Handbook of Amazonian Languages Volume 1-4 3110102579, 311011495X, 3110128365, 3110149915

Citation preview

Handbook of Amazonian Languages

Handbook of Amazonian Languages Volume 1 edited by Desmond C. Derbyshire and Geoffrey K. Pullum

Mouton de Gruyter Berlin · New York · Amsterdam

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter & Co., Berlin.

Library of Congress Cataloging in Publication Data

Handbook of Amazonian languages. 1. Indians of South America-Brazil-Languages. 2. Amazon River Region-Languages. I. Derbyshire, Desmond C. II. Pullum, Geoffrey K. PM5151.H36 1986 498 86-12692 ISBN 0-89925-124-2 (alk. paper)

CIP-Kurztitelaufnahme der Deutschen Bibliothek

Handbook of Amazonian languages / Desmond C. Derbyshire and Geoffrey K. Pullum ed. Berlin ; New York ; Amsterdam : Mouton de Gruyter NE: Derbyshire, Desmond C. [Hrsg.] Vol. 1 (1986). ISBN 3-11-010257-9

Printed on acid free paper.

© Copyright 1986 by Walter de Gruyter & Co., Berlin. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced in any form — by photoprint, microfilm, or any other means — nor transmitted nor translated into a machine language without written permission from Mouton de Gruyter, a Division of Walter de Gruyter & Co., Berlin. Printing: Ratzlow-Druck, Berlin. — Binding: Dieter Mikolai, Berlin. - Printed in Germany.

PREFACE

This volume, the first of a planned series of three, is the result of nearly ten years of collaborative work. The editors worked together on documenting the languages of Amazonia on grants from the U.K. Social Science Research Council to University College London during 1977-1981 (grant nos. HR5114 and HR6990), and are grateful for that support, which provided some of the foundations for the preparation for the present volume. The first-named editor is also grateful to the British Academy for grants of travel funds mat were essential in making this project possible, and to the Summer Institute of Linguistics and its members (particularly those in Brazil and Peru, at the University of North Dakota summer sessions, and in the Academic Publications and Printing Arts departments in Dallas) for their constant cooperation, assistance and support. Both editors thank the Syntax Research Center at the University of California, Santa Cruz, for other assistance and support during 1981-1985. The aim of the work is to increase the accessibility of descriptive and interpretive material relating to the grammatical structures of the languages· of Amazonia. We have both been struck by the paucity of information about the indigenous languages of South America in the mainstream of linguistics. Even relatively wide-ranging comparative surveys of linguistic phenomena often fail to have even a single Amazonian language in their sample, and to a large extent this is not because of deliberate exclusion but rather because of a lack of available information. To the extent that this work is successful in improving the situation, the majority of the credit must go to the authors of the various contributions that we have elicited for inclusion here. These authors have spent significant portions of their lives working in Amazonia and learning the languages they discuss and describe in these pages. The appearance of this collection is primarily a tribute to them. Many other scholars have assisted us in various ways, often giving generously of their expertise and their time. We cannot list every name, but we want to take this opportunity to express our gratitude to every one of them.

SOUTH AMERICA

Approximate locations of some Amazonian language groups 1 2 3 4 5 6 7 8 9

Amuesha Apalai Apurinä Ashaninca Asheninca Canela-Kraho Caquinte Culina Deni

10 11 12 13 14 15 16 17 18 19 20 21 22 23

Guajajara Jamamadi Machiguenga Nomatsiguenga Pajonal Campa Palikur Parecis Paumari Piraha Piro Terena Urubu-Kaapor Waurä Yagua

CONTENTS

Preface

v

Map-South America

vi

Abbreviations

ix

Introduction Desmond C. Derbyshire and Geoffrey K. Pullum

Part I: Grammatical Sketches Outline of Contents for each chapter in Part I Apalai Edward and Sally Koehn

Part

I

31 33

Canela-Kraho Jack and Jo Popjes

128

Pirahä Daniel L. Everett

200

Urubu-Kaapor James Kakumasu

326

: Word Order and Typological Studies Verb Prominence, Verb Initialness, Ergativity and Typological Disharmony in Guajajara Carl H. Harrison

407

Basic Constituent Order in Yagua Clauses: Implications for Word Order Universals Doris L. Payne

440

viii

CONTENTS

Part III: Comparative Arawakan Studies Comparative Survey of Morphology and Syntax in Brazilian Arawakan Desmond C. Derbyshire

469

Grammatical Characteristics of PreAndine Arawakan Languages of Peru Mary Ruth Wise

567

ABBREVIATIONS

ABLAT ABS ACCOMPL ACT ADJ ADJVZR ADLAT ADVERS FF AFFIRM AG ANIM ASP.AUX ASSOC ΑΤΓΝ AUG AUX BEN CAUS CERT CESS CLASS CLOS CMPL CNCT CNJ COLL COMIT COMPAR COMPL COMPLMTZR COND CONT CONTRADES CONTRAEXP CONTRAFACT COP CUST DAT

ablative absolutive accomplished actual adjective adjectivizer adlative adversative affirmative affirmation agent animate aspect auxiliary associative attention augmentative auxiliary, auxiliary verb benefactive causative certainty cessative class, classification, classifier closure completive connective conjunction collective comitative comparison complement complementizer conditional continuative contradesiderative contraexpective contrafactual copula customary dative

ABBREVIATIONS

DEDUCT DEM DENOM DEP DERW DESC DESID DETRANSVZR DIMIN DIR DISC.PRT DISTRIB D1TRANS DITRANSTVZR DL DP DUB DUR EMB EMPH EP ERG EXCLM EXPANS EXPR FEM/f FICT FOC FRUST FRUST.IN1T FRUST .TERM FUT FUT.DEF.PURP GEN

deduction, deductive demonstrative denominalizer dependent derivational descriptive desiderative detransitivizer diminutive direction, directional discourse particle distributive ditransitive ditransitivizer dual distant past dubitative durative, duration of action embedded emphasis epenthetic ergative exclusive exclamative expansion expressive feminine, female name marker fictitious focus frustrative frustrated initiative frustrated termination future future definite purpose genitive general prefix gerundive group, collective habitual historic (tense) hortatory hearsay

ABBREVIATIONS

1C IDE PH JMMED IMP IMPERF INAN INCEPT INCL INCOMPL INDEP INF INGR INST INTER DMTNSF BMTRANS INTROD IO IP IRR ITER JUX LOC LOG. PROG MAL MAN MASC/m MGN MOD MOT MOVE N NASLZN NEG NF NOM NOMLZN NOMLZR NONFOC NONFUT NONSPECFC NONTHEME

immediate constituent ideophone immediate imperative imperfect inanimate inceptive inclusive incomplete independent infinitive ingressive instrument, instrumental interrogative intensifier intransitive, introducer indirect object immediate past irrealis iterative juxtaposition locative logical progression malefactive manner masculine margin modifier motion movement noun nasalization negative nonfinite nominative nominalization nominalizer nonfocus nonfuture nonspecific nonthematic

