, omabi? bitõ nãnõgẽ ẽña-bi ĩ-mi õma-bi ĩ-mi õma-bi 2sg spouse have-2sg cop-2sg lack-2sg cop-2sg lack-2sg ‘Do you have a wife or not?’ (31Oct1001-Tiri-07:26) (106) Boto kii omabo mani beye aye enke anewemo. botõ ki õma-bo mãni=beyẽ aye ẽ-ke ã-ne-wẽ-mo 1sg thing lack-1sg dem1=because then have-sub say-pfv-?-1sg ‘I don’t have one (a dog), that’s why I have gone around saying I want to get one.’ (07Jul1602-Mintape-08:51) In response to polar questions, one can use ba ‘no’, which is therefore, in conjunction with ã ‘say’, a strategy for negative desideratives, shown in (107) and (108).

1228

Alexia Z. Fawcett

(107) Ba anani webin? ba ã-nãni we-bi-ĩ neg say-3pl cry-2sg-infr ‘You cried because they didn’t want you to (told you “no”)?’ (Daila-1:01:44)

(108) Manomain dorani anamai inani ba anani dorani ikene. mãnõ-maĩ do-rãni ã-nãmaĩ ĩ-nãni ba ã-nãni do-rãni dem.manner-sim before-3pl say-neg cop-3pl neg say-3pl before-3pl ikẽne in.contrast ‘On the other hand, the ancestors didn’t speak like that, it was prohibited to speak in that way.’ (10Dec1204-Nene-38:41)

7 Clause combining Functional equivalents of complement clauses are expressed using forms that are not unique to complementation contexts. For example, in (109), clauses are juxtaposed with no overt marker of complementation.

(109) E eñeminipa okone me kewetabopa wantepiye. e ẽñẽ-mĩni-pa õnkõ-ne mẽ kẽwẽ-ta-bo-pa wãntepiye intrj know-2pl-decl house-loc stay live-pst-1sg-decl extended.time ‘You all know that I lived in (this) house a long time.’ (25Mar1301-Nanto-52:04)

The remainder of this section outlines the structure of relative clauses (Section 7.1), the converb (Section 7.2), simultaneous suffix (Section 7.3), adverbial clause linkers (Section 7.4), and conditional constructions (Section 7.5).

7.1 Relative Clauses The Wao Terero relative clause is achieved via a relative pronoun strategy. The relative pronoun nee relativizes human subjects (110); mono relativizes non-human subjects (111)–(112); and nano relativizes non-subjects (113)–(116).

27 Wao Terero

1229

(110) Ebano kete wenogi añomoni go tome nee nani bee tenke goringa, Ewete. ebano ke-te wẽnõ-ŋĩ ã-ñõ-mõni go tõmẽ nẽẽ nãni be how do-cvb escape-fut say-simul-1pl go dem2 rel rel.3pl contact tẽ-ke go-rĩ-ŋã Ewete meet-intn go-pfv-3m prop.n ‘How are we going to distance ourselves from here, the one who went to make contact with them, Ewete.’ (30Oct1002-Omare-29:07) (111) De an mono ñan ñan kero. de ãã mõno ñãn ñãn ke-ro neg exist rel ideo.light ideo do-nmlz? ‘There isn’t anything, that which (usually) shows the signals.’ (07Feb1105-Ramon-00:30) (112) Ompore amo, mono epene keworo… ompore a-mo, mono epẽ-ne ke-wo-ro otter say-1sg rel water-loc do-float-nmlz? ‘The otter I say, those that live in the water…’ (04Dec0501-Cawia-05:23) (113) Eronke bito nano tere. erõnke bitõ nãno tere clear 2sg rel talk ‘What you speak is clear.’ (07Feb1105-Ramon-52:00) (114) Pokoro poni pon wane gi oremo tenewenko nano bakoo gonoñomo iyomo gi ten. pokoro=põni põ wa-nẽ gi orẽmo tẽnẽwẽ-koo nãno fast=very come other-loc enter door/opening firewood.clf:tree rel bakoo gõ-nõ-ñõmo ĩ-yõmo gi tẽ many be.upright-caus-place dem.prox-place enter stand ‘He, on the other hand, entered really quickly to the side of the door where there was a lot of firewood stacked, and there he stayed standing.’ (15Aug0801-Ñay-10:03) (115) Mani one mama ena bara nano eniñomo in mani. mãnĩ õne mama ẽ-nã barã nãno ẽ-nĩ-ñõmo ĩ mãnĩ dem1 chambira.palm mom.sp get-3f mother rel get-pfv-place cop dem1 ‘This here is where mom got the chambira palm.’ (07Feb2001-Enqueri-05:02) While most of the relative clauses identified have been headless, there are instances of both head-initial, as in (116), and head-final relative clauses.

1230

Alexia Z. Fawcett

(116) Kempe nano ano owotapa. kẽmpe nãno ã-no õwo-ta-pa lice rel say-nmlz? be.suspended-clf:shell-decl ‘That which they call lice is hanging (on the leaf).’ (07Feb1105-Ramon-4:45) There are arguably instances of head-internal relative clauses if we take the markers for time, space, or classifiers that can occur on a verb to be the referential nominals for the relative clause. These cases would benefit from further analysis, especially considering discussions of variable headedness (Citko 2004, Epps 2012).

7.2 Converbs Verbs marked with -te conform to Haspelmath’s (1995: 3) definition of a converb “as a nonfinite verb form whose main function is to mark adverbial subordination.” These forms do not take person marking morphology but tend to share a subject with the following matrix verb – examples can be found in (117) and (118). (117) Iñomo, ate ne gongente, gawaramai. ĩ-ñõmo a-te nẽ gõ-gẽ-te gã-wa-ramaĩ dem-place see-cvb stop be.upright-refl-cvb press-clf:foot-neg ‘Here, carefully stand, but don’t step.’ (07Feb1105-Ana-50:47) (118) Pete yakongaranin? pe-te ya-kõ-ga-rãni-ĩ sing-cvb peel-ipfv-rem.pst-3pl-infr ‘They would sing and take out the chambira palm?’ (20Feb1001-Laura-13:56) Since -te appears with non-finite verbs in adverbial clauses and does not co-occur with TAM markers, the relationship (causal, temporal, etc.) between the converb and the more finite verb is not always explicit. This relationship is expressed through clause-linkers such as aye ‘then/more’, and the converb contributes further contextual information to, or a contextual frame for, the more finite verb (as shown in Section 7.4). The suffix is also used in clause-chaining constructions,11 leading to an understanding of events as sequential as in (119) (see Ross (1988) for discussion of “chronological and logical discourse chains”). 11 An issue that arises in appealing to Haspelmath’s definition of a converb is that Wao Terero verbs marked with -te appear to fulfill the functions of both converbs and medial verbs (clause-

27 Wao Terero

1231

(119) Be tente, awenka ente yitamoni. be tẽ-te ãwẽnka ẽ-te yi-ta-mõni. contact contact-cvb hatchet have-cvb fell-pst-1pl ‘When we made contact, we began to get hatchets and cut down (trees).’ (07Ene1101-Awa-01:58) This clause-chaining function of the -te suffix as well as tail-head linkage are common in Wao Terero narratives (further discussed in Section 8.1). Finally, many clauses carry the converb morphology and not necessarily modify a finite verb.

7.3 Simultaneous In contrast with the converbal -te forms, which tend to refer to same subject as the matrix clause, simultaneous -yõ is most often found when subjects differ, as in (120). (120) Mani kete aye ire ekano tano poni keweñomini ponte be tengaranin? mãnĩ ke-te aye ire ekãno tãno=poni kẽwẽ-ñõ-mĩni põ-te dem1 do-cvb more but who first=very live-simul-2pl come-cvb be tẽ-ŋa-rãni-ĩ contact meet-rem.pst-3pl-infr ‘(They were) doing this, but who were the very first who came to make contact, when/while you were living (like this)?’ (07Ene1101-Ramon-05:01) This clause combining strategy is unique in that the morpheme occurs in a slot where one would expect to find a tense or aspectual suffix, and the whole verb form resembles independent clause, as it takes person morphology, yet it occurs in context of another clause. Referencing a temporal relationship between the eventuality of the two clauses, this function lines up with Cristofaro’s (2003: 1555) definition of adverbials as “link[ing] two [states of affairs] such that one of them (the dependent [state of affairs]) corresponds to the circumstances under which the other one (the main [state of affairs]) takes place;” however, this differs from stand-alone adverbial clause linkers introduced in Section 7.4.

chaining). He claims that, between the two, “the key difference lies in the fact that prototypical converbal clauses are subordinate (in the sense of ‘embedded’) while prototypical medial clauses … are not subordinate but are cosubordinate” (Haspelmath 1995: 23). Unlike Haspelmath, Nedjalkov (1995: 98–99) treats this clause-chaining type of construction as a subtype of the converb or a “coordinate converb” that is different from “converbs proper” that are adverbial and embedded and “conjunctional converbs” that have a different subject. Therefore, they are here treated as different functions of a single converb form.

1232

Alexia Z. Fawcett

7.4 Adverbial clause linkers Major Wao Terero adverbial clause types are organized in this section by semantic subtype (Dixon 2011): temporal (I), consequence (II), addition (IV), and manner (VI).12 The converb -te is found to mark all types and is typically used in cases where the subordinated and matrix verbs have the same subject (see Section 7.2). The following data provide examples of linkers for each semantic subtype. Mainly functioning to mark temporal succession (see Holman 1988 for other possible functions), the clause-linker aye translates to ‘more’, ‘then’, or ‘still’ and precedes the focal clause, as in (121). (121) Ongi ongi omo, onte ate. Aye digintai ñanomo abi, manomain onte. õ-ŋi õ-ŋi õ-mo õ-te a-te aye chambira-clf:hair chambira-clf:hair spin-1sg spin-cvb see-cvb then digĩntai ñãnõ-mo a-bi mãnõ=maĩ õ-te net.bag weave-1sg see-2sg dem.manner=sim spin-cvb ‘I am spinning chambira (palm) fiber. Once I have spun the chambira, then I am going to weave a net bag, you see, spinning like this.’ (07Jul1601-Mintape-22:20) In terms of consequence clauses (II), we see use of causal clause-linker =beyẽ, which occurs NP- or clause-finally delimiting the cause of or reason for the event in the matrix clause, as in (122) and (123). (122) Tenonani beye põnamãĩ ĩ. tẽnõ-nãni=beyẽ põ-nãmaĩ ĩ spear-3pl=because come-neg cop ‘Because they killed spearing, (the outsiders) did not come.’ (07Jul1601-Mintape-32:06) (123) Tomẽngã dooye õnõwa ẽwate beye taa ware goka. tõmẽ-ŋã dooye õnõwa ẽ-wa-te=beyẽ ta ware go-kã. dem2–3sg long foot have-clf:foot-cvb=because come.out ideo go-3sg ‘Since he has very long feet, he exits (the bottle) and escapes.’ (19Jun1601-Linda-02:49) Additionally, the causal clause linker ãnte ‘because, therefore, given, since’ is built from the verb ã ‘say’ with the suffix -te, as seen in (124).13

12 The omission here of possible consequence (III) and alternative (V) clause linking does not preclude these clause types from existing in Wao Terero. 13 Haspelmath (1995: 37) notes a common diachronic pathway of the grammaticalization of converbs to subordinating conjunctions.

27 Wao Terero

1233

(124) Wiñenani pekeranimpa ante, wa pete kewengaranimpa ante wa kewegarani. wĩñẽ-nãni pe-ke-rani-ĩ-pa ãnte wa pe-te young-3pl grow-intn-3pl-irr-decl therefore well grow-cvb kẽwẽ-ŋa-rãni-ĩ-pa ãnte wa kẽwẽ-ŋa-rãni-ĩ live-rem.pst-3pl-irr-decl therefore well live-rem.pst-3pl-irr ‘So that the children can grow and live better, and therefore they will live well.’ (07Jul1601-Mintape-30:33) When there is a contrast (IV) between the information presented in the focal and supporting clause, one uses contrast clause linkers inkete or weete, as in (125) and (126). Like ãnte, these are likely formed from a verb with the converbal suffix, but their path to grammaticalization is not yet clear. (125) Bito tere. Ẽñemo inkete ĩñomo. Taramai ĩ webo. bitõ tere ẽñẽ-mo ĩnkete ĩ-ñõmo ta-rãmaĩ ĩ we-bo 2sg speech hear-1sg but dem-place come.out-neg cop worry-1sg ‘I can hear what you say, but here (in the camera) the signal isn’t showing.’ (07Feb1105-Ramón-00:40) (126) Ñowo weete teyamo kowore bai, ba ñowo omena ñimpo kerani dee ba. ñõwõ weete teyãmo kowore=baĩ ba ñõwõ õmẽna ñĩ-po now in.turn rifle outsider=sim neg now blow.gun cease-clf:hand ke-rãni dee ba do-3pl neg neg ‘But today they use rifles like the outsiders, and they’ve abandoned use of the blowgun.’ (14Aug0801-Coba-1:53) Supporting clauses of manner (VI) often include a nominal element with the similative clitic =baĩ, as in (127), or they can simply be marked with -te. (127) Manomain ente mamonani ponani be tentamoni. mãnõ=maĩ ẽ-te mãmõ-nãni põ-nãni be tẽ-ta-mõni. dem.manner=sim have-cvb bring-3pl come-3pl contact contact-pst-1pl ‘Carrying like this, they brought (supplies) and came for us to make contact (with the outsiders).’ (07Ene1101-Awa-08:23)

7.5 Conditional The similative enclitic =baĩ marks the dependent clause of conditional constructions. In possible conditional constructions, the same similative morpheme also oc-

1234

Alexia Z. Fawcett

curs in the focal clause where it is instead affixed to the matrix verb, as in (128) and (129). (128) …Ononke renuncia keka bai tres días anoro weete doobe emaimonipa. Õ̃nõke renuncia ke-kã=baĩ tres días anoro weete suppose resignation.sp do-3sg=cond three.sp days.sp again in.turn doobe ẽ-maĩ-mõni-pa already have-cond-1pl.excl-decl ‘If he resigns, in three days we would already have it (the nomination).’ (29Jun1301-Moi-57:35) (129) Omepobo inte bain colobora kebaimopa… õme-po-bo ĩ-te=baĩ colobora have.nothing-clf:hand-1sg cop-cvb=cond collaborate.sp ke-baĩ-mo-pa do-cond-1sg-decl ‘If I didn’t have work, I would collaborate…’ (25Ene1303-Gabamo-45:50) While this morpheme occurs in both the protasis and apodosis of the possible conditional, the apodosis of counterfactual conditionals employ other affixation strategies – using either a combination of future -kĩ and perfective -rẽ marking as in (130)–(131) or intentional -ke and hypothetical -rõ, as in (132). (130) Wi ana bain, de tenomoni wenginenani. wĩĩ ã-nã=baĩ de tẽnõ-moni wẽ-ŋĩ-nẽ-nãni neg say-3f=cond neg spear-1pl.excl die-fut-pfv-3pl ‘If they hadn’t said this, we would have speared them to death.’ (07Ene1101-Awa-15:34) (131) Wi tenonani gote bai aye tomenani Napairi i inginenani. wĩĩ tẽnõ-nãni go-te=baĩ aye tõmẽ-nãni Nãpa-iri ĩ neg spear-3pl go-cvb=cond still dem2-3pl prop.n-coll dem.prox ĩ-ŋĩ-nẽ-nãni cop-fut-pfv-3pl ‘If they hadn’t speared (them), Napa’s group would still be here.’ (04Jun1301-Ana-37:32) (132) Pomo bain, aye kewenkeronanimpa gininkoiri. põ-mo=baĩ aye kẽwẽ-ke-rõ-nãni-ĩ-pa gĩnĩnko-iri come-1sg=cond still live-intn-hyp-3pl-infr-decl gringo.sp-coll ‘If I had come, the gringos would still have been alive today.’ (14Aug0801-Dayuma-23:33)

27 Wao Terero

1235

Finally, supporting clauses of conditionals can also be marked using -te, as in (133). (133) Peene kente tepe enemaimo. peẽnẽ kẽ-te tepẽ ẽ-nẽ-maĩ-mo plantain eat-cvb chicha have-clf:stomach-cond-1sg ‘Having eaten ripe plantain, I would have like chicha in my stomach.’ (07Ene1101-Awa-31:59)

8 Discourse 8.1 Narrative Wao Terero speakers of all ages use ideophones, and whether a speaker uses ideophones is more likely a question of the content of the speech act than the context or genre; however, they are particularly useful for depicting an experience and therefore important in telling a good story. How one uses ideophones varies, but their use in narratives often exhibits prosodic foregrounding (represented with ↑ in (134)) – via devices like pause, pitch, non-modal phonation, and repetition. (134) ↑Ei ei ei ei↑. Maniñomo poke doobe, wena, wiñenga dee. ei ei ei ei ideo.bit ideo ideo ideo mãnĩ-ñõmõ pokẽ doobe wẽ-nã wĩñẽ-gã dee dem-here bite already die-3f young-3sg neg ‘The boa bit her all over until she was dead.’ (15Aug0801-Ñay-04:40) Repetition can indicate aspectual qualities, not just in ideophones but also in (135) where the repetition of the adverb ĩĩmo ‘yesterday’ intimates that the event took place repeatedly over an extended period. (135) Obe tate wa a wao-ra ome ↑iimo iimo iimo iimo iimo iimo iimo↑ wira. obe ta-te wa’ a wao-rã õmẽ iĩmo iĩmo boa come.out-cvb ideo.look see person-3f land yesterday yesterday iĩmo iĩmo iĩmo iĩmo iĩmo wi-rã yesterday yesterday yesterday yesterday yesterday clear.land-3f ‘The boa came out and watched the woman every day that she worked.’ (15Aug0801-Ñay-00:22) One robust feature of Wao Terero discourse, often occurring in narratives or procedurals, is tail-head linkage. Minimally, the verb of a previous (‘tail’) clause, see

1236

Alexia Z. Fawcett

(136a), is repeated in the first (‘head’) clause of the next sentence, see (136b), usually without TAM or person marking, but carrying the converbal -te suffix. (136) a. Kenenko ebano yate, manomain yarani. Yate mamo. kẽnẽ-koo ebãno ya-te mãnõ=mãĩ ya-rãni manioc-qnt how peel-cvb dem.manner=sim peel-3pl b. ya-te mãmõ peel-cvb bring ‘Just like how one peels manioc, they peeled. Then having peeled, (they) brought (it).’ (07Jul1601-Mintape-04:11) This same linkage strategy can be manifested with a more general verb such as ke ‘do’, in (137). Both formats are used to create discourse cohesion, and act to mark given information. (137) a. Ñeene oko menoinkeronanipa. Kete tomaa ogipo ee wote waponi. ñẽẽnẽ õnko mẽnõ-ĩ ke-rõ-nãni-pa large house construct-ptcp do-caus-3pl-decl b. ke-te tõmaa ogipo ee wo-te wa=põni do-cvb all dirt ideo dig-cvb good=very ‘They constructed the house very large. Doing this, digging out all the dirt, (they made it) very nice.’ (07Jul1601- Mintape-06:18)

8.2 Interaction While often translated as ‘yes’, ao is an affirmative response more akin to ‘okay’. In response to polar interrogatives, one uses õ ‘yes’ or ba ‘no’. For example, see the interaction in (138). (138) A: Doobe emi? ‘Do you have it yet?’ B: Õ doobe emo ‘Yeah, I’ve got it.’ A: Ao ‘Okay’ (20Feb1001-Enqueri/Laura-21:02) For repair, ã ‘what?/huh?’ is used with rising intonation. These interjections, while generally phonologically and morphosyntactically independent, are found in constructions such as, ao ã ‘okay say; be in agreement’ and ba ã ‘no say; prohibit’.

27 Wao Terero

1237

Two non-modal phonation types that have not yet been systematically studied but observed in interaction are ingressive speech and hollow voice. First, pulmonic ingressive speech occurs in contexts of backchanneling, interjections/affirmative responses, reading to oneself or thinking aloud, as well as floor-holding. These patterns are much like what has been noted in Scandinavian and other languages (Eklund 2008, Hill & Zepeda 1999). Second, Saint and Pike (1962: 4) find “faucalization” – elsewhere referred to as “hollow voice” or “yawny voice” and characterized by stretched vocal folds, lowered larynx, and expanded pharynx (Edmondson & Elsing 2006) – to be “a signal of displeasure, disgust – or at times of vague joy, sadness, exuberance, and the like.”

9 Conclusions This chapter is the first survey of basic grammatical elements of Wao Terero, an under-described isolate of the Ecuadorian Amazon region, since early work of SIL linguists beginning in the 1950s, and it thus establishes an important foundation for further work with the language. The description and analysis of Wao Terero is ongoing, and additional research on phonological and morphophonological processes is required – especially in regards to nasalization and vowel harmony. As is common in the region, Wao Terero exhibits contrastive nasalization of vowels and nasal spreading; however, this nasalization differs from other languages of Amazonia in that nasalization is not a property of the syllable. Suprasegmental phenomena such as creak, faucalization (also referred to as “hollow voice” or “yawny voice”), and the interaction between lexical and phrase level prominence are particularly interesting avenues for future research. Wao Terero is agglutinative, predominately suffixing, and the verb is the locus of greatest morphological complexity. The language exhibits morphosyntactic features shared by languages of the region such as a small number of oblique cases and an inclusive-exclusive contrast in the first-person plural. A subset of morphemes central to the tense-aspect system are in some cases optional as has been described for other languages of the region; however, the contexts in which they are considered optional and their interaction with temporal adverbs is not yet well understood. Other typologically interesting phenomena include a multifunctional/ multiple classifier system, object marking developed from a copular construction, and extensive use of ideophones that can form complex predicates.

10 Acknowledgments My sincerest thanks are due to all the Waorani who have welcomed me into their homes and communities. Moreover, this work would not have been possible without

1238

Alexia Z. Fawcett

generosity and patience of those whose words are reflected here in the examples and their source texts, as well as Uboye Gaba, Linda Enkere, Carlos Enqueri, Marcia Enqueri, Nemonte Nenquimo, and Opi Nenquimo, who shared their language expertise via the transcription and translation of these texts. I hope to have accurately portrayed their insights and knowledge. Most of the data reproduced in this synthesis was documented with the support of the Endangered Languages Documentation Programme (No. MDP0224). Thanks to the PIs Connie Dickinson and Casey High for allowing my participation in and access to their continued work with this corpus. My research has also been supported by the National Science Foundation Graduate Research Fellowship Program (No. 1650114). I would like to thank Marianne Mithun, Eric Campbell, Simon Peters, Martin Kohlberger, and the editors and reviewers for comments and guidance on earlier drafts that have greatly improved this chapter, but all remaining errors are my own.

11 References Adelaar, Willem F. H. & Pieter C. Muysken. 2004. The languages of the Andes. Cambridge: Cambridge University Press. Aikhenvald, Alexandra Y. 2000. Unusual classifiers in Tariana. In Gunter Senft (ed.), Systems of nominal classification, 93–113. Cambridge: Cambridge University Press. Aikhenvald, Alexandra Y. 2006. Serial verb constructions in typological perspective. In Alexandra Aikhenvald & R. M. W. Dixon (eds.), Serial verb constructions, 1–68. Cambridge: Cambridge University Press. Alianza Ceibo. 2018. Waorani. https://www.alianzaceibo.org/alianza/waorani/ (15 March 2020). Berlin, Brent & Paul Kay. 1969. Basic color terms. Berkeley: University of California Press. Borman, M. B. 1991. Listas comparativas de palabras en diez idiomas autóctonos ecuatorianos. Cuadernos Etnolingüísticos 13. Quito, Ecuador: Summer Institute of Linguistics. Campbell, Lyle. 2012. Classification of the indigenous languages of South America. In Lyle Campbell & Verónica Grondona (eds.), The indigenous languages of South America: A comprehensive guide. Vol. 2, 59–166. Göttingen: Mouton de Gruyter. Cabodevilla, Miguel Ángel, Randy Smith & Alex Rivas. 2004. Tiempos de guerra: Waorani contra Taromenane. Quito, Ecuador: Abya-Yala. Citko, Barbara. 2004. On headed, headless, and light-headed relatives. Natural Language & Linguistic Theory 22(1). 95–126. Crevels, Mily & Pieter Muysken. 2005. Inclusive-exclusive distinctions in the languages of centralwestern South America. In Elena Filimonova (ed.), Clusivity: Typology and case studies of the inclusive-exclusive distinction, 313–339. Amsterdam: John Benjamins. Cristofaro, Sonia. 2003. Subordination. Oxford: Oxford University Press. Dickinson, Connie, Uboye Gaba, Oswando Nenquimo, Nemo Nenquimo, Mery Nenquihui & Casey High. 2013. Documentation of Wao Terero. London: SOAS, Endangered Languages Archive. https://elar.soas.ac.uk/Collection/MPI1029685 (15 March 2020). Dingemanse, Mark. 2012. Advances in the cross-linguistic study of ideophones. Language and Linguistics Compass 6(10). 654–672. Dixon, R. M. W. 2011. The semantics of clause linking in typological perspective. In R. M. W. Dixon & Alexandra Aikhenvald (eds.), The Semantics of clause-linking: A cross-linguistic typology. Oxford: Oxford University Press.

27 Wao Terero

1239

Edmondson, Jerold A. & John H. Elsing. 2006. The valves of the throat and their functioning in tone, vocal register and stress: Laryngoscopic case studies. Phonology 23. 157–191. Eklund, Robert. 2008. Pulmonic ingressive phonation: Diachronic and synchronic characteristics, distribution and function in animal and human sound production and in human speech. Journal of the International Phonetic Association. 235–324. Epps, Patience. 2012. Between headed and headless relative clauses. In Bernard Comrie & Zarina Estrada-Fernández (eds.), Relative clauses in languages of the Americas: A typological overview, 191–211. Amsterdam: John Benjamins. Fawcett, Alexia Z. 2018. Ideophone integration and expressiveness in Wao Terero. Santa Barbara, CA: University of California MA Thesis. Fiddler, Michael. 2011. Bound nominal roots in Waorani. GIALens 5(1). Fitzgerald, Colleen. 1999. A reanalysis of bidirectionality in Auca. Western Conference on Linguistics (WECOL) 11. 106–118. Gaba, Uboye. 2011. Idéofonos en el wao terero. Conference on Indigenous Languages of Latin America (CILLA) 5, 6–8 October 2011. Gondecki, Philip. 2011. Entre retirada forzosa y autoaislamiento voluntario: Reflexiones sobre pueblos indígenas aislados y estrategias de evitación en el manejo de conflictos en la Amazonía occidental. INDIANA 28. 127–152. DOI: https://doi.org/10.18441/ind.v28i0.127-152. Gondecki, Philip. 2015. Wir verteidigen unseren Wald: Vom lokalen Widerstand zum globalen Medienaktivismus der Waorani im Konflikt zwischen Erdölförderung und Umweltschutz im Yasuni im ecuadorianischen Amazonastiefland. Bonn, Germany: University of Bonn dissertation. Greenberg, Joseph Harold. 1987, Language in the Americas. Redwood City, CA: Stanford University Press. Halle, Morris & Michael Kenstowicz. 1991. The Free Element Condition and cyclic versus noncyclic stress. Linguistic Inquiry 22(3). 457–501. Haspelmath, Martin. 1995. The converb as a cross-linguistically-valid category. In Martin Haspelmath & Ekkehard König (eds.), Converbs in cross-linguistic perspective: Structure and meaning of adverbial verb forms – adverbial participles, gerunds, 1–55. Berlin: Mouton de Gruyter. Hayes, Bruce. 1995. Metrical stress theory. Chicago: The University of Chicago Press. Heine, Bernd & Kyung-An Song. 2010. On the genesis of personal pronouns: Some conceptual sources. Language and Cognition 2(1). 117–147. High, Casey Ray. 2006. From enemies to affines: Conflict and community among the Huaorani of Amazonian Ecuador. London: University of London dissertation. Hill, Jane H. & Ofelia Zepeda. 1999. Language, gender, and biology: Pulmonic ingressive airstream in women’s speech in Tohono O’odham. Southwest Journal of Linguistics 18(1). 15–40. Holman, Mary E. 1988. The function of ayæ̃ ‘then’ in Waorani. In Evelyn Pike & Rachel Saint (eds.), Workpapers concerning Waorani discourse, 17–22. Dallas: Summer Institute of Linguistics. INEC. 2010. VII Censo de Población y VI de Vivienda. Quito, Ecuador: Instituto Nacional de Estadísticas y Censos. (15 March 2020). Kaufman, Terrence, with help from Brent Berlin. South America. In R. E. Asher & Christopher Moseley (eds.), Atlas of the world’s languages, 61–93. 2nd edn. Abingdon, OX & New York: Routledge. Kay, Paul, Brent Berlin, Luisa Maffi, William R. Merrifield & Richard S. Cook. 2009. The world color survey. Stanford, CA: CSLI Publications. Kelley, Patricia M. 1988. Issues for literacy materials development in a monolingual Amazonian culture: The Waodani of Ecuador. Vancouver: University of British Columbia MA thesis. Kohlberger, Martin. 2020. A grammatical description of Shiwiar. Amsterdam: LOT dissertation.

1240

Alexia Z. Fawcett

Korak, Christina. 2015. Indigenous multilingualism and translation: The creation of intra-and intersocial hierarchies in the communities of the Waorani people of Ecuador. Tusaaji: A Translation Review 4(1). 60–81. DOI: https://doi.org/10.25071/1925-5624.40317. Krasnoukhova, Olga. 2012. The noun phrase in the languages of South America. Utrech: Netherlands Graduate School of Linguistics dissertation. Lester, Peter F. 1994. Altered states: A study in Waorani (Auca) stress. Arlington, TX: University of Texas MA thesis. Loukotka, Čestmir. 1968. Classification of South American Indian languages (Reference Series 7). Los Angeles: University of California, Latin American Center. Lu, Flora. 1999. Changes in subsistence patterns and resource use of the Huaorani Indians in the Ecuadorian Amazon. Chapel Hill, NC: University of North Carolina dissertation. Lu, Flora. 2007. Integration into the market among Indigenous peoples: A cross-cultural perspective from the Ecuadorian Amazon. Current Anthropology 48(4). 593–602. Mithun, Marianne. 1984. The evolution of noun incorporation. Language 60. 847–879. Nedjalkov, Vladimir P. 1995. Some typological parameters of converbs. In Martin Haspelmath & Ekkehard König (eds.), Converbs in cross-linguistic perspective: Structure and meaning of adverbial verb forms – adverbial participles, gerunds, 97–136. Berlin: Mouton de Gruyter. Orr, Carolyn & Nancy S. Pearson. 1979. Nombres de aves en cinco idiomas. Cuadernos Ethnolinguísticos 7. Quito, Ecuador: Summer Institute of Linguistics. Peeke, Catherine. 1962. Divisive criteria for Auca word classes. Bloomington, IN: Indiana University MA thesis. Peeke, Catherine. 1973. Preliminary grammar of Auca. (Summer Institute of Linguistics Publications in Linguistics and Related Fields 39.) Norman, Oklahoma: Summer Institute of Linguistics. Peeke, Catherine. 1979. El idioma Huao: Gramática pedagógica. Vol. 1 (Cuadernos Etnolingüísticos 3.) Quito, Ecuador: Summer Institute of Linguistics. Peeke, Catherine M. 1991. Vocabulario huaorani. In Gustavo A. Fierro (ed.), Listas comparativas de palabras en diez idiomas autóctonos ecuatorianos. (Cuadernos Etnolingüísticos 13.) Quito, Ecuador: Summer Institute of Linguistics. Peeke, Catherine. 1994. Waorani. In Peter Kahrel & René van den Berg (eds.), Typological studies in negation, 267–290. Amsterdam: John Benjamins. Peeke, Catherine M. 2015. Waorani dictionary. In Mary Ritchie Key & Bernard Comrie (eds.), The Intercontinental Dictionary Series. Leipzig: Max Planck Institute for Evolutionary Anthropology. Pike, Evelyn & Rachel Saint (eds.). 1988. Workpapers concerning Waorani discourse. Dallas: Summer Institute of Linguistics. Pike, Kenneth L. 1964. Stress trains in Auca. In David Abercrombie, Dennis Butler Fry, Peter MacCarthy, Norman Carson Scott & J. L. M. Trim (eds.), In honour of Daniel Jones: Papers contributed on the occasion of his eightieth birthday, 425–431. London: Longmans. Proaño García, José & Paola Colleoni. 2008. Taromenane Warani Nani: Pueblos Indígenas en aislamiento voluntario Tagaeri-Taromenane, en la Amazonía Ecuatoriana. Quito, Ecuador: Abya-Yala. Regier, Terry, Paul Kay & Naveen Khetarpal. 2009. Color naming and the shape of color space. Language 85(4). 884–889 Rival, Laura. 1992. Huaorani y petróleo. In Giovanna Tassi (ed.), Náufragos del marverde: La resistencia de los Huaorani a una integración impuesta, 125–179. Quito, Ecuador: Ediciones Abya-Yala. Rival, Laura. 1996. Hijos del sol, padres del jaguar: Los huaorani de ayer y hoy. Quito, Ecuador: Ediciones Abya-Yala. Rival, Laura. 2002. Trekking through history: The Huaorani of Amazonian Ecuador. New York: Columbia University Press.

27 Wao Terero

1241

Rival, Laura. 2004. Ecuador: The Huaorani people of the Amazonia, self-isolation and forced contact. World Rainforest Movement Bulletin 87. https://www.forestpeoples.org/sites/fpp/ files/publication/2011/04/indigenous-peoples-their-forests-struggles-and-rights.pdf (Accessed 15 March 2020). Ross, Deborah. 1988. Chronological and logical discourse trains. In Evelyn Pike & Rachel Saint (eds.), Workpapers concerning Waorani discourse, 9–14. Dallas: Summer Institute of Linguistics. Ruhlen, Merritt. 1987. A guide to the world’s languages, vol 1. Stanford: Stanford University Press. Saint, Rachel & Kenneth L. Pike. 1962. Auca phonemics. In Catherine Peeke & Benjamin Elson (eds.), Studies in Ecuadorian Indian Language I, 2–30. Norman, OK: Summer Institute of Linguistics. SIL. 1978. Lista de frases útiles en los diferentes idiomas de la selva y costa ecuatorianas. Cuadernos Ethnolinguísticos 1. Quito, Ecuador: Summer Institute of Linguistics. SIL. 1995. Cuentos de lectura en el idioma Waodäni del Ecuador. Quito, Ecuador: Summer Institute of Linguistics. Van Gijn, Rik. 2014. The Andean foothills and adjacent Amazonian fringe. In Loretta O’Connor & Pieter Muysken (eds.), The native languages of South America. Origins, development, typology. Cambridge: Cambridge University Press, 102–125. Wamöñe, Camilo Tademö & Catherine Peeke. 2009. Palabras y frases útiles castellano-huaorani. SIL Language and Culture Documentation and Description 4. Dallas, Texas: SIL International. Wetzels, W. Leo & Andrew Nevins. 2018. Prenasalized and postoralized consonants: The diverse functions of enhancement. Language 94(4). 834–866. Wilkendorf, Patricia. 1988. Notes on referential elements in three texts. In Evelyn Pike & Rachel Saint (eds.), Workpapers concerning Waorani discourse, 47–69. Dallas: Summer Institute of Linguistics. Yeti Caiga, Cawetipe. 2012. Gramática Waodani Tededo para nivel básico. Cuenca, Ecuador: Universidad de Cuenca BA thesis. Yost, James A. 1981. Twenty years of contact: The mechanisms of change in Wao (“Auca”) culture. In N. E. Witten (ed.), Cultural transformations and ethnicity in modern Ecuador, 678–704. Urbana: University of Illinois Press. Ziegler-Otero, Lawrence. 2006. Resistance in an Amazonian community: Huaorani organizing against the global economy. New York: Berghahn Books.

Andrés Romero-Figueroa and Konrad Rybka

28 Warao 1 2 3 4 5 6 7

Ethnohistorical and sociolinguistic background Phonology Word classes and their morphosyntactic properties Clause structure Summary Acknowledgments References

1 Ethnohistorical and sociolinguistic background Warao is a language isolate of northeastern Venezuela and adjacent parts of Guyana (Forte 1990; Heinen 1988a; wara1303). Today, the Warao territory covers the Venezuelan state of Delta Amacuro, a network of caños, or channels, forming the delta of the Orinoco, and adjoining parts of the states of Monagas, Sucre, and Bolívar, as well as the Barima-Waini region of Guyana. Until recently, the language was spoken as far east as the Maratakka River in Suriname (Staffeleu 1975), and in former times, on the island of Trinidad, visited by the Warao for trading and ritual purposes. Although an isolate today, Warao may have once belonged to a larger linguistic grouping. Deltaic groups known from historical sources as Siawani (Chaguanes), Veriotaus (Farautes), and Tiuitiuas (Tibitíbis) spoke Warao or related languages. The extinct Aricari and Pirao from Cayenne and Guaiqueri from the island of Margarita may have spoken related languages as well (Boomert 2000: 90). Granberry and Vescelius (2004: 75) also claim that Macorís toponyms of the Greater Antilles are of Waraoan origin, though the evidence is inconclusive (cf. Carlin & Hofman 2010). The evidence of a larger linguistic grouping, dependence on marine resources, and Warao oral traditions that remember the time when Trinidad was connected to the mainland by a land bridge suggest that the Warao may descend from archaic people who occupied the coastal areas from Margarita to Cayenne, displaced by the Arawakan and Cariban expansion (Wilbert 1979). In the resultant sphere of interaction, surrounding the Warao, lived speakers of several Arawakan and Cariban languages, of which all but the Lokono and Kari’na are now extinct. When the Europeans arrived in the area, they portrayed the Warao as barterers of canoes – trade items some Warao contingents were celebrated for – which were exchanged for goods absent in the delta (e.g., tobacco and quartz, Boomert 2000: 424). The Warao

Andrés Romero-Figueroa, Universidad Católica Andrés Bello, Caracas, Venezuela Konrad Rybka, Leiden University Centre for Linguistics, Leiden University, Netherlands https://doi.org/10.1515/9783110432732-015

1244

Andrés Romero-Figueroa and Konrad Rybka

28 Warao

1245

developed oral traditions shared with the Lokono and Kari’na and may have adopted ritual elements from populations as distant as the Ye’kuana (Roth 1915; Heinen & García-Castro 2000). These contacts are reflected in loans from Arawakan and Cariban languages (e.g., kareko ‘quartz’, mawari ‘spirit’, cf. Lokono khaleko, Kari’na imawari), most conspicuous in the Warao spoken closest to the Lokono and Kari’na territories (Lorenzano et al. 2009). The autonym Warao has two possible etymologies, based on a possessive phrase with arao ‘people of’: ‘people of the canoes’ (with wa ‘canoe’, Barral 1979: 460) or ‘people of the swamps’ (with waha ‘swamp’, Heinen 1998). Both etymologies are supported by the Warao and speak to the riverine nature of the people. It is thus not surprising that the Warao are often stereotyped as fishermen who extract starch from the swampy moriche palms (Mauritia flexuosa) and deprecate hunting due to food taboos on mammals (Heinen 1973; Heinen & and Ruddle 1974). The fabric of their society is a polygynous, matrilocal family group which relocates following seasonal changes in the availability of resources. Warao spiritual life is in turn characterized by an animistic system of beliefs, regulated by three types of spiritual leaders. However, while this model is found in remote parts of the delta, in other areas the Warao adapted to different socioecological niches (Heinen 1982, 1988b). Guyanese Warao, like most of their modern Lokono and Kari’na neighbors, are monogynous sedentary cultivators of manioc (Manihot esculenta) who represent various Catholic orders and define themselves in opposition to ‘moriche eaters’. Lately, Warao society has undergone further changes. While the Warao population has increased, reaching 32,000 in Venezuela (Simons & Fennig 2018), the traditional socioeconomic model is waning (Barreto & Mosonyi 1980; Heinen, Lizarralde & Gómez 1994). The introduction of taro (Colocasia esculenta) and rice in the 20th century fostered sedentism and horticulture. Logging, mining, and the palm heart industry led to migrations toward often ethnically mixed economic centers. This has had an adverse effect on language vitality. Language shift to Spanish in Venezuela is conspicuous, particularly among younger generations in mixed communities. As a result, Warao monolinguals include all inhabitants of the isolated parts of the delta (e.g., caños Winikina and Sakobana), but only the elderly people elsewhere. In the easternmost communities in Guyana, furthest removed from the Orinoco, only a few elderly speakers remain, all fluent in Warao, Guyanese Creole English, and often Guyanese English. This geographic pattern is undergoing further changes, as the ongoing economic crisis in Venezuela drives the Venezuelan Warao to migrate to Guyana and Brazil in search of food, safety, and employment. Venezuelan Warao shows little dialectal variation, as evidenced by studies from distant parts of the delta (Olea 1928; Osborn 1966a; Barral 1979, but see Herrmann 2018). Guyanese Warao, on the other hand, differs structurally from the Venezuelan variant, which qualifies it as a distinct dialect. Unless stated otherwise, the present analysis applies to both dialects, but to keep track of the differences, the examples from Venezuelan Warao are cited as coming from Romero-Figeroa’s materials, while

1246

Andrés Romero-Figueroa and Konrad Rybka

those from Guyanese Warao are indicated as coming from Rybka’s corpus. Variation has also been observed on the level of discourse. Several speech genres, lexically and structurally distinct from everyday speech, have been identified in Venezuelan Warao. Examples include myths, incantations, and several genres of songs, some of which are exclusive to the ritual practices of spiritual leaders (e.g., Barral 1964; Olsen 1974, 1996; Lavandero & Heinen 1986). Linguistic descriptions of Warao include articles (e.g., Osborn 1966a, 1966b, 1967), grammar sketches (e.g., Olea 1928; Vaquero 1965; Romero-Figueroa 1997), and dictionaries (Barral 1979; Edwards 1980), but there is no comprehensive grammar. Literature on oral traditions, both in Warao and the official languages, is abundant, particularly for Venezuela (e.g., Roth 1915; Wilbert 1970; Daisy & Mosonyi 1980). There are also several detailed ethnographic studies (e.g., Heinen & Ruddle 1974; Heinen 1973, 1988b; Wilbert 1993; Olsen 1996). Publicly available primary data are limited to recordings by Heinen and colleagues (Heinen n. d.) for Venezuelan Warao and a corpus of Guyanese Warao by Rybka (2018). Notably, a movie by Crespo (2015), spoken in Warao by a cast of Warao actors, was selected as the Venezuelan entry for the Best Foreign Language Film at the 88th Academy Awards.