χι

XU

NP O/OBJ OBL OBSERV OCC OM PASS PAST1 PAST2 PAST3 PERF PERFCTV PERSPBCT PL POSSN POSSR POSTP POT PPP PRED.COMP PREP PRES PRO PROG PROX PROX1 PROX2 PRSLZR PRCPL PRT PSEUDOPASS PUNCT PURP Q QM REAS RECEPT RECIP REDUP REFLX REFRNT REL

ABBREVIATIONS

noun phrase object oblique observation occurring object marker passive several days ago several months ago distant or legendary past perfect perfective perspective plural possession possessor postposition potential postpositional particle predicate complement preposition present tense pronoun, pronominal progressive proximate proximate past (less than one day ago) proximate past (one day ago) personalizer participle particle pseudopassive punctiliar purpose question question marker reason receptive reciprocal reduplication reflexive referential relativizer

ABBREVIATIONS

REL.CL RELTR REPET REPORT RES RESL RESTRICT RP S SEQ SG SIMULT SM SPECFC SPECFR STAT SUBJ SUBORD SUBSTVZR SUPERL SUPPOS TEMP TEMPRY TERM THEME TOP TRANS TRANSTVZR UNACCOMPL UNCERT UNSPEC V,VB V VQ VBLZR VERIF VOC 0 1 2 3 4

relative clause relator repetition reportive result resolved or regressive restrictive recent past subject sequence, sequential singular simultaneous subject marker specific specifier Stative subjunctive subordinate, subordinating, subordinator substantivizer superlative supposition temporal temporary state terminate theme, thematic, topic marker topicalization transitive transitivizer unaccomplished uncertain, uncertainty unspecified person verb (in syntax) vowel (in segmental phonology) quotative verb verbalizer verification vocative zero, null realization first person second person third person fourth person

Xlll

INTRODUCTION Desmond C. Derbyshire and Geoffrey K. Pullum 1 Amazonia and its people This volume is one of a planned series that treats the languages spoken by the indigenous people of the area known as Amazonia. These languages are genetically diverse and spread over a vast geographical area, so it is not immediately apparent that the volume treats a homogeneous topic. We believe, however, that it does, as will be explained in this chapter. The term "Amazonia" is used here roughly as defined in Meggers and Evans (1983:292). It denotes a region including the Guianas, most of the Amazon basin of Brazil, and some eastern portions of Bolivia, Colombia, and Peru. This region constitutes a natural ecological subregion of lowland South America. Throughout Amazonia, elevations are not great (under 1600 meters), relative humidity is high (usually over 80%), temperatures are fairly constant (and in a comfortable range, contrary to the myth of the "green hell of the Amazon"), and rain falls on at least 130 days a year. Amazonian languages share a broad cultural and ethnic context. For the most part they are spoken by groups of Amerindian hunters or cultivators who are the survivors of a major demographic catastrophe, and who in some cases live today under conditions of great deprivation. Contact with a civilization of European origin, recently initiated in many cases, and in some instances forcibly imposed, has been leading rapidly to extinction of traditional lifestyles tens of thousands of years old, and in some cases (though by no means all) the traumatic effects of such contacts have led to complete societal collapse. Amazonian languages share a wide range of similarities in grammatical and phonological structure that transcend currently conjectured genetic groupings. Such similarities may ultimately indicate that those groupings are too conservative in the extent of the relatedness they postulate. Alternatively, it may be that Amazonia provides a very striking illustration of how areal similarities can develop in geographically contiguous but unrelated languages. A further thing that the languages of Amazonia have in common is that they are still almost entirely unknown to the majority of general linguists. For example, there appears to be not a single Amazonian language mentioned in the Syntax volume of Greenberg's four-volume Universals of Human Language (Greenberg 1978). Amazonian languages provide an important pool of new data against which to test hypotheses in general linguistic theory. There is a small and scattered literature on some of them, but many are already extinct, some

2

Derbyshire and Pullum

are rapidly heading for extinction, and very few are in the position that would most readily permit further linguistic study: that of having a stable and accessible community of native speakers together with a substantial amount of basic descriptive materials available in published form. The river whose basin essentially defines the geographical area we are concentrating on is the largest body of moving fresh water in the world. Two hundred miles out to sea it is possible to detect the fresh water pushed out into the Atlantic Ocean at the Amazon's mouth—a fifth of the river-water discharge in the entire world. The river is navigable by ocean-going ships not only from Belem at the mouth to Manaus a thousand miles inland, but more than a thousand miles further. It is possible to travel in a large ocean-going ship from the Atlantic Ocean up to Iquitos in Peru, and in a smaller ship one can continue for nearly 500 miles further up to the Pongo de Manseriche, about 2,786 miles from the Atlantic. This immense river and the system of tributaries and subtributaries that feed into it define a terrain unique in the world. The main part of the Amazon basin is covered with the largest continuous mass of tropical rain forest on earth, and has been the habitat of groups of human beings for an unknown length of time, certainly thousands of years and very probably tens of thousands. The river system defines a geographical dichotomy of crucial importance to the ecology of Amazonia: the low-lying flood plains adjacent to the rivers (yafzea) and the higher interfluvial forested land (terre firme). The alluvial flood plains are extremely fertile, providing an abundance of terrestrial and aquatic animal protein, and productive land for growing crops such as manioc and corn. The interfluvial areas are much less attractive for human settlement. The dense tropical forest is rooted in surprisingly shallow unconsolidated sediments which do not offer good prospects for growing crops. The ground is often quite hilly and difficult to traverse, and when the forest cover is stripped the land turns out to be fragile and rather infertile. It erodes rapidly from wind and rain, and this soon exposes the underlying soil, which in many areas is of a red lateritic type (containing plinthite) that has often been characterized as agriculturally useless (but see below). It has been stressed by ecologically concerned investigators that removal of the lush yet vulnerable forest can turn an area into desert within a relatively short time span (see Goodland and Irwin 1975; Davis 1977). It is extremely hard to determine what the population of Amazonia was prior to contact with Europeans. A number of attempts to make the necessary calculations are reviewed by Hemming (1978:487-501). The consensus is mat the figure was in the millions, however. Guesses in the region of two to four million are probably not far wide of the mark. But the present population of

INTRODUCTION

3

Amazonian Indians is lower by perhaps as much as two orders of magnitude. Estimates of the Indian population of Brazil today are often below 100,000. The figure of 155,000 for native speakers of indigenous languages in Brazil recently supplied by Aryon Rodrigues (p.c.) is a relatively high estimate. The calamitous decline in numbers among the aboriginal groups in Amazonia has been caused, directly or indirectly, by European-originating conquest and colonization since 1500. One thing that makes it hard to tell the exact magnitude of the effect of the European invasion is that it still is not clear whether Steward (1949:665) was right in saying that "the population may have been expanding at the time of conquest." Whether or not the people of the area were still moving out into the terre firme parts of Amazonia may never be known. There is something of a dispute in the literature between Betty J. Meggers (see e.g., Meggers 1971) and Donald W. Lathrap (see Lathrap 1970). Meggers sees the Amazon jungle as a "counterfeit paradise," providing only meagre resources for human survival. Lathrap, on the other hand, generally views both the Amazonian environment and the achievements of its tropical forest culture as having been underestimated. These characterizations are of course oversimplified. The situation is very complex, and more recent research has added many new dimensions to the study of Amazonian ecology (see the papers in Hames and Vickers 1983, especially the editors' introduction). Some additional support has been amassed recently for the view that Lathrap takes. Van Wambeke (1978), for example, stresses the great diversity of soil types in Amazonia, and claims that "plinthite, often wrongly described as a major barrier to agricultural development in the Amazon, is in fact only found to be a problem in 2% of the soils." Whether or not the interfluvial regions may be capable of supporting substantial populations (and may have been beginning to support greater numbers at the time of first contact), it is uncontroversial that along the flood plain areas adjacent to the rivers there have been settlements of people for thousands of years. There is archaeological evidence of inhabitation by tropical forest cultures all the way down the shores of the Amazon from Yarinacocha, a lake in eastern Peru (Lathrap 1970), to Marajo, the huge island at the mouth of the Amazon (Meggers and Evans 1957). Yarinacocha sites have been dated at between 2000 and 1600 BC, and sites on the lower Amazon between 950 and 600 BC. The ceramic finds have been similar to those found in the region around the Caribbean, and it is generally believed that many of the Amazonian cultural groups may have originated in that region. There is evidence to suggest the existence of populous, settled groups having a greater degree of social and political organization than that encountered by the first Europeans to arrive in the sixteenth century (Galväo 1979). One hypothesis proposed to

4

Derbyshire and Pullum

explain the disappearance of these more complex societies is that the subsistence base was unable to support the growing populations. However, even after 1500, it is clear that many vafzea settlements were quite populous. Francisco de Orellana, who in 1542 sailed from the upper Napo in Ecuador down the whole length of the Amazon to the Atlantic, wrote of sailing for weeks past a continuous riverside conurbation of villages (Hemming 1978, chap. 9). There have also been small groups of hunters and shifting cultivators in the depths of the interfluvial forests for many centuries. It may be that at an earlier stage in their history, some of these were agriculturalists, forced by other groups into hunting lifestyles in the inhospitable forest areas (Lathrap 1968), but this cannot be said to be definitely established. In general, we know far too little of most of the cultures Amazonia has nourished, and of many we know nothing and never will. Archaeological research, in particular, has still hardly touched most parts of Amazonia. Meggers (1975:155-158) reports mat information about the extinct cultures mat flourished all along the flood plain areas at the beginning of the sixteenth century "comes mainly from collections of surface samples of potsherds obtained by Nimuendaju in the 1920's from some 85 sites on the middle Amazon." She also refers to a small amount of more recent work (by Hubert during the fifties and by Simoes during the seventies) west of Manaus. The Madeira, Jurua, Purus, Negro, and Tapajos rivers "have never been surveyed," and the interfluvial forest land that separates them "is equally unexplored." (Note, however, that Meggers and Evans (1983:289) indicate that the situation is rapidly improving.) The matter of archaeological record takes on a sombre significance in view of the persistence in recent years of the European and North American view that a people without a solid archaeological record of achievement in tangible objets d'art and significant architecture scarcely have a right to exist at all. Of course, there are many people in Europe and North America who are active in work aimed at ensuring rights for the aboriginal peoples of South America and elsewhere. But when John Hemming's important history Red Gold: The Conquest of the Brazilian Indians appeared in 1978, the review by J. W. Lambert that appeared in the London Sunday Times (April 16, 1978) began as follows: A week or two ago I greatly enjoyed a virtually unknown opera—"Spinalba," written by the Portuguese composer Almeida for performance at the Royal Palace in Lisbon during the Carnival of 1739, given its first hearing in Britain during this year's Camden Festival.

INTRODUCTION

At the same time I was in the midst of digesting John Hemming's new excursion into South American history: Red Gold—the Conquest of the Brazilian Indians (Macmillan £9.95 pp677). Though it may stand as a stark epitome of the European explosion over the face of the earth, that particular conquest was of course largely carried out by the Portuguese. The Portuguese opera, too, funny, touching, melodious, musically prescient though unpretentious, may stand as a modest epitome of European culture. Under royal patronage its composer was trained, and its performance paid for, by money garnered from the lands, and the unwilling labours, of several million bewildered Indians. Can the opera, and the culture for which it stands, excuse the carnage, rapacity, and blind stupidity spelt out in Mr. Hemming's relentless chronicle? Astonishingly, Lambert is not leaving this rhetorical question to be answered negatively by the reader. He answers it himself, in the affirmative. In Hemming's catalog of ruthless destruction, he finds evidence that these unhappy folk were, and are, mysteriously lacking in whatever stimulus enabled his Incas1 to advance as far as they did, and Europeans to go a great deal further. No Dante or Shakespeare for them, no Bach or Beethoven." Lambert passes very quickly over the details of what the Brazilian Indians did with their lives. They lived in "hideous tribal discord," he reports, and "did not care for work". Unlike the Celts in Europe, he declares, after being subjugated by their invaders they still remained totally unable to adjust to any but the deeply uninteresting life they had lived for several thousand years before the Portuguese arrived, and in sadly ever-diminishing numbers have lived ever since. Lambert, not convinced "that we have much to learn from these people," finally arrives at this conclusion:

6

Derbyshire and Pullum

Mr. Hemming's cumulatively distressing story leaves only one answer to the question I asked in my fourth paragraph: Yes. Dreadful as its conduct has too often been, European civilisation does in the end stand justified. Lest it be thought that Lambert is just an isolated eccentric in openly placing a higher value on a comic opera than on the lives and cultures of millions of people, we should note that the London Sunday Telegraph also published a review of Hemming's book (on April 9, 1978), and in it, Colin Thubron likewise contrasted the Brazilian Indians with the Incas: Ultimately, of course, the tragedy of a civilisation's demise is commensurate with the value of what it achieved. Unlike the Incas, the Brazilian Indian created nothing durable in building or in art. Again the implication seems to be that a people who cannot exhibit at least a few elegant art galleries or put on an entertaining opera can scarcely expect to be allowed to live. Perhaps views of this sort should be passed over without comment; some colleagues have suggested to us that they should. Certainly, the right of Amazonian Indians to exist is not something we have to argue, and we do not need to take seriously a judgement on the Brazilian Indians' achievements by reviewers who know nothing of those peoples' languages or oral literary traditions (see Villas Boas and Villas Boas 1973 for an introduction to the topic). But students of aboriginal languages and cultures cannot afford to ignore, either as scholars or as human beings, such signs of the resurgence of the kind of mentality that condoned the societal destruction and even extinction of aboriginal peoples in previous periods of history. The remarks we have quoted appeared as late as 1978 in two of the most important newspapers for educated people in the United Kingdom. It would be reasonable to wonder whether some bias in Hemming's book suggested to both the quoted reviewers that the Brazilian Indians were so badly behaved as to be scarcely worth preserving. Our reading of Hemming's important work does not give us that impression. The reviewers seem to take a highly selective view of the Brazilian Indians' ways. Lambert speaks, not quoting Hemming, of The frightful Aimore, the aggressive Tupi, the singular Guaricuru, whose women objected to bearing children and regularly aborted their foetuses, and whose men kept and coddled transvestite catamites.

INTRODUCTION

7

It is interesting to compare what Hemming actually says with what Lambert reports. Hemming reports that the Guaicum (we correct the spelling; Lambert confuses their name with that of a mission settlement in another part of the country) were "conservative [about] maintaining their freedom" but able to "adapt and evolve...better than any other Brazilian tribe" in military skills, taking to horse riding "with the skill and passion of North American plains Indians..." Francisco Mendes, a Franciscan friar, reported that they were extremely skilled veterinarians, treating the diseases of their Andalusian horses "with such skill that they generally recover." They were "tall, ...well muscled and with an almost indescribable capacity for enduring hunger, thirst, and sustained effort" according to Rodrigues de Prado. Extremely proud, they regarded themselves as self-evidently superior to all other tribes, "and do not even judge themselves inferior to Spaniards and Portuguese," Prado reported. Colonel Ricardo Franco de Almeida Serra, a Portuguese officer, reported that the young men "spend most of the day in the arms of their wives, and are so inseparable that, even for corporal necessities, for walking or other actions, they do nothing unless the wife and husband do it together," which suggests marital relationships that modem cultures could envy. The women were "pale and very good looking; ...eloquent and of a happy and joyful nature" according to Mendes, and "very careful to preserve their chastity and honour until someone of equal rank solicits them as wives". Lambert silently passes over all this. Looking only for scandal, he finds what he is looking for: Hemming mentions evidence of nontherapeutic abortion (to delay child-bearing years and prolong sexual freedom and attractiveness); male homosexuality; use of elaborate cosmetics and jewelry; alcohol abuse and sexual promiscuity at social gatherings; sexually open marriage; a high divorce rate; belief in a god without organized religious ceremonies; permissive child-rearing and educational practices; and somewhat exploitative (though generally not cruel) use of another ethnic group as a convenient source of domestic servants. Of course, all these aspects of sixteenth-century Guaicum society, without exception, are also represented in the present-day culture of, say, Los Angeles. One wonders whether Sunday newspaper reviewers would excuse the conquest and enslavement of California by, say, the Soviet Union, under conditions that cause the death of ninety-five per cent of its population within a few hundred years, citing the Bolshoi Ballet as evidence that Soviet civilization "does in the end stand justified." Somehow, it seems unlikely. We are led to wonder whether racism is not the motivating force behind the distaste that people like Lambert and Thubron evince for aboriginal Amazonian societies. We have dwelled on this point, perhaps for too long, because we feel that there is a great danger inherent in the propagation of ill-informed and

8

Derbyshire and Pullum

dismissive reports about Amazonian Indians. The impression can easily be created that the Indians of Amazonia can be consigned to the past: a vanished people with few accomplishments, now mostly extinct, about whom little needs to be said. Reference books will not in general suffice to correct this impression; the 1966 Encyclopaedia Britannica article on the Amazon devotes seventeen words to the entire aboriginal human population of the area; the river turtle gets three times as much discussion! As it happens, the Guaicuru people, just discussed, are reported as extinct by both Loukotka (1968:48-49) and by Voegelin and Voegelin (1977:149), and spoken of entirely in the past tense by Lambert. But they are not extinct. At least some of their descendants, in particular from the subgroup of them called the Kadiweu, still live in the Mato Grosso. Hemming (1978:493) lists their estimated numbers as 200 (compared with a probable 15,000 five hundred years ago), and Hanbury-Tenison (1973:120-130) describes his meeting with them on a journey across Brazil. There can be little hope of promoting fair treatment for a people, or adequate scholarly attention to their language and culture, if they are not only stripped of respect through denigration of their past but incorrectly written off as dead in the present.2 In speaking of Brazilian Indians, we are of course speaking of hundreds of different groups with hundreds of different cultural traditions, so it would be unwise to make general judgements about the quality of their societies or their cultural level. But it should perhaps be pointed out that we are talking about people who were not quite as lacking in artistic abilities as Lambert and Thubron depict them as being. The Omagua people of the great settlements along the central Amazon flood plain, for example, produced ceramics that "seemed to Orellana's men to be equal to the finest porcelain in the world, better than Malaga-ware and comparable with Roman work. It was 'thin and smooth, glazed and with colors shading off into one another in the style of that made in China', according to Toribio de Ortigüera" (Hemming 1978:189). It is quite true that the Indians of Amazonia have not created any architectural monuments. But with a tropical climate and plenty of appropriate building materials like wood and thatch locally available, there was little reason to stud the forest with edifices of stone. The abandoned ruins of Machu Picchu and Angkor Wat stand in mute witness to the fact that buildings will not guarantee the survival of a society. The relationship of the society to its environment is all-important: if natural resources run out, palaces and temples will be left to crumble as people go elsewhere seeking food. The Brazilian Indians concentrated for centuries on achieving a balance with the local environment, and with techniques like shifting cultivation and slash-and-burn agriculture they succeeded in entering into a stable ecological relationship with the Amazonian jungle. One can imagine worse monuments.