2 Phonology The Warao consonant inventory contains eleven consonants. Notable anomalies include the lack of a voiceless bilabial and the presence of a labialized velar in the plosive set, as well as the lack of a lateral. Extra-systemic consonants, such as [l], may appear in onomatopoeia (e.g., lakaatata ‘red-throated caracara’). Most of the allophones described in Table 28.1 are optional, though the speakers tend to credit them to dialectal variation. Morphophonemic rules are only mentioned in the following sections if relevant to the understanding of the presented data. Consonantal phonemes are represented by the following allophones. The bilabial /b/ may be realized as a labio-velar [w] in some words (e.g., [wu.ˈɾa.ka] ‘sloth species’). The voiceless allophone [p] is limited to ad hoc borrowings (e.g., [ˈpot] ‘pot’) in Guyana but is in free variation with [b] in some words in Venezuela. Loans

Tab. 28.1: Warao consonants and their allophones. bilabial plosive fricative nasal flap glide

b [p, b, w] m

alveolar

palatal

velar

labio-velar h

t [t, tɕ] s [s, ɕ, z] n [n, ɲ] ɾ [d, dʑ, ɺ]

k [k, kʲ k ]

w

w

glottal

w
> > > >

amahi(a)kaoria-

example

translation

amoho mamoho hi(a)moho kamoho oriamoho

‘hand(s) of’ ‘my hand(s)’ ‘your hand(s)’ ‘our hand(s)’ ‘each other’s hands’

ble 28.5). The possessive a- is inconspicuous in the first-person, where the prefixes are a-final and lose the /a/ when attached to the possessive. It is visible, however, with the second-person hi- or reflexive ori-, though many speakers drop the /a/ in hia-. The prefix is also visible in the third-person which has no corresponding prefix and when the possessor is expressed with a noun. The combinations of prefixes and the possessive are further phonologically conditioned when attached to nouns: /a/ is deleted before a- and e-initial roots. In the third-person, a- alternates thus with Ø on nouns in /a/ or /ɛ/, as in the equative clauses in (6), contrasting the a-initial aru ‘manioc’ with the consonant-initial koko ‘coconut’. (6)

tamaha Ø-aru tamaha a-koko-tuma dem:prox poss-manioc dem:prox poss-coconut-coll ‘This is [his] bitter manioc, these are his coconuts.’ (Rybka, 2018)

The possessive a- appears on the possessed in possessive phrases, even when the possessor is not explicit, in which case a third-person possessor is assumed. It merits a mention that the order of constituents in possessive phrases, where the possessor precedes the possessed, differs from that of other noun phrases, as modifiers typically follow the head noun. Finally, body part terms rarely appear without the possessive a-, but there is no alienability distinction, as shown in (7), where hara ‘forearm’ is the subject of a predicate consisting of an ideophone and the auxiliary ta, a response to a video showing a hand drop an apple into a bowl.

1252 (7)

Andrés Romero-Figueroa and Konrad Rybka

hara bohiii ta-n-ae forearm ideo.appear aux-tel-pst ‘[A] forearm appeared.’ (Rybka, 2018)

Nouns can take the reflexive prefix ori-, which is mutually exclusive with person prefixes. The reflexive is only compatible with (often implicit) plural possessors in combination with which it has a reciprocal reading, as in (8), where it appears on horu ‘pot’, the object of hohara ‘wash with water’. A reflexive meaning is achieved by using possessive marking on the object which cross-references the subject. (8)

oko ori-a-horu hobara-ya 1pl.sbj refl-poss-pot wash.with.water-prs ‘We are washing each other’s pots with water.’ (Rybka, 2018)

Nouns are transnumeral; when informationally salient, collectivity is marked with -tuma, or its allomorphs -ma and -mo, both restricted to certain nouns denoting people (e.g., nobotomo ‘children’, from noboto ‘child’). The transnumeral nature of nouns is illustrated with the equative clauses in (9), from a story in which missionaries ask the Warao if they have individual names, as opposed to an autonym, hence the collective on the noun wai ‘name’. The reply is a negative existential clause unmarked for collectivity, as collectivity is already clear from the context.2 (9)

A: katukane a-wai-tuma? how poss-name-coll ‘What are your names?’ B: k-a-wai ekida 1pl-poss-name neg.exist ‘We have no names.’ (Lit. ‘Our names do not exist.’) (Rybka, 2018)

The singular and plural agent nominalization in -tu and -mo respectively, are the only number-marked nouns. The collective can appear on plural agent nominalizations for emphasis (e.g., nahoromotuma ‘all those who eat’, from the verb nahoro ‘eat’). The collective functions also as an associative plural (e.g., Mariatuma ‘Maria and her family’).

2 In this story, the Warao are presented through the Western gaze as a poor and uncivilized people, who even lacked proper names, and whom the missionaries “saved” from their plight. This image is highly contested today.

28 Warao

1253

To function as locative adjuncts encoding location and goal, nouns are followed by spatial postpositions or the locative enclitic =yata, both of which can take the distributive -ya and the source -mo (Section 3.4). The distributive signals that the event is distributed through the location or goal, while the source -mo marks the phrase as the source of motion. However, select nouns appear unmarked as locations and goals and can take the distributive and source suffix directly. The speakers’ judgments of such forms vary, but all such nouns encode places (e.g., locative nominalizations in -noko, toponyms, hanoko ‘house’). Finally, nouns can attach the non-subject suffix, which disambiguates (di)transitive clauses with third-person human core arguments (Section 4.2).

3.2.1 Attributive nouns Attributive nouns are nouns that usually modify another noun. Like other nouns, they can receive the possessive a-, which allows them to function as core arguments, as in (10), with the attributive hoko ‘white’ and simo ‘red’ as the subjects of equative clauses. Unlike non-attributive nouns, their roots are bound and reduplicate for collectivity. The distinction between the types is fuzzy, as evinced in pairs such as non-attributive muhu ‘bone’ and attributive muhumuhu ‘skinny’. (10) a-hoko tamaha a-simo amaha poss-white dem:prox poss-red dem:dist ‘White [color] is this one, red [color] is that one.’ (Rybka, 2018) Attributive nouns modify another noun in either of the following ways. First, the attributive root can be the N2 in a left-headed compound with a non-attributive head noun, as in (11), where kaba ‘short’ forms a compound with mekoro ‘Creole’, functioning as the subject of the intransitive nao ‘come’. As compounds, such forms have one primary stress, attach the collective -tuma, and often become lexicalized (e.g., domuana ‘dusky parrot’, from domu ‘bird’ and ana ‘black’). As opposed to structurally similar possessive phrases, there is no possessive prefix in the compounds. Right-headed compounds with attributive nouns are also attested (e.g., dau-hanoko ‘wooden house’), but the constructional difference between the compound types remains unclear. (11) mekoro-kaba m-a-hanoko=wata nao-ae creole-short 1sg-poss-house=loc come-pst ‘The short Creole [man] came to my house.’ (Rybka, 2018)

1254

Andrés Romero-Figueroa and Konrad Rybka

Attributive nouns can also form relative clauses postposed to the head noun. To this end, attributive nouns usually attach the approximative -u, which adds the meaning ‘somewhat, quite’ (e.g., anau ‘blackish’, from ana ‘black’, basau ‘flattish’, from basa ‘flat’). Once suffixed, the noun is followed by the tense-marked form of the auxiliary ta, completing the attributive clause (Section 4.3). The clause can be relativized by attaching a relativizer, as in (12), where simo ‘red’ forms a relative clause with the present form of the auxiliary. Less commonly, attributive nouns that do not normally appear in attributive clauses with ta form equative clauses with the copula ha (Section 4.3). These can also be relativized and postposed to the head noun, as in the equative clause with hido ‘new’ (13). The contrast between attributive and equative clauses is discussed in Section 4.3, but it is unclear whether it corresponds to the contrast between left- and right-headed compounds mentioned above and whether either of these two constructional differences maps onto the difference between qualitative and classifying uses of attributive nouns. (12) tamaha diana hiaka simo-u tia-ha batoko-bu-ae dem:prox already cloth red-aprx aux.prs-rel:indf hang-ints-pst ‘This red cloth is hanging’ (Lit. ‘this cloth, which is reddish, is already hung’) (Rybka, 2018) (13) tamaha hota hido ha-kore dem:prox earth new cop-cond ‘When this earth was new […]’ (a common opening of traditional stories) (Rybka, 2018)

3.3 Determiners There is some variation in the form and function of determiners, in particular demonstratives, the focus of this section, which may be the result of both languageinternal change and language contact. However, some generalizations can be made (Table 28.6). There are three transnumeral pronominal demonstratives, with collective equivalents, as well as three general locative demonstrative adverbs, from which another set of locative adverbs with spatially narrower meanings is derived. In Venezuelan (Herrmann 2018) and Guyanese Warao, pronominal demonstratives can be used adnominally. Since /h/ and word-final vowels can be elided (Section 2), tamaha and amaha may surface as shorter forms with ultimate stress (e.g., [ta.ˈmaː] or [ta.ˈma]). This phonetic process may have been reanalyzed in some areas; Romero-Figueroa (1997) distinguishes adnominal forms without the final syllable from the full pronominal forms. The elision of /ha/ does not apply to the demonstrative tai, used pronominally and adnominally, which, however, often surfaces as [ta]. Given the above, the pronominal–adnominal contrast noted by Romero-Figueroa for

28 Warao

1255

Tab. 28.6: Warao demonstratives. pronominal

proximal distal anaphoric

adverbial

transnumeral

collective

general

precise

tamaha amaha tai

tama(ha)tuma ama(ha)tuma tatuma

tamate amate tata

tamatuka amatuka tatuka

proximal and distal forms only is likely a language internal development. Moreover, all three demonstratives may attach the collective -tuma. Romero-Figueroa (1997) reports the transparent forms with /ha/ such as tamahatuma, phonologically reduced to tamatuma in other varieties; the collective form of tai is consistently tatuma. Like nouns, pronominal demonstratives are transnumeral, as in (14), where tamaha appears despite the head noun referring to a plural entity. (14) tamaha karata ori-arai er-ida tobooow ta-n-ae dem:prox book refl-on many-aug ideo.stacked aux-tel-pst ‘These stacks of books are piled on top of one another […]’ (Rybka, 2018) There are also differences in the use of demonstratives. In Venezuelan Warao, tamaha and its derivatives (including the locative forms) are distance-sensitive and apply to entities within the speaker’s reach or those easily brought within it and are usually accompanied by a touching or pointing gesture (Herrmann 2018). The amahaforms are used for objects beyond the search zone of tamaha-forms, as in (15), and contrastively with this search zone. In Guyana, tamaha-forms, accompanied by a gesture, are used at all scales, while amaha-forms are mostly used contrastively. (15) ure amaha-tuma saba ihi kona-n-ae taro dem:dist-coll ben 2sg.sbj bring-tel-pst ‘You brought taro for those yonder.’ (Romero-Figueroa, 1997; Rybka, 2018) The demonstrative tai and its derivatives are not distance-sensitive. Herrmann (2018) writes that they are used when speech participants pay attention to the referent. While this formulation is accurate, the forms refer, more broadly, to previously mentioned entities even when these are not attended to at a particular moment. The anaphoric rather than deictic nature of tai shows in its use as a reference-tracking device, as in (16), which sums up a monologue about manioc. It is the only demonstrative that attaches the non-subject suffix (Section 4.2), which disambiguates

1256

Andrés Romero-Figueroa and Konrad Rybka

(di)transitive clauses with third-person human arguments, a redundant feature for a deictic form. (16) tai omi oko ha-komoni dem:anph without 1pl.sbj cop-neg.pot ‘We cannot [live] without [manioc].’ (Rybka, 2018) Locative demonstrative adverbs are derived from the pronominal forms with -te and -ta, most likely the reduced forms of the locative enclitic =yata (Section 3.4), and, like the collective pronominal forms, may have undergone phonological reduction (e.g., /tamaha=yata/ > /tamatɛ/). From the locative triad, adverbs indicating a smaller search zone are formed with the diminutive -uka (or -ika). The contrast between the two sets is shown in (17), where tatuka and tata indicate a smaller and a larger area, respectively. Both sets are used to locate entities in space and time. (17) tat-uka duuu ta-n-ae takore tata isia anph.loc-dim ideo.appear aux-tel-pst but anph.loc about kokotuka=witu namina-naka ta-n-ae all=rest know-neg aux-tel-pst ‘I reached that place, but I did not get to know all about the whole place.’ (Rybka, 2018) Locative demonstratives of both types do not require additional morphology to function as adjuncts encoding the location or goal, as in (18). They do, however, require the source suffix -mo when encoding the source, as in (18), where tatukamo is used to refer to a ‘place’ in time. (18) tat-uka-mo i-yakara-t-ubu-ne tane diana nahoro anph.loc-dim-src caus-clearing-vrb-ints-conj then already food aba-ya put-prs ‘After that, they cleaned [the meat] and then they were preparing food.’ (Rybka, 2018) Determiners include the indefinite and definite enclitics =ha and =hakotai, numerals, quantifiers, and three question words that may be used adnominally: sina ‘who’, kasikaha ‘which (thing)’, and katukaya ‘which (kind)’. These modifiers and the demonstratives follow the head noun in Venezuelan Warao, as in (19). In Guyana, however, the adnominally used question words and demonstratives precede the head, as in (14) above. Given that in Guyanese Creole English determiners appear before the head and that the Guyanese Warao are bilingual, this constituent order may have been borrowed.

1257

28 Warao

(19) Warao-tuma sina yaota rubuhi-te-ra? Warao-coll who work hurry-fut-int ‘Which Warao will hurry [their] work?’ (Romero-Figueroa, 1997)

3.4 Postpositions Warao has various postpositions which form postpositional phrases that function as locative or temporal adjuncts or introduce non-core arguments, for example, the beneficiary (Table 28.7). Postpositions form phrases with nominal expressions. Alternatively, they take person prefixes, mutually exclusive with the explicit postpositional object or the reflexive ori-. The postpositional object need not be expressed at all, while the prefixes are attached without the intervening possessive a-, which contrasts postpositional and possessive phrases. Both features of postpositions are illustrated in (20), where the addressee introduced by the benefactive saba is known from the context and remains implicit. In (21) and (22), the reflexive, which has a reciprocal meaning on postpositions, and the second-person prefix, respectively, attach to the benefactive. (20) tane seke saba dibu-ae=yama then sure ben speak-pst=hrs ‘Then of course [the missionaries] spoke to [the Warao], it is said.’ (Rybka, 2018) (21) nebu-tuma tuatane ori-saba wara-e young.man-coll like.this refl-ben talk-pst ‘The young men talked to each other like this.’ (Rybka, 2018) (22) koko Roni hi-saba ma-moa-e coconut Ronnie 2sg.obj-ben 1sg.obj-give-pst ‘Ronnie gave me a coconut for you.’ (Rybka, 2018)

Tab. 28.7: A selection of Warao postpositions. postposition

meaning

postposition

meaning

postposition

meaning

arai autu awere diboto ebe eku

above, on top in the middle near against in front inside

isia isiko isimo kaika kawana kware

by, about with from together among towards

kwarika monika noika nokaba omi saba

above like under behind without for

1258

Andrés Romero-Figueroa and Konrad Rybka

To function as locations or goals, most nouns form postpositional phrases or are followed by the enclitic =yata, treated here because of its locative function. Unlike postpositions, the locative =yata (or =wata if the preceding word ends in /o/) is an enclitic that does not attach to prefixes but follows nouns and subject pronouns. While postpositions express specific spatial relations, such as kwarika ‘above’ in (23), =yata has a general locative or goal meaning, as in (24). (23) naba kwarika kwai yaburu-ae river above high climb-pst ‘He climbed high above the river.’ (i.e., ‘high on a branch overhanging the river’) (Rybka, 2018) (24) ine diawara-e Waromuri=yata 1sg.sbj be.born-pst Waramuri=loc ‘I was born in Waramuri.’ (Rybka, 2018) To function as locative adjuncts encoding source or path, both =yata and postpositions take the source suffix -mo, as in (25) with noikamo ‘from under’, or the distributive -ya, signaling that the event is distributed through the space, as in (26) with =yataya ‘through’. (25) ha noika-mo beroro ehobo-n-ae hammock under-src dog come.out-tel-pst ‘The dog came out from under the hammock.’ (Romero-Figueroa, 1997) (26) oko waromuri=yata-ya naru-ae 1sg.sbj waramuri=loc-distr go-pst ‘We went through Waramuri.’ (i.e., ‘we passed through Waramuri’) (Rybka, 2018) The distributive -ya is lexicalized in the postposition isia ‘by, about’, while the postposition isimo ‘origin’ lexicalized the source suffix -mo. The latter is worth discussing, as it is felicitous in static scenes and dynamic scenes that do not involve locomotion (i.e., when moving an object across a table but not when carrying it across a room). If isimo appears in a clause expressing locomotion, it is read figuratively, as in (27), where it implies a contribution to an event. The postposition contrasts with =yatamo, the combination of =yata and -mo, used when locomotion is involved.

28 Warao

1259

(27) ka-isimo ine kona-te 1pl.obj-origin 1sg.sbj bring-fut ‘I will bring [this] from us.’ (i.e., ‘[This] will be our contribution.’) (Rybka, 2018) Several postpositions derive from nouns that lexicalized the possessive a-, which otherwise does not appear on postpositions (e.g., arai ‘above’, from dai ‘back’). Several other postpositions contain the diminutive -uka (or -ika), which narrows their spatial or temporal scope, as it does with demonstratives (Section 3.3), as illustrated in (28), with ebe ‘in front’ and ebika ‘right in front’. (28) m-eb-ika yaaa ta-ne naru-ø ote=witu yaaa 1sg.obj-in.front-dim ideo.walk aux-cont go-2sg.imp far=rest ideo.walk ta-naka! aux-neg ‘Walk right in front of me, do not go too far!’ (Rybka, 2018)

3.5 Verbs The Warao verb is a multiplex product of agglutinative morphology. For the purposes of this exposition, its most important features are grouped into several broad themes. Core arguments can be referenced on the verb with subject and object affixes, as described in Section 3.1. Verb formation includes compounding, reduplication, and denominal verbalization (Section 3.5.1). There are several, mostly prefixal, valency-changing operators (Section 3.5.2). The pluractional prefix and the intensifying suffix partly compensate for the lack of number marking on nouns (Section 3.5.3). Tense (Section 3.5.4), telicity (Section 3.5.5), and mood (Section 3.5.6) are closely intertwined and expressed mostly by suffixes. Several modal categories, as well as negation, are expressed by suffixes which block further suffixation, requiring ancillary verbs to form a complete predicate (Section 3.5.7).

3.5.1 Derivation Setting aside valency-changing operations (Section 3.5.2), which are not treated as derivational here, verbs are formed with denominal verbalizers or by compounding or reduplication. The intensifying suffixes -era and -ida, which have an augmentative function with nouns, also derive intransitive verbs encoding states. The verb ahera ‘be painful’ in (29), identical to the attributive noun ahera ‘painful’, for instance, derives from ahi ‘pain’. As verbs, such forms do not require the copula, as

1260

Andrés Romero-Figueroa and Konrad Rybka

in the second clause in (29), which is present in nominal predicates, and they attach verbal suffixes such as the conditional -kore in the first clause of (29). (29) hokohi ah-era-kore yaka-era ha nona-kitane sunshine pain-aug-cond good-aug hammock make-inf ‘When the sun shines hard (Lit. ‘is painful’), [it is] good to make hammocks.’ (Rybka, 2018) From nouns, verbs encoding the action of attaining the state encoded by the attributive noun are derived with the verbalizer e- (or i- if the following syllable contains /a/), which functions also as a causative. The prefix is illustrated in (30), where it derives ehido ‘sharpen’ from the noun hido ‘new’. Both attributive and non-attributive nouns can also be verbalized with the suffix -ta, as in (30), with kabata ‘cut’ (i.e., ‘divide into parts’), derived from kaba ‘part’. Borrowed verbal roots (e.g., start) are also overtly verbalized with -ta, which likely grammaticalized from the auxiliary ta (Section 4.3). (30) dau kaba-ta-ne e-hido-ne tree part-vbz-conj caus-new-conj ‘[He] cut a stick and sharpened it and […] (Rybka, 2018) Verbs are also formed by compounding and reduplication. In compound verbs, a nominal root appears before a verbal root, as in the transitive hohara ‘wash with water’ in (8) above, derived from ho ‘water’ and the transitive hara ‘wash’. Moreover, many verb stems can be reduplicated to express iterativity (e.g., bahibahi ‘circle round’, from bahi ‘turn around’), similarly to ideophones (Section 3.6), and in contrast to attributive nouns, where reduplication signals collectivity (Section 3.2.1).

3.5.2 Valency-changing affixes Valency-decreasing devices are limited to the reflexive ori-. The reflexive appears on nouns, postpositions, and verbs and is incompatible with person prefixes. With (di)transitive verbs, where the prefixes encode the object, ori- reduces valency, resulting in reflexive or reciprocal verbs. With singular subjects, ori- has a reflexive meaning, as in (31) with the ditransitive inamina ‘teach’. (31) aruba nona-kitane ine=witu ori-i-namina-e manioc.press make-inf 1sg.sbj=rest refl-caus-know-pst ‘I alone taught myself how to make a manioc press.’ (Lit. ‘I made myself know.’) (Rybka, 2018)

28 Warao

1261

Valency-increasing morphology includes the causative e-, which functions as a verbalizer with nouns (Section 3.5.1). As a valency-increasing prefix, it appears on intransitives such as masibu ‘be deceived’ in (32), which it renders transitive, and on some transitives, such as namina ‘know’ in (31), which it renders ditransitive. In Venezuela, the causative is productive, but its distribution is limited in Guyana. In both dialects, the causer is expressed by the subject, the causee by the object prefix, and the object of the underlying transitive verb is left out or expressed with a noun phrase. (32) amau ihi ma-i-masibu-ae then 2sg.sbj 1sg.obj-caus-be.deceived-pst ‘Then you deceived me.’ (Lit. ‘You caused me to be tricked.’) (Rybka, 2018) The only valency-increasing suffix is the commandative -moro meaning ‘order’, as in (33), where it appears on the intransitive naru ‘go’. Like the transitivizer, the commandative turns transitive verbs into ditransitive verbs. The subject of the resultant predicate encodes the person giving the order; the object prefix, however, encodes the patient, not the causee, which can in turn be expressed with a noun phrase, as in (34), or inferred from the context. (33) ine tai naru-moro-kore 1sg.sbj dem:anph go-cmnd-cond ‘When I order him to go […]’ (Rybka, 2018) (34) ine Roni hi-hata-moro-ae 1sg.sbj Ronnie 2sg.obj-shoot-cmnd-pst ‘I told Ronnie to shoot you.’ (Rybka, 2018) Warao has also several applicatives, the exact differences among which remain unclear. The applicative se-, for instance, usually promotes a locative oblique to an object, as in (35), where it turns the intransitive dokoi ‘vomit’ into transitive serokoi ‘vomit on’. The applicative te-, while promoting an oblique, intensifies the meaning, as in (36), where it turns the transitive dibu ‘say’, which marks the addressee with the benefactive saba, into transitive teribu ‘speak a lot to’. (35) a-sinoto se-rokoi-ae poss-shirt appl-vomit-pst ‘[He] vomited on his shirt.’ (Rybka, 2018)

1262

Andrés Romero-Figueroa and Konrad Rybka

(36) a-tida te-ribu-ya poss-woman appl-say-prs ‘[He] speaks a lot to his wife.’ (Rybka, 2018)

3.5.3 Number Since number is not obligatorily encoded in the noun (Section 3.2), the verb often helps disambiguate the grammatical number of core arguments, which can be indicated by the pluractional no- or entailed by the intensifier -(u)bu. The former indicates the plurality of an argument, as in (37), where it attaches to the intransitive dia ‘disappear’. With transitives, the pluractional implies a plural subject (e.g., nonati ‘(many) stab’, from nati ‘stab’), or object (e.g., nonamu ‘plant (many)’, from namu ‘plant’). (37) no-ria-e diana pluract-disappear-pst already ‘[They] disappeared already.’ (Rybka, 2018) Number may also be entailed by the intensifier, which often conveys the iteration of the event, and which may imply plural participants, as in (38). The intensifier can co-appear with the pluractional, which brings to the fore their different semantics, as in the forms derived from tuta ‘break’: tutubu ‘break into pieces’, notutu ‘(many) break’, and notutubu ‘(many) break into pieces’. In addition, the intensifier has present and past habitual interpretations. (38) tai boro-bu-te dem:anph cut.open.trunk.of.moriche-ints-fut ‘That [man] cuts open many trunks of moriche.’ (Rybka, 2018)

3.5.4 Tense Three suffixes can be described as tense markers: present -ya, future -te, and past -ae. The present -ya likely developed via insubordination from the combination of the progressive subordinator -i (Section 4.5) and the copula -ha (Section 3.5.6), which explains why it covers progressive aspect and present tense. As progressive, it does not attach to the copula ha, negative existential ekida, or intransitive stative verbs derived from nouns. It is compatible with the past and future suffix if these

28 Warao

1263

appear on an ancillary verb (Section 3.5.6.3) and with conditional -kore (Section 4.5), together with which it implies a time-reference that overlaps with that of the main clause, as in (39) with naria, a phonologically reduced and resyllabified variant of naru-ya. However, -ya is also used to speak of the present in general, as in (40), where it appears on namina ‘know’ and dibu ‘say’. Notice that namina encodes a state suggesting a shift from progressive to present semantics. For these reasons and to distinguish it from the progressive -i, -ya is treated as a present tense suffix. (39) nar-ia-kore tane tai a-ramitu diana diboto-n-ae go-prs-cond then dem:anph poss-husband.of.sister already meet-tel-pst ‘While [they were] going, they met their sister’s husband.’ (Rybka, 2018) (40) takore ine namina-ya Wauta dibu-ya but 1sg.sbj know-prs Wauta say-prs ‘“But I know [them],” says Wauta [a mythical frog].’ (Rybka, 2018) The future -te is typically used with a future time-reference, as in (41), where it appears on naru ‘go’. More generally, it signals the next in a series of steps, as in the instructions in (42). By contrast, the past -ae (or -e after /a/) indicates a past time-reference, as in (43), where it is attached to diawara ‘give birth’. As opposed to the present suffix, which can be followed by conditional -kore, the future and past suffixes always close the verbal string and are mutually exclusive. This paradigmatic relationship is echoed on the level of meaning in the Venezuelan variety studied by Romero-Figueroa, where the future tense suffix is also used with a present time-reference, thus being in fully complementary distribution with the past tense suffix. (41) hake hokooo ta-ne aniaka=witu oko diana naru-te tomorrow white.ideo aux-cont dawn=rest 1pl.sbj already go-fut ‘Tomorrow right at dawn, as it gets bright, we will be gone.’ (Romero-Figueroa, 1997) (42) araho ihi nisa-te horu eku abana ihi nisa-te manioc.juice 2sg.sbj take-fut pot inside froth 2sg.sbj take-fut ‘You take the manioc juice, in the pot, you take the froth.’ (Rybka, 2018) (43) m-a-idamo-tuma ma-riawara-e Waromuri arai 1sg-poss-elderly-coll 1sg-give.birth-pst Waramuri on ‘My parents gave birth to me in Waramuri.’ (Rybka, 2018)

1264

Andrés Romero-Figueroa and Konrad Rybka

3.5.5 Telicity Some verbs may encode a contrast in telicity, where telic verb forms refer to events conceptualized as having endpoints and atelic verb forms refer to events conceptualized as lacking them. However, the difference between clauses with atelic and telic forms is elusive, since telicity status affects the interpretation of the whole clause. This explains why the contrast has also been described as a type of number or aspect marking (cf. Osborn 1959; Romero-Figueroa 1997). The form and function of stem alternation and the operative suffix -n remain, however, the least understood part of Warao grammar, and the analysis presented here may need to be revisited in the future. Atelic forms typically imply collective, unaffected patients, and durativity, while telic forms imply non-collective, wholly affected ones, and punctuality. Telicity also interacts with lexical aspect, as atelic forms are typically the only forms available for stative and activity verbs, which are inherently atelic, while telic forms are typically the only forms available for achievement verbs, which are inherently telic. There are also several verbs whose telicity marking does not match their meanings, which are treated as irregular, though in some cases the irregularity can be explained by diachronic processes of semantic shift and lexicalization. Telicity is marked either by stem alternation if the following suffix starts with a consonant, or the suffix -n, if the following suffix starts with a vowel, which is only the case in the past tense and the 2nd person singular imperative (Section 3.5.6). In the past tense, the telic -n attaches to the stem before the past suffix -ae (or -e after /a/). Some verbs thus have two past forms, an atelic one without -n and a telic one with -n (e.g., tuara ‘stop’, which becomes tuarae ‘rested’ or tuaranae ‘stopped’ in the past). Some verbs have only one past form, dependent on their lexical aspect. Achievement verbs (e.g., nabaka ‘arrive’) typically have only a telic form, while stative and activity verbs (e.g., namina ‘know’ and uba ‘sleep’, respectively), typically have only an atelic form. Examples (44) and (45) contrast the two past forms of nisa ‘take’; notice that here the atelic–telic contrast translates into a contrast between a partly and wholly affected patient. (44) m-a-merehi ihi nisa-e 1sg-poss-cashew 2sg.sbj take-pst ‘He took some of my cashews [from the bag].’ (Rybka, 2018) (45) m-a-merehi ihi nisa-n-ae 1sg-poss-cashew 2sg.sbj take-tel-pst ‘He took [e.g., a bag of] my cashews.’ (Rybka, 2018) If the following suffix is consonant-initial, as in the non-past tenses, telicity is encoded with stem alternation, a process restricted to some verbs that end in a vowel

28 Warao

1265

other than /a/. Such verbs change the final vowel to /a/ or rarely /e/ to form a telic stem (e.g., atelic wiri ‘paddle’ and telic wira ‘paddle a stroke’). That said, many verbs that end in a vowel other than /a/ and all verbs that end in /a/ have only one stem. In such cases, the lack of the telicity contrast in the non-past may correspond to its lack in the past, as predicted by lexical aspect. In some cases, however, the contrast is available in the past, where it is marked by the suffix, but not in the non-past (e.g., tuarate ‘will rest’ or ‘will stop’). Examples (46) and (47) contrast the two stems of duhu ‘sit’ and illustrate the inherently telic accomplishment verb nawaha ‘dry out’ and two inherently atelic activity verbs koita ‘call out’ and yaburu ‘climb’, all of which have only one stem. (46) ine arai duhu-te nawaha-kitane 1sg.sbj on sit-fut dry.out-inf ‘I will sit on [the manioc press] to dry [the manioc].’ (Rybka, 2018) (47) yaburu-kitane diana duha-te tane diana koita-te climb-inf already sit.tel-fut then already call.out-fut ‘After climbing the tree, [the howler monkey] will sit down and howl.’ (Rybka, 2018) Stem alternation and the suffix -n normally do not co-occur. However, there are cases where a verb is double-marked as telic. In some cases, the telic stem may have become lexicalized, as appears to be the case with namu ‘plant (vegetables)’ and nama ‘stick in the ground vertically’. It appears also that the system is regularizing, with -a, the most common marker of telic stems, spreading to contexts where it is not expected, such as the second-person imperative (Section 3.5.6).

3.5.6 Mood Warao has several suffixes that fall into the category of mood, most notably imperative (Section 3.5.6.1) and interrogative (Section 3.5.6.2), as well as several other suffixes that straddle the border between mood and aspect (Section 3.5.6.3). Many of these suffixes block the addition of further verbal suffixes, which must then be carried by an ancillary verb and are therefore treated here as well. The following sections discuss the suffixes that may require the formation of complex predicates with such ancillary verbs. Warao has two such verbs, the auxiliary ta and copula ha, the latter of which has a lexicalized collective equivalent eha (from era ha ‘be plenty’). Both verbs function as auxiliaries with verbal predicates and as copulas with nonverbal predicates. In the former function, the auxiliary is more common; in the latter function, the copula prevails, hence their names. As auxiliaries, the two verbs fol-

1266

Andrés Romero-Figueroa and Konrad Rybka

low a content predicate marked by a suffix that closes the verbal string, pushing further morphology to the ancillary. Often both verbs are acceptable; where semantic differences between the construction with the auxiliary and copula have been detected, the auxiliary implies an action while the copula indicates a state, in keeping with the broader set of their uses in verbal and non-verbal predicates, respectively (Section 4.3). Both ancillaries appear in clauses with many of the suffixes; it is also common for a sentence with and without the ancillary to be acceptable. It is beyond the present scope to delve into the form and function of such clauses, but it is worth signaling that some forms derived with these suffixes bear features of deverbal nouns and adverbs.

3.5.6.1 Imperative Starting with canonical imperatives, the second-person singular imperative is expressed by -u or a morphological zero (postulated here because all verb stems apart from stative verbs derived from nouns, the copula ha, and the negative existential ekida are bound stems). Verbs that end in /a/ form imperatives by the addition of -u; other verbs have unmarked imperative forms. Provided that the telic marking is compatible with the lexical aspect of the verb, the verb can attach the telic -n to encode a telic event, the expected form of telicity marking since the following suffix, the imperative -u, is vowel-initial. The atelic and telic second-person singular imperatives of bere ‘sweep’, a verb that happens to exhibit stem allomorphy, are given in (48) and (49), respectively. However, as mentioned above (Section 3.5.5), the stemfinal /a/, the most common marker of telic stem alternants, may be spreading to such forms as well, as illustrated in (50). (48) borohoro bere-Ø floor sweep-2sg.imp ‘Sweep [the] floor.’ (Rybka, 2018) (49) ha-noko bere-n-u hammock-place sweep-tel-2sg.imp ‘Give the house (Lit. ’hammock place’) a sweep.’ (Rybka, 2018) (50) ha-noko bera-n-u hammock-place sweep.tel-tel-2sg.imp ‘Give the house (Lit. ‘hammock place’) a sweep.’ (Rybka, 2018) The second-person plural -kotu and the remaining imperative suffixes, all of which are consonant-initial, attach to telic and atelic stems and are incompatible with the

28 Warao

1267

telic -n. Hortatives include -ki and -kí, both inclusive, the second of which attracts primary stress and expresses a mandate rather than an exhortation. The contrast is on the wane, with -kí becoming a general hortative. The jussive -kunarae expresses an order addressed to a third person. Worth recalling is also the valency-changing commandative -moro, which has imperative semantics (Section 3.5.2). Negative imperatives are formed with the negator -naka, which closes the verbal string and requires further morphology to be carried by the copula ha or the auxiliary ta. The negator -naka pushes the imperative marking to the auxiliary or, less often, the copula, as in (51) with the jussive -kunarae. In the second-person singular imperative, however, the ancillary verbs are often dropped altogether. (51) ka-nuha-naka ha-kunarae 1pl.obj-be.afraid-neg cop-perm ‘They must not be afraid of us.’ (Rybka, 2018)

3.5.6.2 Interrogative Warao has an explicit interrogative marker whose use depends on the dialects. Both in Guyana and in Venezuela, polar and content questions can be left morphologically unmarked and distinguished from the indicative mood by intonation only, as in (52). In the Guyanese variety, the interrogative -ra may appear in polar questions, attached as a final suffix to the finite verb. In Venezuela, however, both types of questions can be explicitly marked by the interrogative -ra, as in (53) and (54) further below. In non-verbal clauses without an overt ancillary verb, -ra is attached to the question word. (52) horu-tuma isa namina-ya? pot-coll about know-prs ‘Do you know about pots?’ (Rybka, 2018) (53) naba muhoko tatuma waba-komoni ta-e-ra? river side dem:anph.coll die-neg.pot aux-pst-int ‘Couldn’t they die on that side of the river?’ (Romero-Figueroa, 1997) Content questions are formed with the question words sina ‘who’, bitu ‘what’, kasikaha ‘which type’, katukaya ‘which kind’, kasaba ‘where’, katukane ‘how’, or katamona ‘how many’, as in (54). It is worth noting that since question words are clauseinitial in Warao, questions about the subject have an SOV word order, whereas OSV is characteristic of most other clauses (Section 4.1).

1268

Andrés Romero-Figueroa and Konrad Rybka

(54) sina ma yehe-bu-te-ra? who 1sg.obj shout-ints-fut-int ‘Who keeps shouting at me?’ (Romero-Figueroa, 1997)

3.5.6.3 Other suffixes that may close the verbal string Warao has numerous other suffixes with modal and aspectual semantics that close the verbal string and call for the ancillaries. This is typically the case if the clause contains tense morphology; if it does not, it is interpreted as present tense and the ancillaries are often not required. A small selection of such suffixes is given below. The antiperfective -moana encodes an action that has not taken place yet. It forms a clause with a present time reference, as in (55), unless specified differently by tense suffixes on the copula. (55) ubu-bu-moana sleep-ints-apfv ‘[They] have not slept yet.’ (Rybka, 2018) The desiderative -turu is usually a closing morpheme, as in (56), and can have frustrative overtones. The suffix has an emphatic equivalent -turuuu, meaning ‘die for’, followed by a pause and the auxiliary ta, a construction typical of ideophones (Section 3.6). (56) osibu ine nahoro-turu ta-te morocoto 1sg.sbj eat-desid aux-fut ‘I want to eat morocoto (Collosama macropomus).’ (Romero-Figueroa, 1997) A frustrative and adversative meaning is also expressed with the suffix -ka, labeled “unintentional” by some authors, as in (57). Its precise meaning may vary per dialect (cf. Osborn 1967; Romero-Figueroa 1997). (57) dura-ka ta-e hit.tel-frus aux-pst ‘It nearly hit him.’ (Rybka, 2018) The facsimile -sita grammaticalized from the verb sita ‘copy’. The suffix indicates that the subject pretends to perform the activity encoded by the verbal root, as in (58), where it follows the intensifier -(u)bu on the verb hobi ‘drink’.

28 Warao

1269

(58) hobi-bu-sita-ya drink-ints-facs-prs ‘[They] are pretending to be drunk.’ (Lit. ‘pretending to be drinking a lot’) (Rybka, 2018) The negative potential -komoni usually expresses the physical inability to perform the action, as in (59), but also moral prohibitions. The opposite potential meaning is encoded by -komo in Venezuela and -kuna in Guyana. (59) masi hata-komoni ta-n-ae deer shoot-neg.pot cop-tel-pst ‘[He] could not shoot the deer.’ (Romero-Figueroa, 1997)

3.5.7 Negation There are two standard negators, -naha and -naka, which may surface as [n] alone, and a few suffixes that encode negation as part of more complex meanings (e.g., negative potential -komoni, anticompletive -moana). Negation is also expressed by the negative existential ekida (Section 4.3), postposition omi ‘without’ (Section 3.4), and the particle =yana used in equative clauses (Section 4.3). The negator -naha appears when there is no tense morphology on the verb. It has a present tense reading, as in (60) and is the negative equivalent of the present -ya. It does not require an ancillary, possibly because it includes the grammaticalized copula ha. (60) ine naru-kore ine homakaba oa-naha 1sg.sbj go-cond 1sg.sbj fish catch-neg.prog ‘When I go [fishing], I do not catch [any] fish.’ (Rybka, 2018) By contrast, the negator -naka appears when additional tense morphology (past or future) is present. The suffix frequently closes the verbal string, pushing further morphology to the copula or the auxiliary, as in (61), but some suffixes may appear before -naka (e.g., commandative -moro). (61) ine namina-naka ta-bu-te 1sg.sbj know-neg aux-ints-fut ‘I certainly do not know.’ (Rybka, 2018) Negative imperatives are a good example of the semantic contrast between the ancillary verbs. The negator -naka often blocks the addition of imperative suffixes on

1270

Andrés Romero-Figueroa and Konrad Rybka

the main verb (Section 3.5.6.1), so that they must be carried by the copula or auxiliary, which entail subtly different meanings: an order to discontinue a state (62) or an action (63). (62) obonobu-naka ha-u! think-neg cop-2sg.imp ‘Do not be worried!’ (63) obonobu-naka ta-kotu! think-neg aux-2pl.imp ‘[You all] stop worrying!’

3.6 Ideophones Warao speakers make extensive use of ideophones, defined as “marked words that depict sensory imagery” (Dingemanse 2011: 25). Their marked status shows, for instance, in long, stressed, word-final vowels, an atypical syllable structure in Warao. Such final vowels, written as triple, are often stretched and followed by a pause. Ideophones typically appear towards the end of the clause, followed by the auxiliary ta, which carries the verbal suffixes, as in (64), with bohiii ‘appear (unexpectedly)’ and bareee ‘splash’. As such, ideophone clauses are almost identical to attributive clauses (Section 4.3). Ideophones may, however, also occur by themselves as single utterances. (64) hara bohiii ta-n-ae horu eku ho naho-n-ae forearm ideo.appear aux-tel-pst pot inside water pour.in-tel-pst naho-ne bareee ta-n-ae pour.in-cont ideo.splash aux-tel-pst ‘A forearm appeared, poured water into a pot, pouring [it] all out with a splash.’ (Rybka, 2018) Ideophones are reduplicated for collectivity (65), as well as for iterativity or intensification (e.g., hihiii ‘move from side to side’, from hiii ‘move’). Collectivity is also inherent in the meaning of some ideophones (e.g., yaaa ‘(one) walk’ and houaaaw ‘(many) walk’). (65) eku-mo ibure bohi~bohiii ta-n-ae diana inside-src white.lipped.peccary coll~ideo.appear aux-tel-pst already tora~toraaa ta-n-ae coll~ideo.stand.still aux-tel-pst ‘The white-lipped peccaries came out one by one, and all stood still.’ (Rybka, 2018)

28 Warao

1271

While most ideophones are underived, ideophones can be derived from verbs and attributive nouns by simultaneously lengthening and stressing the root-final vowel. The ideophone bohiii in (64) and (65) is in fact an example of such a derivation, coined from bohi ‘appear (unexpectedly)’; barobarooo ‘thin’ is in turn an example of an ideophone derived from the attributive baro ‘thin’. Moreover, with some attributive nouns and verbs, changing the root-final vowel to a long stressed /i/ or sometimes /ɛː/ results in an ideophone with an emphatic collective meaning (e.g., aniii ‘plenty black’, from ana ‘black’).