INTRODUCTION

9

It may be that the most important achievements of the Brazilian Indians are in science radier than the arts. As the forests of the Amazon are reduced to provide land for ranching or mineral exploitation, the Indians lose their homeland, but the world also loses a storehouse of genetic diversity whose extent may never be known but which the Indians have made the basis of a rich chemical and pharmacological technology. The issue of what drugs and other chemicals may be derivable from irreplaceable Amazonian trees and plants is important enough that at least one company, Biotechnical Research and Development Services of London, has been set up to expand and develop the potential of Amazonian ethnobotanical knowledge. For example, there is much therapeutic potential in some of the chemical defenses developed by certain trees against viral or insect attack. In many cases the Indians of Amazonia have been found to be already using plant-derived substances for a wide range of purposes such as contraception, healing, and ecologically safe poisoning of aquatic life (a common fishing technique). Far from having a clear understanding of what the Amazonian Indians know about such matters, we are in danger of destroying the environment that is the basis for that knowledge, and losing it for ever (cf. Lewis 1979). Amazonia's Indian population today is scattered across a huge area in small numbers, and exists in a wide variety of circumstances and conditions. There are groups who thrive in such protected areas as the Xingu national park. Some are defending their rights through adroit political activity, and preserving their ethnic identity with pride, while having little difficulty in assimilating those aspects of European-originated culture—particularly in the domain of material goods—that impact on their lives; the Xavante are a good example (see Moynahan 1978 for an interesting journalistic report). Other groups, however, have been demoralized through their experience with the sudden impact of "pacification" programs and needless and preventable epidemics of diseases they could have been vaccinated against. Groups contacted for the first time in the 1970's such as the Surui (see Lafer 1981) and the Kreen-Akrore (Baruzzi et al. 1978; Heelas 1978) provided sad examples for a time (but see below). The history of the Indians of Brazil since 1900 (not dealt with by Hemming) has been summarized by Ribeiro (1957). He distinguishes four types of relationship which indigenous groups have had with the national society: (1) isolation; (2) intermittent contact; (3) permanent contact; and (4) integration. His case study comprises 230 groups (not necessarily whole tribes, some tribes being subdivided where different subgroups had very different experiences with the dominant Brazilian culture). He compares the situation of these 230 groups as they were in 1900 with what they had become in 1957. The most striking

10

Derbyshire and Pullum

result of the study was the need to add one more category to the original four types of relationship: (5) extinction. Of the 230 tribes still in existence in 1900, 87 had to be placed in the extinct category only half a century later (though in at least one case, Ribeiro was mistaken; the Karitiana are not extinct, and their language is mentioned below). It is vitally important not to assume from a reading of the shocking history of repeated genocide in Amazonia that the disappearance of the Indian cultures of the area is inevitable. It is not. There will be Indians in Brazil and the surrounding countries in the twenty-first century as there are now. It may be that the members of recently contacted groups will continue to be permitted to die from lack of the basic medical care that Europeans and Americans take for granted, and it may be that the rights in law of those who continue to live will be violated by non-Indians hungry for land; but there is nothing inevitable about it. It can be prevented, with very little more than the cheap and easy modern techniques of vaccination against nonindigenous diseases and some semblance of attention to civil rights. Moore (1981:49ff) describes the case of the entry of the Zoro people into contact with Brazilian society, which was achieved without any significant ill-effects to the tribe through responsible health care and vaccination procedures. And even in cases where a community does fall into a prolonged decline in health, morale, and numbers, a turnaround is always a possibility. Schwartzman (1984) reports on the status of the Kreen-Akrore (Krenakore) in the early 1980's, and describes a dramatic upswing in their fortunes beginning in 1977. Agriculture, building, collective labor, traditional rituals, food production, trade, and even the birthrate picked up in an interrelated set of developments involving a renewal of community life. Life and laughter have returned to Kreen-Akrore society, and Schwartzman provides examples of a new vein of sociolinguistically sophisticated humor that has sprung up on the multilingual post where the Kreen-Akrore now live. It would be a fortunate by-product of the linguistic research represented in this book if the dissemination of information about Amazonian languages brought about an increased scholarly awareness of the conditions under which the people who speak them have to survive, and of the possibilities that still exist for seeing the languages in question survive and flourish.

INTRODUCTION

11

2 Genetic classification of Amazonian languages The languages of Amazonia are probably less well understood in terms of genetic grouping than those of any öfter area in the world, except perhaps New Guinea. Although there are some clearly established major families, many assignments to families are based on very scanty information, and some are based merely on geographical contiguity, travelers' tales, and conjecture. Little confidence can be placed in the currently proposed phylum-level relationships. It is quite likely that wider connections than are currently apparent will ultimately be established, but on the other hand it is probable that some currently accepted phylum assignments are based on faulty data and will have to be revised. The standard source for the classification of the world's languages is Voegelin and Voegelin (1977), which synthesizes most of what was known about language classification up to the early seventies. However, as far as South America is concerned, it is often unreliable. It depends chiefly on the proposals of a four-page paper by Greenberg (1960), which presents an ambitious classification scheme for the entire Central and South American continent. Greenberg's is certainly an important and stimulating effort, but it is based on data that have not been published. For many of the languages Greenberg was doubtless able to identify cognates from reliable materials, but for some he cannot have had anything to go on but speculation and circumstantial evidence, since too many South American aboriginal languages simply have no usable descriptive literature. The Voegelins do not cite Loukotka (1968), the fourth edition of one of the most important books on die subject, though they cite a 1935 precursor of it. The reason may be that Loukotka relies to a considerable extent on geographical and cultural considerations in setting up his groupings, and includes in his classification languages for which no data were available to him and languages known only from early historical accounts. Greenberg, on the other hand, works strictly from broad-ranging collections of cognate forms, ignoring areal and historical considerations and previously proposed classifications. The Greenberg and Loukotka proposals differ at even the most general and inclusive levels, as seen in Table 1.

12

Derbyshire and Pullum

Table 1: Two phylum-level classifications of South American languages. I. Greenberg (1960) MACRO-CHIBCHAN: CHIBCHA PROPER PAEZAN ANDEAN-EQUATORIAL: ANDEAN JIVAROAN MACRO-TUCANOAN EQUATORIAL GE-PANO-CARIB: MACRO-GE-BORORO MACRO-PANOAN MACRO-CARIB Huarpe Nambicuara Taruma 2. Loukotka (1968) PALEO-AMERICAN: SOUTHERN CHACO CENTRAL BRAZE, NORTHEASTERN NORTHWESTERN TROPICAL FOREST: NORTH CENTRAL SOUTH CENTRAL CENTRAL NORTHEASTERN ANDEAN: NORTHERN NORTH CENTRAL SOUTH CENTRAL SOUTHERN Some examples may help to give an impression of the conflict between the two classifications, the inadequacy of the factual basis for both of them, and, incidentally, of the riotous confusion of names and spellings that the classifier has to deal with. 1. Amuesha. Loukotka (p. 147) has "Amoishe or Amisha or Amuescha or Amage or Lorenzo" in the Lorenzo group of the Arawak stock in the North Central division of the Tropical Forest category, and incorrectly implies ("once spoken") that it is extinct. The Voegelins place Amuesha (or "Amuexa") as a single language branch of the Arawakan family of the Equatorial division in the Andean-Equatorial phylum, coordinate with groupings such as Maipuran within Arawakan. (Wise

INTRODUCTION

2.