3.7 Adverbs Adverbs are a heterogenous class. Defined as terms that modify the predicate or the clause, Warao adverbs include a number of simplex forms as well as potentially unlimited expressions with verbs, nouns, and ideophones. Simplex forms include temporal (e.g., ama ‘now’, hake ‘tomorrow’, and atehe ‘long time ago’), spatial (e.g., kwai ‘high’, ito ‘far’), and other adverbs (e.g., atae ‘again’, kuwana ‘hard (with effort)’, nome ‘really’). There are also several sentence adverbs, most of which are lexicalized combinations of the auxiliary ta and one of the verbal subordinating suffixes (e.g., taitane ‘then’, takore ‘when, but, yet’, tiarone ‘however’, tane ‘then’, tiame ‘therefore’, see Section 4.5). Adverbs usually precede the predicate, as in (66), with takore, which often expresses counter-expectation vis-à-vis the discourse, and ama ‘now’. (66) takore ama diana ine nona-komuni yet now already 1sg.sbj make-neg.pot ‘Yet now I cannot make [hammocks] anymore.’ (Rybka, 2018) Subordinate verb forms with the continuative -ne also often modify the predicate (Section 4.5). The continuative is exploited to form minimal subordinate clauses of manner, as in (67), with dihisane ‘secretly’. Such subordinate forms can bear some of the verbal suffixes, such as the negative -naka, in which case they are followed by tane, the auxiliary ta, required by -naka, suffixed with the continuative (e.g., maretanaka tane ‘seriously’, from mareta ‘joke’). (67) dihisa-ne ya-nabaka-n-ae hide-cont appl-arrive-tel-pst ‘[I] met up with [her] secretly.’ (Rybka, 2018) In Venezuelan Warao, nouns form such minimal clauses of manner as well. The noun waku ‘turtle’, for instance, may be the complement of the auxiliary ta marked

1272

Andrés Romero-Figueroa and Konrad Rybka

with the continuative -ne, as in waku tane ‘slowly (i.e., like turtle)’. Some such expressions, including those with content verbs instead of the auxiliary ta, have become lexicalized (e.g., burakabane ‘slowly’, from buraka ‘two-toed sloth’ and aba ‘put’). In Guyanese Warao, speakers use the postposition monika ‘like’ in such expressions instead. Related to these forms is the non-transparent adverb tuatane ‘like this’, which stands for a previous demonstration or description of manner; accompanying an actual demonstration is simply tane, as in (68). (68) ta-ne ihi esinabu-ya aux-cont 2sg.sbj sift-prs ‘Like this you sift.’ (esinabu is an idiolectal variant of esinamu ‘sift’) (Rybka, 2018) Ideophones also commonly modify the predicate. However, in the ideophone clause (Section 3.6), the ideophone is technically subordinate to the auxiliary ta, not a modifier of the content verb. The clause can be extended to the inclusion of a content verb marked by the continuative -ne preceding the ideophone, as in (69). As such, the content verb is subordinate to the auxiliary. (69) hiwara-ne sodooow ta-n-ae pull.out-cont ideo.come.out.with.effort aux-tel-pst ‘[She] pulled it out with an effort.’ (Rybka, 2018)

4 Clause structure The discussion of clause structure focuses first on simplex clauses, addressing in particular basic word order (Section 4.1), strategies for disambiguating third-person core arguments in (di)transitive clauses (Section 4.2), and several types of copular clauses with non-verbal predicates (Section 4.3). Subsequently, three types of complex clauses are discussed, namely relative clauses, subordinate clauses, and reported speech (Sections 4.4–4.6). Complementation strategies are not covered in this chapter under a separate section. The Warao equivalents of the more common complement-taking predicates are expressed either as suffixes (e.g., desiderative, Section 3.5.6.3), as verbs which take a nominal complement modified by a relative clause (e.g., namina ‘know’, mi ‘see’, Section 4.4), or as verbs that take a whole clause as a complement, as is the case with verbs of speech (Section 4.6).

4.1 Word order As illustrated with numerous clauses above and in (70) below, the basic constituent order in Warao is OSV in transitive and SV in intransitive clauses, including copular

28 Warao

1273

clauses and polar questions. In Guyana, SOV order is also common, likely a structural influence of contact languages. Adjuncts typically appear before the OSV complex but may also appear after it or interrupt it. (70) tai a-idamo yakera-era ine mi-ae dem:anph poss-elderly good-aug 1sg.sbj see-pst ‘I saw that very good chief.’ (Romero-Figueroa, 1997) OSV word order is violated in content questions and focus constructions involving the subject of (di)transitive clauses. Since question words are clause-initial in Warao, questions about the subject have an SOV word order (Section 3.5.6.2). Bringing a constituent into focus involves fronting it, which also runs against the OSV word order, as in (71). Note that if the object is in contrastive focus and it is expressed with a pronoun, it is the subject pronouns that are used, as in (72). (71) tai atono saba yasi yak-era nisa-te dem:anph antonio ben hat good-aug take-fut ‘It is he (who) will get a better hat for Antonio.’ (Romero-Figueroa, 1985: 124) (72) ine ma-basi-n-ae 1sg.sbj 1sg.obj-bite-tel-pst ‘[The dog] bit me [not you].’ (Rybka, 2018)

4.2 Third-person argument marking As a general rule, core arguments and the indirect object are morphologically unmarked (i.e., there are no obligatory case suffixes) while most other arguments are introduced by postpositions. Since Warao predicates are intransitive and (di)transitive, theoretically there may be up to three unmarked arguments in a clause. Any number of arguments may also be left out. Both the lack of case-marking and the possibility of eliding arguments create the potential for ambiguity in transitive and ditransitive clauses. Two mechanisms serve to counteract this problem. First, firstperson and second-person singular core arguments can be unequivocally marked on the verb with object prefixes and subject suffixes (Section 3.1). Second, thirdperson arguments are disambiguated with a dedicated non-subject suffix. The suffix, which has at least two allomorphs, -ma and -si, attaches to the anaphoric demonstrative tai and nouns. The suffix is only felicitous in (di)transitive clauses, as in (73), the only scenarios when ambiguity of third-person constituents may arise. The suffix disambiguates the third-person human constituents by marking one as

1274

Andrés Romero-Figueroa and Konrad Rybka

non-subject, as is illustrated in (73); if atida ‘his wife’ appeared without the nonsubject suffix, it could be interpreted as the subject of the clause in focus. (73) tane diana a-tida-si neburatu na-e then already poss-woman-nsbj young.man kill.one-pst ‘Then the young man killed his wife.’ (Rybka, 2018)

4.3 Copular clauses with non-verbal content predicates Several types of clauses require ancillary verbs. Apart from clauses with verbal content predicates that require auxiliaries because of suffixes that block further suffixation (see Section 3.5.6.3), clauses with non-verbal predicates, which cannot carry verbal morphology, require copulas. Both the auxiliary ta and copula ha can function as copulas. The fact that the copula is a stative verb that functions as the predicate (uninflected in the present tense) and that the auxiliary is an active verb that must be inflected has repercussions for the form and function of different clause types. The copula is the default in equative, existential, possessive, and locative clauses and yields a stative predicate, usually unmarked for tense and aspect. The auxiliary ta is employed in attributive and ideophone clauses, in which the concepts encoded by attributive nouns and ideophones become part of an active predicate, which must be tense-marked. In clauses with the copula ha, the subject is followed by the nominal or postpositional complement of the clause-final copula. Equative clauses, in which the complement is a noun, differ from existential, possessive, and locative clauses in three ways. First, in equative clauses, the copula ha is often elided, as in the second equative clause of (74). Second, in negative polarity contexts, the copula is replaced by the negative particle =yana (or wana if the preceding word ends in /o/), as in (75). Third such clauses express a state with a present time-reference since they are unmarked for tense. If the clause requires tense morphology, the auxiliary ta may be used instead. If negated, ta requires the negator -naka (Section 3.5.7), which, in its turn, calls for another auxiliary ta to bear tense marking, as in (76), in which the speaker discusses the design of the house he is building. (74) ine Warao ha yatu hotarao 1sg.sbj Warao cop 2pl.sbj non.Warao ‘I am Warao and you (are) a non-Warao.’ (Rybka, 2018) (75) tamaha m-a-rani=yana dem:prox 1sg-poss-mother=neg ‘This is not my mother.’ (Rybka, 2018)

28 Warao

1275

(76) tamaha isaba-noko ta-naka ta-te dem:prox cook-place aux-neg aux-fut ‘This will not be the kitchen.’ (Rybka, 2018) Existential, possessive, and locative clauses also employ the copula in positive polarity contexts. In contrast to equative clauses, the copula is retained in other tenses and is rarely deleted. Moreover, in cases of negative polarity, the copula is replaced by the negative existential ekida. The positive and negative polarity existential clauses are shown in (77) and (78), respectively. (77) a-nibora ha a-tida ha poss-husband cop poss-woman cop ‘There was a husband and there was a wife.’ (Rybka, 2018) (78) hokohi ekida-kore temuku-bu-ya sunshine neg.exist-cond become.tangled-ints-prs ‘When there is no sunshine, [the thread] will become tangled up.’ (Rybka, 2018) The morphosyntactic rules described for existential clauses apply to possessive clauses, which have some additional features. In such clauses, the subject encodes the possessor while the nominal complement encodes the possessed, which must cross-reference the possessor with a person prefix. The subject noun phrase is, however, often elided, as in (79) and (80). (79) m-a-beroro ha 1sg-poss-dog cop ‘I have a dog.’ (Rybka, 2018) (80) a-noboto-mo=hakotai a-hiyaka ekida poss-child-coll.def=rel:def poss-clothes neg.exist ‘[There are] children who have no clothes (of their own).’ (Rybka, 2018) In locative clauses, which are also used to locate entities in time, the subject encodes the entity to be located, while the entity with respect to which it is located and the spatial relation between them are encoded by a postpositional phrase, as in (81), or a locative adverb.

1276

Andrés Romero-Figueroa and Konrad Rybka

(81) yaburu-koina noika mesi ha climb-ins.nmlz under cat cop ‘The cat is under (the bridge from the river to the house).’ (Romero-Figueroa, 1997) By contrast, attributive and ideophone clauses require the auxiliary ta (Sections 3.2.1 and 3.6). The only difference between the two is the approximative -u on attributive nouns in the former type and the presence of a pause after ideophones in the latter. It is worth drawing attention to the fact that the auxiliary is always inflected. The concepts encoded by attributive nouns and ideophones are therefore expressed as part of a predicate marked for verbal categories, as in (82). This contrasts with clauses with the copula, which are usually unmarked. (82) a-hiyaka simo-u tia poss-dress red-aprx aux.prog ‘Her dress is red.’ (i.e., ‘She is wearing a red dress now’). (Rybka, 2018) While clauses such as (82) are common, attributive nouns can avoid the temporal implications of ta by forming compounds with the head noun or functioning as a nominal predicate in equative clauses, a possibility limited to some attributive nouns (Section 3.2.1). Alternatively, both attributive and ideophone clauses can be subordinated with the continuative -ne to the copula ha, as in (83). Finally, as mentioned earlier, ideophones can also occur by themselves as single utterances. In (84), two ideophones, bohiii ‘appear (unexpectedly)’ and hoaaa ‘make hair stand on end as if the skin of your head is coming off’, each occur by themselves as single utterances.3 (83) a-hiyaka simo-u ta-ne ha poss-dress red-aprx aux-cont cop ‘Her dress continues to be red.’ (i.e., ‘Her dress has a stain that does not go away.’) (Rybka, 2018) (84) tobe bohiii jaguar ideo.appear m-a-kwa a-horo hoaaa 1sg-poss-head poss-skin ideo.hair.stand.on.end ‘[When] the jaguar suddenly appeared, my hair stood on end.’ (Barral 1979: 205) 3 Example (84), taken from Barral (1979) who marks the stressed and long vowels of ideophones rather inconsistently (using several diacritics and vowel strings of different length), was standard-

28 Warao

1277

4.4 Relative clauses Warao has three relativizers, the most common being the indefinite -ha, as in (84) and the definite -hakotai, as in (85). The former is historically equivalent to the copula, the latter, often still treated as -ha kotai, is a combination of the copula with the definite kotai. As a free form, hakotai is a definite article (Section 3.3). The relativizers may lose the /h/, fusing with the verb resulting in long vowels. The definite -hakotai may be replaced by -hametai, but the difference between the two is unclear. Warao relative clauses are internally headed. As shown in (86), in Venezuelan Warao the definite relativizer may be reduced to -kotai (in keeping with the general trend to lose hV syllables in the variety); no such process occurs in Guyanese Warao. (85) noko-kore hanoko eku diana esemoi koita-ya-ha tia=yama hear-cond house inside already music play-prs-rel:indf aux.prs=hrs ‘[…] when [they] heard music that was playing in the house, it is said.’ (Rybka, 2018) (86) warao-tuma nabaka-ya-hakotai mi-ae warao-coll arrive-prs-rel:def see-pst ‘They saw the people who were arriving.’ (Rybka, 2018) (87) ima-ya domu nari-te-kotai mi-kitane noa-kotu night-distr bird fly-fut-rel:def see-inf come-2pl.imp ‘Come see the bird that flies at night!’ (Romero-Figueroa, 1997) Instead of relative clauses, singular and plural agent nominalizations, -tu and -mo, respectively, can be used to modify a noun. Nominalizations of transitive verbs can include possessive morphology, as in (88). The nominalization strategy has, however, the semantic implication of an actor performing the activity habitually or as a profession. (88) m-a-raisa murako a-nahoro-tu nao-ya 1sg-poss-partner avocado poss-eat-ag.nmlz.sg come-prs ‘My friend who eats avocados is coming.’ (Rybka, 2018)

ized to the orthography used in this chapter and parsed in keeping with the analysis presented therein. The original reads tobe bojí; ma kua a joro joáa translated by Barral into Spanish as ‘se presenta el tigre, y los cabellos se me pusieron de punta y sentía como si se me fuera el cuero de la cabeza.’

1278

Andrés Romero-Figueroa and Konrad Rybka

4.5 Adverbial clauses Subordinate clauses of the temporal, manner, concession, purpose, and cause types are marked by suffixes, a selection of which is given below. Temporal subordinators include several suffixes. The conditional -kore covers anterior, simultaneous, posterior, and conditional clauses, depending on the context. A more specific temporal subordinator is the anterior -itane. The conjunctive -ne appears related to the anterior -itane and infinitive -kitane, and clauses in which it appears are often followed by a temporally posterior main clause, which suggests it has a subordinating function. However, given the syntactic freedom of clauses with -ne, which appear to be able to function as main clauses, the suffix is treated here as a conjunctive device that joins two main clauses and has the meaning ‘and then’. All three suffixes are illustrated in the passage in (89). (89) tai mi-tane diana atae=hese nona-e nona-ne tane dem:anph see-prior already again=exactly make-pst make-conj then e-buta-ne ta-kore diana yak-era aniako hokona caus-float.tel-conj aux-cond already good-aug early.morning dawn ha diana butu-ne cop already float-conj ‘After finding this [clay], he made [a boat] again the same way; he made it, he launched it, then it was, early in the morning, at dawn, it was floating already.’ (Rybka, 2018) The progressive -i marks simultaneous temporal clauses specifically, as in (90), which may sometimes be quite similar to manner clauses. Manner clauses are formed with the continuative -ne, discussed in Section 3.7 on adverbs, many of which are minimal subordinate clauses with the continuative. The meaning of the continuative is contrasted with that of the progressive -i in (90), where the subordinate clause with the progressive foregrounds simultaneity, while that with the continuative expresses manner (notice that when -i is suffixed to an i-final esemói it merely shifts the stress). (90) hanoko=wata nabaka-kore diana neburatu ha mi-ae ha house=loc arrive-cond already young.man rel:indf see-pst hammock eku hewere-bu-i diana esemoí diana yai-ne inside rock-ints-prog already play.music.prog already lie.down-cont ‘When [the girls] arrived home, [they] saw a man in the hammock, who was lying down while swinging and playing music.’ (Rybka, 2018) Finally, concessive clauses are marked by the suffix -rone, illustrated in (91). Purposive clauses are marked by the infinitive suffix -kitane, as in (92), though there

28 Warao

1279

are also other suffixes with similar functions (e.g., resultative -moroi). Clauses of reason or cause are in turn marked by the suffix -me, as in (93), where it is attached to the copula of a complex predicate and where a borrowed coordinator bikad (from English because) appears as well. (91) deta-e-rone diana nanaka-n-ae be.afraid-pst-conc already come.down-tel-pst ‘Although he was afraid, he came down [from the treetop].’ (Rybka, 2018) (92) kono nao-ae=yama a-raisa isiko uba-kitane crested.oropendola come-pst-hrs poss-partner with sleep-inf ‘Crested oropendola came to sleep with his friend, they say.’ (Rybka, 2018) (93) bikad ine diana yaota-komuni ha-me kuhu-ya because 1sg.sbj already work-neg.pot cop-reas walk-prs ‘Because I cannot, I walk [about visiting friends]’ (Rybka, 2018)

4.6 Reported speech There is no indirect speech in Warao, and speech events are therefore reported verbatim. The complement is presented first, followed by the verb of speech or the auxiliary ta, in keeping with the OSV basic word order, as in (94). If the quote is reported from hearsay, the hearsay enclitic =yama can be attached to the verbum dicendi, as in (94), which is an utterance from a story about the culture hero Haburi. The hearsay marker is, however, not obligatory, as in (95) from the story about the origin of the Waramuri village. The hearsay marker can attach to other verbs as well to indicate the nature of evidence for a given statement; it remains unclear what other categories of evidentiality are distinguished in Warao. (94) dani Haburi k-omi nari-ya tia=yama mother Haburi 1pl-without go-prs aux.prs=hrs ‘“Mother, Haburi (culture hero) is going without us,” (the parrot) said, they say.’ (Romero-Figueroa, 1997) (95) diko-naka eha-kotu tia tai Warao nibora=hakotai shake-neg coll.cop-2pl.imp aux.prs dem:anph Warao man=rel:def ‘“All of you don’t move,” said the man who was Warao.’ (Rybka, 2018)

1280

Andrés Romero-Figueroa and Konrad Rybka

5 Summary Warao is an understudied language isolate of Venezuela and Guyana, spoken by the traditionally riverine and seafaring indigenous people of the Orinoco delta and the surrounding area, whose material and immaterial culture is relatively well described. The language is remarkable for its morphological complexity, the frequent composite predicates, a multiplex system of tense, aspect, and telicity, omnipresent ideophones, and the typologically rare OSV basic word order. There remain, however, numerous grammatical domains that require further study, as occasionally indicated in the chapter. Little is also known about the structure of the lexicon. In particular, there is hardly any linguistic work on cross-linguistically well-studied semantic domains such space, kinship, colors, fauna, flora, or emotions.

6 Acknowledgements This chapter is the result of collaboration with the Warao speakers from Waramuri, Yakariyene, and caño Buha, whom the authors want to thank for sharing their language. The authors are also grateful to Allegra Robertson for collecting data on imperatives, to Dr. Astrid de Wit (University of Antwerp) for comments on telicity and to the editors for their feedback on the first draft. The authors also wish to express their support to the funding institutions. The research on Guyanese Warao was funded by the Netherlands Organization for Scientific Research (project number 446–15–012) and a postdoctoral grant from the Musée du Quai Branly–Jacques Chirac. The research on Venezuelan Warao was funded by Consejo de Desarrollo Científico, Humanístico y Tecnológico, and Centro de Investigación y Formación Humanística de la Universidad Católica Andrés Bello, Caracas.

7 References Barral, Basilio. 1964. Los indios guaraúnos y su cancionero (Biblioteca Missionalia Hispánica 15). Madrid: Consejo Superior de Investigaciones Científicas. Barral, Basilio. 1979. Diccionario warao-castellano, castellano-warao. Caracas: Universidad Católica Andrés Bello. Barreto, Daisy & Esteban Mosonyi. 1980. Literatura warao. Caracas: Ediciones del Consejo Nacional de la Cultura. Boomert, Arie. 2000. Trinidad, Tobago and the lower Orinoco interaction sphere. An archeological / ethnohistorical study. Alkmaar: Cairi Publications. Carlin, Eithne & Corinne Hofman. 2010. The ever-dynamic Caribbean: Exploring new approaches to unraveling social networks in the Pre-Colonial and early colonial periods. In Eithne Carlin & Simon van de Kerke (eds.), Linguistics and archaeology in the Americas: The historization of language and society (Brill’s studies in the indigenous languages of the Americas 2), 107–122. Leiden: Brill.

28 Warao

1281

Crespo, Mario, director. 2015. Dauna. Lo que lleva el río. Distributed by Asociación Civil Yakarí; Alfarería Cinematográfica; CasAmérica (2016, Spain, theatrical), Cines Unidos (2015, world-wide, all media). Dingemanse, Mark. 2011. The meaning and use of ideophones in Siwu. Nijmegen: Radboud University dissertation. Edwards, Walter F. 1980. A short dictionary of the Warau language of Guyana. Georgetown: University of Guyana & American Languages Project. Forte, Janette. 1990. The population of Guyanese Amerindian settlements in the 1980s. Timehri 46. Granberry, Julian & Gary Vescelius. 2004. Languages of the Pre-Columbian Antilles. Tuscaloosa: University of Alabama Press. Heinen, Dieter. 1973. Adaptive changes in a tribal economy: a case study of the Winikina-Warao. Ann Arbor: University of Michigan Press. Heinen, Dieter. 1982. Estructura social tradicional y mecanismos de desintegración en la sociedad warao. Acta científica venezolana 33. 419–423. Heinen, Dieter. 1988a. Los warao. In Walter Coppens (ed.), Los aborígenes de Venezuela vol. 3, 585–689. Caracas: Fundación La Salle & Monte Ávila. Heinen, Dieter. 1988b. Oko Warao: Marshland people of the Orinoco Delta. Münster: Lit Verlag. Heinen, Dieter. 1998. History, kinship and the ideology of hierarchy among the Warao of the central Orinoco Delta. Antropológica 89. 25–78. Heinen, Dieter. n. d. Warao language and culture collection of H. Dieter Heinen and Werner Wilbert, ailla.utexas.org. Access: public. PID ailla: 119510. Accessed 10 January 2018. Heinen, Dieter & Álvaro García-Castro. 2000. The multiethnic network of the Lower Orinoco in early colonial times. Ethnohistory 47 (3–4). 561–579. Heinen, Dieter, Roberto Lizarralde & Tirso Gómez. 1994. El abandono de un ecosistema: el caso de los morichales del delta del Orinoco. Antropologica 81. 3–36. Heinen, Dieter & Kenneth Ruddle. 1974. Ecology, ritual, and economic organization in the distribution of palm starch among the Warao of the Orinoco Delta. Journal of Anthropological Research 30(2). 116–138. Herrmann, Stephanie. 2018. Warao demonstratives. In Stephen Levinson, Sarah Cutfield, Michael Dunn, Nick Enfield, Sergio Meira & David Wilkins (eds.), Demonstratives in cross-Linguistic perspective, 282–302. Cambridge: Cambridge University Press. Lavandero, Julio & Dieter Heinen. 1986. Cancionero y bailes del ritual de la nouara. Montalbán 17. 199–243. Lorenzano, Antonio, Dieter Heinen, Segundo Rivero & Tirso Rivero. 2009. The Kanobo cult of the Warao Amerindians of the central Orinoco Delta: The Nahanamu sago ritual (Ethnologische Studien 41). Zürich & Piscataway, NJ: LIT Verlag. Olea, Bonifacio. 1928. Ensayo gramatical del dialecto de los indios guaraúnos. Caracas: Empresa Gutenberg. Olsen, Dale. 1974. The function of naming in the curing songs of the Warao Indians of Venezuela. Anuario interamericano de investigacion musical 10. 88–122. Olsen, Dale. 1996. Music of the Warao of Venezuela: song people of the rain forest. Gainesville: University Press of Florida. Osborn, Henry. 1959. Singular-Plural in Warao Verbs. Antropológica 6: 1–6. Osborn, Henry. 1966a. Warao I: phonology and morphophonemics. International Journal of American Linguistics 32(2). 108–123. https://doi.org/10.1086/464890. Osborn, Henry. 1966b. Warao II: Nouns, relationals, and demonstratives. International Journal of American Linguistics 32(3). 253–261. https://doi.org/10.1086/464910.. Osborn, Henry. 1967. Warao III: verbs and suffixes. International Journal of American Linguistics 33(1). 46–64. https://doi.org/10.1086/464939.

1282

Andrés Romero-Figueroa and Konrad Rybka

Romero-Figueroa, Andrés. 1985. OSV as the basic order in Warao. Lingua 66(2–3). 115–134. Romero-Figueroa, Andrés. 1997. A reference grammar of Warao. Munich: Lincom Europa. Roth, Walter Edmund. 1915. An inquiry into the animism and folk-lore of the Guiana Indians (Annual report, Smithsonian Institution Bureau of American Ethnology 30). Washington: U.S. Government Printing Office. Rybka, Konrad. 2018. Warao Archive. In The Language Archive. https://hdl.handle.net/1839/ 65ed97e4-d40c-406d-8587-0fa8818a7a34 (10 January 2018). Simons, Gary & Charles Fennig. 2018. Ethnologue: languages of the world, twenty-first edition. Dallas, Texas: SIL International. Staffeleu, P. 1975. Surinaamse zoogdiernamen. Zoologische bijdragen 18. 1–74. Vaquero, Antonio. 1965. Idioma warao: morfología, sintaxis, literatura. Caracas: Estudios Venezolanos Indígenas. Wilbert, Johannes. 1970. Folk literature of the Warao Indians. Narrative material and motif content. Los Angeles: University of California. Wilbert, Johannes. 1979. Geography and telluric lore of the Orinoco Delta. Journal of Latin American lore 51(1). 129–150. Wilbert, Johannes. 1993. Mystic endowment: Religious ethnography of the Warao Indians (Religions of the World). Cambridge, MA: Harvard University Center for the Study of World Religions.

Esteban Emilio Mosonyi and Raoul Zamponi

29 Yaruro (Pumé) 1 2 3 4 5 6 7 8 9

Overview Phonology Word classes and morphological structure Phrases Clauses Clause-linking Conclusion Acknowledgments References

1 Overview Yaruro, or Pumé (still better, pũmɛ̃́ mãɛ̃́ ‘language of the Yaruro/Pumé’; Glottocode: pume1238, ISO 639–3: yar), is spoken actively and fluently by about 9,500 people (Instituto Nacional de Estadística 2015: 31) who inhabit the central sector of the Apure Llanos of western Venezuela, mainly near the Arauca, Cunaviche, Capanaparo, and Cinaruco rivers, with the notable exception of the Capuruchano subdivision who traditionally reside at a distance from any fluvial ecosystems. The present description of some basic grammatical features of Yaruro is based on data from the variety spoken in the area of the upper Capanaparo River (Apure State).1 Yaruro has long been considered an isolate (see Gilij 1782: 205), and this is also our opinion, but some authors have proposed that it is related to different families and individual languages. Among the attempts to establish a genealogical relationship for this language, we briefly mention the following. Seler (1902: 62–64) linked Yaruro to the extinct Esmeraldeño, a presumed language isolate of the coastal region of Ecuador, based on a few lookalikes in grammatical morphemes and lexicon. Loukotka (1968: 233) also grouped Yaruro with Esmeraldeño, classifying both as “Paleo-Chibchan” languages. Kaufman (1994: 62) did the same by subsuming the two languages in a stock called “Takame-Jarúroan”. Swadesh (1962) tried to include Yaruro in a “Macro Quechua” phylum together with several other languages and families, including Zuni (western New Mexico) and Huarpean (an extinct language

1 All of the initial data gathering and the preliminary analysis of the data was done by Esteban Emilio Mosonyi. Raoul Zamponi is responsible for the revisions and the final write-up of the chapter. Esteban Emilio Mosonyi, Caracas, Venezuela Raoul Zamponi, Macerata, Italy https://doi.org/10.1515/9783110432732-016

1284

Esteban Emilio Mosonyi and Raoul Zamponi

Yaruro

29 Yaruro (Pumé)

1285

family of western Argentina). Greenberg (1987: 384) assigned the language to a “Jibaro-Kandoshi” subgroup, within a much wider “Equatorial-Tucanoan” stock, together with A’ingae (Cofán) (southwestern Colombia and northeastern Ecuador), Esmeraldeño, the Chicham (Jivaroan) family (southern Ecuador and northern Peru), and Kandozi-Chapra (northern Peru). Finally, Pache (2016) recently proposed a new relationship for Yaruro, with the Chocoan family (Pacific coast of Colombia and eastern Panama). Although dialectal differences are reported in the literature (Mosonyi 1966: 126– 127, Mosonyi, Mosonyi & García 2000: 545, Obregón Muñoz 1981: 10, 16), these differences do not appear to impinge on mutual intelligibility, nor do they affect the Yaruro/Pumé people’s perception of their own cultural and linguistic affinity. Yaruro is still an understudied language. No complete grammar of the language has been compiled so far. The most significant works include Álvarez and Guerreiro de Pirela (2015), on conditional sentences; Mosonyi (1966) on verb morphology; two basic morphological descriptions (Obregón Muñ oz & Díaz Pozo 1989; Castillo, Díaz Pozo & Obregón Muñoz 2003); two grammar sketches (Mosonyi, Mosonyi & García 2000; Krisólogo 2002); and a basic grammar (Guerreiro de Pirela 2016). Obregón Muñoz, Díaz Pozo and Pérez (1984) is a preliminary basic vocabulary. The earliest available attestations of Yaruro date back to the last quarter of the 18th century and consist of wordlists (Gilij 1782: 212; Hervás y Panduro 1787a: 163– 219; Lucena 1788), a list of numerals (Hervás y Panduro 1786: 105–106), the text of the Pater Noster (Hervás y Panduro 1787b: 109–110), and grammar notes (Hervás y Panduro n. d.). Despite the approval of an official alphabet in 1982 and the publication of teaching materials (see Rivas 2013) and, recently, colloquial and mythological texts, literacy in Yaruro is still developing. In this chapter we use phonemic transcription based on the IPA.

2 Phonology Yaruro has one of largest phonological systems in the Orinoco region. It includes 19 consonants and 15 vowels that make a total of 34 phonemes.2 The consonant system is given in Table 29.1. The voiced oral stops commonly occur before an oral vowel (e.g., ɟʝ͡ abá ‘parrot’, dadé ‘he sees (♀ addressee)’) and only rarely before a nasal vowel (e.g., ɟʝ͡ abã́ ‘now’, dãdé ‘he saw (♀ addressee)’).3 Nasal consonants, on the other hand, commonly occur

2 Only two languages of the Sáliban family have a larger phonemic inventory: Piaroa, with 38 segments, and Mako, with 35 segments (see Rosés Labrada, Volume 2). 3 The voiced affricate /ɟʝ͡ / seems to occur exclusively before an oral vowel.

1286

Esteban Emilio Mosonyi and Raoul Zamponi

Tab. 29.1: Consonant phonemes.

stop

affricate

bilabial

dental

plain

p

t

aspirated

pʰ

b

alveolar

palatal

d

velar

k

tʰ

glottal

g

kʰ

plain

cç͡

aspirated

ɟʝ͡

ʧʰ

fricative

β

nasal

m

tap

postalveolar

h n

ɲ

ŋ

ɾ

before a nasal vowel (e.g., mɔ̃ ɛ ̃́ ‘crush’ (v.), nɛ̃ ‘louse’, ɲɔ̃dé ‘he says (♀ addressee)’) and only rarely before an oral vowel (e.g., mɔɛ́ ‘sunlight’, ne ‘jet (lignite)’, ɲodé ‘he copulates (♀ addressee)’). The most remarkable phonologically conditioned alternations concerning the consonantal segments are the following. Intervocalically, stops are usually long after an oral vowel and pronounced with a (full) homorganic nasal after a nasal vowel: ɟʝ͡ ipɛ́ [ɟʝ͡ iˈpːɛ] ‘plant sp. (Dracontium sp.)’, ɔdɛ́ [ɔˈdːɛ] ‘that (nf)’, ĩbú [ĩmˈbu] ‘nose’, tãdɛ́ [tãnˈdɛ] ‘stone’, tãĩkʰiá [tãĩŋkʰiˈa] ‘walk’. The latter realization of stops is also possible across a word boundary: hũĩ ́ daβɤ́ [hũˈĩ ndaˈβɤ] ‘Look at this (♀ addressee)!’, kãɛ̃ã́ kɛ̃́ [kãɛ̃ˈã ŋkɛ̃] ‘my name’. In addition, voiced stops and the voiced affricate /ɟʝ͡ / are prenasalized in utteranceinitial position: baɤ́ [mbaˈɤ] ‘go’, do [ndo] ‘sun, day’, ɟʝ͡ a [ɲɟʝ͡ a] ‘mouth’. Affricates and nasals are usually long intervocalically, and the voiced velar stop /g/ is sometimes weakened to [ɣ] in the same position: ãnã́ [ãˈnːã] ‘big’, gɛɾɛgɛɾɛ́ [gɛɾɛɣɛˈɾɛ] ‘parakeet’. Some words, however, do not seem to permit the fricativization of intervocalic /g/: agoɾé [agoˈɾe] ‘defecate’, aɾiguɾí [aɾiguˈɾi] ‘anteater’, baguɾá [baguˈɾa] ‘run’. This phenomenon does not correlate with stress placement, but, at present, we are not able to supply further details about it. The voiced alveolar tap /ɾ/ is realized as an alveolar nasal tap between nasal vowels but not between an oral vowel and a nasal vowel or a nasal vowel and an oral vowel: ãɾɔ̃ ɾɛ̃́ [ãɾ̃ɔˈ̃ ɾ̃ɛ]̃ ‘we’, hudiɾɔ̃́ [hudiˈɾɔ̃] ‘those (nvis, nf)’, tãɾɛ́ [tãˈɾɛ] ‘hear’ (cf. tãɾã́ [tãˈɾ̃ã] ‘ear’). Finally, in spontaneous/informal speech, the voiceless aspirated bilabial stop /pʰ/ occasionally occurs as a voiceless bilabial fricative [ɸ] (e.g., pʰaɛpá [ɸaɛˈpa]

29 Yaruro (Pumé)

1287

Tab. 29.2: Vowel phonemes. oral front

close close-mid open-mid open

central

i e ɛ

back unrounded

rounded

ɯ ɤ

u o ɔ

ɯ̃

ũ ɔ̃

a nasal

close open-mid open

ĩ ɛ̃ ã

‘wind’), and the voiceless aspirated palatal affricate /ʧʰ/ may be realized as an aspirated or unaspirated dental fricative [s̪] or [s̪ʰ] (e.g., ʧʰĩ [s̪ʰĩ] ‘heart’). A non-phonemic glottal stop occurs in most dialects before vowel-initial words, including word-initial words in compounds (or lexicalized noun phrases): i [ʔi] ‘skin’, adó [ʔaˈdo] ‘again’, ŋɔ̃ ɛ-̃ í [ŋɔ̃ɛˈ̃ ʔi] ‘paper (Lit. ‘mark-skin’)’. Occasionally, it can also be heard in intervocalic position within unanalyzable words: páa [ˈpaʔa] ‘yes’. The vowel system, shown in Table 29.2, consists of nine oral vowels and six nasal vowels. The only allowed syllable shapes are V, CV, and CVV: u.í ‘year’, ɔ.ã.ĩ ́ ‘man’, a.ta.ɾa.pá ‘count’ (v.), ha.dɔ.cç͡ ĩ.mĩ ́ ‘four’, á.mãĩ ‘father’. Stress is phonologically unpredictable. Although primary stress usually falls on the final syllable of a word, there are some words with the primary stress on the penultimate syllable: ámãĩ ‘father’, cç͡ ɛrámɛ̃ ‘yam’, ɔkaɾáɾo ‘hen’, ɔtɛ́ bo ‘all’. Unstressed vowels and non-final primary-stressed vowels are usually short. Monosyllabic CV roots in words including a stressed suffix receive secondary stress, and their vowel is lengthened: da-dé [ˌdaːˈdeˑ] ‘he sees (♀ addressee)’. Word-final primary-stressed vowels are normally half-long. Almost all suffixes carry lexical stress: either primary (in final position) or secondary (in non-final position).

3 Word classes and morphological structure The following word classes can be identified in Yaruro: nouns, pronouns, adjectives, verbs, adverbs, postpositions, conjunctions, and interjections. We have no data suggesting the presence of ideophones.

1288

Esteban Emilio Mosonyi and Raoul Zamponi

3.1 Nouns Nouns in Yaruro are characterized by the categories of number and gender. Both categories are discernible via the agreement of nouns with personal pronouns, demonstrative pronouns (except the plural oblique ones), and verbs; see (1) and (2). (1)

ɔãĩ ́ hudí haɾa-dí man dem.nvis.nf.nom drink-3sg.nf.sbj.♀addr ‘This man drinks (♀ addressee).’

(2)

kɔdɛ́ da-ɾekɛ̃hɛ̃́ iãĩ ́ hĩ 1sg.pro.nom see-1sg.a.3sg.f.o.♂addr woman dem.nvis.sg.f.obl ‘I see the woman (♂ addressee).’

The numbers distinguished are singular and plural. Only very few animate nouns have a plural form, marked by an infix , shown in (3). (3)

a. ɔãĩ ́ b. ũĩ ́

‘man’ ‘son’

c. ɔaí d. ũĩ ́

‘men’ ‘sons’

The term for ‘woman’ has a fully suppletive plural form, shown in (4b). (4)

a. iãĩ ́

‘woman’

b. iβí

‘women’

Yaruro distinguishes feminine and non-feminine (or masculine) gender. For animates, gender relates to sex distinctions. Nouns with inanimate referents (including abstract nouns) are non-feminine. Non-feminine, functionally, is therefore the unmarked gender in Yaruro. Two of the three available agentive nominalizing suffixes show a gender distinction: -ɲĩ,́ feminine singular (in (5)), and -mɛ̃,́ non-feminine singular (in (6)). (5)

ŋɔ̃ a-dɛ-ɲĩ ́ have-neg-nmlz.sg.f ‘poor woman’

(6)

cç͡ ɔ̃ mũĩ-ɾɛ̃-mɛ̃́ fish fish-hab-nmlz.sg.nf ‘fisherman’

The third agentive nominalizer, -hĩɾĩ,́ shown in (7), expresses plural number and is gender-neutral.

29 Yaruro (Pumé)

(7)

1289

daba-hĩɾĩ ́ know-nmlz.pl ‘wise people’

Aside from with verbs, as in (5)‒(7), the three agentive nominalizers are also used with adjectives, numerals, and adverbials, as in (8)–(10), respectively. (8)

ʧʰakʰia-ɲĩ-nɛ̃́ iãĩ=ɾɛ̃́ ĩnɛ̃́ dem.dist.f.nom woman=nom beautiful-nmlz.sg.f-3sg.f.sbj.♂addr ‘That woman is (one who is) beautiful (♂ addressee).’

(9)

ɲɔ̃ ãɾĩ-mɛ̃́ two-nmlz.sg.nf ‘the second one (nf)’

(10) ʧʰia=ɾi-hĩɾĩ ́ west=abl-nmlz.pl ‘people of the west’ (ʧʰiá ‘tail; west’) The same suffixes -ɲĩ,́ -mɛ̃,́ and -hĩɾĩ ́ also function as markers of gender and/or number with some human nouns, for example: o-ɲĩ ́ ‘sister, female cousin; sibling/cousin-sg.f’, o-mɛ̃́ ‘brother, male cousin; sibling/cousin-sg.nf’, o-hĩɾĩ ́ ‘sisters, brothers, cousins; sibling/cousin-pl’. Note that an unstressed variant -mãĩ of the non-feminine singular suffix normally occurs in this context, as in kɛɾá-mãĩ ‘brother-in-law’ and hadé-mãĩ ‘mother’s brother’. Other nominal derivational suffixes in Yaruro that we are aware of include a non-agentive (basically instrument/object/place) nominalizer -ɾɛã́ (after an oral vowel) ~ -ɾɛ̃ã́ (after a nasal vowel), as in (11); a homophonous (class-maintaining) deceased morpheme, as in (12); augmentative -tĩ,́ as in (13); diminutive -ĩ ́ and -bũĩ ́ (cf. buíʧʰĩ ‘small, little’), as in (14) and (15), respectively; and depreciative -kotĩ,́ as in (16). None of these suffixes interact with gender. ‘paint’ (v.) ‘sew’

b. hɔa-ɾɛã́ ‘paint’ (n.) d. kaɾa-ɾɛã́ ‘needle’

e. kɛ̃nã́ g. huɾá

‘think, believe’ ‘eat’

f. kɛ̃nã-ɾɛ̃ã́ ‘religion’ h. huɾa-ɾɛã́ ‘food, meal’

i. mɔ̃ ã́ k. cç͡ a

‘sleep’ ‘sit’

j. mɔ̃ ã-ɾɛ̃ã́ l. cç͡ a-ɾɛã́

(11) a. hɔá c. kaɾá

(12) a. á-mãĩ ‘father’ c. kɛɾá-mãĩ ‘brother-in-law’

‘bed’ ‘chair’

b. a-mãĩ-ɾɛ̃ã́ ‘deceased father’ d. kɛɾa-mãĩ-ɾɛ̃ã́ ‘deceased brother-in-law’

1290

Esteban Emilio Mosonyi and Raoul Zamponi

(13) a. ʧʰidó c. tʰɔbɛ́

‘capybara’ ‘head’

b. ʧʰido-tĩ ́ d. tʰɔbɛ-tĩ ́

‘big capybara’ ‘big head’

(14) a. ɔtɛ́ c. to

‘chief’ ‘tree’

b. ɔtɛ-ĩ ́ d. to-ĩ ́

‘little chief’ ‘little tree’

(15) a. pũmɛ̃́ c. ɛ̃ké

‘Yaruro’ ‘crab’

b. pũmɛ̃-bũĩ ́ d. ɛ̃ke-bũĩ ́

‘little Yaruro’ ‘little crab’

(16) a. cç͡ aɾá c. hɔ̃

‘canoe’ ‘house’

b. cç͡ aɾa-kotĩ ́ d. hɔ̃ -kotĩ ́

‘beat-up canoe’ ‘bad house’

3.2 Pronouns Personal pronouns distinguish the categories of person (first and second), number (singular and plural), and case (nominative and oblique); see Table 29.3. Oblique personal pronouns occur as objects, as in (127) and (135); extensionto-core arguments, as in (57) and (59); and are the personal pronouns used with postpositions, as in (70) and (71). Note in Table 29.3 that the second-person plural nominative form directly derives from the corresponding singular form by the addition of a plural suffix -ɾɔ̃,́ while the remaining plural forms are idiosyncratic and show no relationship to the corresponding singular forms. Since predicates and (usually) complements of non-verbal clauses are marked for subject person and number, personal pronouns are not necessary as subjects, as shown in (17). (17) baɤ-ɾekodé go-1sg.sbj.♂addr ‘I go (♂ addressee).’