3.

4.

5.

6.

7.

13

(this volume) holds that Amuesha is located within Pre-Andine Maipuran in the Arawak family.) Arara. Loukotka (p. 122) mentions Arara as a Tupian language "spoken at the mouth of the the Preto and Gi-Parana Rivers," but the Voegelins (p. 336) list it as an extinct Tupian language. (Moore (1981) reports 77 Arara living with the Gaviäo.) Gaviäo. Loukotka (p. I l l ) assigns "Kaapor or Urubu or Gaviäo" to the Northern group of the Tupi stock in the North Central division of the Tropical Forest category, correctly identifying it as a Tupian language. The Voegelins appear to miss Gaviäo altogether. They cite Arua and Monde as extinct, which is incorrect (Moore 1981:46), and do not mention the Gaviäo, the Zoro, or the Cinta Larga, who speak closely related languages. (Moore (1981) reports more than 140 Gaviäo and more than 170 Zoro at the end of the 1970's.) Hixkaryana. Loukotka lists "Hishcariana or Ishkariyana or Tucano" in the Waiwai group of the "Karaib" (Carib) stock in the Northeastern division of the Tropical Forest category (p. 206). The Voegelins cite "Hishkaryana" as of unknown affiliation (p. 158). (Loukotka cites Derbyshire (1961), which the Voegelins seem to have missed.) Kadiweu. As mentioned above, Loukotka (p. 49) has "Caduveo or Kadiueo" as a dialect of Guaicuru, in the Guaicuru stock of the Chaco division of the Paleo-American category, still spoken in the "state of Mato Grosso...in a few villages in the vicinity of Miranda." The Voegelins record "Caduveu" as a dialect of the "Mbaya-Guaicuru" language in the "Guaycuru" family, but mark the whole family as extinct. (Bible translators have in fact been working on Kadiweu for several years, and have published articles on it; Griffiths and Griffiths 1976 is one example.) Karitiana. Loukotka (p. 122) lists "Caritiana" in the Arikem group of the Tupi stock in the North Central division of the Tropical Forest category. The Voegelins (p. 336) place Karitiana in the Arikem group of the Tupi family in the Equatorial division in the Andean-Equatorial phylum, but again, like Ribeiro (1957), assert wrongly that it is extinct. (David and Rachel Landin are working on Karitiana, which is still spoken by about seventy people; see Landin 1980.) Nadeb. Loukotka (p. 192) gives "Nadöbö" as a dialect spoken by an independent tribe within the Macu stock in the Central division of the Tropical Forest category. The Voegelins (p. 334) list "Nadobo" as one of the dialects of a language called "Macu" in the Puinave (or Macu or Guaipunavo) family within Macro-Tucanoan in the Andean-Equatorial phylum, but warn that "Macu may represent several languages, some of which may not even belong to the Puinave family." (Helen Weir, who

14

Derbyshire and Pullum

works on this language, reports (p.c.) that there is no clear evidence yet of Nadeb being related to any of the Puinavean languages it is commonly bracketed with.) 8. Surui. Loukotka lists "Suruim" among "Unclassified or Unknown Languages of the Area of the North Central Division" of the Tropical Forest category. The Voegelins have nothing. (Surui is a Tupian language with about 300 speakers, who were not contacted until 1969; cf. Lafer (1981).) Notwithstanding the many individual problems of classification and failures to coordinate information such as those summarized above, it should be emphasized that there is a solid core of well-established facts concerning certain major linguistic stocks and families that are represented in the major classifications. We shall refer to these language families throughout this work. Some introduction to them is therefore in order. The two largest stocks, both in geographical spread and in number of languages, are Arawakan and Tupian. Both contain various peripheral subfamilies and individual languages whose relationship to the rest of the group is in doubt, but within both there is a central family containing many languages whose close relationship to each other is not in doubt: Maipuran in Arawakan and Tupi-Guarani in Tupian. Arawakan. The Arawakan stock is spread from Belize in Central America to Southern Brazil, Bolivia, and, according to one recent report, even into Chile, and from the mouth of the Amazon to the foothills of the Andes. A number of the Caribbean islands had Arawakan-speaking populations at first contact, and the "Island Carib" that survived in Dominica and St Vincent until about 1920 was in fact an Arawakan language, the same language (modulo certain dialect differences) that survives today in Central America as "Black Carib" (Taylor 1958). This volume contains papers by Derbyshire and by Wise which survey in detail some syntactic and morphological characteristics of a few of the Brazilian and Peruvian Arawakan languages. They both briefly discuss intra-Arawakan genetic relationships in the light of their studies, and suggest that modifications to earlier classifications may be needed, especially to that of the Voegelins. Tupian. The Tupian languages are nearly all located south of the Amazon (an exception being Oiampi, spoken by the descendants of eighteenth-century migrants to what is now French Guiana). They were originally found along the south bank of the Amazon, but later spread widely southward and were securely established all the way down the Brazilian coast. The Tupian family contains the languages spoken by the Tupinamba, a very populous tribe in eastern Brazil before the Portuguese conquest, and many related groups. One subpart of Tupian is the Tupi-Guarani family, which includes what the

INTRODUCTION

15

Voegelins regard as a single language called Tupi that comprises not only the modern dialects descended from the language of the Tupinamba but also Chiriguano, spoken by about 20,000 people in Bolivia and Argentina. The Tupi-Guarani family also includes Guarani, spoken natively by more than a million people in Paraguay (where it serves as the only Amerindian language with national-language status, being—with Spanish—one of the two national languages) and also spoken in southern Brazil. In the present volume we include a description of one Tupian language, Urubu and a paper reporting typologically significant facts about another, Guajajara. In a future volume we plan to publish a comparative survey article on Tupian similar to the Arawakan surveys in this one. Ge. Another well-established linguistic grouping, not as big as Arawakan or Tupian, is the Ge (or Je) stock. These languages are spoken mainly in inland central Brazil. The Voegelin and Voegelin classification of Ge and Macro-Ge seems more secure than that of some other stocks and families, being largely based on the solid comparative phonological work of Davis (1966, 1968; cf. also Hamp 1969). Ge-speaking groups are often spread over a wide area, and have attracted much anthropological interest for their complex social organization and kinship structures. Their societies are in many cases politically and militarily very robust, and in some cases (e.g., the Cayapo, the Xavante) have resisted outside influence on their affairs very successfully. The Ge languages are represented in this volume by a description of Canela. Carib. Another frequently referenced group of related languages is the Carib family, which is primarily associated with the Guiana Highlands of eastern Venezuela, Guyana, Suriname, French Guiana, and northern Brazil (see Durbin 1977 for a recent proposal for internal classification of the group). Carib is represented today by languages spread between north-western Colombia and southern Brazil, but mostly north of the Amazon, the southern groups being the result of an isolated migration southward along the Xingu river. The Caribs expanded aggressively between 1150 and 1450 into the Caribbean islands (which the Spaniards named after them), conquering the Arawakan inhabitants who had arrived in prehistoric times.3 The Carib family is represented in this volume by a description of Apalai (northern Brazil). Panoan. Finally, there is a linguistically well established stock called Panoan, containing languages spoken mostly in Peru, with a few in Brazil and Bolivia. This volume does not include a description of a Panoan language, but it is intended that a later volume should remedy this. When we consider proposed groupings larger than the stocks we have just mentioned, we find that we are dealing with hypotheses of extremely dubious status. For example, while Greenberg proposes a relationship between Tupian and Arawakan, we find that Aryon Rodrigues, the foremost scholar of South

16

Derbyshire and Pullum

American languages working today, is completely unable to support any such connection (see Rodrigues 1967). On the other hand, Rodrigues (1974:56) claims he has been able to find evidence of a fairly close relationship between Tupian and Carib. This is incompatible with Greenberg's proposal to group Carib, Ge, and Panoan into a single "Macro-Ge-Pano-Carib" phylum that does not include Tupian or Arawakan. The only published support for linking Tupian and Arawakan is the work of Noble (1965), in which evidence of shared vocabulary between proto-Arawak and proto-Tupian is presented. But in the light of the probable early distributions of the two groups (see Lathrap 1970, chap. 4) it is perfectly possible that lexical items could have been borrowed from one to the other. Rodrigues1 view (1974:51-52) is that Noble's work is based almost entirely on deficient and fragmentary material requiring real daring from the author in the postulation of many of his results and says further (p. 56) of his own work that it was impossible for me to find any evidence whatever favorable to a closer affinity between Tupi and Arawak. In addition to the major stocks and families we have discussed, there are many languages in Amazonia that remain unclassified. Some belong to smaller families which in the past consisted of a number of languages but today are represented by only one or two extant languages. One such is Pirahä, which is the subject of a paper in part I of this book. In future volumes we plan to include descriptions of Nadeb, Nambikuara, and Sanuma (one of the Yanomama group), as well as a more complete description of Yagua. 3 Areal typological similarities in Amazonia Amazonia, with its enormous extent (four million square miles) and its variety of types of habitats, contains, as would be expected, a rich variety of types of linguistic structure. Nevertheless, there are area-wide typological tendencies in a number of specific directions. One of the typological features that has most engaged the interests of the editors of this volume is the tendency among many Amazonian languages to favor an order of major constituents in the sentence that puts the direct object earlier than the subject in a transitive clause with two full noun phrases. It remains a relatively little-known fact that VOS, OVS, and OSV languages exist, and assertions that they do not are scattered throughout the literature of linguistics. Even since 1980, we find, for example, Williams (1981:25) claiming that a "glaring failure" of a certain theory is that it does not have as

INTRODUCTION

17

a consequence "the fact that languages are all either VSO, SVO, or SOV—in other words, S precedes O." Even where the existence of VOS languages is recognized (Keenan (1978) presents a review of the syntactic properties of such languages), the existence of OVS and OSV languages often is not. Thus Fromkin and Rodman (1983:227), in a popular textbook, say that "No language has been discovered that has the preferred word order OVSub," and Givon (1984:198) refers to evidence for OVS order but says nothing whatever about OSV. All of the O-before-S possibilities seem to be realized in Amazonian languages. We have reported in Derbyshire and Pullum (1981) on the Object-Verb-Subject (OVS) and Object-Subject-Verb (OSV) languages of Amazonia that had come to our attention by 1978. Pullum (1981:151-152) provides a slightly fuller catalog of some currently known OVS, OSV, and VOS languages. A brief summary of Amazonian cases of O-before-S languages that we currently know of is given in Table 2. VOS languages are found elsewhere among the world's languages (particularly in Mexico, Madagascar, and the Western Pacific), but the object-initial word orders are not definitely known to occur anywhere outside Amazonia, which could make the area a particularly important source of typologically unusual data. This volume includes information about constituent order in five of the languages in Table 2. Derbyshire discusses Apurinä, Jamamadi, and Terena in his comparative Arawakan survey paper, Koehn and Koehn contribute a description of Apalai, and Kakumasu offers a descriptive overview of Urubu. It should be noted that in the description offered here, Kakumasu questions the correctness of his own earlier claim (1976) that Urubu is strictly OSV. Here, having conducted further study of Urubu constituent order, Kakumasu stresses that SOV clauses are more frequent and may represent the basic type (hence the asterisk against Urubu in Table 2). We shall not take a view here on whether his present claim or his earlier one is correct. We feel that the issue currently remains open. OSV is still a relatively common order in Urubu and until the exact details of the grammatical use of the postpositional marker -ke are clearer to us than they are at present we shall not feel confident about the status of the language vis ä vis constituent order typology. For Apalai, on the other hand, Koehn and Koehn now give rather stronger evidence supporting an OVS basic order than was available to us in our paper on object-initial basic orders (Derbyshire and Pullum 1981). It is important to note that there is some evidence of languages in Amazonia undergoing syntactic change toward object-initial constituent orders. Derbyshire (1981) discusses a shift toward OVS from original SOV in three Carib languages.

Derbyshire and Pullum

18

Table 2: Some OS (or probably OS) languages in Amazonia Language

Family

ovs *Amahuaca Apalai Arecuna Asurini *Bacairi Barasanos (Southern) Hianacoto Hixkaryana *Macushi *Oiampi *Panare Teribe Urarina *Wayana

Panoan Carib Carib Tupi-Guarani Carib Eastern Tucanoan Carib Carib Carib Tupi-Guarani Carib Chibchan (unaffiliated) Carib

osv Arawakan (unaffiliated; probably related to Nadeb) Arawakan (unaffiliated) Tupi-Guarani Ge

Apurinä Hupda *Jamamadi Nadeb *Urubu Xavante

vos Baure Kaiwa Terena

Arawakan Tupi-Guarani Arawakan

* Basic constituent order type still a matter of dispute.