Tab. 29.3: Personal pronouns. person

number

case

form

1

sg

nom obl

kɔdɛ́ kɔá

pl

nom obl

ãɾɔ̃ɾɛ̃́ ibɛ́

sg

nom obl

mɛ̃nɛ̃́ mɛ̃ã ́

pl

nom obl

mɛ̃nɛ̃ɾɔ̃́ dibɛ́

2

29 Yaruro (Pumé)

1291

Tab. 29.4: Possessive pronouns. singular 1 2 3

plural

kãɛ̃ã ́ ~ kã ibɛ̃ã ́ nãɛ̃ã ́ ~ nã dibɛ̃ã ́ hibɛ̃ã ́ ~ hɛbɛ̃ã ́

In information structurally-marked constructions, personal pronouns can, however, be doubled and placed both before and after the predicate of a verbal clause or, as in (18), the complement of a non-verbal clause. (18) kɔdɛ́ piɟʝ͡ i-ɾekodé kɔdɛ́ 1sg.pro.nom here-1sg.sbj.♂addr 1sg.pro.nom ‘I am here (♂ addressee).’ Possessive pronouns distinguish singular and plural forms in first and second person, but not in third person, as evident in Table 29.4. Possessive pronouns can be used as modifiers preceding a nominal head as well as nominal heads of their own, independently of a possessum, as in (19) and (20). (19) kãɛ̃ã́ hadɔ́ hudí 1sg.poss friend dem.nvis.nf.nom ‘my friend’ (Lit. ‘the my friend’) (20) kãɛ̃ã́ hũĩ ́ 1sg.poss dem.nvis.sg.nf.obl ‘mine’ (Lit. ‘the mine’) Demonstrative pronouns, given in Table 29.5, categorize their referent or the referent of the noun to which they refer to with respect to its accessibility in space. Two visual demonstrative pronouns are used when the referent is visible to the speaker at a greater or lesser distance (distal vs. proximal). The non-visual demonstrative pronoun is used when the referent is outside the speaker’s visual field. All three demonstrative pronouns have distinct forms for number (singular and plural); case (nominative, non-nominative, and oblique); and, except in oblique plural, gender (feminine and non-feminine). As shown in Table 29.5, the nominative plural forms of demonstrative pronouns are directly derived from their singular counterparts by the addition of the same plural suffix -ɾɔ̃́ found with the second-person plural pronoun in the nominative.

1292

Esteban Emilio Mosonyi and Raoul Zamponi

Tab. 29.5: Demonstrative pronouns. case

nom

obl

number

gender

visual

non-visual

proximal

distal

sg

f nf

ɲĩnɛ̃́ ɟʝ͡ ɔdɛ́

ĩnɛ̃́ ɔdɛ́

hĩnĩ ́ hudí

pl

f nf

ɲĩnɛ̃ ɾɔ̃́ ɟʝ͡ ɔdɛɾɔ̃́

ĩnɛ̃ ɾɔ̃́ ɔdɛɾɔ̃́

hĩnĩɾɔ̃́ hudiɾɔ̃́

sg

f nf

ɲĩã ́ ɟʝ͡ ɔá

ĩã ́ ɔá

hĩ hũĩ ́

pl

–

ɲĩɾã ́

ĩɾã ́

hĩɾĩ ́

The oblique plural forms look like the corresponding feminine singular ones pluralized by the infix (Section 3.1).4 Interestingly, demonstrative pronouns have three different uses in Yaruro: (i) as third-person pronouns, as in (21) and (22); (ii) as demonstrative determiners, shown in (23) and (24); and (iii) as markers of definiteness (definite articles), as in (25) and (26). When used as demonstrative determiners, demonstrative pronouns precede the head; when used as definiteness markers, they follow the head noun. (21) ɔdɛ́ hãnã-dé dem.dist.nf.nom come-3sg.nf.sbj.♂addr ‘He comes (♂ addressee).’ (22) hĩnĩ-ɾṍ nĩbɤ-nĩɾɛ̃́ dem.nvis.f.nom-pl speak-3pl.f.sbj.♂addr ‘They (f) speak (♂ addressee).’ (23) ɟʝ͡ ɔdɛ́ nĩβɛ́ dem.prox.nf.nom non-indigenous.person ‘this non-indigenous person’ ɟʝ͡ ae (24) ɔdɛ-ɾṍ dem.dist.nf.nom-pl horse ‘those horses’

4 hĩ seems to be a contraction of an ancient *hĩĩ ́ (which differs from its non-feminine counterpart hũĩ ́ for having a front vowel instead of a back vowel, like all the other forms of Yaruro demonstrative pronouns).

29 Yaruro (Pumé)

1293

(25) bɤɾé ɟʝ͡ ɔdɛ́ hammock dem.prox.nf.nom ‘the hammock (near the speaker)’ (26) bɤɾé ɔdɛ́ hammock dem.dist.nf.nom ‘the hammock (far from the speaker)’ The non-visual demonstrative pronoun is also used anaphorically, as in (27). (27) ɟʝ͡ abã́ buíʧʰĩ ɔdɛ-ɾṍ mɔ̃ ã-mɔ̃ ã́ hɤɾɤtɤ́ hãdí taɾá now child dem.dist.nf.nom-pl sleep-sleep tired in.this.way because hudi-ɾɔ̃́ ʧʰaɔkʰia-diɾɛ̃́ dem.nvis.nf.nom-pl play-3pl.nf.sbj.♂addr ‘As the children are tired of sleeping, they are playing (♂ addressee).’ Its nominative forms are normally reduced and encliticized when it occurs in postnominal position; the full forms are information structurally-marked. hĩnĩ ́ or pũmɛ̃-ñĩ=(h)nĩ ́ (28) pũmɛ̃-́ ñĩ Yaruro-sg.f dem.nvis.f.nom ‘the Yaruro woman’ (29) nĩβɛ́ hudí or nĩβɛ=(h)dí non-indigenous.person dem.nvis.nf.nom ‘the non-indigenous man’ (30) aboéa hĩnĩ-ɾɔ̃́ or aboea=nɔ̃́ pig dem.nvis.f.nom-pl ‘the sows’ or aboea=dɾɔ̃́ (31) aboéa hudi-ɾɔ̃́ pig dem.nvis.nf.nom-pl ‘the pigs’ The non-feminine oblique form of the non-visual demonstrative pronoun hũĩ ́ has an optional postnominal variant form =hṼ́ with a (nasal) vowel that assimilates to the features [back], [round], [high], and [low] of the preceding vowel.5

5 Ergo, the vowel of this variant form is ɛ̃ after e, ɛ, and ɛ̃, ɔ̃ after ɔ, ɔ̃ , and o, and ɯ̃ after ɤ, ɯ, and ɯ̃. Recall from Table 29.2 that there are no close-mid nasal vowels in Yaruro.

1294

Esteban Emilio Mosonyi and Raoul Zamponi

(32) cç͡ aɾá hũĩ ́ or cç͡ aɾa=hã́ canoe dem.nvis.sg.nf.obl ‘the canoe’ (33) edé hũĩ ́ or ede=hɛ̃́ pumpkin dem.nvis.sg.nf.obl ‘the pumpkin’

3.3 Adjectives Adjectives are words denoting properties or attributes that functionally modify nouns in noun phrases (uninflected) and can fill the complement slot (without an accompanying free noun) in non-verbal clauses (usually inflected for subject person-number). They constitute a large class of forms belonging to – as far as we could observe – the vast majority of the semantic types listed in Dixon (2004: 3–5), for example: dimension (anã́ ‘big, large, wide’, buíʧʰĩ ‘small’), age (bo ‘young’), value (ʧʰadɛ́ ‘good’), color (kãɾã́ ‘black’), physical property (ʧʰuá ‘hard’), speed (hɔ̃ ɛɾ̃ ã́ ‘swift, fast, quick’), difficulty (badɛɾĩ ́ ‘easy’), qualification (tʰamɔ̃́ ‘true’), and quantification (hĩɾã́ ‘many’). The presence of reduplicated forms among color and physical property adjectives is noteworthy; see, for example, bɛɾɛbɛɾɛá ‘light, white’, buɾɛbuɾɛá ‘dark, black’, tɔtɔá ‘blue’, ʧʰuʧʰuʧʰuʧʰuá ‘transparent’, boboá ‘soft’, tatá ‘smooth’, cç͡ acç͡ á ‘heavy’, and kɯakɯá ‘warm’.

3.4 Numerals Numerals constitute a separate class of modifiers in Yaruro. While adjectives (including those expressing quantification) follow the head noun they modify (e.g., cç͡ ɛɾɛ́ hĩɾã́ ‘a lot of money’), numerals precede it (Section 4.1). Only the numerals ‘one’ to ‘five’ are single phonological words. (34) a. b. c. d. e.

́ ɛ̃ kʰãɾɛ̃m ɲɔ̃ ãɾĩ ́ tʰaɾãɾĩ ́

‘one’ ‘two’ ‘three’ hadɔcç͡ ĩmĩ ́ ‘four’ kʰãɾicç͡ ímɔ̃ ‘five’

For all numerals, a synchronic or diachronic analysis is possible. The numeral ‘one’ appears to include a root kʰãɾɛ̃́ also found in the adjective kʰãɾɛ̃mɔ̃́ ‘other’ and the (non-feminine singular) agentive nominalizer -mɛ̃́ (usually stressed but unstressed ́ ɛ̃). The numeral ‘two’ seems to be from the verb ɲo ‘copulate’ suffixed by in kʰãɾɛ̃m perfective past -ã́ (Section 3.5.2.1) and the adverbializer -ɾĩ.́ The numeral ‘three’ seems

29 Yaruro (Pumé)

1295

to be from the noun ʧʰaɾá ‘twin’ also suffixed by -ɾĩ.́ ‘Four’ is derived from the expression ‘companion to the hand’ (presumably for ‘companion to the fingers’): hadɔ́ ‘friend, companion’, icç͡ í ‘hand’, =mĩ ́ benefactive. The numeral ‘five’ is a complex item in which we can identify the numeral ‘one’, the noun icç͡ í ‘hand’, and the enclitic =mɔ̃ ‘alone, by oneself’. The numerals ‘ten’ and ‘twenty’ are lexicalized noun phrases: icç͡ í ʧʰɯ̃nĩ ́ ‘ten (Lit. ‘hand complete’)’, ta ʧʰɯ̃nĩ ́ ‘twenty (Lit. ‘foot complete’)’. Numerals ‘six’ to ‘nine’ are additive base-five formations composed of kʰãɾicç͡ í ‘five’ and the allative postposition =ɾɤpɛ̃́ followed by ‘one’, ‘two’, ‘three’, or ‘four’, as set out in (35). (35) a. b. c. d.

́ ɛ̃ kʰãɾicç͡ i=ɾɤpɛ̃́ kʰãɾɛ̃m kʰãɾicç͡ i=ɾɤpɛ̃́ ɲɔ̃ ãɾĩ ́ kʰãɾicç͡ i=ɾɤpɛ̃́ tʰaɾãɾ̃ĩ ́ kʰãɾicç͡ i=ɾɤpɛ̃́ hadɔcç͡ ĩmĩ ́

‘six’ ‘seven’ ‘eight’ ‘nine’

3.5 Verbs A distinction between finite and non-finite verb forms can be established based on the inflectional features that the verb require. Finite forms express person and number of the subject, while the non-finite forms are characterized by the absence of subject-marking affixes. Finite verbs are also inflected for various TAME categories, discussed in Section 3.5.2, and morphologically marked for negation (see Section 5.4). Non-finite verbs typically appear in subordinate clauses and may occur as bare stems (but see the exception mentioned in Section 6.2.2 concerning the nonfinite predicates in subordinate reason clauses).

3.5.1 Subject and object agreement Finite Yaruro verbs, as well as bearing subject agreement, bear object agreement if they have an animate object; this marking is obligatory if the object argument is human and optional if it is non-human. The personal markers used with non-imperative verbs (Section 3.5.1.1) are different than those used with imperative verbs (Section 3.5.1.2). Non-finite verbs without subject-marking affixes are sometimes also used as predicates in non-imperative independent clauses with an overtly expressed subject, as in (36) and (37). (36) mɛ̃nɛ̃́ ʧʰaɔkʰiá 2sg.pro.nom play ‘You (sg) play.’

1296

Esteban Emilio Mosonyi and Raoul Zamponi

(37) hakó baɤ-pá ɔtɛ=(h)dí tomorrow go.away-nr.fut chief=dem.nvis.nf.nom ‘Tomorrow the chief will go away.’

3.5.1.1 Non-imperative verbs Intransitive verbs and transitive verbs used intransitively mark subject person and, with a third-person subject, gender of the subject (S), by means of suffixes that also signal the gender of the addressee: female (♀) or male (♂). (38) baguɾa-kɛ́ run-1sg.sbj.♀addr ‘I run (♀ addressee).’ (39) da-ɾenɛ̃́ see-2sg.sbj.♂addr ‘You (sg) see (♂ addressee).’ (40) iãĩ ́ hĩnĩ ́ haɾa-nĩ ́ woman dem.nvis.f.nom drink-3sg.f.sbj.♀addr ‘The woman drinks (♀ addressee).’ (41) ɔãĩ ́ hudí haɾa-dé man dem.nvis.nf.nom drink-3sg.nf.sbj.♂addr ‘The man drinks (♂ addressee).’ The same suffixes also mark the subject of a transitive verb (A) when its object (O) argument is inanimate, as shown in (42) and (43). (42) kɔdɛ́ haɾa-ɾekodé ɛɾɔ́ ʧʰadɛ=á 1sg.pro.nom drink-1sg.sbj.♂addr carato good=objv ‘I drink a good carato drink (♂ addressee).’ (43) mɛ̃nɛ̃́ bɛ hũĩ ́ gatɛ-ã-ɾenɛ̃́ 2sg.pro.nom house dem.nvis.sg.nf.obl buy-pst.pfv-2sg.sbj.♂addr ‘You (sg) bought the house (♂ addressee).’ These suffixes are optionally employed to mark the A argument of a transitive verb when its O argument is animate but non-human. (44) aboéa ŋũnã-ɾekodé pig kill-1sg.sbj.♂addr ‘I kill a pig (♂ addressee).’

29 Yaruro (Pumé)

1297

Tab. 29.6: Subject agreement markers of non-imperative verbs (common allomorphs). subject 1

–

number

addressee

form

sg

♀ ♂ ♀ ♂

-kɛ́ -ɾekodé -ɾɔ̃́ -ɾɛɾɛ̃́

♀ ♂ ♀ ♂

-mɛ̃́ -ɾenɛ̃́ -nĩɾɔ̃́ -ɾenĩɾɔ̃́

♀ ♂ ♀ ♂ ♀ ♂ ♀ ♂

-nĩ ́ -nɛ̃́ -nĩɾɔ̃́ -nĩɾɛ̃́ -dí -dé -diɾɔ̃́

pl

2

–

sg pl

3

f

sg pl

nf

sg pl

-diɾɛ̃́

(45) cç͡ ɔ̃ a-dé fish throw.an.arrow.to-3sg.nf.sbj.♂addr ‘He throws an arrow to a fish (♂ addressee).’ The complete set of the subject agreement suffixes used with non-imperative verbs is shown in Table 29.6. Note the interaction of grammatical gender with indexical addressee’s gender in the third person. Transitive and extended transitive non-imperative verbs cross-reference their A and animate O or extension-to-core (E) arguments with portmanteau suffixes that also signal the addressee’s gender. (46) ɔãĩ ́ hũĩ ́ da-ɾekɔ̃ nɛ̃́ kɔdɛ́ man dem.nvis.sg.nf.obl see-1sg.nf.a.3sg.nf.o.♂addr 1sg.pro.nom ‘I see the man (♂ addressee).’ da-ɾekɔ̃ nɛ̃́ puá ɔá (47) hacç͡ í far.away see-1sg.nf.a.3sg.nf.o.♂addr bird dem.dist.sg.nf.obl ‘I see that bird from a distance (♂ addressee).’ tãɾɛ-diɾɔ̃ koé (48) hudi-ɾɔ̃́ dem. nvis.nf.nom-pl hear-3pl.nf.a.1sg.o.♂addr ‘They hear me (♂ addressee).’

1298

Esteban Emilio Mosonyi and Raoul Zamponi

(49) pʰuɛcç͡ ó ɟʝ͡ ɔɾɔ-dɛ-ã-nĩɾɔ̃ koá mɛ̃nɛ̃-ɾɔ̃́ maize give-neg-pst.pfv-2pl.a.1sg.e.♀addr 2sg.pro.nom-pl ‘You (pl) did not give me maize (♀ addressee).’ (50) ɔdɛ́ ɲɔ̃ -dɛ̃hɛ̃ɾɛ̃́ ɔaí hĩɾĩ ́ dem.dist.nf.nom say-3sg.nf.a.3pl.e.♂addr man dem.nvis.pl.obl ́ ɛ̃ mãɛ̃́ kʰãɾɛ̃m one word ‘He says one word to the men (♂ addressee).’ Specifically, the suffixes in question supply the following information: (i) A person, number, and, for third person, grammatical gender; (ii) O or E person, number, and, for a third-person singular O or E, if the A is not third-person singular, grammatical gender; and (iii) gender of the addressee (invariably) (see Table 29.7). When O is possessed, a transitive verb must also mark the person and number of the possessor of the O argument. This, with non-imperative verbs, is done by means of a portmanteau form that expresses: (i) person, number, and, if third person, grammatical gender of the A argument; (ii) number and grammatical gender of the O argument, but only if the A argument is first-person singular and the O argument is (third-person) singular; (iii) person and number of the possessor of the O argument, distinguishing a third-person possessor that is coreferential with the A of the verb from a third-person possessor that is not; and (iv) indexical gender of the addressee. (51) ɛa-dɛ-ɾekĩhĩkɛ̃́ kãɛ̃ã́ iãĩ ́ love-neg-1sg.a.3sg.f.o.1sg.possr.♂addr 1sg.poss woman hĩ dem.nvis.sg.f.obl ‘I do not love my woman (♂ addressee).’ (52) kɔdɛ́ gaɾa-ɾekãnãnãɛ̃́ nãɛ̃ã́ aboéa 1sg.pro.nom steal-1sg.a.3sg.nf.o.2sg.possr.♂addr 2sg.poss pig hũĩ ́ dem.nvis.nf.obl ‘I steal your (sg) pig (♂ addressee).’ (53) kãɛ̃ã́ hɔ̃ hũĩ ́ bɛɾɛbɛɾɛa=ɾã́ 1sg.poss house dem.nvis.nf.obl white=ins hɔa-ɾekũnĩkɛ̃́ paint-1sg.a.3sg.nf.o.1sg.possr.♂addr ‘I paint my house white (♂ addressee).’

3

2

nf

f

pl

sg

pl

sg

pl

sg

pl

♀ ♂ ♀ ♂

♀ ♂ ♀ ♂

♀ ♂ ♀ ♂

♀ ♂ ♀ ♂

1

sg

addr.

agent

-dikoá -dikoé -diɾɔ̃koá -diɾɔ̃koé

-nĩkoá -nĩkoé -nĩɾɔ̃koá -nĩɾɔ̃koé -diβɛ́ -diβé -diɾɔ̃βɛ́ -diɾɔ̃βé

-nĩβɛ́ -nĩβé -nĩɾɔ̃ɾĩβɛ́ -nĩɾɔ̃ɾĩβé

-mĩβɛ́ -ɾemĩβé -nĩɾɔ̃βɛ̃́ -ɾenĩɾɔ̃βé -nĩɾiβɛ́ -nĩɾiβé -nĩɾɔ̃ɾĩβɛ́ -nĩɾɔ̃ɾĩβé -diɾiβɛ́ -diɾiβé -diɾɔ̃ɾĩβɛ́ -diɾɔ̃ɾĩβé

-nĩmã ́ -nĩmɛ̃́ -nĩɾɔ̃mã ́ -nĩɾɔ̃mɛ̃́ -dimã ́ -dimɛ̃́ -diɾɔ̃mã ́ -diɾɔ̃mɛ̃́

-kiɾiβɛ́ -ɾekiɾiɾiβé -ɾɔ̃ɾiβɛ́ -ɾeɾɔ̃ɾiβé

-kãmã ́ -ɾekodemɛ̃́ -ɾɔ̃mã ́ -ɾeɾɔ̃mɛ̃́

-mãkoá -ɾenĩkoé -nĩɾɔ̃koá -ɾenĩɾɔ̃koé

plural

singular

singular

plural

second

first

object/extension to core

Tab. 29.7: A-O/E agreement markers of non-imperative verbs (common allomorphs).

-dĩhĩ ́ -dɛ̃hɛ̃́ -diɾɔ̃hĩ ́ -diɾɔ̃hɛ̃́

-dĩhĩ ́ -dɛ̃hɛ̃́ -diɾɔ̃hũĩ ́ -diɾɔ̃hɔ̃ɛ ̃́

-nĩhĩ ́ -nɛ̃hɛ̃́ -nĩɾɔ̃hũĩ ́ -nĩɾɔ̃hɔ̃ɛ ̃́

-nĩ ́ -ɾenɛ̃́ -nĩɾɔ̃hũĩ ́ -ɾenĩɾɔ̃hɔ̃ɛ ̃́

-mĩhĩ ́ -mɛ̃hɛ̃́ -nĩɾɔ̃hɛ̃́ -ɾenĩɾɔ̃hɛ̃́ -nĩhĩ ́ -nɛ̃hɛ̃́ -nĩɾɔ̃hĩ ́ -nĩɾɔ̃hɛ̃́

-kũnĩ ́ -ɾekɔ̃nɛ̃́ -ɾɔ̃hũĩ ́ -ɾeɾɔ̃hɔ̃ɛ ̃́

nf

-kĩhĩ ́ -ɾekɛ̃hɛ̃́ -ɾɔ̃hĩ ́ -ɾeɾɔ̃hɛ̃́

f

singular

third

-dĩhĩɾĩ ́ -dɛ̃hɛ̃ɾɛ̃́ -diɾɔ̃hĩɾĩ ́ -diɾɔ̃hɛ̃ɾɛ̃́

-nĩhĩɾĩ ́ -nɛ̃hɛ̃ɾɛ̃́ -nĩɾɔ̃hĩɾĩ ́ -nĩɾɔ̃hɛ̃ɾɛ̃́

-ɾenĩɾɔ̃hɛ̃ɾɛ̃́

-nĩhĩɾĩ ́ -ɾenɛ̃hɛ̃ɾɛ̃́ -nĩɾɔ̃hĩɾĩ ́

-kĩhĩɾĩ ́ -ɾekɛ̃hɛ̃ɾɛ̃́ -ɾɔ̃hĩɾĩ ́ -ɾeɾɔ̃hɛ̃ɾɛ̃́

plural

29 Yaruro (Pumé)

1299

3

2

nf

f

♀ ♂

♀ ♂

pl

♀ ♂

pl

sg

♀ ♂

♀ ♂ ♀ ♂

♀ ♂

sg

pl

sg

pl

♂

♀

1

sg

addr.

agent

3

— —

— —

— —

— —

— — — —

— —

sg — pl sg — pl

object

— —

— —

— —

— —

— — — —

— —

f nf — f nf —

singular

plural

-kĩhĩbɛ̃ã ́ -kũnĩbɛ̃ã ́ -kĩhĩɾĩbɛ̃ã ́ -ɾekĩhĩbɛ̃́ -ɾekũnĩbɛ̃́ -ɾekĩhĩɾĩbɛ̃́ -ɾɔ̃bɛ̃ã ́ -ɾɔ̃bɛ̃́ -nĩbɛ̃ã ́ -ɾenĩbɛ̃́ -nĩɾɔ̃bɛ̃ã ́ -ɾenĩɾɔ̃bɛ̃́ -nĩbɛ̃ã ́ -nĩbɛ̃́ -nĩɾɔ̃bɛ̃ã ́ -nĩɾɔ̃bɛ̃́ -dibɛ̃ã ́ -dibɛ̃́ -diɾɔ̃bɛ̃ã ́ -diɾɔ̃bɛ̃́

singular

-kĩhĩkã ́ -kũnĩkã ́ -kĩhĩɾĩã ́ -ɾekĩhĩkɛ̃́ -ɾekũnĩkɛ̃́ -ɾekĩhĩɾĩkɛ̃́ -ɾɔ̃kã ́ -ɾɔ̃kɛ̃́ -nĩkã ́ -ɾenĩkɛ̃́ -nĩɾɔ̃kã ́ -ɾenĩɾɔ̃kɛ̃́ -nĩkã ́ -nĩkɛ̃́ -nĩɾɔ̃kã ́ -nĩɾɔ̃kɛ̃́ -dikã ́ -dikɛ̃́ -diɾɔ̃kã ́ -diɾɔ̃kɛ̃́

-diɾɔ̃nã ́ -diɾɔ̃nãɛ̃́

-dinã ́ -dinɛ̃́

-nĩɾɔ̃nã ́ -nĩɾɔ̃nãɛ̃́

-nĩnã ́ -nĩnɛ̃́

-nãnã ́ -ɾenãnãɛ̃́ -nĩɾɔ̃nã ́ -ɾenĩɾɔ̃nãɛ̃́

-ɾɔ̃nãnã ́ -ɾɔ̃nãnãɛ̃́

↑ -kãnãnã ́ ↓ ↑ -ɾekãnãnãɛ̃́ ↓

2

1

possessor

-diɾɔ̃diɾibɛ̃ã ́ -diɾɔ̃diɾibɛ̃́

-didiɾibɛ̃ã ́ -didiɾibɛ̃́

-nĩɾɔ̃ɾĩbɛ̃ã ́ -nĩɾɔ̃ɾĩbɛ̃́

-nĩɾĩbɛ̃ã ́ -nĩɾĩbɛ̃́

-nĩɾĩbɛ̃ã ́ -ɾenĩɾĩbɛ̃́ -nĩɾɔ̃nĩbɛ̃ã ́ -ɾenĩɾɔ̃nĩbɛ̃́

-ɾɔ̃nãnã ́ -ɾɔ̃nãnãɛ̃́

↑ -kãnãnã ́ ↓ ↑ -ɾekãnãnãɛ̃́ ↓

plural

Tab.29.8: A-O-POSSR agreement markers of non-imperative verbs (common allomorphs).

-dabã ́ -dabãɛ̃́

-nãmã ́ -nãmãɛ̃́

-dĩɾɔ̃hũĩhã ́ -dĩɾɔ̃hũĩhɛ̃́

-dihã ́ -dihɛ̃́

-nĩɾɔ̃hũĩhã ́ -nĩɾɔ̃hũĩhɛ̃́

-nĩhã ́ -nĩhɛ̃́

-nĩhĩhã ́ -ɾenĩhĩhɛ̃́ -nĩɾɔ̃hũĩhã ́ -ɾenĩɾɔ̃hũĩhɛ̃́

-ɾɔ̃hã ́ -ɾɔ̃hãɛ̃́

↑ -kãhãhã ́ ↓ ↑ -kãhãhãɛ̃́ ↓

-dababã ́ -dababãɛ̃́

-nãmãmã ́ -nãmãmãɛ̃́

coref

coref

non-core

plural

singular

3

-nĩɾɔ̃hĩɾĩhã ́ -diɾɔ̃hĩɾĩhɛ̃́

-dihĩhã ́ -dihĩhãɛ̃́

-nĩɾɔ̃hĩɾĩhã ́ -nĩɾɔ̃hĩɾĩhɛ̃́

-nĩhĩhã ́ -nĩhĩhɛ̃́

-nĩhĩɾĩhã ́ -ɾenĩhĩɾĩhɛ̃́ -nĩɾɔ̃hĩɾĩhã ́ -ɾenĩɾɔ̃hĩɾĩhɛ̃́

-ɾɔ̃hã ́ -ɾɔ̃hãɛ̃́

↑ -kãhãhã ́ ↓ ↑ -kãhãhãɛ̃́ ↓

non-core

1300 Esteban Emilio Mosonyi and Raoul Zamponi

29 Yaruro (Pumé)

1301

(54) hĩnĩ ́ ɛa-nãmã́ hibɛ̃ã́ dem.nvis.sg.f.nom love-3sg.f.a.3.o.3sg.coref.possr.♀addr 3.poss ũĩ ́ hĩɾĩ ́ son dem.nvis.pl.obl ‘Shei loves heri children (♀ addressee).’ The full set of personal markers used with transitive verbs when the (third-person) O argument has a possessor is given in Table 29.8. 3.5.1.2 Imperative verbs Imperative verbs have a second-person subject in Yaruro. Intransitive verbs and transitive verbs used intransitively mark their (second-person) subject by means of two suffixes that also signal gender of the addressee:6 -hé, feminine, and -βɤ́, masculine. (55) mɔ̃ ã-hé! sleep-2.sbj.♀addr.imp ‘Sleep (sg, ♀ addressee)!’ A plural subject is also marked by placing the suffix -ʧʰĩ before the personal marker. (56) mɔ̃ ã-ʧʰĩ-βɤ́! sleep-pl-2.sbj.♂addr.imp ‘Sleep (pl, ♂ addressee)!’ As shown in Table 29.9, transitive and extended transitive imperative verbs use suffixes to mark the gender of the addressee and the person and number of animate O or E arguments, as well as gender with a third-person singular O or E.

Tab. 29.9: A-O/E agreement markers of transitive imperative verbs. a

addr

o/e 1 sg

2

♀ ♂

-koá -koé

3 pl

-iβɛ̃́ -iβé

sg

pl

f

nf

-hĩ ́ -hɛ̃́

-hũĩ ́ -hɔ̃ɛ ̃́

-hĩɾĩ ́ -hɛ̃ɾɛ̃́

6 The addressee necessarily coincides with the subject of the verb in this case.

1302

Esteban Emilio Mosonyi and Raoul Zamponi

Tab. 29.10: A-O-POSSR agreement markers of imperative verbs. a

2

addr

♀ ♂

o

3

possr 1 sg

pl

-kã ́ -kɛ̃́

-ibɛ̃ã ́ -ibɛ̃́

(57) cç͡ ɛɾɛ́ ɟʝ͡ ɔɾɔ-koé kɔá! money give-2.a.1sg.e.♂addr.imp 1sg.pro.obl ‘Give me some money (sg, ♂ addressee)!’ (58) ipa-hɔ̃ɛ ̃́ aboéa ɔá! skin-2.a.3sg.nf.o.♂addr.imp pig dem.dist.sg.nf.obl ‘Skin that pig (sg, ♂ addressee)!’ Plurality of an A argument is marked by placing -ʧʰĩ between the stem of the verb and the A person marker, as in (59). (59) cç͡ ɛɾɛ́ ɟʝ͡ ɔɾɔ-ʧʰĩ-koé kɔá! money give-pl-2.a.1sg.e.♂addr.imp 1sg.pro.obl ‘Give me some money (pl, ♂ addressee)!’ If the (third-person) O argument of a transitive imperative verb has a first-person possessor, the verb agrees with it, using the suffixes reported in Table 29.10. Also in this case, plurality of an A argument is marked by the suffix -ʧʰĩ; compare (60) and (61). (60) kɔnɛɾĩkʰia-dɛ-ibɛ̃ã́ ibɛ̃ã́ ũĩ ́ hĩɾĩ!́ mistreat-neg-2.a.3.o.1pl.possr.♀addr 1pl.poss son dem.nvis.pl.obl ‘Don’t mistreat our children (sg, ♀ addressee)!’ (61) kɔnɛɾĩkʰia-dɛ-ʧʰĩ-ibɛ̃́ ibɛ̃ã́ ũĩ ́ mistreat-neg-pl-2.a.3.o.1pl.possr.♂addr 1pl.poss son hĩɾĩ!́ dem.nvis.pl.obl ‘Don’t mistreat our children (pl, ♂ addressee)!’

3.5.2 Tense, Aspect, Mood, and Evidentiality The Yaruro TAME system is relatively simple, containing a limited number of values for each category.

29 Yaruro (Pumé)

1303

3.5.2.1 Tense and aspect Tense in Yaruro shows a three-fold past-present-future tense distinction, with a perfective/imperfective aspectual contrast in the past and two degrees of remoteness in the future: near and distant. Perfective past is expressed by the suffix -ã.́ (62) ɔdɛ́ dãdé (da-ã-dé) dem.dist.nf.nom see-pst.pfv-3sg.nf.sbj.♂addr ‘He saw (♂ addressee).’ Imperfective past is marked by the suffix -ɾé and requires the agentive nominalization of the verb. ɟʝ͡ aé ãnã=ɾã́ pɛaɾɤ́ (63) ʧʰaɔkʰia-ɾe-ɲĩ-ɾekodé play-pst.ipfv-nmlz.sg.f-1sg.sbj.♂addr horse big=ins at.the.beginning hudí dem.nvis.nf.nom ‘At the beginning, I (f) was playing with the big horse (♂ addressee).’ Present is not morphologically marked. (64) ɔdɛ́ da-dé dem.dist.nf.nom see-3sg.nf.sbj.♂addr ‘He sees (♂ addressee).’ Near future and distant future are signaled by the suffixes -pá and -bɛá, respectively. (65) ɔdɛ́ da-pa-dé dem.dist.nf.nom see-nr.fut-3sg.nf.sbj.♂addr ‘He will see (♂ addressee).’ ãdɛ́ ʧʰɤtɤ́ ãɾɔ̃ ɾɛ̃́ (66) ʧʰadɛ́ ɾĩ do-pa-bɛa-ɾeɾɛ̃́ well sun-vbz-dist.fut-1pl.sbj.♂addr 1pl.pro.nom sky pass dabu=ɾɤ́ hudí world=loc dem.nvis.nf.nom ‘We will live spiritually in the other world (♂ addressee).’ (dopá ‘live spiritually’) Besides perfective and imperfective, expressed with past tense, there are no other aspects marked by inflectional morphology in Yaruro.

1304

Esteban Emilio Mosonyi and Raoul Zamponi

3.5.2.2 Mood and modality In addition to declarative and imperative sentential moods (see Section 5), Yaruro has three verbal moods: optative, potential, and hypothetical. The optative mood, marked by the suffix -tɤ́, expresses a wish or desire and may also be used to issue a command to a first person or a third person. (67) mãnã-tɤ-dé nãɛ̃ã́ hadɔ́ ɔdɛ́ ! come.back-opt-3sg.nf.sbj.♂addr 2sg.poss companion dem.dist.nf.nom ‘May your companion come back (♂ addressee)!’ or ‘Let your companion come back (♂ addressee)!’ kaɾaka=ɾɤpɛ̃́ ãɾɔ̃ ɾɛ̃!́ (68) baɤ-tɤ-ɾɛ̃7 go-opt-1pl.nf.sbj.♂addr Caracas=all 1pl.pro.nom ‘May we go to Caracas (♂ addressee)!’ or ‘Let’s go to Caracas (♂ addressee)!’ The potential and hypothetical moods, unlike the optative mood, require the agentive nominalization of the verb. Potential mood, marked by -é, indicates that the situation referred to by the verb is likely to happen. It occurs both in independent clauses and in the apodoses of open conditionals, shown in (69) and (70), respectively. ́ ́ hũɾĩ ́ bɛá ãnã=ɾɤpɛ̃ hɤ-e-mɛ̃-ɾenɛ̃́ (69) mɛ̃nɛ̃́ 2sg.pro.nom go-pot-nmlz.sg.nf-2sg.nf.sbj.♂addr river big=all there ‘You (sg) will have to go to the Capanaparo River (♂ addressee).’ (70) nĩbɤ=ɾɤ=(h)dí kãɛ̃ã́ á-mãĩ speak=loc=dem.nvis.nf.nom 1sg.poss relative-sg.nf ́ ́ hũĩ=di ɛba-e-ɲĩ ́ mɛ̃ã=di dem.nvis.nf.obl=com have-pot-nmlz.sg.f 2sg.pro.obl=com ‘If you speak with my father, I will marry you.’ The hypothetical mood, marked by the suffix -tɤɾé (apparently from -tɤ́ optative + -ɾé imperfective past), expresses situations that are considered hypothetical or counterfactual by the speaker. It occurs in the apodoses of hypothetical conditionals, as in (142), and counterfactuals, as in (71), (143), and (144). (71) ɔdɛ́ nĩbɤ-tɤɾe-mɛ̃-dé dem.dist.nf.nom speak-hyp-nmlz.sg.nf-3sg.nf.sbj.♂addr kɔá=di kɔdɛ́ piɟʝ͡ í kʰɔ̃ ɾɔ̃ =ɾɤ́ hudí 1sg.pro.obl=com 1sg.pro.nom here live=loc dem.nvis.nf.nom ‘He would speak with me if I lived here (♂ addressee).’ 7 -ɾɛ̃ is a variant of -ɾeɾɛ̃.́

29 Yaruro (Pumé)

1305

Other modalities are expressed periphrastically in Yaruro. Desiderative, for example, is expressed by a main clause including the verb ‘stand’ and a non-finite adverbial subordinate clause (describing the event wished for) in the optative mood (without personal marking) including the similative postposition =mĩ;́ see (72). (72) kɔdɛ́ ʧʰidó ŋɔ̃ ã́ huɾa-tɤ=mĩ ́ kʰia-ɾekodé 1sg.pro.nom capybara meat eat-opt=sim stand-1sg.sbj.♂addr ‘I want to eat capybara meat (♂ addressee).’ The verb ‘stand’ is also used for expressing negative desiderative; in (73) it appears together with a negative clause adverbialized by means of adverbializing -ɾĩ.́ (73) kɔdɛ́ tɔpaɾá huɾa-dɛ-ɾĩ ́ kʰia-ɾekodé 1sg.pro.nom hare eat-neg-advr stand-1sg.sbj.♂addr ‘I do not want to eat hare (♂ addressee).’ Obligation is expressed though a main clause with the verb ‘carry’ and a non-finite adverbial subordinate (describing the event of the imposed obligation) marked by the purposive postposition =pɛ̃ã,́ as in (74). (74) hãbepa=pɛ̃ã́ cç͡ ia-ɾekodé work=purp carry-1sg.sbj.♂addr ‘I must work (♂ addressee).’ We do not have information on how other epistemic modalities are expressed in Yaruro, nor on the expression of ability and intentions.

3.5.2.3 Evidentiality Yaruro has an evidentiality system with two choices: reportive versus everything else. The reported term is marked by the suffix -βe while the non-reported term is unmarked. Examples (75) and (76) provide two examples of reportive marking. (75) ɔdɛ́ da-βe-dé dem.dist.nf.nom see-rpt-3sg.nf.sbj.♂addr ‘They say he sees (♂ addressee).’ mitɛ-dɛ-ã-βe-diɾɛ̃́ (76) pũmɛ̃=dɾɔ̃́ Yaruro=dem.nvis.pl.nf.nom seize-neg-pst.pfv-rpt-3pl.nf.sbj.♂addr ‘They say the Yaruro did not seize (anything) (♂ addressee).’

1306

Esteban Emilio Mosonyi and Raoul Zamponi

3.5.3 Derivation Yaruro exhibits a semi-productive verbalizer -pá that shows up in a number of morphologically transparent verbs: haʧi-dɛ́ ‘near; far-neg’, haʧi-dɛ-pá ‘approach, draw near’; ɔtɛ̃́ ‘tall’, ɔtɛ̃-pá ‘grow’; ŋɔ̃ ɛ ̃́ ‘mark’ (n.), ŋɔ̃ ɛ-̃ pá ‘write; study’; tatá ‘light’ (n.), tata-pá ‘shine’; do ‘sun’, do-pá ‘dawn’. Various suffixes espress Aktionsart. Avertive -tɛ̃ã,́ in (77), for example, means something like ‘nearly, almost’. (77) ɔdɛ́ da-tɛ̃ã-ɾe-mɛ̃-dé dem.dist.nf.nom see-avrt-pst.ipfv-nmlz.sg.nf-3sg.nf.sbj.♂addr ‘He was almost seeing (♂ addressee).’ The suffix -ɾiɤ́ gives the verb an iterative sense, as in (78) and (79). (78) tãɾɛ-ɾiɤ-pa-ɾeɾɛ̃́ hear-iter-nr.fut-1pl.sbj.♂add r ‘We will hear (it) constantly (♂ addressee).’ huɾa-ɾiɤ-ɾe-mɛ̃-dé (79) baɤ-mɛ̃=ɾɛ̃́ go-nmlz.sg.nf=nom eat-iter-pst.ipfv-nmlz.sg.nf-3sg.nf.sbj.♂addr ‘He walked (Lit. ‘went’) eating and eating (♂ addressee).’ The suffix -ɛ̃nɔ̃́ has a terminative (or cessative) value, demonstrated in (80). (80) nĩbɤ-ɛ̃nɔ̃-pa-ɾekodé kɔdɛ́ speak-term-nr.fut-1sg.sbj.♂addr 1sg.pro.nom ‘I will finish speaking (♂ addressee).’ Delimitative -diɤ́, shown in (81), indicates that the subject of the verb does something a little bit. (81) huɾa-diɤ-ɾe-mɛ̃-ɾekodé ʧʰidó ŋɔ̃ ã́ eat-dlm-pst.ipfv-nmlz.sg.nf-1sg.sbj.♂addr capybara meat ‘I was eating capybara meat for a little while (♂ addressee).’ The combination of terminative -ɛ̃nɔ̃́ and delimitative -diɤ́ gives rise to a pausative marker -diɛ̃nɔ̃́ indicating that a given event stopped for a while; see (82). (82) ɛɾɔ́ gitó haɾa-diɛ̃nɔ̃ -ɾeɾɛ̃́ carato pain drink-paus-1pl.sbj.♂addr ‘We stopped drinking carato for a while (♂ addressee).’

29 Yaruro (Pumé)

1307

Finally, -ɾɛ́ (after an oral vowel) ~ -ɾɛ̃́ (after a nasal vowel, as in (6)) refers to habitual events. See (83) and (84) for two examples. (83) ɔdɛ́ da-ɾɛ-dé dem.dist.nf.nom see-hab-3sg.nf.sbj.♂addr ‘He always sees (♂ addressee).’ tiotĩ ́ hĩ-ɾĩ ́ (84) hãdikʰia-ɾɛ-ɾe-mɛ̃-kedé8 dream-hab-pst.ipfv-nmlz.sg.nf-1sg.sbj.♂addr spirit dem.nvis.obl-pl ‘I (nf) used to dream about the spirits (♂ addressee).’ (tiotĩ ́ ‘God, spirit’) There is only one valence-changing mechanism in Yaruro: the causative. It involves the suffixation of -tɯnɛ̃́ (a grammaticalization of the verb tɯnɛ̃́ ‘cause, let, permit’) or -pãɾɛ̃́ to the verb stem, and it increases valency. It converts intransitive verbs into transitive verbs, as in (85), or transitive verbs into ditransitive verbs, as in (86) and (87). (85) ɟʝ͡ ɔdɛ́ ɔãĩ=ɾɛ̃́ dem.prox.sg.nf.nom man=nom bɛdɔ-tɯnɛ̃-ã-dɛ̃hɛ̃́

bo hũĩ ́ wake.up-caus-pst.pfv-3sg.nf.a.3sg.o.♂addr child dem.nvis.sg.nf.obl ‘The man woke the child up (♂ addressee).’

(86) ibɛ̃ã́ ɔtɛ́ hudí ʧʰɤnɛ̃ĩ=ã́ da-tɯnɛ̃-diβé 1pl.poss chief dem.nvis.nf.nom all=obl see-caus-3sg.nf.a.1pl.e.♂addr dabú ɟʝ͡ oɾó ɔá earth border dem.dist.sg.nf.obl ‘Our chief makes us see all, the borders of the earth.’ (87) kɔdɛ́ huɾa-pãɾɛ̃-ɾekoné ʧʰidó buíʧʰĩ 1sg.pro.nom eat-caus-1sg.nf.a.3sg.nf.e.♂addr capybara child hũĩ ́ dem.nvis.sg.nf.obl ‘I make the child eat capybara (♂ addressee).’ Non-verbal stems can also host the suffix -pãɾɛ̃́ with a verbalizing effect: nɔ̃ ‘path’, nɔ̃ -pãɾɛ̃́ ‘make a path’; ɔtɛ́ ‘adult’, ɔtɛ-pãɾɛ̃́ ‘rear’; gaɔ́ ‘dry’ (adj.), gaɔ-pãɾɛ̃́ ‘dry out’.

8 -kedé is a variant of -ɾekodé.

1308

Esteban Emilio Mosonyi and Raoul Zamponi

3.6 Adverbs, conjunctions, and interjections Adverbs modify predicates or entire clauses. The main types involve time (e.g., ɟʝ͡ abã́ ‘now’), space (e.g., ɟʝ͡ i ‘here’), and manner (e.g., hãdí ‘this way’). Adverbs of space and, at least, the manner adverb hãdí ‘this way’ can also occur as complements in non-verbal clauses. (88) kɔdɛ́

ɟʝ͡ i-ɾekodé 1sg.pro.nom here-1sg.sbj.♂addr ‘I am here (♂ addressee).’

(89) hãdi-dé hudí this.way-3sg.nf.sbj.♂addr dem.nvis.nf.nom ‘He is like that (♂ addressee).’ We observed just one coordinating conjunction with a disjunctive/adversative meaning and a few subordinating conjunctions in the language (see Sections 6.1 and 6.2.2). Interjections are bare roots behaving like whole utterances. They encode a number of different meanings, some of which listed in (90). (90) a. b. c. d. e.

aɟʝ͡ á ahá eé páa ́ kʰɯ̃ɾĩma

speaker feels pain hearer follows/understand speaker attention-getter ‘yes’ ‘no’

4 Phrases The following subsections deal with noun phrases (Section 4.1) and postpositional phrases (Section 4.2).

4.1 Noun phrases Possessive pronouns and numerals precede the head noun, while adjectives follow it. Examples (91) and (92) demonstrate possessive pronouns and numerals, while (93) shows an adjectival construction. (91) hibɛ̃ã́ dabú 3.poss land ‘his/her/its/their land’

29 Yaruro (Pumé)

1309

(92) ɲɔ̃ ãɾĩ ́ oɾé two dog ‘two dogs’ ́ (93) mĩʧʰĩ buíʧʰĩ cat little ‘little cat’ As indicated in Section 3.2, demonstrative pronouns precede the head noun when functioning as demonstrative determiners and follow it as markers of definiteness; see (94) and (95), respectively. (94) ɟʝ͡ ɔdɛ̃́ nĩβɛ́ dem.prox.nf.nom non.indigenous.person ‘this non-indigenous man’ ɟʝ͡ ɔdɛ̃́ (95) nĩβɛ́ non.indigenous.person dem.prox.nf.nom ‘the non-indigenous man’ Relative clauses can also occur in pre-head and post-head position (see Section 6.2.1). In possessive noun phrases with a full noun phrase possessor, the possessed noun directly follows the possessor marked by the genitive postposition =hã́, as in (96). dabú (96) pũmɛ̃́ hũĩ=hã́ Yaruro dem.nvis.nf.obl=gen land ‘land of the Yaruro’

4.2 Postpositional phrases All adpositions are postpositional in Yaruro. Most of the postpositions are used to mark peripheral roles like locative =ɾɤ́ (97) and instrumental =ɾã́ (98), but nominative =ɾɛ́ (after an oral vowel) ~ =ɾɛ̃ (after a nasal vowel) (100) and objective =βá (after /a/) ~ =á (after any other oral vowel) ~ =βã́ (after /ã/) ~ =ã́ (after any other nasal vowel) mark core participants (see (101) and (107); respectively, subject and object or extension-to-core argument (indirect object)). (97) puɛ=ɾɤ́ hole=loc ‘in a hole’

1310

Esteban Emilio Mosonyi and Raoul Zamponi

(98)

pũmɛ̃́ mãɛ̃=ɾã́ Yaruro language=ins ‘in Yaruro language’

(99)

́ ɛ̃ mãɛ̃duɾí nĩbɤcç͡ a-diɾɛ̃́ ɔaí ɔdɛɾɔ̃́ kʰãɾɛ̃m man dem.dist.pl.nf.nom discuss-3pl.nf.sbj.♂addr one story ʧʰiaɾí about ‘Those men are discussing a story (♂ addressee).’

ɟʝ͡ aɾɯka-dé (100) ɟʝ͡ ɔdɛ́ ɔãĩ=ɾɛ̃́ dem.prox.sg.nf.nom man=nom shaman-3sg.nf.sbj.♂addr ‘This man is a shaman (♂ addressee).’ (101) ʧʰaɔdɛĩ ́ dabadɛtɤ-kedé oɾɛ=á suddenly lose-1sg.sbj.♂addr consciousness=objv ‘I suddenly lost (my) consciousness (♂ addressee).’ Besides postpositions that are both synchronically and diachronically unanalyzable, Yaruro has postpositions that appear to be etymologically complex, like =ɾɤpɛ̃́ allative, locative (‘in a wide place’) (from =ɾɤ́ + pɛ̃ ‘place’) and =ɾɤpɔ̃ mṍĩ ‘toward’ (from =ɾɤ́ + pɛ̃ + -mɔ̃ ‘only’ + -ĩ diminutive). Some postpositions are also synchronically complex like, for example, =ɾɤpɛ̃́ ɟʝ͡ oɾó ‘until’ (with ɟʝ͡ oɾó ‘shore, bank’). See (102) for an example with =ɾɤpɛ̃́ ɟʝ͡ oɾó. (102) nãe͂ ã́ tʰado=ɾɤpɛ̃́ ɟʝ͡ oɾó 2sg.poss garden=all shore ‘until your (sg) garden’ The nominative and objective postpositions are not used with pronouns. The other enclitic postpositions may give rise to idiosyncratic formations together with demonstrative pronouns; compare (103a) and (103b), and see also (144). (103) a. cç͡ ɛɾɛɾã́ hũnĩɾã́ ‘with the money’

b. *cç͡ ɛɾɛ=ɾã́ hũĩ=ɾã́ money=ins dem.nvis.sg.nf.obl=ins

As (103a) shows, when a noun is followed by a demonstrative pronoun (as a definiteness marker), the same enclitic postposition can be attached both to head noun and the modifying demonstrative pronoun.

29 Yaruro (Pumé)

1311

5 Clauses This section deals with the three main types of independent clauses in Yaruro and their negation (Section 5.4): declarative (Section 5.1), imperative (Section 5.2), and interrogative (Section 5.3). There is not a special interrogative mood in the language. Declarative clauses typically occur in the indicative mood, which is is unmarked, as in (104). (104) kʰɔ-dé rain-3sg.nf.sbj.♂addr ‘It is raining (♂ addressee).’

5.1 Declarative clauses The order of both core and peripheral constituents of declarative clauses is relatively free. However, the following orders for the core constituents appear basic and unmarked. Intransitive clauses have SV order, as in (105), while transitive clauses have AOV order, as in (106). (105) hĩnĩ ́ hãnã-nɛ̃́ dem.nvis.f.nom come-3sg.f.sbj.♂addr ‘She comes (♂ addressee).’ (106) mɛ̃nɛ̃́ bɛ hũĩ ́ 2sg.pro.nom house dem.nvis.nf.obl gatɛ-ã-ɾenɛ̃́ buy-pst.pfv-2sg.nf.a.3sg.nf.o.♂addr ‘You bought the house (♂ addressee).’ In ditransitive clauses, the E argument appears after the verb, resulting in AOVE order, shown in (107). ́ ɛ̃ ɔ̃ ãĩ=ã́ (107) kɔdɛ́ cç͡ ɛɾɛ́ ɟʝ͡ ɔɾɔ-ɾekɔ̃ nɛ̃́ kʰãɾɛ̃m 1sg.pro.nom money give-1sg.nf.a.3sg.nf.e.♂addr one man=obctv ‘I give money to a man (♂ addressee).’ Non-verbal clauses have subject-complement order, as in (108) and (109). (108) ɟʝ͡ ɔdɛ̃́ ɔãĩ=ɾɛ̃́ pũmɛ̃-dé dem.prox.nf.nom man=nom Yaruro-3sg.nf.sbj.♂addr ‘This man is a Yaruro (♂ addressee).’

1312

Esteban Emilio Mosonyi and Raoul Zamponi

(109) huɾa-ɾɛã́ hudí ʧʰadɛ-dé eat-nmlz dem.nvis.nf.nom good-3sg.nf.sbj.♂addr ‘The food is good (♂ addressee).’ The possible semantic relationships between the two core arguments in non-verbal clauses include identity, as in (108); attribution, as in (109); and location, as in (90). The complement usually (but not obligatorily) takes the same (intransitive) subject agreement suffixes used with non-imperative intransitive verbs (see Section 3.5.1.1). The encoding of TAME on complements of non-verbal clauses appears to be analogous (but not identical) to that on verbal predicates; see (110) through (112). The details of TAME marking in non-verbal clauses are, however, unclear to us.9 (110) tɔ̃ heŋɔ̃ ã-mɛ̃-ɾe-mɛ̃-ɾekodé dance-nmlz.sg.nf-pst.ipfv-nmlz.sg.nf-1sg.sbj.♂addr ‘I (nf) was a musician (♂ addressee).’ (111) ɟʝ͡ aɾɯka-tɤɾe-mɛ̃-dé evil.spirit-hyp-nmlz.sg.nf-3sg.nf.sbj.♂addr ‘He could be an evil spirit (♂ addressee).’ (112) cç͡ iɾikoa-βe-dirɛ̃́ Guahibo-rpt-3pl.nf.sbj.♂addr ‘They say they are Guahibo (♂ addressee).’

5.2 Imperative clauses In clauses used for commands directed to a second person, verbs are marked (for imperative mood) with special person suffixes, as we saw in Section 3.5.1.2. These clauses tend to be short. Second-person personal pronouns rarely occur in them, and final syllables usually exhibit high pitch. See (113) and (114) for two examples. (113) mɔ̃ ã-βɤ́! sleep-2.sbj.♂addr.imp ‘Sleep (sg, ♂ addressee)!’ (114) cç͡ ɛɾɛ́ ɟʝ͡ ɔɾɔ-koé! money give-2.a.1sg.e.♂addr.imp ‘Give me some money (sg, ♂ addressee)!’

9 In this regard, we can only indicate that none of the non-verbal clauses in our data contain perfective past, habitual, or delimitative markers.

29 Yaruro (Pumé)

1313

For commands directed to a first person or a third person, verbs are marked for optative mood (Section 3.5.2.2).

5.3 Interrogative clauses Polar questions, shown in (115) and (116), have the same structure as a declarative clause and are marked by a rise in pitch on the penultimate syllable of the last word. (115) nĩβɛ́ mãɛ̃́ nibɤ-ɾɛã́ non-indigenous.person word speak-nmlz daba-mɛ̃-ɾenɛ̃?́ know-nmlz.sg.nf-2sg.nf.sbj.♂addr ‘Can you (sg) speak Spanish (♂ addressee)?’ (116) mɛ̃nɛ̃́ nĩbɤ-ã-nɛ̃?́ 2sg.pro.sg.nom speak-pst.pfv-2sg.nf.sbj.♂addr ‘Did you (sg) speak (♂ addressee)?’ Reduced variants of the verbal personal markers beginning in ɾe are normally used in polar questions. In such variants, the syllable ɾe does not occur; compare -ɾenɛ̃́ in (117a), which is a declarative, with -nɛ̃́ in the polar question in (117b). (117) a. mɛ̃nɛ̃́ huɾa-ɾenɛ̃́ 2sg.pro.nom eat-2sg.nf.sbj.♂addr ‘You (sg) eat (♂ addressee).’ huɾa-nɛ̃?́ b. mɛ̃nɛ̃́ 2sg.pro.nom eat-2sg.nf.sbj.♂addr ‘Do you (sg) eat (♂ addressee)?’ In constituent questions, the interrogative pronoun occupies the first position, as shown in (118). mɛ̃nɛ̃?́ (118) tʰanã=ɾɤpɛ̃́ baɤ-nɛ̃́ what=all go-2sg.nf.sbj.♂addr 2sg.pro.nom ‘Where are you (sg) going (♂ addressee)?’

5.4 Negation Both verbal and non-verbal clauses are negated by means of the negative suffix -dɛ́ . In verbal clauses, the suffix occurs in the verb and (usually) precedes its TAME markers and personal endings, as in (119) and (120).

1314

Esteban Emilio Mosonyi and Raoul Zamponi

(119) ɔdɛ́ da-dɛ-ã-dé dem.dist.nf.nom see-neg-pst.pfv-3sg.nf.sbj.♂addr ‘He did not see (♂ addressee).’ (120) gua-dɛ-koé! prick-neg-2.a.1sg.e.♂addr.imp ‘Don’t prick me (sg, ♂ addressee)!’ In non-verbal clauses, -dɛ́ is added to the complement, as in (121) and (122). (121) kɔdɛ́ pũmɛ̃-dɛ-ɾekodé 1sg.pro.nom Yaruro-neg-1sg.sbj.♂addr ‘I am not a Yaruro (♂ addressee).’ ɟʝ͡ ɔá (122) katíβai hudí cç͡ iɾi=ɾɤ-dɛ́ Kativai dem.nvis.nf.nom dem.prox.sg.nf.obl earth=loc-neg ‘Kativai (a mythological being) is not in the world.’

6 Clause-linking 6.1 Coordination There is no copulative coordinating conjunction in Yaruro. The adverbs pɛ̃ãhãdí ‘also’ and adó ‘again’ can connect (coordinate) independent clauses, as shown in (123) and (124). (123) huɾa-ɾekodé haɾa-ɾekodé pɛ̃ãhãdí eat-1sg.sbj.♂addr drink-1sg.sbj.♂addr also ‘I eat and drink (♂ addressee).’ (124) kãɛ̃ã́ ó-mɛ̃ hudí ŋɔ̃ ã-dí 1sg.poss sibling-sg.nf dem.nvis.nf.nom sing-3sg.nf.sbj.♀addr kucç͡ ikucç͡ ikʰia-dí adó ɛnã-dí dance-3sg.nf.sbj.♀.add again cry.out-3sg.nf.sbj.♀addr ‘My brother sings, dances, and cries out (♀ addressee).’ As indicated in Section 3.6, the language has a disjunctive/adversative coordinating conjunction whose form is habó, illustrated in (125) (see Guerreiro de Pirela 2016: 133, 157–159).

29 Yaruro (Pumé)

1315

(125) ɔdɛ́ nĩβɛ=ɾɛ́ ʧʰadɛ-dé dem.dist.nf.nom non-indigenous.person=nom good-3sg.nf.sbj.♂addr habó kɛ̃nã-dɛ-mɛ̃́ but think-neg-nmlz.sg.nf ‘That non-indigenous man is good but stupid (♂ addressee).’

6.2 Subordination Subordination has not yet been examined in depth. Only relative clauses and adverbial clauses are briefly discussed here. We have too little data to describe, even summarily, the grammatical structure of complement clauses.

6.2.1 Relative clauses There is no single strategy or process of relativization in Yaruro, but rather, the language uses different relativization strategies constrained by the role of the noun phrase head within the relative clause. In the material we gathered, we observed four strategies involving the use of nominalizer suffixes, postpositions, or no marker at all. The first strategy involves agent nominalization. When the role of the noun phrase head within the relative clause is that of subject argument, the relative clause is nominalized by attaching one of the three agentive nominalizers mentioned in Section 3.1 to the predicate: -ɲĩ (sg.f); -mɛ̃ (sg.nf), as in (126); or -hĩɾĩ (pl), as in (127). ɔãĩ ́ pɛã́ hudí (126) [oɾé hũĩ ́ ŋũnã-mɛ̃]́ dog dem.nvis.nf.obl hit-nmlz.sg.nf man young dem.nvis.nf.nom ɛnã-ɾɛ̃ã-dé cry.out-nmlz-3sg.nf.sbj.♂addr ‘The boy who hit the dog cried out (♂ addressee).’ [ibɛ́ tiotĩ ́ hudi-ɾɔ̃́ (127) nɔ̃ ĩ-dɛ-diɾɛ̃́ few-neg-3pl.nf.sbj.♂addr spirit dem.nvis.nf.nom-pl 1pl.pro.obl hida-hĩɾĩ]=ɾɛ̃́ protect-nmlz.pl=nom ‘The spirits who protect us are numerous.’ The second strategy is similar, involving agent nominalization and postposition. When the role of the head noun phrase inside the relative clause is that of themative (‘about something’), as in (128); comitative, as in (129); or “thanks to” adjunct, as in (130), the relative clause is also nominalized by means of an agentive nominalizer

1316

Esteban Emilio Mosonyi and Raoul Zamponi

and marked by a postposition that specifies the role of the head noun phrase within the relative clause. Note that, to support an enclitic adposition, the non-visual demonstrative pronoun (in the oblique) must be used. This demonstrative agrees in gender and number with the head noun, as shown in (129) and (130). ʧʰiaɾí] (128) mãɛ̃duɾí hudí [pũmɛ̃-hĩɾĩ=ɾɛ̃́ nĩbɤcç͡ a-hĩɾĩ ́ story dem.nvis.nf.nom Yaruro-pl=nom discuss-nmlz.pl about ʧʰakʰia-mɛ̃-dé beautiful-nmlz.sg.nf-3sg.nf.sbj.♂addr ‘The story which the Yaruro people are discussing is beautiful (♂ addressee).’ ́ (129) [hũĩ=di Pédɾo hudí hãbepa-ɾɛ-mɛ̃]́ dem.nvis.sg.nf.obl=com Peter dem.nvis.nf.nom work-hab-nmlz.sg.nf ɔãĩ ́ pɛã́ ɔdɛ́ pũmɛ̃-maɛa-dé man young dem.dist.nf.nom Yaruro-really-3sg.nf.sbj.♂addr ‘The boy with which Peter works is a true Yaruro (♂ addressee).’ ́ (130) [pió hũĩ=di mãcç͡ ia-dɛ-ɾĩ ́ kɔdɛ́ same dem.nvis.sg.nf.obl=com sad-neg-advr 1sg.pro.nom hãbepa-mɛ̃]́ pũmɛ̃=(h)dí nãɛ̃ã́ work-nmlz.sg.nf man=dem.nvis.nf.nom 2sg.poss aɟʝ͡ i-mãĩ-dé older.sibling-sg.nf-3sg.nf.sbj.♂addr ‘The person thanks to which (Lit. ‘with which’) I happily work is your (sg) older brother (♂ addressee).’ A third relativization strategy involves non-agentive nominalization. When the role of the noun phrase head within the relative clause is that of instrumental adjunct, the predicate of the relative clause is nominalized by the non-agentive nominalizer -ɾɛã́ ~ -ɾɛ̃ã́ (see Section 3.1), as in (131). oɾé hũĩ ́ ŋũnã-ɾɛ̃ã]́ to hudí (131) [ɔãĩ ́ pɛã=ɾɛ̃́ man young=nom dog dem.nvis.nf.obl hit-nmlz stick dem.nvis.nf.nom cç͡ acç͡ a-ɾe-mɛ̃-dé heavy-pst.ipfv-nmlz.sg.nf-3sg.nf.sbj.♂addr ‘The stick that a boy was hitting the dog with was heavy (♂ addressee).’ Finally, we noted that, when the role of the noun phrase within the relative clause is that of object argument, as in (132) and (133), or locative adjunct, as in (134), the predicate of the relative clause is marked by -ã,́ which is identical in form to the perfective past marker (Section 3.5.2.1).

29 Yaruro (Pumé)

1317

(132) hudí-dé [paɾiape-ɲĩ ́ kumã́ dem.nvis.nf.nom-3sg.nf.sbj.♂addr create-nmlz.sg.f Kuma hĩnĩ ́ ɲɔ̃ -ã]́ mãɛ̃́ hudí dem.nvis.f.nom say-rel word dem.nvis.nf.nom ‘The word that Kumá (f) the creator says is this.’ ́ nĩβɛ=ɾɛ́ (133) [kɔdɛ́ dã (da-ã)] 1sg.pro.nom see-rel non-indigenous.person=nom hãbepa-dé work-3sg.nf.sbj.♂addr ‘The non-indigenous person I see works (♂ addressee).’ (134) [ãɾɔ̃ ɾɛ̃ hɤkʰia-dɛ-ã]́ hɔ̃ ãí 1pl.pro.nom run-neg-rel forest ‘a forest where we do not run’ As examples (126)–(134) show, relative clauses may either precede or follow the head noun in Yaruro.

6.2.2 Adverbial clauses We have limited data on adverbial clauses but can briefly describe purpose, reason, concessive, temporal, and conditional clauses. Due to limited data, we cannot with confidence specify their positions relative to the main clause. Four of the five types of adverbial clauses we recorded are non-finite and do not take agreement morphology. Apparently, only clauses indicating cause can mark TAME. Purpose clauses are marked by the conjunction pɛ̃ãhũĩ ́ ‘in order to’, illustrated in (135), which appears to be composed of the purposive postposition =pɛ̃ã́ and the oblique non-feminine singular form of the non-visual demonstrative pronoun. hãnã́ mɛ̃ã́ tãɾɛ́ pɛ̃ãhũĩ ́ (135) ĩnɛ̃́ dem.dist.f.nom come 2sg.pro.obl listen in.order.to ‘She comes to listen you (sg).’ Reason clauses are marked by the conjunction taɾá, as in (136). ɟʝ͡ ɔa=iɤ́ (136) kɔdɛ́ hɔ̃ =ã́ kʰɔ̃ ɾɔ̃ -ɾekodé 1sg.pro.nom dem.prox.nf.sg.obl=loc house=obl live-1sg.nf.sbj.♂addr ɟʝ͡ ɔɾɔ-ã́ kɔá taɾá 1sg.pro.obl give-pst.pfv because ‘I live in this house, because they gave it to me (♂ addressee).’

1318

Esteban Emilio Mosonyi and Raoul Zamponi

Concessive clauses are formed with the conjunction rɤbokʰĩ,́ demonstrated in (137). hɔ̃ ɾɛ̃tá rɤbokʰĩ ́ habɛ́ (137) mɛ̃nɛ̃́ 2sg.pro.nom think although still kɛ̃nã-dɛ-mɛ̃-ɾenɛ̃́ think-neg-nmlz.sg.nf-2sg.sbj.♂addr ‘Although you are thinking, you are still stupid (♂ addressee).’ When the situation denoted by the adverbial clause occurs simultaneously to the situation of the main clause, the adverbial temporal clause is marked by locative postposition =ɾɤ́, as in (138). haɾa-ɾenɛ̃́ (138) kɔdɛ́ hɔ̃ ɾɛ̃ta=ɾɤ́ mɛ̃nɛ̃́ 1sg.pro.nom think=loc 2sg.pro.nom drink-2sg.sbj.♂addr ‘While I think, you drink (♂ addressee).’ Protases of non-counterfactual conditional sentences have the predicate marked by the locative postposition =ɾɤ́ followed by the nominative singular non-feminine form of the non-visual demonstrative pronoun (hudí ~ =(h)dí). Apodoses are variously formed. In this regard, it is convenient to define three types of non-counterfactual conditional sentences based on the way in which the apodosis is expressed: (i) “general truths”, (ii) “open”, and (iii) “hypothetical”. General truths are those where the speaker refers to a situation which always happens if certain conditions are met. Open conditions have a chance of being fulfilled in the real world in particular instances. Hypothetical conditions describe a situation that would occur if a condition were met which, for one reason or another, is unlikely or unreal. In conditional sentences expressing a general truth, the apodosis may occur with a predicate in the habitual aspect or the future tense or, as in (139), it may be nominalized by an agentive nominalizer (Section 3.1). See Álvarez and Guerreiro de Pirela (2015: 215–216) for further examples. (139) lola=ɲĩ ́ moã=ɾɤ=(h)dí hĩnĩ ́ Lola=dem.nvis.f.nom sleep=loc=dem.nvis.nf.nom dem.nvis.f.nom kʰɔnɔ̃ -kʰia-ɲĩ ́ snore-stand-nmlz.sg.f ‘If Lola sleeps, she snores.’ In open conditionals, such as (140)–(141), the predicate of the apodosis is in the future tense with agentive nominalization, or it is in the potential mood (whose expression requires an agentive nominalization of the same predicate, as seen in Section 3.5.2.2); see Álvarez and Guerreiro de Pirela (2015: 217–218) for further examples.

29 Yaruro (Pumé)

1319

hũĩ ́ gãbi=á kʰiɛtaɾa=ɾɤ́ (140) ãɾɔ̃ ɾɛ̃́ 1pl.pro.nom dem.nvis.nf.obl tobacco=objv smoke=loc hudí cç͡ ũɾĩcç͡ ia-kʰia-pa-hĩɾĩ-ɾeɾɛ̃́ dem.nvis.nf.nom be.in.trance-stand-nr.fut-nmlz.pl-1pl.sbj.♂addr ‘If we smoke this tobacco, we will be in a trance (♂ addressee).’ (141) ibɛ́ =di kʰiadiɤ=ɾɤ́ hudí cç͡ ɔ̃ 1pl.pro.obl=com remain=loc dem.nvis.nf.nom fish mũĩ-e-hĩɾĩ-ɾeɾɛ̃́ fish-pot-nmlz.pl-1pl.sbj.♂addr ‘If you remain with us, we will fish (♂ addressee).’ In hypothetical conditionals, the predicate of the apodosis is in the hypothetical mood (also expressed by an agentive nominalization of the same predicate). (142) hãbepa-dɛ=ɾɤ́ hudí ʧʰaɔkʰia-tɤɾe-mɛ̃-ɾekodé work-neg-loc dem.nvis.nf.nom play-hyp-nmlz.sg.nf-1sg.nf.sbj.♂addr ‘If I was not working, I would play (♂ addressee).’ In contrast to non-counterfactual conditionals, the protases of counterfactual conditional sentences are also marked by the locative postposition =ɾɤ́ and the demonstrative pronoun hudí ~ =(h)dí, but their predicate is in the perfective past. The apodosis may be expressed in the hypothetical mood like that of hypothetical conditionals (as in (143)), or it may be marked for both perfective past and hypothetical mood (as in (144)). Again, see Álvarez and Guerreiro de Pirela (2015: 219–220) for further examples. kɔá ãɾɔ̃ nɔ̃ -ã=ɾɤ=(h)dí (143) hudi-ɾɔ̃́ dem.nvis.nf.nom-pl 1sg.pro.obl teach-pst.pfv=loc=dem.nvis.nf.nom kɔdɛ́ daba-tɤɾe-mɛ̃́ ŋɔ̃ ɛ-̃ nĩbɤ-ɾɛã́ 1sg.pro.nom know-hyp-nmlz.sg.nf mark-speak-nmlz hũĩ ́ dem.nvis.sg.nf.obl ‘If someone had taught me, I would know how to read.’ (Lit. ‘know the reading’) (144) kʰiadiɤ-ã=ɾɤ=(h)dí ɔa=iɤ́ remain-pst.pfv=loc=dem.nvis.nf.nom dem.dist.nf.obl=loc hĩdɛ́ hãbo-ã-tɤɾe-mɛ̃-dé ʧʰɯʧʰa=ɾɤ́ a.little.longer die-pst.pfv-hyp-nmlz.sg.nf-3sg.nf.sbj.♂addr cold=loc ɔdɛ́ buiʧʰĩ=ɾɛ̃́ dem.dist.nf.nom child=nom ‘If he had remained there a little longer, that child would have died of cold (♂ addressee).’

1320

Esteban Emilio Mosonyi and Raoul Zamponi

7 Conclusion This chapter has presented a short grammatical sketch of Yaruro, an isolate language spoken in the Apure Llanos of Venezuela. The language is remarkable for its complex verbal agreement inflection. In addition to marking – according to their valence, person, number, and (if third person or, specifically, third-person singular) gender of their S, A, O, and E arguments – finite verbs in Yaruro index the gender of the addressee. This type of gender indexicality, concretely manifested by 238 pronominal suffixes 126 of which also express grammatical gender of the subject or both the subject and the O/E argument), is particularly rare in South America. The only other languages in which such a system is found (without interacting with grammatical gender) are the isolate Cholón of central Peru, and Southern Nambikwara in the Brazilian state of Mato Grosso (Rose 2015: 505–507). The agreement of finite transitive verbs with possessors internal to the O argument is another noteworthy phenomenon of Yaruro. This is also extremely uncommon in South America, with a parallel, as far we know, only in the isolate Chimané (Chimané-Mosetén) of the western Bolivian lowlands (Ritchie 2019), and this strategy is also rare in the rest of the world (see Bárány, Bond & Ritchie 2019). Finally, the various relativization strategies that we observed, which depend on the relativized syntactic/semantic position, are a further unusual trait of this typologically fascinating language.

8 Acknowledgements We are grateful to the editors of this volume and an anonymous reviewer for all their excellent advice. Our sincere thanks also go to our informants in the upper Capanaparo River, and especially to Jorge Ramón García, Luis Pérez, Cleto Castillo, and Milián Ruiz for their cooperation, insights, and patience.

9 References Álvarez, José & Yandira Guerreiro de Pirela. 2015. El período condicional en pumé/yaruro: la morfología de la irrealidad y la contrafactualidad. Opción 78. 203–223. Bárány, András, Oliver Bond & Sandy Ritchie. 2019. Database of prominent internal possessors. Guildford: University of Surrey. DOI: https://doi.org/10.15126/surreydata.8952959.v3 Castillo, Cleto, Jorge Díaz Pozo & Hugo Obregón Muñoz. 2003. Morfología yarura. Caracas: Ministerio de Educación, Cultura y Deportes. Dixon, R. M. W. 2004. Adjective classes in typological perspective. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Adjective classes: A cross-linguistic typology, 1‒49. Oxford: Oxford University Press.

29 Yaruro (Pumé)

1321

Gilij, Filippo Salvadore. 1782. Saggio di storia americana; o sia, Storia naturale, civile e sacra de’ regni, e delle provincie spagnuole di Terra-ferma nell’America Meridionale, vol. 3: Della religione, e delle lingue degli Orinochesi, e di altri Americani. Rome: L. Perego. Greenberg, Joseph H. 1987. Language in the Americas. Stanford: Stanford University Press. Guerreiro de Pirela, Yandira. 2016. Lengua indígena yaruro: una gramática tipológica-referencial. Córdoba: Universidad de Córdoba dissertation. Hervás y Panduro, Lorenzo. 1786. Idea dell’Universo, vol. 19: Aritmetica nelle nazioni e divisione del tempo fra l’Orientali. Cesena: Gregorio Biasini. Hervás y Panduro, Lorenzo. 1787a. Idea dell’Universo, vol. 20: Vocabolario poligloto con prolegomeni sopra più di CL. lingue dove sono delle scoperte nuove, ed utili all’antica storia dell’uman genere, ed alla cognizione del meccanismo delle parole. Cesena: Gregorio Biasini. Hervás y Panduro, Lorenzo. 1787b. Idea dell’Universo, vol. 21: Saggio pratico delle lingue con prolegomeni, e una raccolta di orazioni Dominicali in più di trecento lingue, e dialetti con cui si dimostra l’infusione del primo idioma dell’uman genere, e la confusione delle lingue in esso poi succeduta, e si additano la diramazione, e dispersione delle nazioni con molti risultati utili alla storia. Cesena: Gregorio Biasini. Hervás y Panduro, Lorenzo. n. d. Elementi grammaticali della lingua yarura. Manuscript in Opera Nostrorum 342 [folios 202r–209v], Archivio Romano della Compagnia di Gesù, Rome. Instituto Nacional de Estadística. 2015. Censo nacional de población y vivienda 2011. Empadronamiento de la población indígena. Caracas: Instituto Nacional de Estadística. Kaufman, Terrence. 1994. The native languages of South America. In Christopher Moseley & Ronald E. Asher (eds.), Atlas of the world’s languages, 46–89. London: Routledge. Krisólogo, Pedro Juan. 2002. Manual glotológico del idioma hapotein (versión del area del río Cunaviche). Caracas: Ministerio de Educación, Cultura y Deportes. Loukotka, Čestmír. 1968. Classification of the South American Indian languages. Johannes Wilbert (ed.). Los Angeles: Latin American Center, University of California. Lucena, Jerónimo José de. 1788. Traducion de la leng.a española en la de ynd.o yarura. Manuscript no. 2927 [folios 8r–10v], Biblioteca del Palacio Nacional de Madrid. Mosonyi, Esteban Emilio. 1966. Morfología del verbo yaruro: estudio de los sufijos personales. Caracas: Universidad Central de Venezuela. Mosonyi, Esteban Emilio, Jorge Carlos Mosonyi & Jorge Ramón García. 2000. Yaruro (Pumé). In Esteban Emilio Mosonyi & Jorge Carlos Mosonyi (eds.), Manual de lenguas indígenas de Venezuela. Vol. 2, 544–593. Caracas: Fundación Bigott. Obregón Muñoz, Hugo. 1981. La variabilidad de las lenguas indígenas venezolanas y algunos problemas de planificación lingüística. Antropológica 56. 3–24. Obregón Muñoz, Hugo & Jorge Díaz Pozo. 1989. Morfología yarura. Maracay: Instituto Universitario Pedagógico Experimental de Maracay. Obregón Muñoz, Hugo, Jorge Díaz Pozo & Luis Jesús Pérez. 1984. Lexico yaruro-español, españolyaruro. San Fernando de Apure: Gobernación del Estado Apure. Pache, Matthias. 2016. Pumé (Yaruro) and Chocoan: Evidence for a new genealogical link in northern South America. Language Dynamics and Change 6. 99–155. Ritchie, Sandy. 2019. Disjoint and reflexive prominent internal possessor constructions in Chimane. In András Bárány, Oliver Bond & Irina Nikolaeva (eds.), Prominent internal possessors, 107–130. Oxford: Oxford University Press. Rivas, Pedro. 2013. Aproximación a la escrituralización del idioma pumé: reflexiones acerca de la evolución reciente de los sistemas de trascripción empleados y las estrategias para optimizar la funcionalidad de su escritura. Boletín de lingüística 25. 153–186. Rose, Françoise. 2015. On male and female speech and more: A typology of categorical gender indexicality in indigenous South American languages. International Journal of American Linguistics 81. 495–537.

1322

Esteban Emilio Mosonyi and Raoul Zamponi

Seler, Eduard. 1902. Gesammelte Abhandlungen zur Amerikanischen Sprach- und Alterthumskunde, vol. 1: Sprachliches. – Bilderschriften. – Kalender und Hieroglyphenentzifferung. Berlin: Asher & Co. Swadesh, Morris. 1962. Afinidades de las lenguas amerindias. In Herbert Baldus (ed.), Akten des 34. Amerikanistenkongresses, Wien, 18.–25. Juli 1960, 729–738. Vienna: Berger.

Rik van Gijn

30 Yurakaré (Yurújare) 1 2 3 4 5 6 7 8 9 10

Background Phonology Morphology The nominal constituent The verbal constituent Clauses and sentences Information structure Conclusions Acknowledgements References

1 Background 1.1 Classification In spite of several attempts to connect Yurakaré1 (yura1255) to larger genealogical units, none of these proposals have had lasting success. Yurakaré is therefore generally classified as an isolate language in modern language classifications (Hammarström et al. 2022, Eberhard et al. 2022), Swadesh (1959, 1962) considered Yurakaré to be part of what he termed the Macro-Quechuan network, including inter alia neighboring languages such as Quechua, Itonama, Mosetén, and Cayuvava. Greenberg (1987) classifies Yurakaré as an equatorial language, in the Andean-Equatorial stock, along with, among others, Cayuvava, Arawakan, and Tupian languages. Suárez (1974) links Yurakaré to Mosetén and Chonan languages, as well as to Panoan and Tacanan in a stock he calls Macro-Pano-Tacanan. An important reason that these proposals to link Yurakaré to other languages have not led to consensus is that they are based on scant data, and some use methodologies that today are considered unreliable, such as mass comparison. Lewis et al. (2015) assumes two dialects, but there seems to be no linguistic evidence supporting this.

1 The name Yurakaré [juɾakaˈɾe] is based on the Hispanicized pronunciation, which violates the prosodic and morphophonological patterns of the language. The name Yurújare [juˈɹ̟uhaɹ̟e] is the same term, but in accordance with the prosodic and morphophonological rules. In the literature, the spelling Yuracaré or Yuracare is also often found. Following common practice in the literature, I will use the name Yurakaré in the remainder of this chapter, also because it follows the modern spelling conventions (in contrast to Yuracaré).

Rik van Gijn, Leiden University, Netherlands https://doi.org/10.1515/9783110432732-017

1324

Rik van Gijn

Yurakaré

30 Yurakaré (Yurújare)

1325

1.2 Society Yurakaré is spoken by an estimated 2,000–3,000 speakers in the drainage area of the upper Mamoré River in central Bolivia. Two major zones of habitation can be distinguished. A first group of Yurakaré settlements can be found along the Chapare and Ichilo Rivers in the Cochabamba department. Linguistic neighbors, apart from the Spanish-speaking mestizo community, are mainly Quechua-speaking communities and the Tupí-Guaranian Yuki. A second major group lives mainly along the Isiboro and Sécure Rivers in the Isiboro-Sécure National Park. Linguistic neighbors here are Chimanes (Mosetenan), Mojeño Trinitario and Mojeño Ignaciano (two closely related Arawakan languages). Yurakaré communities are generally small, consisting of only a few families, although larger villages (of over a hundred inhabitants) also exist. Today the Yurakaré largely live fully sedentary lives; the more traditional lifestyle can be characterized as semi-nomadic in the sense that the Yurakaré used to periodically relocate their houses within a certain area. The permanent settlement in legally recognized territories that characterizes the modern Yurakaré society has facilitated access to education for Yurakaré children. However, since education is offered in Spanish and teachers do not speak Yurakaré, this has had the consequence of further endangering the already endangered language. Today it is hard to find a child with active command of Yurakaré, although passive knowledge is still there.

1.3 Previous work and data sources Previous work on the language includes historical records by La Cueva (Adam 1893) and Lasinger (1915). The former is a dictionary and grammar compiled by Adam, based on material collected in the beginning of the 19th century by Franciscan missionary La Cueva. The latter is a short grammatical sketch based on data collected in the beginning of the 20th century. In the second half of the 20th century, the evangelical US-based New Tribes Mission settled in Yurakaré territory. Their presence lasted almost half a century and resulted in a grammatical tutorial (New Tribes Mission, n. d.) and an extensive wordlist (New Tribes Mission 1991). Ribera et al. (1991) is a highly interesting example of Bolivian-based participatory fieldwork that has resulted in a very useful vocabulary. New work in the form of educational material is being published by the Confederación de los Pueblos Indígenas de Bolivia (CIDOB). The data in the present overview mainly come from two sources: (i) the author’s PhD project (2001–2006), where data was collected in the villages of La Misión in the Chapare area, as well as the villages of Tacuaral, Nueva Canaan, and Loma del Masi in the Isiboro-Secure National Park), (ii) A DoBeS documentation project (2006–2011 with V. Hirtzel and S. Gipper), where data was collected in the villages of La Misión (Hirtzel), San Pablo (Gipper), and Nueva Lacea and Tres de Mayo (Van Gijn).

1326

Rik van Gijn

2 Phonology 2.1 Phonemes and graphemes Yurakaré has 17 phonemic consonants and 7 phonemic vowels (Tables 30.1 and 30.2). Most consonants can appear in lengthened form, though never contrastively. The appearance of geminate consonants is analyzed as facilitating building foot structure and thus creating a proper structure for accentuation (see 2.3). There is some free variation between [dj] and [j] for underlying /dj/. In the vowel inventory, the presence of the low front vowel /æ/ is noteworthy as it seems to be uncommon in the region. Mid vowels /e/ and /o/ and the low vowel /a/ receive a more closed pronunciation in closed syllables: [ɛ], [ɔ], and [ɑ], respectively. Vowels have no contrastive lengthening or nasalization. Yurakaré syllable structure avoids complexity in all basic syllabic positions, allowing for (C)V(C) structure and word-internal consonant clusters of at most two consonants. The few exceptions mostly concern loanwords, like kwentu ‘tale’, treseti ‘thirteen’, and some variants in the pronunciation of individuals such as pwede ‘can’, profesor ‘teacher’. All consonants appear in coda position, except /b/, and the alveolar and palatal plosives. At a convention held in Santa Cruz de la Sierra, Bolivia, July 25–26, 2007, a delegation representing the Yurakaré people determined the orthographic alphabet of Yurakaré, which will also be used in this chapter. The orthographic representations in Tables 30.3 and 30.4 correspond to the sounds in the tables below:

Tab. 30.1: Yurakaré phonemes: Consonants.

plosives

labial

alveolar

palatal

velar

glottal

p b

t d

ʧ dj

k

ʔ

s

ʃ ɹ̟

n

ɲ

fricatives nasals

m

laterals

h

l

semivowels w

j

Tab. 30.2: Yurakaré phonemes: Vowels. front high mid low

back

i

ɨ e

æ

u o

a

1327

30 Yurakaré (Yurújare)

Tab. 30.3: Consonant phonemes and graphemes. phonetic symbol

p

t

ʧ

k

b

d

dʲ

s

ʃ

h

m

n

ɲ

l

ɹ̟

w

j

orthographic symbol

p

t

ch

k

b

d

dy

s

sh j

m

n

ñ

l

r

w

y

Tab. 30.4: Vowel phonemes and graphemes. phonetic symbol

i

ɨ

u

e

o

æ

a

orthographic symbol

i

ü

u

e

o

ë

a

Stress is variable (see below) but falls on the penultimate syllable in the majority of cases. In the orthography, stress is indicated with an acute accent whenever it does not fall on the penultimate syllable. Should non-penultimate stress fall on one of the two diacritically marked vowels or , this is indicated with a circumflex accent.

2.2 Morphophonological processes Yurakaré has few morphophonological processes. There is a productive process of lenition of /k/ to [h] after a vowel, and of /ʧ/ to [ʃ] and /b/ to [p] in coda position. Nasals in coda position assimilate in place to following consonants. Vowel elision or reduction of /u/ and /i/ to [w] and [j], respectively, sometimes occurs when morphological processes yield two adjacent vowels. Some of the elision of vowels and/ or syllables, as well as consonant lengthening, is conditioned by accentuation rules discussed in the next section.

2.3 Stress The Yurakaré stress pattern follows from an interaction between prosodic rhythm and morphological interference. The basic rhythmic system for word-level accentuation is based on an iambic pattern constructed from left to right, with sensitivity to syllabic weight and an extrametrical final syllable. Morphological interference includes stress-attracting prefixes which alter the start of the rhythmic pattern. This can be seen in the contrastive examples in (1) where the prefix /ti/ is a stress-neutral prefix, and /ta/ a stress-attracting prefix (i.e., it requires at least secondary stress). The root is /sibæ/ in 1a–b, /pohore/ in 2. Weight sensitivity is shown in (3). For ease of processing, the prefixes are in boldface.

1328

Rik van Gijn

(1)

a. tiˈsibæ b. ˈtasibæ

(tisi)Ft (ta)Ftsi

‘my house’ ‘our house’

(2)

a. tiˈpohoɹ̟e b. ˌtapoˈhoɹ̟e

(tipo)Ftho (ta)Ft(poho)Ft

‘my canoe’ ‘our canoe’

(3)

a. ˈkuk:ulæ b. taˈhuk:ulæ

(kuk)Ftku (tahuk)Ftku

‘plantation, chaco’ ‘our plantation, chaco’

In (1b) and (2b), the stress-attracting prefix /ta/ causes the first syllable to be stressed (minimally secondarily). From there an iambic pattern is created towards the right edge, as far as possible. The rightmost head of a foot receives primary stress. (3a) shows that the heavy first syllable (geminate consonants create preceding heavy syllables) of /kuk:ulæ/ acts as a foot in its own right, and (3b) shows that – possibly in order to avoid stress clash – stress-attracting /ta/ and the heavy first syllable of the root together form a foot. As can be seen, left-to-right foot construction cannot be completed in (1b) because of the extrametrical final syllable, leading to unfooted material. In (2a), regular left-to-right footing also leads to unfooted material. Often these non-footed syllables are partly elided in trisyllabic roots (e.g., [tiˈpɔhɹ̟e] rather than [tiˈpohoɹ̟e] can be heard regularly); this elision process takes place obligatorily if the root consists of two identical syllables, even in disyllabic roots (4a)–(4b) and in (arguably) older loanwords such as chapapa (4c). (4)

a. /ta/ + /pæpæ/ → /ˈtɑp:æ/ ‘our grandfather’ b. /ta/ + /meme/ → /ˈtɑm:e/ ‘our mother’ c. /ti/ + /tʃapapa/ → /titʃɑp:a/ ‘my skewer (for grilling)’

The majority of disyllabic roots, such as /sibæ/ in (1), when appearing without any prefixes, have a geminated middle consonant [sib:æ]. This can be regarded as resulting from the requirements of the metrical system, which requires minimally one bimoraic foot (excluding the final syllable) to form an iambic pattern. Given the quantity-sensitivity mentioned above, we may assume that closed syllables have two morae associated with them. Lengthening the middle consonant creates a bimoraic structure before the last (extrametrical) syllable, thus fulfilling the minimal requirements for the stress system: (sib)Ft(μμ). Without the gemination, the word form would have provided only a monomoraic syllable which is not enough for a well-formed foot in Yurakaré. With syllabic prefixes, the gemination is not necessary because a bisyllabic structure of two open syllables is also bimoraic: (tisi)Ft(μμ). Curiously, disyllabic verbs and adjectives seem to be exempt from this operation, although they do have a minimality requirement of two syllables. (For more information on the accentual system, see Van Gijn 2014a).

30 Yurakaré (Yurújare)

1329

3 Morphology 3.1 Morphological units and processes Yurakaré is a synthetic, concatenating language, with a preference for head-marking. One of the salient morphological characteristics of Yurakaré is that it has a relatively well occupied prefix field, which includes aspectual marking, imperative marking, two referential positions, and an applicative system for verbs (see Section 5), and possessive prefixes for nouns (see Section 4). Another productive morphological operation is reduplication. There are three types of reduplication in Yurakaré: full reduplication, partial prefixed reduplication, and partial suffixed reduplication. Full reduplication is associated with forming deadjectival or denominal descriptive adjectives, often with a connotation of spatial distribution; prefixed reduplication generally copies the first two moras of the base, and adds a final [h] which replaces the final consonant in cases of closed syllable reduplication; suffixed partial reduplication generally copies the CV of the first syllable of the root, and is functionally associated either with a distributive reading or with causative. The latter form of reduplication no longer seems to be productive. Examples of each type are given in (5), for more information, see Van Gijn (2014b).

(5)

a. pujshi~pujshi ‘furry’ (Lit. ‘fur~fur’) b. ñuj~ñuñujulë ‘really small’ c. i-bëjtu~bë ‘see everywhere, repeatedly’

There are few further non-concatenative morphological processes. One set of person prefixes referring to the voluntary comitative (see Section 5) can be argued to be formed by ablaut, from /a/ and /i/ in the default paradigm to /u/ and /æ/ in the voluntary comitative paradigm, respectively. Some verbs have suppletive singular and plural stems, according to the number of the absolutive participant, as in (6).

(6)

a. mala ‘go.sg.s’ b. dele ‘fall.sg.s’ c. chitta ‘throw.sg.p’

bali ‘go.pl.s’ ñeta ‘fall.pl.s’ kokko ‘throw.pl.p’

Apart from affixing, reduplication, and further non-concatenative operations, Yurakaré allows for endocentric N-N compounding, as well as for exocentric N-N compounding, the latter type is marked with an interfix -n- which probably derives from the demonstrative/article na (7b). Endocentric compounds are unmarked and rightheaded, as shown in (7a).

1330

(7)

Rik van Gijn

a. wajá-pëlë cow-skin ‘cow hide’

b. talipa-n-chata chicken-exco-food ‘opossum’ (Didelphis albiventis) (Lit. ‘chicken is its food’)

3.2 Major lexical classes A combination of morphological and syntactic criteria yields the following word classes: verbs, nouns, adjectives, adverbs, interjections, ideophones, and particles. Most classes have subdivisions, but limitations of space do not allow going into this topic any deeper here. The reader is referred to Van Gijn (2006) for more precise information. Verbs, nouns, and adjectives can all function as predicates without any additional marking, and in that function, all of these parts of speech can take verbal morphology as well, such as evidentiality or tense/modality marking, in (8a) and (8b), respectively. (8)

a. maj~matata-ya ints1~big-rpt ‘It was really big.’

b. poropeso-shta=la teacher-fut=comm ‘He will be a teacher.’

Verbs, however, generally cannot function as arguments or modifiers without some additional marking (a few verbs of creation like nënë ‘cook’, dula ‘make’ are exceptions). Adjectives can marginally function as arguments, but only in elliptic constructions. Predicative adjectives are extremely frequent. Table 30.5 presents the prototypical characteristics of verbs, nouns, and adjectives, bearing in mind that the predicative function is open to all three word classes. The adverb class seems to be relatively small, consisting of a handful of synchronically unanalyzable roots. Manner adverbs appear directly before the predicate. Monomorphemic manner adverbs (including spatial ones) include kalantata ‘noisily’, kaminshë ‘carefully, slowly’, kónsono ‘easily, calmly’, límeye ‘last(ly)’, liwsha ‘first, up front’, and willë ‘far’. Yurakaré also has pro-adverbial forms anu ‘like this’ and achu ‘like that’. Apart from that there is a productive category-changing derivational process deriving manner adverbs from adjectives with -sh and from

Tab. 30.5: Prototypical characteristics of verbs, nouns, and adjectives in Yurakaré.

syntax morphology (exclusive to function)

verbs

nouns

adjectives

predicate Agr, TAM, valency, evaluation

argument possession, restrictive, number

modifier distribution, collective

30 Yurakaré (Yurújare)

1331

nouns and verbs with -shkuta. Many of the manner-adverbial functions, moreover, are performed by ideophones (see below). Adverbs of time (monomorphemic examples include tishilë ‘now’, shëy ‘yesterday’, and shínama ‘before, in the old days’) tend to appear at the periphery of the clause, especially at the beginning. A few other adverbs with other functions (e.g., yosese ‘again’, nentaya ‘maybe’) are distributed more freely in the clause. Time and space adverbs can alternatively be formed with the case enclitics. Interjections and ideophones often have phonologically deviant forms (bearing final stress, allowing for otherwise disallowed coda consonants, etc.). Their syntax, however, is quite distinct. Interjections (which are associated with a variety of emotions and experience like surprise, admiration, understanding, fear) can form independent utterances by themselves and generally appear at the beginning of sentences. Ideophones (associated with sounds, postures and positions, and movement) are dependent elements, which require an accompanying (directly following) verb. Morphologically, interjections and ideophones differ in that the former do not take any morphology, the latter can take middle voice -ta or causative -che (or partial reduplication). If they take one of these markers, ideophones function as verbs. Particles, finally, are more difficult to define, but their position is generally restricted, their phonology is deviant in that they are often monosyllabic, and many of them are enclitic particles that require a phonological host.

4 The nominal constituent 4.1 Noun phrase structure The syntactic template of noun phrases in Yurakaré can be schematized as in Figure 30.1; see listed sections for further description. Noun phrases obligatorily contain a referential item: this is prototypically a noun, but it can also be a (demonstrative) pronoun or – marginally – an adjective in an elliptic construction. Usually the elements of a noun phrase occur together, but Yurakaré also allows for discontinuous noun phrases in which a quantitative or qualitative modifier occurs in preverbal position, when it is focused, as in (9). (9)

lëshie ma-bëjti ta-ëshëta-w two 3pl-see.1sg.sbj 1pl-friend-pl ‘I see our two friends.’

dem (4.4)

qnt (4.4)

adj (4.4)

possr (4.3)

Fig. 30.1: Schematic template of the Yurakaré NP.

n (4.3)

adj (4.4)

RelCl (6.2)

=case (4.2)

1332

Rik van Gijn

In the next subsections, I will discuss each of the elements of Figure 30.1 in more detail, starting with case markers, which operate over the entire NP (Section 4.2); then nouns (Section 4.3),; and then modifiers, subsuming adjectives, quantifiers, and attributively used demonstratives (Section 4.4). Section 4.5 contains a brief discussion of noun classification. Relative clauses will be discussed in Section 6.2.

4.2 Phrasal enclitics: case (and a note on number) Case markers appear on the last element of the noun phrase. I choose to call these elements case markers because they do not demonstrably govern any complement, nor do they take agreement markers. Yurakaré has the following case enclitics, given in (10). (10) a. b. c. d. e.

=tina =la =y =chi =jsha

comitative instrument/path/reason locative direction ablative

The phrase-finality of these markers can be shown by contrasting NPs where the order of two elements can be reversed, as in (11). (11) a. [inele elle]=y inside earth=loc ‘inside the ground’

b. [elle inele]=y earth inside=loc ‘inside the ground’

The element inele is a fixed form (probably related to elle ‘earth’); it functions as a relator element 2 which can appear on either side of the noun it modifies. Whichever side of the noun it appears on, the locative marker =y remains in place. There is one other element with the same position in Yurakaré, which is moreover in complementary distribution with the markers in (10): the marker =ja. The precise conditions for the appearance of this marker are still somewhat unclear, but it seems to be restricted to marking subject NPs. The enclitic is glossed theme because it tends to appear in situations where the subject is emphasized or marked as sentence topic. More research is necessary, however, to determine the precise function of =ja. The plural marker -w seems to be losing its phrasal enclitic status. In New Tribes Mission (n. d.) it is reported to fully behave like the case markers, which it precedes. However, in Van Gijn (2006) the plural marker is reported to show variable behav-

2 Relator elements are usually possessed nouns in Yurakaré; in that sense inele is an exception.

30 Yurakaré (Yurújare)

1333

ior, tending to stay on the noun, rather than on the noun phrase, although phrasal behavior is still possible.

4.3 (Pro)nouns Pronouns are either personal or demonstrative. Demonstrative pronouns are also used as personal pronouns in that they can refer to third persons (undifferentiated for gender). Demonstrative pronouns differ from both nouns and personal pronouns in that they can be used attributively. Both demonstrative and personal pronouns differ from common nouns in that they cannot take possessive prefixes. Possessive prefixes in Yurakaré are for the most part related to the personal pronouns, as can be seen in Table 30.6, which also shows that the object markers found on verbs (Section 5.1) are closely related to the possessive markers. All person paradigms are based on the same 3-person × 2 number values format, although object markers partially encode collectivity (see Sections 4.5 and 5.1). Nonpredicative possessive constructions in Yurakaré consist minimally of a possessed noun carrying one of the person prefixes, and an optionally overtly expressed possessor, usually preceding the possessed noun (as in (12)). Possession in Yurakaré includes, among other relations, part-whole relations, like in the case of so-called relator nouns, which specify a more general location in combination with one of the locational case enclitics. In some cases, these relator nouns take on a special meaning, as is the case with doj(j)o, which normally means ‘body’, but as a relator noun it means ‘topside’ giving the interpretation ‘on top of’, as in (12). (12) lëtta wowore=ja tütü matata tuwë a-dojo=y one snake=thm sit big trunk 3sg.possr-body=loc ‘One snake lies on top of a big tree trunk.’ Apart from possessive and number inflectional marking, nouns can take derivational morphology, which can be divided into category-preserving derivational morphology and category-changing derivational morphology. An overview of the deri-

Tab. 30.6: Possessive markers, free pronouns, and object markers. gloss

free pronoun

possr. marker

object marker

1sg 2sg 3sg 1pl 2pl 3pl

sëë mëë (l)ati, ana, naa tuwa paa (l)atiw, anaw, naaw

timiatapama-

timiø/katapama-

1334

Rik van Gijn

Tab. 30.7: Nominal derivational morphology in Yurakaré. category-preserving

category-changing

marker

gloss

marker

gloss

-nñu -tebe -shama -ni -tA -jti

diminutive function, use nominal past in process, to-be evocative limitative ‘only’

-kka -j -jta -shkuta i-

measure (> adj) material (> adj) smell (> v) similative (> adv) have X (> v)

vational morphemes is given in Table 30.7, and a description of each morpheme follows. The marker -nñu is used to indicate small size and/or endearment. The intended use (or purpose) marker -tebe indicates that an object is used or will be used in a particular function, expressed in the base of the suffix. (13) bëbë-n-tu yutiche ta-chata-tebe search.for-intn-1pl curassow 1pl-food-func ‘Let’s look for curassow bird (Mitu tuberosum) for our food.’ The suffix -shama (also the plural root of ‘to die’) on a person-denoting noun indicates that the person has passed away, whereas on object-denoting nouns it indicates that the object is no longer usable because it has decayed (e.g., pojore-shama ‘canoe-wreck’). The marker -jti on verbs indicates habitual (see Section 5.2), but on (pro)nouns it means ‘only, just’ (e.g., së-jti ‘only, just me’). The evocative marker -tA (vowel harmonizes with preceding vowel) is found on a handful of nouns only, where it has a variety of semantic effects, loosely described as metaphorical or metonymic extension of the base meaning (e.g., sibë-të ‘house-evo’ is used to refer to, for example, cigarette cases or cassette cases; meye-te-w ‘ear-evo-pl’ refers to earrings). The marker -kka is frequently found on demonstratives, where it may indicate a measure of time (usually in combination with the locative case marker, e.g., ati-kka=y ‘in that time’) or physical size (e.g., ana-kka ‘this big’), but can also be found on common or proper nouns, to indicate the size of a person or object (e.g., Karina-kka ‘the size of Karina’). The resulting category is an adjective. Demonstratives marked with -kka enable a further derivation with the rare marker -shbache, yielding the interpretation ‘ever since that time’. There is a verb bache ‘send’ which is likely the source for this marker, which seems to be formed in an analogous way to -shkuta, where ku-ta is an (inflected) verb meaning ‘he says to him’, but with the grammaticalized function ‘act as if’ or ‘act like’. Marginal suffixes are -j, to indicate material from which an object is made, and -jta meaning ‘smell like X’ (e.g., chajmu-jta ‘it smells

30 Yurakaré (Yurújare)

1335

Tab. 30.8: Demonstrative pronouns and their morphological derivations. basic form

gloss

topical reference

emphatic reference

verbal derivation

adverbial derivation

ana ati naa

proximate medial/distal distal

l-ana l-ati –

bë-na bë-ti –

anama achama –

anu(ta) achu(ta) –

like dog’). A final, much more frequent operation is the formation of a possessive or acquisition verb with the prefix i-, as in i-tanti ‘to have eyes’, i-sibë ‘make/build a house’, and i-sama ‘get water’. Yurakaré has three basic demonstrative pronouns (ana proximate, ati medial, naa distal). It seems that the former two form the original demonstrative system, since only these two can take the full morphological set of derivations associated with demonstratives, as shown in Table 30.8. Yurakaré furthermore has three question words: a nominal question word ama, an adjectival question word tëtë, and an adverbial question word ëshë ‘why’. Most other question words are derived from ama by means of the case markers (e.g., ama-jsha ‘from where’, am-chi ‘whereto’) or from the nominalization tëtë-pshë ‘what, lit. what-thing’.

4.4 Modifiers Qualitative adjectives are generally used as predicates (as in (14)), although they can be used attributively as well (as in (12)). A distinctive characteristic (both in attributive and predicative function) of adjectives and quantifiers is that they can take collective markers -ima/yma, sg and -ibe/-ybe, pl (as in, e.g., ‘a bunch of bananas’ vs. ‘several bunches of bananas’, respectively). Collectivity in Yurakaré either refers to a collection, or to a bounded mass. A lake is conceptualized as a bounded mass of water. (14) maj~matat-ima-ya ati kudawa ints1~big-coll-rpt dem lake ‘That lake was really big.’ Example (14) also shows the intensity reduplication that adjectives share with verbs (see Section 3.1). Other evaluative-like morphology that can be found on adjectives are the diminutive -nñu, hedging -mashi, and amplificational -lë. The diminutive -nñu on verbs and adjectives generally indicates that the quality or action applies only to a very limited extent. The function of -mashi is similar in that it diminishes the extent to which an action or quality accurately describes reality. Intensifier -lë is not very frequent, but it is often found in comparative constructions.

1336

Rik van Gijn

(15) së=ja pëpë-li më=jsha 1sg.pro=thm old-ints2.1sg.sbj 2sg.pro=abl ‘I am older than you.’ Category-changing derivation includes -sh to form adverbs (e.g., shudyulë ‘beautiful’, shudyulësh ‘beautifully’) and -pshë – from bëshëë ‘shape, form, thing’ (e.g., matata ‘big’ matata-pshë ‘the big one’) to refer to the subject participant, the latter derivation is shared with verbs. Quantifiers share all of the morphology discussed above as long as it is semantically compatible. The adverb-forming suffix -sh does not occur with quantifiers, but there are two further markers: -cha completive and -ju(ta)/-jta frequentative that are found with quantitative adjectives. The former derives phrases like ‘both’ from the numeral ‘two’, and the latter forms adverb-like elements such as ‘once’ from the numeral ‘one’. Yurakaré has four native numerals left: lëtta ‘one’, lëshie ‘two’, liwi ‘three’, and lëpsha ‘four’. There is evidence that there used to be more native numerals (Adam 1893), but it is unclear whether they formed a lexicalized system or rather an ad hoc system. Interestingly, numerals can take verbal person prefixes (see Section 5.1) to form predicates: the benefactive prefixes in combination with the frequentative and the plural (16a), and the involuntary comitative (16b). (16) a. ti-n-lëpshu-jta 1sg-ben-four.pl-frq ‘I do it four times.’

b. ka-shie 3sg.com.nvol-two ‘It takes him/her two days.’

The fact that the numerals all start with lë/li, added to the fact that the construction in (16b) takes the second half of the numeral, suggests that the numerals are diachronically morphologically complex; they probably include the delimitative prefix li- (see Section 5.3). Numerals higher than five are borrowed from Spanish and usually take the amount-marker -ti (sinkuti, seysti, etc.) which is not very frequently used in other contexts, except to derive the question word amti ‘how much’ from the question nominal ama. Demonstratives can also be used attributively. In attributive function, they may be procliticized by an=, at=, and na=. It is at this point unclear what type of discourse functions the different proclitic forms encode, although na= seems to mark highly activated participants, and the other two, less activated participants.

4.5 Classification Yurakaré has a peculiar system of classifying nouns based on whether or not their denotata are conceptualized as inherently singular or not, as well as on whether they are inherently collective or not.

30 Yurakaré (Yurújare)

1337

Tab. 30.9: Four nominal classes based on plurality and collectivity.

non-inherently collective inherently collective

non-inherently singular

inherently singular

shunñe ‘man’ ewete ‘broom’

püü ‘road’ kudawa ‘lake’

The different possibilities result in four types of common nouns, as indicated in Table 30.9. Inherently collective nouns are either conceptualized as bounded aggregates (in which case they are inherently plural) or as bounded pieces of mass (in which case they are inherently singular). Inherently collective nouns trigger collective morphology on adjectives and quantifiers (see Section 4.4), and singular collectives (bounded masses) trigger special object agreement (Section 5.1). In this respect, the system could be seen as a gender system in the sense of Corbett (1991). On the other hand, the system has characteristics that are not gender-like, such as the fact that nouns can be recategorized (e.g., palanta ‘banana’ is a non-collective pluralizable noun like shunñe, but the same word treated as a [+plural, +collective] noun refers to ‘bunch of bananas’. In that respect, collectivity seems to be derivational in nature. The [+/− inherently singular] parameter is not the same as a count versus mass distinction since many of the inherently singular nouns (like e.g., püü ‘road’) can be counted without problems and without any apparent semantic effects. For more information on the classification system, see Van Gijn (2004).

5 The verbal constituent The (somewhat simplified) verbal template can be roughly schematized as in Figure 30.2. persobj

persobj

appl

asp

ROOT

voi

asp

mod

perssubj

evid/epist

Fig. 30.2: The verbal template.

Verbs can be modified directly by manner adverbs, ideophones, and a few particles. In the next few subsections (Sections 5.1–5.3) I sketch the morphological potential of verbs; in Section 5.4 I discuss possible verb modifiers.

5.1 Person marking and applicatives Verbal roots in Yurakaré can generally be divided into intransitive and transitive. Intransitive roots generally do not take object agreement markers and if they do –

1338

Rik van Gijn

Tab. 30.10: Verbal person paradigms. gloss

direct object

applicative object

goal

voluntary comitative

subject

1sg 2sg 3sg 1pl 2pl 3pl

timiø/katapama-

timikatapama-

tëmëkatapama-

tëmëkutupumu-

-y -m Ø -tu -p -w

unlike transitive roots – they require an applicative construction. Subjects (S/A) are suffixed; objects are prefixed. Before we move on to the functions and morphotactics of the different types of objects, Table 30.10 shows the five verbal person paradigms of the language. The applicative object paradigm is used for the involuntary comitative, benefactive, and malefactive (see below). As can be seen, the object person paradigms are all closely related but slightly different. In what follows I briefly discuss each of the participant types and refer back to the different paradigms of Table 30.10. One way of viewing the Yurakaré system of object marking is in terms of types of objects, based on their function and their place in the template. As seen in Figure 30.2, there are two positions for object person markers. There are morphotactic rules that determine the relative position of two object person markers if they appear on a verb at the same time. One can therefore speak of inner objects (appearing closer to the root) and outer objects (appearing further away). There are three inner object types: direct objects, goal or purpose objects, and involuntary comitative objects. Direct objects are formed by the direct object paradigm and no additional marking. The optionality in the third-person singular form of the direct object correlates with the collective-non-collective distinction in the inherently singular group (ka- is used for singular collective nouns, Ø for all other cases). This optionality is not present in the other paradigms. The goal object is formed by the goal paradigm and an applicative marker y- directly following it. One could say that the deviant first- and second-person markers in this paradigm are due to dissimilation of the /i/ in order to preserve the distinctness of the applicative marker y-, which would reduce the number of object person paradigms to three, although dissimilation is not a commonly occurring process in Yurakaré. Goal applicatives are largely restricted to verbs of movement. Involuntary comitatives are formed by the default applicative paradigm and no additional marking. They indicate that the participant undergoes the same event that is actively performed by the subject participant, not out of free choice, but because it is manipulated by the subject. The goal object and involuntary comitative object are illustrated by (17b) and (17c), respectively.

30 Yurakaré (Yurújare)

1339

(17) a. mala ‘S/he went.’ b. ka-y-mala ‘S/he went for it (fetched it).’ c. ka-mala ‘S/he took it.’ The outer objects are the benefactive, malefactive, and voluntary comitative. The benefactive is formed by the default applicative paradigm and an applicative marker n- which appears after the person marking cluster. Its semantics are in fact more general than beneficiary and also include for example, recipient, (animate) goal, and experiencer. The malefactive is formed on the basis of the default applicative paradigm and applicative -la. Its semantics also extend beyond pure maleficiary (e.g., source of movement, as in (18b)). The voluntary comitative, finally, is formed by the voluntary comitative paradigm and no further marking. The outer objects can also be illustrated with the verb mala ‘to go (sg)’. (18) a. ka-n-mala ‘S/he went to him/her.’ b. ka-la-mala ‘S/he went away from him/her.’ c. ku-mala ‘S/he went with him/her, or s/he followed him/her.’ Combinations of an inner and outer object have the order: outer object person marker, inner object person marker, applicative marker. (19) ka-ma-n-kaya 3sg-3pl-ben-give ‘He gave them to him/her.’ For more information on the person marking system and the applicatives, see Van Gijn (2005, 2011a, 2011b).

5.2 Tense, aspect, modality, and evidentiality Yurakaré is best analyzed as a future versus non-future language: future tense is marked by the suffix -(i)shta; present and past are unmarked. (20) a. ayma ma-ch-ishta fire 3pl-eat-fut ‘The fire will burn (Lit. ‘eat’) them.’

b. ayma=ja ma-che fire=thm 3pl-eat ‘The fire burnt (ate) them.’

The future marker -(i)shta may mark absolute or relative future. This latter point is particularly clear in realis subordinate clauses (Section 6.2) which are generally interpreted as being completed at the time of the main event.

1340

Rik van Gijn

(21) amala-ya shínama ta-ppë-shama-w ma-ch-ishta-ja come-rpt before 1pl.possr-grandfather-nml.pst 3pl-eat-fut-ss.real He came (they say) when he was going to burn (eat) our ancestors.’ Most predicates in Yurakaré do not carry any aspectual marker. For predicates denoting non-states (i.e., actions, achievements, processes and the like) the default aspectual interpretation seems to be completed, but there is some flexibility. Aspectual readings can be forced by using morphological means. The progressive prefix a- marks an event as ongoing or unfinished (this marker is not available for stative predicates, which have a default non-telic reading). (22) a. nënë cook ‘He cooked it.’

b. a-nënë prog-cook ‘He is cooking.’

The perfective marker i- is largely restricted to a combination with the distributive marker -uma/-wma (or partial reduplication) which indicates repeated events. The prefix i- binds the repeated events together as one episode. (23) i-shudyuj-ta-wma winani-ja (…) pfv-hidden-mid-distr walk-ss.real ‘After she walked hiding every now and again (…)’ Stative (adjectival) predicates with the distributive marker do not require the perfective marker, which is the main reason not to consider i- … -uma as a circumfix. There is also an alternative expression for pluractionality, by means of partial reduplication (lexically determined). (24) i-bëbë~bë-tu-ja amala-tu pfv-search.for~distr-1pl.sbj-ss.real come-1pl.sbj ‘We looked for him everywhere and then came here.’ There is another suffixed aspectual opposition between recent completive (marked with -lë) versus anticipatory aspect (marked with -nishi). Both are not very common, and especially the recent completive seems to be largely restricted to perfective temporal clauses, marked with -ja ss or -ti ds, to be discussed below. The remainder of the TAME morphology can be regarded as a main division between event-oriented modality suffixes, which precede the subject person markers, and epistemic/evidential markers, which follow the subject person markers. The event-oriented markers are given in Table 30.1. Although functionally these markers seem to form a heterogeneous bunch, they are all treated as irrealis by the switch-reference system (see Section 6.2, also Van

30 Yurakaré (Yurújare)

1341

Tab. 30.11: Event-modality markers. marker

gloss

-(i)shta -jti -ni -ta -nta -cha -iba ~ -yba

future habitual intentional hypothetical desiderative jussive obligative

Gijn & Gipper 2009, Van Gijn 2011c, 2014c). In fact, the irrealis markers also include the future tense marker -(i)shta (see above in this section) and the imperative markers (Section 6.2). Because they seem to form subcategories of irrealis, I treat them as modal markers. The evidential-epistemic markers are given in Table 30.12. The reason to regard this as a system is partly based on function, partly on the distribution of the markers in the verbal template (although some combinations are possible). Some markers have related, context-dependent, evidential and epistemic readings, illustrated in (25) for reportative -ya, which has an additional possibility interpretation; and in (26) for inferential -tiba, which has an additional mirative function (examples from Gipper 2011): (25) a. kayashi-w-ya wejshe shoot-3pl-rpt peccary ‘They say they shot the peccary.’

b. a-destino-ya comadre 3sg.possr-destiny-rpt comadre ‘It probably is his destiny, comadre.’

(26) a. ka-yle-jti-tiba 3sg-know-hab-infr ‘He must know [how to perform traditional mourning chants, given that he is very old].’ b. achama-tiba-la?!? be.like.that-infr-comm ‘Really?!?’ (reacting in surprise to a previous statement) Frustrative -chi is somewhat external to the system, but it is mentioned here since it has uses that fall outside the core area of frustration and into a more epistemological one when used with questions and perhaps also with commands, as in (27). (27) a. li-pi-ujwa-chi! del-imp.pl-look-frus ‘(Go and) look there!’

b. am=chi tütü-chi? q=dir sit;be-frus ‘Where is she?’

1342

Rik van Gijn

Tab. 30.12: Evidential-epistemic markers. marker

gloss

evidential setting

-ya -jtë -tiba -laba -chi

reportative/possible fact conjectural inferential/mirative subjective frustrative

external source no available evidence external evidence hearer is assumed to have no evidential access (no possible evidence)

The interpretation of -chi in these contexts is not quite clear yet, but one interpretation is that it marks heightened curiosity on the part of the speaker. Gipper (2011) interprets these uses as politeness marking. Table 30.12 lists the markers as well as the evidential situation type in which the markers can be applied. For more information on these markers the reader is referred to Gipper (2011).

5.3 Other verbal morphology Some verbs have a morphological opposition between middle voice, marked with -tA, and a causative form, marked either with -che or with a partial reduplication. (28) a. posojto ‘explode’ b. duta ‘burn (intr)’

vs. vs.

posojpo ‘cause to explode’ duche ‘burn (tr)’

The way the causative is formed is lexically determined, although the fact that some stems have both types at their disposal suggests a more complex situation in the past. Some of these verbs are synchronically derived from ideophones (e.g., posój ‘boom’), others have roots that are no longer meaningful (e.g., du) but can be suspected to have an ideophonic past. The middle voice marker -tA has spread to other parts of the lexicon and can also indicate reflexivity and reciprocity on transitive verbs, but the opposition middle-causative is specific to (former) ideophones. Verbs share most evaluative morphology with adjectives (see the description in Section 4.4): -nñu for ‘a little’ or endearment, -mashi for ‘somewhat’, and the intensifying reduplication. The suffix -lë on verbs, as mentioned previously, means recent completive rather than intensification. Before all other prefixes (excluding the intensity reduplication), verbs can have the marker li-, which can have one of three different functions: it indicates that the event takes place in a delimited space (29a); that a state has come into being, or has become worse (29b); or that an event happened before (29c).

30 Yurakaré (Yurújare)

1343

Tab. 30.13: Category changing morphology for verbs. marker

function

-bëshëë ~ -pshë -shku(ta) -ilë ~ ylë -uta ~ -wta -tA

subject nominalizer adverbializer (acting as if) habitual subject nominalizer (generally used as nominal predicate) place nominalizer instrument nominalizer (marginal)

(29) a. li-busha a-petche a-werta=y del-lay 3sg.possr-fish 3sg.possr-basket=loc ‘He puts his fish into his basket.’ b. li-lojo-tiba del-crazy-infr ‘He must have gone crazy.’

c. li-mashi del-rain ‘It rained again.’

The marker -ni, homophonous with the intentional marker, marks associated motion ‘go and do’, and seems to be lexicalized in stems like winani ‘walk’ and temeni ‘go around’. Although intentional -ni and associated motion -ni most likely stem from the same diachronic source, they can combine, and so must be treated as separate morphemes.3 (30) ma-bëbë-ni-ni ta-tiba-ni-w 3pl-search-am:go&do-intn.1sg.sbj 1pl-pet-intn-pl ‘I am going to go and look for our pets-to-be.’ Finally, verbs can take category-changing morphology. The markers are given in Table 30.13.

5.4 Verb modifiers Verbs can be modified by adverbs, ideophones, and particles. Modifiers mostly directly precede the predicate. Whereas underived adverbs form a small class (though they can be formed on the basis of adjectives, nouns, and verbs), ideophones are abundantly present, and cover semantic fields like sound, posture, orientation, motion, and speed. They stand out as a separate word class because they generally violate the phonotactic and accentuation patterns of the language, they require a

3 Note that the marker -ni can also be found on nouns, where it indicates future X, or X to-be, as can be seen in example (30).

1344

Rik van Gijn

predicate, and they usually have the possibility to carry the middle voice and causative morphology discussed in Section 5.3. Ideophones mostly combine with semantically relatively empty verbs ta ‘say’ or mala/bali- ‘go’, but they can also combine with verbs that have more semantic content, as in (31), where ideophones are in boldface. (31) a-mashtu=la shüj tom puppu-jti-ya 3sg.possr-machete=ins movement.through.air snap cut.off-hab-rpt ‘With his machete, he cut it [whisker of a mythological snake] off.’ Example (31) also shows that more than one (though rarely more than two) ideophones can be combined to encode a complex event (in this case the machete going through the air and hitting the target). The preverbal particles are partly formally related to each other: nish (allomorph nij) ‘not’, kani ‘not yet’, kaj ‘still’, and si ‘almost’. These particles can be stressed independently, and they can take more elements than just the verb in scope, as in (32). (32) nish awëwë-w-ya binta neg cry-3pl-rpt loud ‘They did not cry loudly (but rather softly).’

6 Clauses and sentences 6.1 Simple clauses The rudimentary structure of the simple declarative clause in Yurakaré can be schematized as in Table 30.14. Simple clauses minimally contain a predicate and optional argument NPs and modifiers at different levels. A count of over a thousand sentences in Van Gijn (2006) revealed that the A argument is most likely to be omitted (ca. 80 % of the cases) followed by the S argument (60 %) and the P argument (50 %). Applied ob-

Tab. 30.14: Schematic template of the simple clause. preclausal field

preverbal field

Setting expressions Thematic/focused (Sentence adverbs, arguments, connectors, tailadjuncts head-linkage)

VP

postverbal field

post-clausal field

Predicate + modifiers

Arguments, adjuncts

Backgrounded information, afterthoughts, clarifications

30 Yurakaré (Yurújare)

1345

jects are left unexpressed in about 90 % of the cases. There is a preference for overtly expressed subjects to follow the verb; for the P argument there is no clear preference. Generally speaking, Yurakaré seems to avoid the overt expression of more than one argument per predication, conforming to Du Bois’ (2003) one lexical argument constraint. Polar questions do not differ syntactically from declaratives; the difference is mainly expressed in terms of pitch (rising for questions). Wh-questions contain a question word, which appears in the preverbal field. There are three basic question words: nominal ama ‘who/what’, adjectival tëtë ‘what/which’, and adverbial ëshë ‘why’ (Section 4.3). Often, though not always, questions are marked with the frustrative marker. (33) a. ama=ja mi-n-dyuju-chi ushta ati? who/what=thm 2sg-ben-tell-frus before dem ‘Who told you that before?’ b. tëtë ululche mibej~mi-bene-jti-chi? which/what wasp ints1~2sg-hurt-hab-frus ‘Which wasp is most harmful to you?’ c. ëshë adyindyi-m-chi? why sad-2sg-frus ‘Why are you sad?’ Imperatives are associated with special morphology that replaces subject person marking: -ma for singular imperatives (marginally m-), pi- for plural imperatives (marginally -pa), and -yu for prohibitive, which replaces the singular imperative, but combines with the plural, as shown in (34). (34) a. shinojo-yu breathe-proh ‘Don’t breathe (sg).’

b. pi-ushpë-yu! imp.pl-bathe-proh ‘Don’t bathe (pl)!’

A handful of enclitic particles have clause-based distributions in that they can appear at the end of the clause, optionally still followed by post-clausal information. They are given in Table 30.15. The reconciliative can receive many interpretations, often in combination with the other clausal enclitics (apart from the resignative with which it is incompatible) such as surprise (35a), soliciting confirmation of the hearer (35b), and correction of hearer assumption (35c). (35) a. oyyo=naja=ye stink=new.sit=reconc.♀ ‘It (suddenly) stinks here.’ (looking for source of smell)

1346

Rik van Gijn

Tab. 30.15: Clausal enclitics and their approximate functions. marker

gloss

comment

=bë

momentaneous/attention

indicates the momentariness of a situation, but also seems to be used to attract the hearer’s attention

=bëla

still

indicates that a situation existed previously

=naja

new situation

indicates that the proposition marks a new situation that replaces a previous situation

=se (=si)

recordative

element that recalls an entity or situation that happened or was discussed before

=la

commitment

indicates that the speaker commits to the truth of the proposition or asks for such commitment from the hearer in questions

=ri/=yu

resignative (male/female speaker)

indicates that the situation is beyond the speaker’s (and hearer’s) sphere of influence or possible knowledge state

=ra/=ye

reconciliative (male/ female speaker)

indicates that the speaker attempts to reconcile what s/he perceives to be a temporary mismatch between his/her internal knowledge state and the outside world

b. bentiyuno=ye? twenty-one=reconc.♀ ‘It was the twenty-first, right?’ c. l-an=chi li-tütü shëy=si=ye del-dem=dir del-be yesterday=rcrd=reconc.♀ ‘But he was here yesterday.’

6.2 Complex clauses The tightest type of complex sentence is a multi-verb construction without any marking of conjunction (36). The verbs are immediately adjacent, share at least one participant (usually, but not necessarily the subject) as well as tense, modality, evidentiality, and polarity. These categories (subject and TMEP) are expressed only once for the entire complex (negation to the left, all other categories to the right of the complex). Aspect can be expressed on the left-most verb, as can be seen in, for example (23). (36) ajanta ashebu-m sing sow-2sg ‘You are singing while you sow.’

30 Yurakaré (Yurújare)

1347

Tab. 30.16: The switch reference system. realis

irrealis

identity (ss)

-ja

-ya

non-identity (ds)

-ti

Multi-verb constructions often receive a simultaneity interpretation, but they can also be used to encode functional equivalents to complementation, as in (37). The “complement-taking” verbs in these constructions are always on the right, and often have non-subject experiencers. As can be seen, although the experiencer on the left-most verb and the subject of the rightmost verb are shared, this does not lead to the non-expression of the experiencer person marking, since it is not a subject. (37) ti-m-pëlë ashuu-shti 1sg-ben-be.finished weed-fut.1sg ‘I am going to stop weeding.’ Causative constructions can also be expressed by a multi-verb construction. They are like the “complementation” construction, except that the “complement-taking” verb comes at the end. This construction is unique in the language in that the benefactive object of the verb ibëbë ‘treat’, shortened to -bë- governs subject deletion under identity on the leftmost verb. (38) awëwë ti-m-bë cry 1sg-ben-treat ‘He made me cry.’ Semantics associated with these serial-verb-like constructions are manner constructions (or cotemporal events), and constructions expressing phasality, ability, difficulty, and causation (see Van Gijn 2010 for a more detailed description). Most other types of complex clauses are marked for switch reference. The basic design of the switch reference system in Yurakaré is as in Table 30.18. Irrealis identity (same-subject) clauses express purposive, conditional, irrealis temporal, and irrealis coordinate events. The predicate marked for switch reference is not inflected for subject or evidentiality, and to a limited extent for modality. All verbs that are marked with one of the event-modality suffixes from Table 30.11 or one of the imperative markers count as irrealis. (39) lëtëmë=chi wita-ya a-nënë-cha-m jungle=dir arrive.sg-ss.irr prog-cook-juss-2sg.sbj ‘When you arrive in the jungle, you should cook.’

1348

Rik van Gijn

Realis identity clauses are inflected for subject and can express realis temporal and realis coordinate clauses. They also marginally express relative clauses and complement clauses (see below). They cannot take any of the evidential/epistemic suffixes. (40) mi-bëjti-ja mi-la-shuyuj-ta-y 2sg-see.1sg.sbj-ss.real 2sg-mal-hidden-mid-1sg.sbj ‘When I saw you I hid myself from you.’ Clauses marked with -ti express most complement clauses, relative clauses, as well as (realis and irrealis) adverbial clauses. The verbs marked with -ti are fully inflected except for evidentiality/epistemic modality. Complement-taking verbs in Yurakaré are often impersonal verbs with a non-subject experiencer, and so in that sense are examples of different-subject constructions. And in fact, when for some verbs an opposition between same-subject and different-subject can be contrasted, it can be shown that the switch-reference system also applies to complement clauses. (41) a. ayajta mala-m-ti mi-bëjta-y fast go.sg-2sg.sbj-ds 2sg-see-1sg.sbj ‘I saw you running.’ b. bëjta-ta-y ti-manchijsha ti-buybu ka-n-dyuju-y-ja see-mid-1sg.sbj 1sg-self 1sg-language 3sg-ben-inform-1sg.sbj-ss ‘I saw myself teaching my language (on video).’ Relative clauses are somewhat deviant, in that what determines the appearance of either -ja or -ti is not subject coreferentiality, but coreferentiality between the relativized argument and the subject of the main clause.4 (42) a. dele na shunñe danda-ja kummë=la fall.sg dem man go.up-ss tree=ins ‘The man that went up the tree fell.’ b. lulë-ni na yutiche bobo-y-ti pluck-intn.1sg.sbj dem razor.billed.curassow kill-1sg.sbj-ds ‘I am going to pluck the mutún bird I killed.’ Some adverbial clauses are formed with -ti and one of the case enclitics, independently of whether or not their subjects are the same as the main clause subject:

4 An alternative interpretation of (42a) is as a temporal clause (the man fell when he climbed up the tree) but the point is that this type of sentence would not be expressed with a -ti clause, even though it is interpretable as a relative clause.

30 Yurakaré (Yurújare)

1349

-ti=la (reason), -ti=jsha (concession), and more generally with the local case markers to form locational clauses. A final type of complex clause is one where two clauses are juxtaposed asyndetically (coordination) or flagged by a phonologically free connective: achaya (interclausal, different-subject purpose), kusuti (bisyndetic, beginning of each clause, disjunction), lachamatijsha (interclausal, adversative) are among the coding possibilities. These connectors are relatively transparently related to (pro)verbs acha(ma) ‘do like that’ and kusu ‘want’.

7 Information structure Many aspects of information status and the way Yurakaré grammar deals with it remain unclear. In this section I give the current state of knowledge, and I will also indicate where more information is needed. Yurakaré is a pro-drop language, which generally does not express full noun phrases. Expression of NPs or independent pronouns often indicates emphasis or new or weakly activated information, although NPs and pronouns may also be expressed to disambiguate, and possibly for other purposes. Although the article system is still poorly understood, it is clear that the proclitic form na= is found with NPs expressing activated referents. As mentioned in Section 6.1, positions in the clause tend to be associated with information structural status, and the preclausal field is reserved for connective and setting expressions, which create cohesion in a series of connected utterances, especially in narratives. Elements found here include location expressions, temporal expressions, but also repeated clauses in dependent form (tail-head linkage). The preverbal field seems to be a position in which elements are focused. Interestingly, Yurakaré allows for extraction of elements from a noun phrase to this position in order for them to be focused, as in (43), where the quantifier bëmë is found in preverbal position, but the head noun it pertains to (titëbë) in postverbal position. (43) bëmë wojto ti-tëbë much/many pour.out 1sg-blood ‘Much of my blood poured out.’ The postclausal position seems to be mainly reserved for afterthoughts, repairs, etc. and is usually separated from the rest of the clause by a pause.

8 Conclusions A few aspects of Yurakaré make it look like a fairly typical Amazonian language as described by Aikhenvald (2012: 383–391), for example, the presence of a central high

1350

Rik van Gijn

vowel, synthetic morphology, prefixes, bound person markers for several roles, the presence of a frustrative marker, evidentiality, a fairly elaborate system of applicatives, and multi-verb constructions. Other hallmarks of Amazonian languages are lacking, such as nasal spread, tone, clusivity, alienability, or ergative patterns. Some remarkable aspects of Yurakaré from an areal (and sometimes global) perspective are the following: With respect to phonology, the phoneme system of Yurakaré contains a few areally uncommon sounds like /dj/, /ɹ̟/, and /æ/, and the iambic-based metrical stress system with complex interactions with morphology is also uncommon in the area. In morphology we find the elaborate use of several reduplication patterns, many prefix positions (even for an Amazonian language), and an elaborate word class of ideophones. Nouns display an unusual type of noun classification based on inherent collectivity and plurality, noun phrases are separable and subconstituents can be fronted. Numerals display uncommon morphological potential. The verbal constituent shows a unique applicative system based on ‘inner’ and ‘outer’ object types, a typologically unusual evidential-epistemic system, and causatives that are marked by reduplication. In terms of clausal syntax, Yurakaré has post-verbal subjects, which are globally underrepresented, there are unusual types of multi-verb constructions involving experiencer verbs and their notional complements, and the language has a switchreference system that extends into relative clauses and complement clauses.

9 Acknowledgements First and foremost, I would like to thank the speakers of Yurakaré for sharing their knowledge. I would furthermore like to thank the editors and an anonymous reviewer for highly useful comments on earlier drafts of this chapter. Finally, the analysis reported here has benefited greatly from numerous discussions with Vincent Hirtzel and Sonja Gipper, Jeremías Ballivian, Mily Crevels, and Pieter Muysken†. Remaining errors are mine. This publication is part of the ERC Consolidator project South American Population History Revisited, funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 818854 – SAPPHIRE). This support is gratefully acknowledged.

30 Yurakaré (Yurújare)

1351

10 References Adam, Lucien. 1893. Principes et dictionnaire de la langue yuracaré ou yurujure composé par le R. P. La Cueva et publié conformément au manuscrit de A. D’Orbigny. Paris: Maisonneuve. Corbett, Greville G. 1991. Gender. Cambridge: Cambridge University Press. Eberhard, David M., Gary F. Simons, and Charles D. Fennig (eds.). 2022. Ethnologue: Languages of the World. Twenty-fifth edition. Dallas, Texas: SIL International. Online version: http:// www.ethnologue.com Du Bois, John W. 2003. Discourse and grammar. In Michael Tomasello (ed.), The new psychology of language: Cognitive and functional approaches to language structure, vol. 2, 47–89. Mahwah & London: Lawrence Erlbaum. Gipper, Sonja. 2011. Evidentiality and intersubjectivity in Yurakaré. Nijmegen: Radboud University PhD dissertation. Greenberg, Joseph H. 1987. Language in the Americas. Stanford: Stanford University Press. Hammarström, Harald, Robert Forkel, Martin Haspelmath & Sebastian Bank. 2022. Glottolog 4.6. Jena: Max Planck Institute for the Science of Human History. http://glottolog.org (accessed 17 November 2015). Lasinger, Fulgencio. 1915. El idioma yuracarés. Archivo de la Comisaría Franciscana de Bolivia, año VII, 47–49, 117–119, 158–159, 204–205, 234–236, 400–401. New Tribes Mission. 1991. Diccionario yuracare-castellano, castellano-yuracare. Cochabamba: New Tribes Mission. ms. New Tribes Mission. no date. [Yuracaré teaching material]. Cochabamba: New Tribes Mission. ms. Ribera, Julio, Asencio Rocha & Walter Rivero. 1991. Vocabulario yuracare-castellano, castellanoyuracare. Trinidad, Bolivia: Eparu & Misereor. Suárez, Jorge A. 1974. Macro-Pano-Tacanan. International Journal of American Linguistics 37(3). 137–154. Swadesh, Morris. 1959. Mapas de clasificación lingüística de México y las Américas. Mexico: Universidad Nacional Autónoma de México. Swadesh, Morris. 1962. Afinedades de las lenguas amerindias. Akten des 34 Internationalen Amerikanistenkongresses, 729–738. Vienna: Fred Berger. Van Gijn, Rik. 2004. Number in the Yurakaré noun phrase. In Cornips, Leonie & Jenny Doetjes (eds.), Linguistics in the Netherlands 2004, 69–79. Amsterdam & Philadelphia: John Benjamins. Van Gijn, Rik. 2005. Head marking and dependent marking of case relations in Yurakaré. In Amberber, Mengistu & Helen de Hoop (eds.), Competition and variation in natural languages: The case for case, 41–72. Oxford: Elsevier. Van Gijn, Rik. 2006. A grammar of Yurakaré. Nijmegen: Radboud University PhD dissertation. Van Gijn, Rik. 2010. Multi-verb constructions in Yurakaré. In Aikhenvald, Alexandra Y. & Pieter C. Muysken (eds.), Multi-verb constructions: A view from the Americas, 255–282. Leiden: Brill. Van Gijn, Rik. 2011a. Subjects and objects in Yurakaré: A semantic account. International Journal of American Linguistics 77(4). 595–621. Van Gijn, Rik. 2011b. Pronominal affixes, the best of both worlds: the case of Yurakaré. Transactions of the Philological Society 109. 41–58 Van Gijn, Rik. 2011c. Grammatical and semantic integration in Yurakaré subordination. In van Gijn, Rik, Katharina Haude & Pieter Muysken (eds.), Subordination in native South American languages, 169–192. Amsterdam & Philadelphia: John Benjamins. Van Gijn, Rik. 2014a. Word accent and mapping rules in Yurakaré. Morphology 24(3). 223–244. Van Gijn, Rik. 2014b. Reduplication in Yurakaré. In Danielsen, Swintha, Katja Hannss & Fernando Zúñiga (eds.) Word formation in South American languages, 143–161. John Benjamins.

1352

Rik van Gijn

Van Gijn, Rik. 2014c. Repeated dependent clauses in Yurakaré. In van Gijn, Rik, Dejan Matić, Jeremy Hammond, Saskia van Putten & Ana V. Galúcio (eds.) Information structure and reference tracking in complex sentences. Amsterdam & Philadelphia: John Benjamins. Van Gijn, Rik & Sonja Gipper. 2009. Irrealis in Yurakaré – a comparative perspective: On the cross-linguistic consistency of an elusive category. In Hogeweg, Lotte, Helen de Hoop & Andrej Malchukov (eds.), The cross-linguistic semantics of tense-aspect-modality, 155–178. Amsterdam & Philadelphia: John Benjamins.

Index A abilitative, Taushiro 1014 ablative – Máku 777 – Taushiro 1007 – Trumai 1088 – Wã́nsöjöt 1167 – Yurakaré 1332 absolutive – Trumai 1082–3, 1088, 1090, 1097, 1099–100, 1102 – Yurakaré 1329 accusative – Kwaza 724, 729–30, 753 – Movima 829 – Muniche 880 – Mỹky 930 – Omurano 944, 945, 946, 947 – Taushiro 1009–10 – Trumai 1098 – Wao Terero 1223 Achuar 948–9 Adam, Lucien 1325 additive – Kanoé 708, 714 – Kwaza 756 – Muniche 861 adessive – Taushiro 1007 – Wã́nsöjöt 1167 adjectives – Kanoé 664, 683 – Kwaza 723, 731–2 – Máku 777, 784 – Movima 823–4 – Muniche 859, 867–8, 882 – Mỹky 908, 918 – Pirahã 970, 977, 985 – Taushiro 1017 – Tinigua 1053, 1054 – Trumai 1081, 1085, 1087, 1094 – Urarina 1115–16, 1120 – Wã́nsöjöt 1157, 1158, 1165–6 – Wao Terero 1199–200, 1205, 1207–8, 1210 – Yaruro 1294, 1308–9 – Yurakaré 1330, 1334, 1335, 1342 https://doi.org/10.1515/9783110432732-018

adjuncts – Movima 820, 829, 833, 834–5 – Tinigua 1044, 1060 – Warao 1256, 1257, 1258, 1273 – Yaruro 1315 admonition, Wao Terero 1217 adpositions see also postpositions; prepositions – Kwaza 730, 736 adverbial systems – Kanoé 675, 683, 689, 690–1, 701, 702, 710, 712–13 – Kwaza 723, 726, 733, 735, 745, 752–3, 756, 757 – Máku 782, 790, 799, 800–2 – Movima 829, 835–7 – Muniche 860, 868–9 – Mỹky 928 – Pirahã 969–70, 988 – Taushiro 1025 – Tinigua 1044, 1048, 1049, 1055, 1056, 1066, 1068 – Trumai 1082, 1083, 1088, 1089, 1090, 1091, 1100–2, 1103 – Urarina 1115 – Wã́nsöjöt 1156, 1158, 1183 – Wao Terero 1205, 1220–1, 1226, 1231, 1232–3 – Warao 1256, 1271–2, 1275, 1278–9 – Yaruro 1308, 1314, 1317–19 – Yurakaré 1330–1, 1336, 1343–4, 1348 adverbializer – Kwaza 726 – Tinigua 1055, 1063 – Trumai 1082, 1083 – Wã́nsöjöt 1183 – Yaruro 1294, 1305 – Yurakaré 1343 adversative – Kanoé 708, 709 – Kwaza 760 – Urarina 1141 – Wã́nsöjöt 1186 – Warao 1268 – Yaruro 1308, 1314 affixation see infixation; prefixation; suffixation agent nouns, Máku 780–1

Index

A

agentive – Kanoé 680 – Movima 830 – Wã́nsöjöt 1151, 1165, 1177, 1179 – Yaruro 1288–9, 1303, 1315, 1318 agglutination – Kanoé 664 – Kwaza 723 – Máku 777 – Movima 812 – Pirahã 964 – Wã́nsöjöt 1158 – Wao Terero 1199, 1200 – Warao 1249, 1259 agreement see also alignment; gender marking; number marking; person marking; predication – Kanoé 664, 677, 691–2 – Máku 785–7, 794 – Movima 812, 826 – Muniche 860–1, 870, 880, 886 – Mỹky 909–12, 918–19 – Pirahã 975 – Taushiro 1010–12 – Tinigua 1047, 1052 – Urarina 1118, 1119 – Wã́nsöjöt 1159 – Yaruro 1295–6 – Yurakaré 1337–8 Aikanã 669n5, 719, 722, 734 Aktionsart – Urarina 1115 – Yaruro 1306 Alicea Ortiz, Nectalí 995–7, 998, 999, 1006, 1025 alienable possesion – Kwaza 734 – Máku 778–9 – Movima 822–3 – Muniche 862, 882 – Omurano 948 – Pirahã 973 – Taushiro 1008 – Tinigua 1052 – Trumai 1082, 1085 – Urarina 1121 – Wã́nsöjöt 1165 – Wao Terero 1203 alignment see also agreement – Kwaza 753 – Movima 829

– Muniche 880 – Mỹky 930 – Trumai 1103 – Wã́nsöjöt 1169, 1176–7, 1178 – Wao Terero 1223–4 allative – Máku 777 – Taushiro 1007, 1022 – Trumai 1088 – Wã́nsöjöt 1150, 1167, 1177 – Yaruro 1295, 1310 anaphora – Kanoé 664, 673 – Kwaza 727 – Mỹky 907 – Trumai 1085 – Warao 1255 animacy – Kwaza 723, 729, 730 – Máku 780 – Movima 826 – Mỹky 914 – Tinigua 1045, 1050, 1052 – Trumai 1082, 1086, 1087 – Urarina 1117, 1121 – Wã́nsöjöt 1160, 1163 – Wao Terero 1206, 1213–14, 1224 – Yaruro 1288, 1296, 1301 animals – Kanoé 667 – Muniche 862 – Trumai 1083, 1087 – Wao Terero 1214, 1224 – Warao 1250 antipassive, Trumai 1083, 1092 antiperfective, Warao 1268 antonymy, Kanoé 706 Apáitisí (Pirahã) 957–94 apocope, Máku 776 appellative, Kwaza 742 applicative – Kanoé 694, 695 – Kwaza 725, 747, 748 – Movima 832, 833 – Warao 1261 – Yurakaré 1329, 1337–8 apprehensive, Urarina 1117 approximation, Pirahã 971 approximative, Warao 1254

B

argumentation – Kanoé 692 – Kwaza 723, 724, 725, 727, 729–30, 742–4, 747, 750, 756 – Movima 826–7, 834, 835, 836 – Muniche 860–1 – Mỹky 910, 930 – Omurano 947 – Pirahã 964, 965, 967, 969, 973, 981n16 – Taushiro 1010 – Tinigua 1045, 1058–61 – Trumai 1082–3, 1092, 1098, 1103 – Urarina 1131, 1132, 1133 – Wã́nsöjöt 1156, 1157, 1159, 1168–9, 1178–9, 1181–2, 1184 – Wao Terero 1221–3 – Warao 1262, 1273–4 – Yaruro 1295–7 – Yurakaré 1330, 1345 article – Kanoé 664 – Kwaza 729 – Movima 816–20, 837 – Pirahã 967 – Yurakaré 1349 Arutani 769, 772 aspect – Kanoé 667, 689–95 – Kwaza 729, 735, 739–41 – Máku 780, 784–5, 787–90, 794 – Movima 812, 818, 821, 825 – Muniche 860, 871–3 – Mỹky 919–20 – Pirahã 964, 969, 974, 975–9 – Taushiro 1010, 1012–13, 1016 – Tinigua 1044, 1056–7 – Trumai 1089, 1091 – Urarina 1115, 1127 – Wã́nsöjöt 1159, 1160–1, 1169–70, 1172–6, 1183–4 – Wao Terero 1213, 1216 – Warao 1262–5, 1268 – Yaruro 1303, 1312 – Yurakaré 1329, 1339–42, 1346 assertive – Pirahã 976, 981 – Urarina 1127, 1128 – Wã́nsöjöt 1154, 1175 – Wao Terero 1216

Index

associated motion markers – Muniche 876 – Urarina 1123 – Wã́nsöjöt 1172 – Yurakaré 1343 associative – Pirahã 982, 989 – Tinigua 1045, 1046 – Urarina 1114 – Wã́nsöjöt 1168 – Wao Terero 1206 – Warao 1252 atelicity see telicity attenuative, Kwaza 740 attributive – Kwaza 723, 726, 727, 732, 733, 753, 756 – Máku 793 – Trumai 1094, 1096 – Wã́nsöjöt 1165, 1166, 1180, 1184 – Warao 1250, 1253–4, 1260, 1271, 1276 – Yurakaré 1335, 1336 attrition, language – Muniche 856 – Tinigua 1039–40, 1056 augmentative – Kanoé 681–2 – Taushiro 1009 – Tinigua 1053 – Wã́nsöjöt 1161 – Wao Terero 1208 – Yaruro 1289 augmented inclusive, Máku 781 auxiliary – Kanoé 667, 688–9, 710n17 – Trumai 1082, 1089, 1090, 1091, 1094 – Urarina 1130 – Warao 1251, 1260, 1265–6, 1267, 1274, 1276 avertive – Muniche 875 – Yaruro 1306

B Becker-Donner, Etta 661 benefactive – Kanoé 685 – Kwaza 736, 747, 752 – Movima 821 – Mỹky 912 – Taushiro 1017 – Warao 1257, 1261

Index

B

– Yaruro 1295 – Yurakaré 1336, 1338, 1339, 1347 beneficiary relationships – Kanoé 684 – Kwaza 725, 731 bereavement, Máku 781 Bible translations – Taushiro 998 – Wã́nsöjöt 1146 bilabial voiceless fricative /ɸ/, Máku 772, 803 body parts – Kanoé 680, 684, 699 – Kwaza 728, 734 – Máku 779, 784 – Movima 821 – Mỹky 896–7, 915 – Trumai 1082, 1085–6 – Urarina 1121, 1123 – Wã́nsöjöt 1164–5 – Wao Terero 1203, 1212n8 – Warao 1251 body posture, Trumai 1089, 1098 Bontkes, Willem 661 borrowings – Máku 769 – Movima 807, 811, 815 – Muniche 856 – Omurano 948–9 – Pirahã 966 – Tinigua 1033, 1041 – Urarina 1116, 1139 – Wao Terero 1198 – Warao 1245, 1256, 1260, 1279 – Yurakaré 1326, 1336 bound stems – Kanoé 664, 665, 680 – Movima 815 – Wã́nsöjöt 1160 – Wao Terero 1209

C calques – Movima 807 – Muniche 882 case marking – Kwaza 723–4, 729–30, 731, 736, 744–5 – Máku 776, 777–8, 782 – Movima 812, 821, 826 – Mỹky 908, 930 – Pirahã 964, 981

– Taushiro 1007 – Trumai 1097–8 – Urarina 1114 – Wã́nsöjöt 1160, 1166, 1178, 1180, 1188 – Wao Terero 1211, 1223 – Warao 1273 – Yurakaré 1332–3, 1334, 1335 Castellví, Marcelino de 1032 cataphora, Pirahã 975 Caudmont, Jean 1146 causative – Kanoé 708 – Kwaza 735, 738, 747 – Movima 832 – Muniche 860, 877 – Mỹky 934 – Taushiro 1010, 1016–17 – Trumai 1082, 1089, 1092 – Urarina 1115, 1127, 1130–1 – Wã́nsöjöt 1151, 1172 – Wao Terero 1218–19 – Warao 1260, 1261 – Yaruro 1307 – Yurakaré 1329, 1331, 1342, 1344, 1347 causativization – Kwaza 730 – Taushiro 1016 – Trumai 1092 centralization of /a/, Máku 773–4 Chicham 999 Christianization 1145 cislocative, Kwaza 726, 731 classifiers – Kanoé 664, 665–6, 676–81, 692, 714 – Kwaza 725, 727–9, 730, 732, 733, 745 – Máku 784 – Muniche 860, 863–6 – Mỹky 916–17 – Pirahã 963 – Tinigua 1049–52 – Wã́nsöjöt 1160, 1161–2 – Wao Terero 1199, 1203, 1209–12, 1214 – Yurakaré 1336–7 clausal structures – Kwaza 732, 752–5 – Máku 795–8 – Movima 826–9 – Muniche 879–87 – Mỹky 930–5 – Pirahã 980–5

C

– Tinigua 1058–66 – Trumai 1093–8 – Urarina 1132–7 – Wã́nsöjöt 1177–82 – Wao Terero 1223–8 – Warao 1272–9 – Yaruro 1311–14 – Yurakaré 1344–6 clause linking – Kanoé 705, 714–15 – Kwaza 756–63 – Máku 798–803 – Muniche 887–8 – Pirahã 985–91 – Taushiro 1020–2 – Tinigua 1066–8 – Trumai 1099–103 – Urarina 1137–41 – Wã́nsöjöt 1176–7, 1184 – Wao Terero 1228–35 – Yaruro 1314–20 clitics see also enclitics; proclitics – Kwaza 760, 761 – Movima 810, 811, 815–16, 820, 822, 826, 834 – Muniche 860–1, 869, 882, 883, 886 – Pirahã 975, 990 – Taushiro 1014, 1019, 1021 – Tinigua 1044, 1045, 1046, 1047, 1054, 1055 – Trumai 1081, 1083, 1085–6, 1088, 1090, 1097, 1102 – Urarina 1126–7 – Wã́nsöjöt 1154, 1159, 1175 – Wao Terero 1233 – Warao 1253, 1256, 1279 – Yaruro 1293, 1310 – Yurakaré 1331, 1332–3, 1336, 1345, 1348, 1349 collective – Kanoé 666, 668, 671, 674 – Kwaza 726, 729 – Tinigua 1046 – Wao Terero 1206 – Warao 1252, 1253, 1265–6, 1270, 1271 – Yurakaré 1335, 1336–7, 1338, 1350 color terms – Kwaza 732 – Pirahã 970, 991 – Urarina 1123, 1125 – Wã́nsöjöt 1165

Index

– Wao Terero 1208–9 – Yaruro 1294 comitative – Kanoé 684, 685, 694 – Kwaza 725, 731, 736, 748 – Máku 777, 780 – Muniche 881 – Taushiro 1007 – Tinigua 1060 – Trumai 1088, 1101 – Wã́nsöjöt 1168, 1182 – Wao Terero 1211 – Yurakaré 1329, 1332, 1336, 1338, 1339 commandative, Warao 1261 commands see imperative comparative – Kanoé 707, 710–11 – Wã́nsöjöt 1168 – Yurakaré 1335 complement clauses – Kanoé 711–12 – Kwaza 756 – Máku 795, 799 – Movima 835–7 – Muniche 887 – Pirahã 985–6, 987, 988 – Tinigua 1067 – Trumai 1098, 1099–100 – Urarina 1134, 1137, 1138 – Wã́nsöjöt 1183 – Wao Terero 1228 – Warao 1271–2, 1275 – Yaruro 1314, 1315 – Yurakaré 1347, 1348 complementization – Pirahã 986n17 – Wã́nsöjöt 1183 completive – Kwaza 752 – Muniche 872 – Mỹky 919, 924, 925 – Urarina 1127 – Wã́nsöjöt 1161, 1174 – Yurakaré 1336, 1340 compound nouns see also noun incorporation – Kanoé 665 – Kwaza 735 – Máku 793 – Movima 814–15 – Pirahã 968

Index

C

– Tinigua 1046 – Trumai 1083 – Urarina 1123, 1124 – Wã́nsöjöt 1159, 1161–2 – Wao Terero 1209 – Warao 1250 – Yurakaré 1329 compound verbs – Kanoé 709 – Kwaza 735, 746, 751–2 – Pirahã 974–5 – Tinigua 1055 – Trumai 1084 – Wã́nsöjöt 1170 – Wao Terero 1221–3 – Warao 1260 concessive – Kanoé 708, 712–13 – Kwaza 753, 756, 760 – Mỹky 934 – Warao 1278–9 – Yaruro 1317, 1318 conditional – Kanoé 702, 712–13 – Kwaza 753, 756, 759 – Máku 789, 800 – Muniche 874, 888 – Pirahã 988 – Taushiro 1021n12 – Trumai 1100, 1101 – Urarina 1117, 1137 – Wã́nsöjöt 1177 – Wao Terero 1233–5 – Warao 1260, 1263, 1278 – Yaruro 1304, 1318 – Yurakaré 1347 conjectural – Kwaza 738 – Yurakaré 1342 conjunction – Kanoé 713 – Kwaza 759 – Pirahã 989 – Urarina 1116, 1137, 1138–9 – Wã́nsöjöt 1185–6 – Wao Terero 1200, 1230 – Warao 1278 – Yaruro 1308 – Yurakaré 1346 connectors, Yurakaré 1349

consonant elision, Warao 1247 consonant insertion see epenthetic consonants constituent order – Kanoé 668, 682–3, 700–7 – Kwaza 730, 750, 751, 752 – Máku 796–7 – Movima 843–6 – Muniche 866, 869, 880 – Mỹky 907 – Omurano 946–7 – Pirahã 980–1, 991 – Taushiro 1009–10 – Tinigua 1058–61 – Trumai 1097 – Urarina 1107, 1132–4, 1142 – Wã́nsöjöt 1169, 1178, 1180–1 – Wao Terero 1213, 1223, 1224 – Warao 1272–4 – Yaruro 1311–12 – Yurakaré 1344–5 content questions – Máku 797 – Muniche 884–5 – Mỹky 933 – Pirahã 981 – Taushiro 1019 – Tinigua 1063 – Urarina 1126, 1134, 1139–40 – Wã́nsöjöt 1175 – Wao Terero 1224 – Warao 1267–8, 1273 – Yaruro 1313 continuative – Pirahã 978 – Urarina 1127 – Warao 1271, 1272, 1276 contrastive – Kwaza 724, 743, 761 – Warao 1273 converb, Wao Terero 1230–1, 1236 coordinating clauses – Kanoé 707–11 – Kwaza 759–60 – Máku 798 – Movima 825 – Mỹky 934–5 – Trumai 1099 – Urarina 1138–9 – Wã́nsöjöt 1185–6 – Yaruro 1314–15 – Yurakaré 1349

D

co-participant marker, Movima 833 copula – Kanoé 673, 688, 713 – Movima 837, 844 – Muniche 881, 888 – Pirahã 968, 969, 970, 982, 983–5 – Taushiro 1023 – Trumai 1082 – Urarina 1122, 1125 – Wao Terero 1210, 1223, 1226 – Warao 1254, 1259, 1262, 1265–6, 1267, 1268, 1269, 1274–6, 1277 coreferentiality – Máku 798, 802 – Muniche 862 – Pirahã 979 – Tinigua 1046, 1067 – Trumai 1098 – Wao Terero 1214 – Yaruro 1298 – Yurakaré 1348 cosubordination – Kanoé 705, 708, 714–15 – Kwaza 757n6 counterexpectational – Wã́nsöjöt 1174 – Warao 1271 counterfactual – Movima 825 – Wao Terero 1234 – Yaruro 1304, 1319 creak, Wao Terero 1195, 1237

D Daggett, J. 853 dative – Mỹky 912 – Trumai 1082, 1083, 1089, 1099–100 – Wã́nsöjöt 1162–4, 1166 deceased – Movima 817 – Wao Terero 1212 – Yaruro 1289 – Yurakaré 1334 declarative marking – Mỹky 926 – Pirahã 976, 981 – Taushiro 1018 – Urarina 1119 – Wao Terero 1216, 1217

Index

deduction, Pirahã 979 definite articles, Movima 817 definiteness morphology – Mỹky 908–9, 912–13 – Yaruro 1292 degrammaticalization, Kwaza 738, 753 deictic classifiers, Wao Terero 1210 deixis see distal; proximal delimitative, Yaruro 1306 demonstrative – Kanoé 673, 683, 701, 716 – Kwaza 723, 725–6, 729, 743 – Máku 782 – Movima 815, 818–20, 828, 835 – Muniche 861–2, 867 – Pirahã 967 – Tinigua 1044, 1048, 1052 – Trumai 1082, 1085, 1086 – Urarina 1113, 1116, 1117, 1119–20, 1139 – Wã́nsöjöt 1162–4 – Wao Terero 1197, 1201, 1202, 1207 – Warao 1254–5, 1256 – Yaruro 1288, 1291–3, 1309, 1310, 1316, 1317 – Yurakaré 1329, 1333, 1334–5, 1336 denominal verbs, Warao 1259 deontic – Kanoé 704 – Muniche 873 – Tinigua 1057 – Trumai 1091 – Wao Terero 1217 dependent-marking – Kwaza 723–4, 732, 733–4, 757–8 – Muniche 880 – Urarina 1114 deprecative – Kwaza 729, 735, 755 – Yaruro 1289 derivational morphology – Kanoé 665–6, 698 – Kwaza 722, 723, 735, 763 – Máku 777, 780, 791 – Movima 812 – Pirahã 969 – Taushiro 1010 – Tinigua 1052, 1055 – Trumai 1083–5 – Urarina 1122, 1125 – Wã́nsöjöt 1161–2 – Wao Terero 1209–11, 1213

Index

D

– Warao 1250, 1259–60 – Yaruro 1289, 1306 – Yurakaré 1330–1, 1333–4, 1336, 1337 desiderative – Kanoé 689 – Kwaza 737, 738, 739 – Máku 791 – Muniche 875–6 – Pirahã 987 – Taushiro 1014 – Trumai 1091 – Wao Terero 1217, 1227–8 – Warao 1268 – Yaruro 1305 – Yurakaré 1341 determiner – Movima 815, 820, 822 – Mỹky 908 – Trumai 1085 – Warao 1254–7 – Yaruro 1292 detransitivization – Guató see antipassive – Movima 825, 834–5, 840, 842, 843 – Taushiro 1018 deverbal constructions – Máku 780–1 – Movima 826 – Muniche 866, 867 – Warao 1251 dialects – Mỹky 898 – Tinigua 1033 – Wã́nsöjöt 1143 – Warao 1245 – Yaruro 1285 dictionaries – Pirahã 959–60 – Yurakaré 1325 diminutive – Kanoé 681–2 – Máku 781, 791 – Muniche 860, 865–6 – Mỹky 908, 930 – Taushiro 1009 – Tinigua 1053 – Urarina 1129 – Wã́nsöjöt 1161 – Yaruro 1289, 1310 diphthongization, Wã́nsöjöt 1148

direct infix, Movima 814, 830, 831, 838 direct objects – Kanoé 692, 701 – Máku 786–7 – Omurano 947 – Taushiro 1021 – Wã́nsöjöt 1163 – Yurakaré 1338 direct speech, Kwaza 741 directional – Kanoé 694, 696–8 – Kwaza 726, 744–6 – Máku 792 – Pirahã 972–3, 975 – Tinigua 1060 – Trumai 1089, 1090 – Urarina 1117 – Wã́nsöjöt 1160, 1167 – Wao Terero 1219–20 – Yurakaré 1332 directive see also imperative; prohibitive – Muniche 885–7 – Taushiro 1020 discontinuous constructions – Kanoé 699 – Taushiro 1009 – Yurakaré 1331 discontinuous/continuous aspect see habitual; iterative discourse particles – Movima 825 – Pirahã 985 – Wã́nsöjöt 1182 discourse structure – Kanoé 715–16 – Kwaza 760–2 – Máku 802–3 – Urarina 1140–1 – Wao Terero 1235–7 disjunction – Pirahã 989 – Yaruro 1308, 1314 distal – Kanoé 673 – Kwaza 725 – Máku 782 – Muniche 869 – Pirahã 967 – Taushiro 1025 – Tinigua 1048

E

– Trumai 1086 – Wã́nsöjöt 1163 – Wao Terero 1201 – Warao 1255 – Yaruro 1291 – Yurakaré 1335 distributive – Kwaza 740, 746–7 – Máku 791 – Urarina 1127 – Wao Terero 1206 – Warao 1253, 1258 – Yurakaré 1329, 1340 double negation, Kanoé 706 dual – Mỹky 909 – Trumai 1086, 1087 – Urarina 1117, 1119 – Wao Terero 1201–2, 1213–14 dubitative – Kanoé 691 – Trumai 1091 – Wao Terero 1217, 1225 durative – Movima 825 – Mỹky 925, 927

E ejective consonants, Trumai 1079 elative, Máku 777, 778n9 ellipsis – Kanoé 710, 714 – Kwaza 742, 759, 762–3 – Yurakaré 1330 embedded clauses – Movima 836–7, 838 – Pirahã 985, 986–7, 991 emphatic marking – Urarina 1136–7 – Warao 1268, 1271 empty root construction, Kwaza 734 enclitics – Máku 797, 800, 801, 802 – Movima 811, 822, 824, 826, 830, 832, 835, 844 – Muniche 869, 870 – Trumai 1081, 1085–6, 1088, 1090, 1097, 1099–100 – Urarina 1114, 1126–7, 1128, 1134, 1139–40 – Wao Terero 1211

Index

– Warao 1253, 1256, 1279 – Yaruro 1293, 1310, 1316 – Yurakaré 1331, 1332–3, 1345, 1348 endangered languages – Kanoé 659 – Movima 807 – Pirahã 959 – Tinigua 1029, 1035 – Trumai 1077 – Yurakaré 1325 endocentric compounds, Yurakaré 1329 endocentric denominal derivation – Taushiro 1009 – Trumai 1083 endoclitics, Tinigua 1045 enumerative, Urarina 1129 epenthetic consonants – Movima 812 – Pirahã 987n18 equative – Máku 795 – Pirahã 983 – Tinigua 1065–6 – Warao 1254, 1274, 1276 ergative – Movima 829 – Trumai 1082–3, 1089, 1092, 1097, 1104 – Wã́nsöjöt 1160, 1166, 1179, 1182, 1188 Esquini, Pietro 941 evaluative, Kwaza 729 eventive, Wã́nsöjöt 1177 evidentiality – Kanoé 691 – Kwaza 741–2 – Máku 785, 787–90, 794 – Movima 818, 825 – Mỹky 910, 918, 922–4 – Pirahã 964, 969, 974, 975–9 – Tinigua 1057 – Urarina 1128, 1140 – Wã́nsöjöt 1160–1, 1169–70, 1172–6, 1183–4 – Wao Terero 1217–18 – Warao 1279 – Yaruro 1305, 1312 – Yurakaré 1330, 1339–42, 1348 evocative, Yurakaré 1334 exclamative – Kwaza 755 – Máku 797–8

Index

E

exclusive – Máku 778 – Urarina 1117, 1119 – Wao Terero 1202 exhaustive, Muniche 863 exhortative, Kanoé 700, 702, 704 existential – Kwaza 734, 745 – Máku 795 – Movima 828, 838–9 – Muniche 883, 884 – Pirahã 983, 984 – Taushiro 1024 – Tinigua 1065–6 – Wao Terero 1227 – Warao 1252, 1269, 1274, 1275 explicative, Kanoé 708, 709 extension-to-core argument – Máku 786 – Yaruro 1297, 1309 extinct languages, Máku 767 extractivist industries 998, 1193

F family terminology see kinship terms Fast, Gerhard 944 faucalization, Wao Terero 1237 Figueroa, Francisco de 855, 1001n3 finality, Kanoé 712–13 firsthand evidential, Máku 787, 795n23 focus marking – Kanoé 672, 701 – Kwaza 752, 760–1 – Máku 802 – Muniche 888 – Pirahã 990 – Trumai 1082, 1088, 1090 – Urarina 1126, 1133, 1139–40, 1141 – Wã́nsöjöt 1179, 1186–7 – Wao Terero 1233–4 – Warao 1273 frequentative – Kwaza 740, 746–7 – Trumai 1081 – Yurakaré 1336 fricativization, Yaruro 1286 fronting – Movima 844–6 – Muniche 884 – Mỹky 933

– Urarina 1133, 1134 – Warao 1273 frustrative – Kanoé 691 – Kwaza 738 – Movima 825 – Muniche 875 – Mỹky 925, 926 – Pirahã 978, 981 – Warao 1268 – Yurakaré 1341, 1342, 1345 fusion – Movima 812 – Wao Terero 1200 future – Kanoé 689–95, 702 – Kwaza 736–7, 744 – Máku 787, 799 – Movima 825 – Muniche 871, 873, 874 – Mỹky 918, 923, 932 – Pirahã 976 – Taushiro 1010, 1012–13 – Trumai 1101 – Urarina 1127, 1135 – Wã́nsöjöt 1173 – Wao Terero 1215, 1234 – Warao 1262–3, 1269 – Yaruro 1303, 1318 – Yurakaré 1339–40, 1341

G gender marking – Kanoé 671, 673, 676, 677, 683 – Kwaza 727 – Máku 785 – Movima 812, 821 – Muniche 860, 863 – Mỹky 908, 909, 913–14, 921, 925 – Pirahã 973 – Tinigua 1045, 1050 – Trumai 1086 – Wã́nsöjöt 1158, 1160–1, 1163 – Wao Terero 1202, 1213–14 – Yaruro 1288, 1289, 1291, 1296, 1297, 1298, 1301 – Yurakaré 1337 genderlects, Pirahã 962–3, 991 generic classifier, Kwaza 756n5 generic nouns, Pirahã 966–7, 975

H

genitive – Trumai 1085 – Yaruro 1309 gesture – Kwaza 726 – Pirahã 967 – Warao 1255 glides – Kwaza 722 – Máku 772–3 – Pirahã 963 – Taushiro 1003, 1011 – Tinigua 1036 – Urarina 1111, 1112 – Wã́nsöjöt 1147, 1148 glottal fricative – Kwaza 722 – Máku 773 – Mỹky 900 – Taushiro 1004–5 – Tinigua 1041 – Urarina 1111 glottal stops – Kanoé 662 – Kwaza 722 – Máku 773, 776 – Movima 810, 811, 816 – Muniche 858, 865 – Mỹky 900 – Pirahã 961, 962, 985 – Taushiro 1005n6, 1011n7 – Tinigua 1040, 1041 – Wã́nsöjöt 1151 – Yaruro 1287 goal – Muniche 880 – Warao 1253, 1256, 1258 – Yurakaré 1338 Goodall, Harold 853 grammars – Pirahã 960 – Urarina 1109 – Wã́nsöjöt 1146 – Wao Terero 1193 – Yurakaré 1325 grammaticalization – Kwaza 746, 747 – Mỹky 930 – Wã́nsöjöt 1166, 1172 – Wao Terero 1200, 1223, 1233

Index

– Warao 1260, 1268 – Yaruro 1307 – Yurakaré 1334 Greenberg, Joseph H. 769, 1077, 1193, 1285 greetings, Urarina 1126 grounds, Kwaza 745 Guahiban family 1032–3 Guaporé cultural complex/Marico cultural complex 661, 719, 721, 734, 897–8 Guaraní 807

H habitual – Kwaza 740 – Movima 825 – Muniche 866, 872n15 – Mỹky 932 – Pirahã 976, 977 – Trumai 1083, 1091 – Urarina 1127 – Wã́nsöjöt 1161 – Wao Terero 1226 – Warao 1262 – Yaruro 1307, 1318 – Yurakaré 1341 Hanke, Wanda 959 haplology, Máku 775 head marking – Kwaza 723–4, 734 – Máku 777 – Muniche 860, 862 – Mỹky 906–7 – Urarina 1114 – Yurakaré 1329 headless noun phrases – Kwaza 733 – Máku 782 hearsay evidential – Pirahã 979 – Urarina 1128, 1140 – Warao 1279 Hervás y Panduro, Lorenzo 855, 939, 942, 1285 hierarchical indexation – Movima 826, 827, 829, 835 – Wao Terero 1203–4 high rounded front vowel /y/, Máku 772, 803 hodiernal past, Movima 817, 825 hollow voice, Wao Terero 1237 honorific, Wao Terero 1204, 1213–14

Index

H

hortative – Kwaza 738, 739, 755 – Máku 789 – Muniche 874 – Taushiro 1020 – Urarina 1117, 1134 – Wã́nsöjöt 1175 – Warao 1267 human versus non-human referents – Kanoé 671, 676, 677, 683 – Movima 817, 819, 836 – Mỹky 914, 921 – Pirahã 973, 974 – Tinigua 1046 – Trumai 1087 – Urarina 1119, 1121 – Wã́nsöjöt 1160 – Wao Terero 1202, 1206, 1213–14, 1223, 1224, 1228–9 – Warao 1273–4 – Yaruro 1289, 1296 hummed speech 960 Hup 793n21 hydronymic, Taushiro 1009 hypotaxis, Kanoé 711 hypothetical see also irrealis – Wao Terero 1234 – Yaruro 1304, 1318, 1319 – Yurakaré 1341

I identificational focus, Kwaza 724, 743, 761 identificative, Mỹky 924, 925 ideophone – Kwaza 723 – Máku 776 – Trumai 1082 – Urarina 1110, 1115, 1139, 1141 – Wao Terero 1194, 1200, 1222–3, 1235 – Warao 1248, 1251, 1260, 1268, 1270–1, 1272, 1274, 1276 – Yaruro 1287 – Yurakaré 1330, 1331, 1342, 1343–4 illative – Kwaza 752 – Taushiro 1007 illocutionary markers, Pirahã 976 imperative see also hortative; jussive; prohibitive – Kanoé 688, 700, 703–4 – Kwaza 743

– Máku 785, 789, 790, 797–8 – Movima 831–2 – Muniche 874, 876, 885–7 – Mỹky 918, 923 – Omurano 947 – Pirahã 978, 981 – Taushiro 1020 – Tinigua 1054, 1064–5 – Trumai 1082, 1093, 1096 – Urarina 1119, 1134–5 – Wã́nsöjöt 1175 – Wao Terero 1217, 1225 – Warao 1264, 1265–7, 1269–70 – Yaruro 1295, 1301, 1304, 1312–13 – Yurakaré 1329, 1341, 1345 imperfective – Máku 788 – Muniche 871, 872, 883 – Mỹky 919, 920 – Pirahã 978 – Trumai 1091 – Wã́nsöjöt 1174 – Wao Terero 1216 – Yaruro 1303 impersonal, Kwaza 749 implosives, Kwaza 721 inalienability – Kanoé 684 – Kwaza 734 – Máku 778–9 – Movima 821, 822–3 – Muniche 862 – Taushiro 1008 – Trumai 1082, 1085 – Urarina 1121 – Wã́nsöjöt 1157, 1160, 1164–5 – Wao Terero 1203 inanimacy see also animacy – Kwaza 753 – Movima 835 – Mỹky 914 – Pirahã 973 inceptive, Wao Terero 1217 inchoative – Kanoé 689 – Muniche 875 – Trumai 1091 – Wã́nsöjöt 1174 inclusive – Trumai 1086 – Urarina 1117, 1119

I

– Wã́nsöjöt 1163 – Wao Terero 1201–2 incompletive – Kwaza 738 – Tinigua 1056 incorporated noun, Kanoé 698–9 incorporating verbs – Kanoé 687, 692 – Kwaza 750 – Wao Terero 1214 indefinite person marker, Kwaza 741 indefinite pronoun – Kanoé 671–2, 701 – Kwaza 725, 729 – Tinigua 1047 – Wã́nsöjöt 1162–4, 1169 indirect objects, Kwaza 730 indirect speech markers see reported speech inessive – Máku 777, 782 – Wã́nsöjöt 1167 inferential – Kwaza 741 – Mỹky 922, 924 – Pirahã 979 – Wao Terero 1215, 1218, 1225 – Yurakaré 1341, 1342 infinitive – Urarina 1138 – Warao 1278–9 infixation – Máku 777 – Movima 813–14, 822, 823 – Yaruro 1288 – Yurakaré 1329 inflectional morphology see also case marking; gender marking; number marking; person marking – Kanoé 664, 667–8, 692 – Kwaza 722, 723, 729 – Máku 777, 785 – Movima 812 – Mỹky 907, 918, 924–6, 928–9 – Taushiro 1010 – Trumai 1081 – Urarina 1116 – Wã́nsöjöt 1156 – Wao Terero 1213 – Warao 1276 – Yaruro 1303 – Yurakaré 1333

information structure – Kanoé 700, 715–16 – Kwaza 760–2 – Máku 802–3 – Muniche 888 – Omurano 946n8, 947 – Pirahã 989–90 – Tinigua 1045, 1059 – Trumai 1090–1 – Urarina 1139–40 – Wao Terero 1201, 1236 – Yaruro 1291, 1293 – Yurakaré 1349 ingressive – Pirahã 972 – Wao Terero 1237 inherency, Máku 788 in-law talk, Urarina 1127, 1128–9 instrumental – Kanoé 685 – Kwaza 726, 731, 736 – Máku 777 – Movima 823 – Muniche 866, 881 – Mỹky 907 – Pirahã 964 – Taushiro 1007 – Tinigua 1060 – Trumai 1088 – Wã́nsöjöt 1168, 1177 – Wao Terero 1211 – Yaruro 1316 – Yurakaré 1332 intensifier – Kanoé 675, 681 – Kwaza 729 – Máku 791 – Tinigua 1058 – Trumai 1089, 1094 – Wao Terero 1221 – Warao 1259, 1262 – Yurakaré 1335, 1342 intensive, Kwaza 740, 741 intentional – Máku 799 – Movima 825 – Mỹky 927 – Wao Terero 1217, 1234 – Yurakaré 1341

Index

Index

I

interjection – Kwaza 755 – Movima 825 – Pirahã 985 – Trumai 1082 – Urarina 1116, 1136 – Wao Terero 1237 – Yaruro 1308 – Yurakaré 1330, 1331 intermarriage 854, 895, 942, 943, 1193 interrogatives – Kanoé 672, 688, 689, 700–2, 703–4 – Kwaza 725, 729, 735, 742, 754, 755 – Máku 797–8 – Movima 843–4 – Muniche 872, 884–5 – Mỹky 918, 923, 932–4 – Pirahã 976, 978, 981–2, 985 – Taushiro 1019, 1023–4 – Tinigua 1062–4 – Trumai 1082, 1097 – Urarina 1128, 1134–5 – Wã́nsöjöt 1162–4, 1175 – Wao Terero 1210, 1224–5 – Warao 1265–6, 1267–8 – Yaruro 1313 – Yurakaré 1335, 1345 intonation see also prosody; stress – Kanoé 663, 702, 703, 704 – Kwaza 762 – Máku 797n27 – Movima 811–12 – Muniche 884 – Mỹky 932 – Pirahã 981, 990 – Tinigua 1062 – Trumai 1099 – Wao Terero 1222, 1224, 1235, 1236–7 – Warao 1267–8 – Yaruro 1313 – Yurakaré 1345 intransitivity see also transitivity; valency – Kanoé 669, 670, 680, 687, 688, 693 – Máku 784, 793, 796 – Movima 816, 821, 823, 827–9, 843 – Muniche 877–8 – Mỹky 910–11 – Taushiro 1010–11, 1016–17 – Trumai 1082–3, 1093 – Urarina 1115, 1122–3, 1125, 1130, 1133

– Wã́nsöjöt 1166, 1179, 1182 – Warao 1259, 1261, 1262 – Yaruro 1296, 1301, 1307, 1312 – Yurakaré 1337–8 intransitivizer, Urarina 1115, 1130 introducers, Urarina 1116, 1117, 1135 inverse, Movima 830, 831, 836, 838 irrealis – Kwaza 738, 753–4 – Movima 814, 821, 838–9 – Muniche 871, 874–5, 876, 883, 886, 888 – Mỹky 910, 918, 923 – Urarina 1117, 1125, 1127 – Wao Terero 1217, 1225 – Yurakaré 1340–1, 1347, 1348 isolating languages, Trumai 1081 iterative – Kanoé 666 – Kwaza 740 – Mỹky 918, 924, 925 – Trumai 1081 – Warao 1260, 1270 – Yaruro 1306

J Jesuits 855, 893, 895, 939, 941–2, 999–1003, 1031, 1033 Judy, Robert and Judith 809 jussive – Urarina 1134 – Warao 1267 – Yurakaré 1341 juxtaposition – Kanoé 665, 707, 709, 710 – Kwaza 723, 732, 733–4, 759 – Movima 816, 824, 845 – Muniche 881 – Mỹky 908, 914–15, 928–9, 934–5 – Pirahã 989 – Taushiro 1021 – Tinigua 1046, 1055, 1065, 1066, 1067 – Trumai 1099 – Urarina 1116, 1121 – Wã́nsöjöt 1161 – Yurakaré 1349

K Kakua-Nukakan 1145 Kandozi-Chapra 939, 942, 948–9 Kanoé 659–718, 719

M

Kichwa 1193 kinship terms – Kanoé 677 – Máku 779 – Movima 822, 823 – Mỹky 921, 922 – Pirahã 973 – Tinigua 1046 – Trumai 1082 – Urarina 1121, 1123, 1128–9 – Wã́nsöjöt 1164–5 – Wao Terero 1204, 1212 Kwaza 719–66

L labialization, Urarina 1110, 1111 lamentative, Kwaza 729 language contact – Máku 769 – Movima 807 – Muniche 855–6 – Mỹky 899 – Pirahã 966 – Tinigua 1034 – Trumai 1086 – Urarina 1107 – Wã́nsöjöt 1145 – Warao 1273 Lasinger, Fulgencio 1325 lative, Máku 782 Les Corts, Estanislao 1031, 1035 lingua francas see Portuguese; Spanish Língua Geral 966 linguistic areality – Kwaza 728 – Máku 803 – Mỹky 897 linguistic diversity – Muniche 855 – Mỹky 897 liquids, Muniche 857 litotic, Kanoé 675, 682, 702, 706–7, 710, 716 loanwords see borrowings location see also distal; proximal – Kwaza 725–6, 730 – Pirahã 972–3 – Taushiro 1023–4 – Urarina 1114 – Wã́nsöjöt 1167 – Wao Terero 1219–20

Index

locative – Kanoé 685 – Kwaza 731, 736, 745 – Máku 777, 782, 795 – Movima 828, 829 – Muniche 868–9, 881 – Mỹky 907 – Pirahã 964, 983, 984 – Taushiro 1007, 1023–4 – Tinigua 1060–1 – Trumai 1086 – Urarina 1117, 1123 – Wã́nsöjöt 1177 – Wao Terero 1211 – Warao 1253, 1254, 1256, 1257, 1258, 1261, 1274, 1275 – Yaruro 1310, 1316, 1318, 1319 – Yurakaré 1332, 1334 Lord’s Prayer translations 1285 Loukotka, Čestmir 853, 948, 957, 959, 1032, 1033, 1283

M Maciel, Iraguacema Lima 770 Macro-Jê 1079 Macro-Tukanoan 854 Magnin, Juan 941 Mako 767n1 Makú 767n1, 1145, 1146 Máku 767–806 malefactive – Kwaza 748 – Movima 833 – Yurakaré 1338, 1339 Manoki 893–6 Manresa, Fructuoso de 1031, 1032, 1035 Marico cultural complex/Guaporé cultural complex 661, 719, 721, 734, 897–8 Meader, Robert E. 895–6, 902 medial clauses, Kwaza 740, 758, 759 meteorological terms, Kanoé 688, 693 middle voice, Yurakaré 1342, 1344 Migliazza, Ernesto C. 770 mirative – Wã́nsöjöt 1155, 1174 – Yurakaré 1341, 1342 missionaries – Kwaza 721 – Muniche 855 – Mỹky 893, 895

Index

M

– Omurano 941 – Pirahã 959–60 – Taushiro 999–1003 – Tinigua 1031, 1032, 1033, 1035 – Wao Terero 1193 – Yurakaré 1325 mitigation, Kanoé 705 mood/modality see also sentential mood – Kanoé 664, 667, 688–9, 691, 700, 702–5 – Kwaza 723, 736, 738–9, 742–4, 757, 758, 759 – Máku 785, 787–90, 794 – Movima 812, 821, 825 – Muniche 860, 873 – Mỹky 918 – Pirahã 964, 969, 974, 975–9 – Taushiro 1014 – Tinigua 1044, 1054, 1056–7 – Trumai 1089, 1091 – Urarina 1115, 1128 – Wã́nsöjöt 1160–1, 1169–70, 1172–6, 1183–4 – Wao Terero 1213, 1217–18 – Warao 1265–6 – Yaruro 1304, 1312 – Yurakaré 1330, 1339–42 Moore, Denny 661 motion semantics – Máku 794 – Wao Terero 1219 Moura, José de 896 Movima 807–49 multilingualism 1107, 1145, 1193, 1245 Muniche 851–91 Muran family 957, 959 Mỹky 893–937

N names – Kwaza 726 – Mỹky 914 – Tinigua 1046 – Wao Terero 1206, 1212 – Warao 1252 nasalization – Kanoé 663–4 – Kwaza 722 – Máku 773 – Muniche 857 – Mỹky 902, 904, 931

– Pirahã 962 – Taushiro 1003–4, 1006 – Urarina 1112, 1113 – Wã́nsöjöt 1149 – Wao Terero 1195–7 – Warao 1247 – Yurakaré 1326 negation – Kanoé 682, 701, 704, 705–6 – Kwaza 750, 754–5, 759 – Máku 775, 790–1 – Movima 825, 837–9, 841, 845 – Muniche 883 – Mỹky 922, 931–2 – Pirahã 981, 982–3 – Taushiro 1010, 1014, 1015–16 – Tinigua 1054, 1057, 1061–2 – Trumai 1081, 1093, 1094–6 – Urarina 1126, 1127, 1135–6 – Wã́nsöjöt 1156, 1170, 1182 – Wao Terero 1200, 1225–8 – Warao 1267, 1269–70, 1274 – Yaruro 1305, 1313–14 negative affect, Urarina 1128, 1140 Nimuendajú, Curt 661 nominal modification see adjectives; classifiers nominalization – Kanoé 664, 666, 682, 690, 712 – Kwaza 723, 725, 726, 727, 729, 731, 732, 733–4, 739–40, 750, 751, 756 – Máku 780, 799 – Movima 821, 836, 838 – Muniche 864, 866–7 – Mỹky 917, 929 – Pirahã 965, 969, 980, 987–8 – Tinigua 1042, 1045, 1046, 1050, 1055 – Trumai 1083 – Urarina 1115, 1122–3, 1129–30 – Wã́nsöjöt 1155, 1161, 1176–7, 1184, 1188 – Wao Terero 1210 – Warao 1250, 1252, 1277 – Yaruro 1288–9, 1303, 1304, 1315, 1316, 1318 – Yurakaré 1335 nominative – Kwaza 753 – Muniche 880 – Omurano 947 – Taushiro 1009–10 – Trumai 1098

P

– Wao Terero 1223 – Yaruro 1290, 1291–2, 1293, 1310 nonfinite marking – Wã́nsöjöt 1157, 1168–9, 1177, 1179 – Wao Terero 1230 non-present demonstrative, Kanoé 673–4, 702, 716 Nordenskiöld, Erland 721 noun incorporation see also compound nouns – Movima 814–15 – Muniche 860, 878–9 – Taushiro 1010, 1016 – Trumai 1082 – Wã́nsöjöt 1159, 1181–2 null affixes – Kanoé 669, 684 – Muniche 872 – Mỹky 907, 913 number marking see also plurality – Kanoé 664, 674 – Kwaza 725, 726–7, 729, 743 – Máku 777, 778, 780, 785, 794 – Movima 812, 821 – Muniche 860–1, 865 – Mỹky 909–12 – Pirahã 973 – Taushiro 1007, 1009 – Tinigua 1045–6, 1049–50, 1054, 1055–6 – Urarina 1114, 1115, 1116, 1118–19 – Wã́nsöjöt 1158, 1160–1, 1177 – Wao Terero 1205–7 – Warao 1252, 1255, 1262 – Yaruro 1288, 1289, 1298 – Yurakaré 1332–3 numerals – Kanoé 674–5 – Kwaza 726–7 – Máku 782–4 – Movima 824 – Muniche 867–8 – Mỹky 908 – Pirahã 971–2, 974, 991 – Taushiro 1009 – Tinigua 1044, 1049–50 – Trumai 1082, 1085 – Urarina 1114, 1116, 1129 – Wã́nsöjöt 1166 – Wao Terero 1205 – Yaruro 1294, 1308–9 – Yurakaré 1336

Index

O Oberg, Kalervo 896 object function – Kwaza 725, 730 – Máku 786–7 – Tinigua 1059 object marking – Kwaza 724, 729, 753 – Muniche 869, 870, 883 – Urarina 1118 – Wao Terero 1210, 1211 – Warao 1250, 1273 – Yaruro 1295, 1299–300, 1309 – Yurakaré 1333, 1338 obligative – Guató see deontic – Yurakaré 1341 oblique – Kanoé 667, 683, 685–6 – Kwaza 730, 752 – Máku 796 – Movima 812, 820, 826, 829 – Muniche 877, 880–1 – Pirahã 964 – Urarina 1131 – Wã́nsöjöt 1160 – Wao Terero 1211 – Warao 1261 – Yaruro 1290, 1291–2, 1293, 1316 obviative, Movima 811, 816, 820, 826, 827–9, 830, 832, 833, 836, 840, 845 Olivares, Antonio 1035 Omurano 939–55, 1002n4 onomatopoeia – Kanoé 667 – Urarina 1115 – Warao 1246 operative, Warao 1264 optative – Movima 825 – Yaruro 1304 orthography – Mỹky 906 – Omurano 950 – Wao Terero 1194 – Yaruro 1285 – Yurakaré 1326, 1327

P palatalization – Máku 772–3 – Muniche 865–6

Index

P

– Mỹky 901, 903, 915 – Omurano 952 – Pirahã 961 – Tinigua 1039 – Urarina 1110, 1111, 1112 – Warao 1247 Pamigua 1032–3 participant marking see also person marking – Pirahã 990 participles – Taushiro 1017, 1018 – Urarina 1137 – Wã́nsöjöt 1157 particles – Kanoé 664, 705, 709, 710, 714 – Kwaza 723, 755, 760, 763 – Máku 797, 802–3 – Movima 821, 825, 838 – Muniche 883, 884 – Pirahã 985, 989 – Taushiro 1020–1 – Tinigua 1065 – Trumai 1082, 1088, 1089, 1090, 1092, 1094, 1096 – Urarina 1136–7, 1140, 1141 – Warao 1249n1 – Yurakaré 1330, 1331, 1343–4, 1345 passive – Kwaza 749 – Muniche 877 – Taushiro 1017 – Trumai 1092 – Urarina 1115, 1129–30 past tense – Kanoé 690 – Kwaza 729, 737 – Máku 787 – Movima 817, 825 – Muniche 871 – Mỹky 910, 918, 919, 932 – Pirahã 976, 977 – Trumai 1091 – Wã́nsöjöt 1161, 1173, 1174 – Wao Terero 1215 – Warao 1248, 1262–3, 1269 – Yaruro 1303 – Yurakaré 1339 patient-like arguments, Movima 829 paucal – Kwaza 729 – Pirahã 971, 974

pausative, Yaruro 1306 Peeke, M. Catherine 1193 Pereira, Adalberta de Holanda 896 perfective – Máku 788 – Muniche 871, 872, 883 – Mỹky 919 – Pirahã 978 – Trumai 1091 – Wã́nsöjöt 1173 – Wao Terero 1216, 1234 – Yaruro 1303, 1319 – Yurakaré 1340 perlative, Wã́nsöjöt 1167 permanency – Máku 776, 780, 788, 789 – Trumai 1094 permansive, Mỹky 925, 926 person hierarchies see hierarchical indexation person marking – Kanoé 664, 667, 669, 670, 686–7, 688–9, 690, 691–2, 715, 716 – Kwaza 723, 724–6, 729, 737–8, 742–4, 759 – Máku 778, 785, 794 – Muniche 854, 856, 860–1, 863, 869–70, 886 – Mỹky 909–12, 918 – Taushiro 1010–12 – Tinigua 1047, 1054, 1055–6 – Trumai 1086, 1090 – Urarina 1115, 1125–6, 1127, 1135, 1141 – Wã́nsöjöt 1182 – Wao Terero 1201–5, 1213–14, 1223, 1227 – Warao 1249–50, 1257, 1275 – Yaruro 1298 – Yurakaré 1333, 1337–8 personal pronouns – Kanoé 692 – Kwaza 724–6, 743 – Máku 781, 783 – Movima 816–20, 835 – Mỹky 909, 931 – Trumai 1086 – Urarina 1117, 1119 – Wã́nsöjöt 1162–4 – Yaruro 1288, 1290–1 persuasive, Kwaza 754 Pirahã (Apáitisí) 957–94 pluractional see also iterative – Kwaza 746–7 – Movima 814

P

– Muniche 873 – Warao 1262 – Yurakaré 1340 plurality see also number marking – Kanoé 666, 671, 674 – Kwaza 726 – Máku 777, 778, 780, 781, 786, 794 – Movima 818, 832 – Muniche 865 – Mỹky 909, 911, 924, 926 – Pirahã 973 – Tinigua 1044, 1045–6, 1048n6 – Trumai 1082, 1085, 1086, 1087 – Urarina 1114, 1116, 1118–19, 1123, 1125, 1127 – Wã́nsöjöt 1151, 1160–1, 1163, 1177 – Wao Terero 1197, 1201–2, 1205–6 – Warao 1252, 1254–5 – Yaruro 1288, 1290, 1291–2, 1301–2 – Yurakaré 1332–3, 1336–7, 1350 polar questions – Máku 797, 798 – Muniche 884 – Pirahã 981 – Taushiro 1019 – Tinigua 1062–3 – Trumai 1097 – Urarina 1126, 1134 – Wã́nsöjöt 1175 – Wao Terero 1224–5, 1227–8, 1236–7 – Warao 1267–8 – Yaruro 1313 – Yurakaré 1345 politeness marking – Urarina 1119, 1128–9 – Yurakaré 1342 polymorphemic verb bases, Wã́nsöjöt 1170–1 polysynthetic languages – Kanoé 664 – Kwaza 723 – Mỹky 906–7, 910, 918, 924, 935 – Urarina 1114, 1127 – Wã́nsöjöt 1159 – Wao Terero 1199 – Yurakaré 1329 portmanteau forms – Muniche 872n16 – Yaruro 1297 Portuguese – Kanoé 700 – Máku 784n17

Index

– Mỹky 895, 899 – Pirahã 959, 966 – Trumai 1086 possession see also inalienability – Kanoé 667–8, 670–1, 680, 684 – Kwaza 723, 725, 733–4, 748 – Máku 777, 778–80 – Movima 815, 816, 818, 820, 822–3, 826, 828, 838, 842 – Muniche 860, 862–3, 866 – Mỹky 903, 909, 914–16 – Omurano 948 – Pirahã 965, 968, 981, 984, 985–6 – Taushiro 1008, 1024 – Tinigua 1052–3 – Trumai 1085, 1102 – Urarina 1114, 1120, 1121–2, 1123 – Wã́nsöjöt 1157, 1160, 1163, 1164–5, 1177 – Wao Terero 1203 – Warao 1250 – Yaruro 1291, 1298, 1308–9 – Yurakaré 1333, 1335 possessive – Kanoé 670–1, 683, 686 – Kwaza 733–4 – Máku 778 – Movima 823, 826, 828, 836 – Muniche 860, 862–3 – Mỹky 915 – Omurano 946 – Pirahã 973 – Taushiro 1008 – Trumai 1099–100, 1101 – Urarina 1114, 1121 – Warao 1251–2, 1253, 1257, 1259, 1274, 1275 – Yaruro 1309 – Yurakaré 1329, 1333 postpositions – Kanoé 676 – Máku 776, 801 – Muniche 880 – Mỹky 907, 909, 914–15, 916, 917–18 – Pirahã 972–3 – Taushiro 1007, 1010, 1023–4 – Trumai 1082, 1088 – Urarina 1116, 1131 – Wao Terero 1200, 1211–12 – Warao 1253, 1257–9, 1272, 1275 – Yaruro 1290, 1305, 1309–10, 1315, 1319

Index

P

potential – Kwaza 738 – Trumai 1101 – Wã́nsöjöt 1174 – Warao 1269 – Yaruro 1304 pragmatics – Kwaza 733, 735–6, 737, 742, 752, 761 – Movima 843–6 – Pirahã 990 – Trumai 1087–8, 1098, 1104 – Wã́nsöjöt 1146, 1155, 1176–7 – Wao Terero 1235 predication – Kanoé 688, 693–4, 715–16 – Kwaza 736 – Máku 800–1 – Movima 811, 818, 821–2, 824, 826, 827–9, 836, 843–6 – Muniche 881–2 – Mỹky 916, 929, 931–2 – Pirahã 969, 983, 987 – Taushiro 1009, 1022–4 – Tinigua 1046, 1065 – Trumai 1082, 1093 – Wã́nsöjöt 1157, 1180 – Wao Terero 1199, 1208, 1212, 1221–3 – Warao 1251, 1260, 1261, 1266, 1271, 1274, 1279 – Yaruro 1290, 1312, 1316, 1318 – Yurakaré 1330, 1335, 1336, 1340, 1343–4, 1347 prefixation – Kanoé 664, 669 – Kwaza 723 – Máku 775, 777, 792 – Movima 812 – Mỹky 906, 918 – Omurano 946, 948 – Taushiro 1010 – Tinigua 1044, 1054, 1061 – Urarina 1114 – Wã́nsöjöt 1155, 1157, 1164–5, 1169, 1178 – Warao 1249, 1260 – Yurakaré 1329 prepositions – Movima 826 – Muniche 880 presentational, Taushiro 1024–5 privative, Trumai 1095

proclitics – Movima 820 – Tinigua 1047, 1054, 1055 – Urarina 1114, 1116, 1117, 1121 – Yurakaré 1336, 1349 procrastinative, Kwaza 738 pro-drop languages, Yurakaré 1349 progressive – Kanoé 667, 690 – Kwaza 739–40 – Máku 788, 789 – Muniche 871, 872, 883 – Mỹky 919, 920 – Pirahã 977 – Tinigua 1056 – Warao 1262, 1263, 1278 – Yurakaré 1340 prohibitive – Kanoé 704, 706 – Kwaza 754 – Máku 775, 789, 790 – Movima 838 – Muniche 861, 885, 887 – Taushiro 1020 – Urarina 1128, 1135, 1136 – Wã́nsöjöt 1175 – Yurakaré 1345 prolative, Máku 777, 778n9 pronominal systems see also demonstrative – Kanoé 668–74, 676, 715 – Kwaza 724–6, 743 – Máku 781, 782, 783 – Movima 816–20, 843 – Muniche 861–2 – Mỹky 907 – Pirahã 965–7, 973, 975, 981, 990 – Taushiro 1007, 1019, 1025 – Tinigua 1047–50 – Trumai 1082, 1085–7 – Urarina 1116, 1117, 1119 – Wã́nsöjöt 1156, 1157, 1162–4, 1169, 1178 – Wao Terero 1201–2, 1203, 1228–9 – Warao 1249–50, 1254, 1256, 1273 – Yaruro 1290–4 – Yurakaré 1333–5 propositional attitude, Pirahã 968, 976 prosody see also intonation; stress; tone – Mỹky 900, 905 – Pirahã 990 – Taushiro 1005–6, 1021

R

– Urarina 1113–14 – Yurakaré 1327–8 prospective – Trumai 1091 – Wã́nsöjöt 1174 proximal – Kanoé 673 – Kwaza 725 – Máku 782 – Movima 811, 815, 818, 826–9, 830, 832, 833, 836, 840, 841, 844, 845 – Muniche 861, 869 – Pirahã 967 – Taushiro 1024–5 – Tinigua 1048 – Wã́nsöjöt 1163 – Wao Terero 1201 – Warao 1255 – Yaruro 1291 – Yurakaré 1335 Puinave see Wã́nsöjöt (Puinave) Pumé (Yaruro) 1283–322 purposive – Kwaza 738, 739 – Máku 789, 799, 800 – Muniche 887 – Pirahã 987 – Taushiro 1022 – Urarina 1123 – Wao Terero 1217 – Yaruro 1317 – Yurakaré 1347

Q quantifiers see also number marking – Kanoé 675, 683 – Kwaza 726–7 – Máku 782–4, 793 – Movima 824 – Muniche 867 – Mỹky 908 – Pirahã 971–2, 974 – Taushiro 1008–9 – Tinigua 1044, 1049–50 – Trumai 1082, 1094 – Urarina 1116, 1118–19, 1121 – Wao Terero 1206–7 – Yaruro 1294 – Yurakaré 1335–6

Index

Quechua – Muniche 855, 868, 882 – Urarina 1116 question formation see interrogatives quotation constructions, Kwaza 722, 738–9, 741 quotative – Kwaza 738–9, 750 – Taushiro 1025

R realis – Kwaza 753–4 – Mỹky 910, 918 – Wao Terero 1216 – Yurakaré 1339, 1348 recent past – Kwaza 737 – Máku 787 – Pirahã 977 – Trumai 1091 – Wã́nsöjöt 1173, 1174 recipient marking, Taushiro 1007 reciprocal – Kanoé 687, 695 – Kwaza 749 – Máku 792 – Movima 830 – Muniche 877 – Pirahã 979 – Tinigua 1057 – Warao 1252, 1257 – Yurakaré 1342 reconciliative, Yurakaré 1345 recursion – Kwaza 741, 750, 763 – Pirahã 985 reduplication – Kanoé 666–7 – Kwaza 735, 737–8, 740, 746–7, 750, 763 – Movima 812, 813, 822, 823, 824, 835 – Pirahã 975 – Trumai 1080–1 – Urarina 1123, 1124 – Warao 1250, 1253, 1259, 1260, 1270 – Yaruro 1294 – Yurakaré 1329, 1335, 1340, 1342 referential indices – Movima 816–20 – Muniche 882

Index

R

– Wao Terero 1210 – Warao 1255 reflexive – Kanoé 687, 695 – Kwaza 749 – Movima 830, 831 – Pirahã 979 – Taushiro 1010, 1011, 1016 – Tinigua 1057 – Trumai 1092, 1098 – Urarina 1114, 1115, 1125 – Wao Terero 1219 – Warao 1251, 1252, 1257, 1260 – Yurakaré 1342 regressive – Pirahã 978 – Taushiro 1013 relative clauses – Kanoé 688, 713–14 – Kwaza 732, 733, 753 – Máku 793, 799, 800 – Movima 838, 840 – Pirahã 985 – Taushiro 1021 – Tinigua 1067 – Trumai 1102–3 – Wã́nsöjöt 1183, 1184 – Wao Terero 1228–30 – Warao 1254, 1277 – Yaruro 1309, 1315–17 – Yurakaré 1348 relativizer – Kwaza 756 – Máku 800 – Movima 825, 829, 839–43 – Pirahã 985 – Taushiro 1021 – Trumai 1082, 1083, 1085, 1088, 1097, 1102 – Wã́nsöjöt 1183, 1184 – Wao Terero 1203 – Warao 1254, 1277 remote future – Máku 787 – Urarina 1127 – Yaruro 1303 remote past – Kwaza 729, 737 – Máku 787 – Movima 825 – Mỹky 923

– Pirahã 977 – Urarina 1127 – Wã́nsöjöt 1173 – Wao Terero 1215 repeater classifier, Pirahã 975 repetitive – Kanoé 667 – Kwaza 740, 741, 746–7 reportative – Mỹky 922 – Urarina 1128 – Wã́nsöjöt 1161, 1176 – Yaruro 1305 – Yurakaré 1341, 1342 reported speech – Pirahã 987, 988 – Urarina 1141 – Warao 1279 resultative – Movima 830, 831, 833, 836 – Pirahã 989 – Wã́nsöjöt 1177 – Warao 1279 resumptive particles, Kwaza 763 resyllabification, Movima 816 revitalization, language – Movima 807 – Mỹky 895, 896 – Trumai 1077 – Wã́nsöjöt 1145 rhetorical questions, Urarina 1128 rhinoglottophilia, Máku 773 Riva Herrera, Martín de la 1002 river-based directionals – Kwaza 744 – Taushiro 1009 – Wao Terero 1219 Rivero, Juan de 1031 Rivet, Paul 1033, 1146 Roamaina 939, 941, 1002n4 Rodrigues, Aryon 770 Rondon, Cândido Mariano da Silva 893, 896 rubber 721, 893, 942, 1034

S Sáliban 1032, 1033 sameness morphology, Mỹky 899, 921–2 Santos, Tomás 1000–1 Sapé 769, 772 Seler, Eduard 1283

S

sentential mood – Kanoé 702 – Kwaza 753–4 – Taushiro 1018 – Urarina 1134–5 – Wã́nsöjöt 1175 sequential markers – Máku 800–1 – Pirahã 988 serial verbs – Kwaza 751–2 – Máku 794 – Mỹky 928–9 – Trumai 1090 – Urarina 1131–2 – Wã́nsöjöt 1170 – Wao Terero 1213, 1221–3 – Yurakaré 1347 sex marking – Kanoé 666, 671, 676, 683 – Kwaza 727 – Movima 817 – Mỹky 921 – Pirahã 962–3 – Tinigua 1050 – Urarina 1128–9 – Yaruro 1288, 1296, 1301 shift, language – Muniche 855 – Pirahã 966 – Tinigua 1029 – Warao 1245, 1264 Shuar 1193 SIL (Summer Institute of Linguistics) organization 809, 944, 959–60, 995, 998, 1193 similative – Wao Terero 1211, 1233 – Yaruro 1305 simultaneous – Wao Terero 1231–3 – Yurakaré 1347 Spanish – Movima 807, 811, 815, 845 – Muniche 855, 856, 867, 868, 882, 888 – Taushiro 998, 1017–18 – Urarina 1107, 1110, 1139 – Wã́nsöjöt 1145, 1166 – Wao Terero 1193 – Warao 1245 – Yurakaré 1325, 1336

Index

spatial indices see also location – Muniche 880 – Taushiro 1007 – Tinigua 1048 – Urarina 1116, 1117, 1123 – Wã́nsöjöt 1160, 1167 – Wao Terero 1200, 1210, 1211–12 – Warao 1253, 1256, 1258 speaker-anchoring versus hearer-anchoring, Kwaza 725 speculative – Mỹky 922 – Wao Terero 1217 speech act participants (SAPs) – Movima 818, 819 – Muniche 882 – Mỹky 921 stative – Kanoé 665, 670, 683, 690 – Kwaza 725, 732, 735 – Máku 784, 788, 791, 793 – Muniche 860, 871, 878, 882 – Mỹky 918, 924, 927–8 – Pirahã 969 – Urarina 1115, 1125 – Warao 1266 – Yurakaré 1340 stress – Kanoé 663 – Kwaza 722, 754, 755 – Máku 774 – Movima 810–11, 826 – Muniche 858–9 – Mỹky 905, 906 – Omurano 946 – Pirahã 964 – Taushiro 1006 – Tinigua 1042–3, 1046, 1050 – Trumai 1080 – Urarina 1113–14 – Wao Terero 1198–9, 1224 – Warao 1248, 1253 – Yaruro 1287 – Yurakaré 1327–8 subject marking – Kanoé 669, 670, 712 – Kwaza 724, 743, 758 – Máku 785 – Muniche 883 – Mỹky 910, 918–19

Index

S

– Omurano 947 – Pirahã 981 – Taushiro 1010, 1022–3 – Tinigua 1059 – Urarina 1119 – Warao 1249, 1273 – Yaruro 1290, 1295, 1309 – Yurakaré 1332, 1340 subject switch reference markers see switch reference subjective – Wao Terero 1217 – Yurakaré 1342 subordinate clauses – Kanoé 708, 711–14 – Kwaza 753, 756–7 – Máku 799 – Movima 825, 843 – Mỹky 934–5 – Pirahã 975, 987 – Taushiro 1004 – Tinigua 1067, 1068 – Trumai 1082, 1098, 1099–101, 1103 – Urarina 1137 – Wã́nsöjöt 1164, 1177, 1180, 1183–5, 1188 – Wao Terero 1226, 1230 – Warao 1271, 1278 – Yaruro 1305, 1315–19 – Yurakaré 1339 suffixation – Kanoé 664 – Kwaza 723, 735 – Máku 775, 777, 792 – Movima 811, 812 – Muniche 860, 866, 872, 880 – Mỹky 898, 906–7, 918, 934 – Omurano 946–7 – Pirahã 964–5, 969 – Taushiro 1007, 1008, 1010 – Tinigua 1042, 1043–4, 1054–5 – Urarina 1112, 1114–15, 1118, 1126–7, 1129–30 – Wã́nsöjöt 1155 – Wao Terero 1199, 1213 – Warao 1249 – Yaruro 1289 superessive, Taushiro 1007 superlatives, Wã́nsöjöt 1173 suppletive – Tinigua 1054 – Yaruro 1288

supplicative, Kanoé 702 Swadesh, Morris 1072–5, 1145, 1283, 1323 switch reference – Kanoé 715 – Kwaza 757–9, 762 – Máku 799 – Urarina 1139 – Yurakaré 1340–1, 1347, 1348 syllable structures – Kanoé 662–3 – Kwaza 722 – Máku 774 – Movima 810 – Muniche 858–9 – Mỹky 900, 905 – Omurano 945, 951 – Pirahã 963, 964 – Taushiro 1005 – Tinigua 1041 – Trumai 1080 – Urarina 1113 – Wã́nsöjöt 1148–9 – Wao Terero 1196, 1197–8 – Warao 1247–8 – Yaruro 1287 – Yurakaré 1326 syncope, Máku 775–6

T tag questions, Kwaza 742 tail-head linkage – Wao Terero 1235–6 – Yurakaré 1349 TAM see aspect; mood/modality; tense Tariana 793n21 Taushiro 939, 948–9, 995–1027 Tejedor, Senén F. 942–3 telicity – Muniche 872 – Mỹky 928 – Warao 1264–5, 1266 – Yurakaré 1340 temporality – Kanoé 667, 685, 689–95, 715 – Kwaza 726, 735, 757 – Máku 777, 787, 801 – Movima 817, 863–7 – Muniche 868–9, 871, 873, 874 – Mỹky 935 – Pirahã 976, 977, 985, 988

V

– Taushiro 1021–2 – Tinigua 1048, 1055, 1056, 1066 – Trumai 1090–2, 1100, 1101 – Urarina 1136 – Wã́nsöjöt 1173 – Wao Terero 1220–1, 1232 – Warao 1256, 1263, 1274, 1275, 1278 – Yurakaré 1331, 1334, 1347 Tenharim 966 tense – Kanoé 689–95 – Kwaza 729, 736–42 – Máku 785, 787–90, 794 – Movima 812, 821, 825, 837 – Muniche 860, 871–3 – Mỹky 918, 935 – Pirahã 964, 969, 974, 975–9 – Taushiro 1012–13 – Tinigua 1044, 1056–7 – Trumai 1088, 1090 – Urarina 1115, 1127 – Wã́nsöjöt 1160–1, 1169–70, 1172–6, 1183–4 – Wao Terero 1213, 1215 – Warao 1262–3, 1268, 1269, 1274 – Yaruro 1303, 1312 – Yurakaré 1330, 1339–42 terminative – Kwaza 752 – Máku 777 – Yaruro 1306 Tessmann, Günther 853, 855, 939, 943, 950–2, 1002n4 Tinigua 1029–75 tone – Mỹky 904–5, 906 – Omurano 945–6, 951–2 – Pirahã 963–4 – Taushiro 1005–6, 1023 – Urarina 1113–14 – Wã́nsöjöt 1152–6, 1157, 1182 topic marking – Kanoé 700 – Kwaza 762 – Pirahã 990 – Urarina 1140 – Wã́nsöjöt 1154–5, 1186–7 – Yurakaré 1332 Tovar, Antonio 948 transitivity see also intransitivity; valency – Kanoé 669, 670, 687, 691–2, 693–4, 700 – Kwaza 730, 747, 749

Index

– Máku 784, 786–7, 792–3, 794, 796, 799 – Movima 811, 815, 822, 827–9, 830, 832–4, 836, 842, 845 – Muniche 870, 877 – Mỹky 910–11, 930 – Omurano 947 – Pirahã 979–80 – Taushiro 1010–12 – Tinigua 1057–8, 1060 – Trumai 1083, 1092, 1093–4, 1097 – Urarina 1115, 1122–3, 1125, 1130, 1133 – Wã́nsöjöt 1160, 1179 – Warao 1260–2 – Yaruro 1296, 1297, 1298, 1307, 1311 – Yurakaré 1337–8 transitivizer – Kanoé 664, 680, 696 – Kwaza 730 – Movima 832–4 translative, Wã́nsöjöt 1167 Trumai 1077–105 Tukanoan family 854

U undirected motion, Máku 792 universal quantifiers, Pirahã 971, 974 Urarina 939, 942, 948–9, 950, 1107–42, 1193

V valency see also intransitivity; transitivity – Kanoé 686–7, 692–5 – Kwaza 747–50 – Máku 792 – Movima 829–35, 836 – Muniche 877–8 – Pirahã 975, 979–80 – Taushiro 1016–18 – Tinigua 1057–8 – Trumai 1092–3 – Urarina 1115, 1129–31 – Wã́nsöjöt 1172 – Wao Terero 1213, 1218–19 – Warao 1259, 1260–2 – Yaruro 1307 Veigl, Francisco Xavier 941–2, 1000n2, 1003 Velasco, Juan de 855, 939, 942 Velasques de Medrano, Salvador 1000, 1002 Velie, Daniel 998 verbalization – Kanoé 664 – Kwaza 723, 735–6

Index

V

– Muniche 879 – Urarina 1122–3 – Warao 1260 – Yaruro 1306 verbless clauses – Máku 795–6, 800 – Tinigua 1061, 1062 Villarejo, Avencio 943, 999 visual/nonvisual evidentiality, Mỹky 923–4 vocative – Mỹky 931 – Pirahã 985 voiceless lateral fricative /ɬ/, Movima 809–10 voiceless velar stop /k/, Máku 772 volitional – Kanoé 689, 700 – Kwaza 738, 754 – Muniche 875 Von den Steinen, Karl 1079 vowel elision, Yurakaré 1327 vowel harmony – Máku 773, 775, 790 – Mỹky 904 – Yurakaré 1334 vowel hiatus resolution, Omurano 945 vowel lengthening (morphological) – Movima 811, 826 – Mỹky 902 – Pirahã 963 – Tinigua 1036, 1042 – Urarina 1112 – Warao 1247, 1271 vowel lowering, Máku 774

W Wanderwörter 897–8 Wã́nsöjöt (Puinave) 1143–89 Wao Terero 1191–241 Waorani 1192 Warao 1243–82 Westernization 721 see also Christianization whistled speech 960 word order see also constituent order – Kanoé 715 – Kwaza 730, 751, 753 – Muniche 884 – Omurano 946, 953 – Pirahã 971, 973, 980–1 – Taushiro 1009, 1026 – Trumai 1085, 1094, 1095–6, 1104 – Warao 1249, 1267, 1272–4

Y Yaruro (Pumé) 1283–322 Yurakaré (Yurújare) 1323–52

Z Zack, Stanislau 661 zero-marked derivation – Kanoé 680, 690, 692 – Kwaza 735, 743 – Máku 779 – Movima 822, 826 – Urarina 1121 – Wao Terero 1215 – Warao 1266