INTRODUCTION

19

There appear to be a number of other syntactic characteristics that are shared by languages of diverse genetic origins in Amazonia besides the tendency toward object-initial constituent order. Any one of them might be expected to turn up in a language anywhere in the world without being considered very surprising. But the repeated occurrence of significant subsets of the following set of features might be indicative of an interesting areal confluence of properties.

1.

2.

3.

4.

5. 6. 7. 8.

9. 10.

Verb agreement with both subject and object (cf. e.g., the Bantu languages) plus null realization of subject and object pronouns, making it extremely common for a sentence to have no full NP subject or object. Predictability of whether subject and object will be realized as full NP's or merely indicated by verbal affixes, depending on whether they represent given or new information (cf. the notions of "topic continuity" and "pragmatically-controlled word-order" in Givon 1983). Use of nominalizations to substitute for finite relative clauses and other types of subordinate clause, so that in many cases no finite subordinate clauses are found at all. A combination of phrase constituent orders which, while not unique, is unusual, and seems to be found regardless of whether main clauses have VO or OV constituent order: N-Adj, Gen-N, and NP-P (i.e. postnominal attributive adjectives, prenominal genitives, and postpositions). No construction identifiable as an agentive passive. No indirect statement form, hence heavy reliance on direct rather than indirect speech constructions. No coordinating conjunctions, hence reliance on juxtaposition to express logical coordination. Extensive use of right-dislocated paratactic constructions, i.e. sequences of NP's, adverbials, or postpositional phrases, in which the whole sequence bears only one grammatical relation in the sentence, and in which each phrase is phonologically dislocated from the others (this being the principal means of coordination and modification in many languages). Extensive use of particles that are phrasal subconstituents syntactically and phonologically but are sentence operators or modifiers semantically. Ergative subject-marking, or relics of such marking (there are signs of a drift away from ergativity in many languages; cf. Harrison, this volume, and Derbyshire 1981).

20

Derbyshire and Pullum

The ten characteristics just listed are suggested as typical ones by an examination of available materials on about twenty languages, together with more intensive study of one of them, Hixkaryana (Derbyshire 1979). The amount of information available to us has clearly not been sufficient to permit any certainty in stating that the above characteristics identify an areal Amazonian linguistic type. But providing further information for comparative study is precisely the purpose of this Handbook. We have attempted to obtain contributions describing languages of differing constituent-order types, and from different linguistic families (according to currently accepted classifications). In the longer term we hope to provide a fund of information that can extend the list given above, and no doubt provide exceptions to the main areal generalizations proposed. In this first volume, Everett's description of Pirahä shows that language to be notably different from other known Amazonian languages in its strong preference for sentences that have the subject and direct object expressed by full noun phrases rather than by free pronouns, clitics, affixes, or zero anaphora. On the other hand, Derbyshire's concluding section in his study of Brazilian Arawakan argues that the languages in that family do, in general, exhibit the patterns that we have tentatively identified as typical for Amazonia, and that furthermore they exemplify to an extreme degree one other pattern widely encountered in languages of the area, namely complex verbal morphology. Most of these characteristics are found also in the other languages discussed in the book. Two of the four languages described at length in Part I (Canela, Pirahä) are fairly rigidly SOV. Another (Urubu) is now claimed by the author of the description to be SOV but with the possibility that the other common order, OSV, may prove to be the more basic one. The remaining language (Apalai) is presented as an OVS language, but with SOV also a commonly occurring constituent order. What now appears to be emerging as an hypothesis with increasingly strong support is that at an earlier stage in their history, all these Amazon area languages were fairly rigidly SOV (see Derbyshire, Harrison, and Wise in the this volume; also Derbyshire 1981). 4 Aims and organization of the present volume The languages of the aboriginal inhabitants of Amazonia, even of those groups that are well adjusted to contact with modern South American societies and are demographically secure, are currently almost completely unknown to linguists in general. It is not uncommon to find that a wide-ranging comparative survey in grammar or phonology has not a single Amazonian language in its sample. Moreover, access to these languages is not at all easy.

INTRODUCTION

21

Contact with the speakers of many of the languages is made difficult or impossible by government restrictions that, in Brazil for example, put Indian areas off-limits to most non-Indians, and make it very hard indeed for anyone to obtain permission to visit those areas. However, to an extent that has varied with the political climates of the countries concerned over the past three decades, members of the Summer Institute of Linguistics (SIL) have been permitted to live with a significant number of Indian groups in Amazonia, in several countries, while engaged in linguistic work relating to literacy and Bible translation projects. We have relied to a large extent on SIL members to assist us in amassing the contributions and background information that are presented in this book. Without their dedicated work over many years, no such volume could have been produced. There are highly encouraging signs of increasing interest among South American linguists in the indigenous languages of their countries (in Brazil, for example, mainly through the influence of Professor Aryon Rodrigues of the University of Campinas). However, in most cases,4 their work has not advanced to the point of their being able to supply outline descriptions of languages, which is what we have chiefly been requesting from SIL linguists and a few others for this project. Sketches and outlines are of course insufficient to permit the kind of intensive work that linguists (especially syntacticians) would like to see done on a wider range of languages. There can be no doubt that the descriptions collected here will be incomplete in many respects, and too superficial to provide a basis for anything more than preliminary theoretical conclusions. Lightfoot (1979:383-384) has warned against the tendency to assume that superficial analysis will suffice for little-known languages—that "the more exotic the language, the less need for precise analysis and the less controversy about correct descriptions." He calls mis assumption the "Ebeling Principle," in honor of Ebeling (1960:43-44), who observed how phonemic analysis seemed to every investigator to be satisfactory only for languages other than his own. The reminder is appropriate, of course. Linguists have often been prepared to make claims about a relatively little-studied language of a sort that they would not dream of making about a well-studied European language on the basis of the same amount of exposure and analysis. It should go without saying that aboriginal languages of far-off continents will have subtleties and deeply complex subregularities of construction as much as any other language, and these will not all be apparent when a language is first analysed. But with Amazonian languages at the present time, we sometimes face an unwelcome

22

Derbyshire and Pullum

choice: having an outline description of a language, with only limited exemplification, or remaining in complete ignorance. It seems sensible, faced with this choice, to opt for the former. We make no pretence of being able to assemble for the languages of Amazonia what thousands of scholars have amassed over two millennia for the languages of Europe. But we have done our best to elicit progress reports from people whose ongoing work occupies them too completely to allow for a definitive statement; we have tried to obtain tentative current statements from linguists still in the field despite the fact that their work may in due course establish that their present views need to be modified; and we have prepared or commissioned articles summarizing and synthesizing the evidence contained in scattered working papers, unpublished draft materials, and archived texts. We believe that this work, as presented in this volume and its planned successors constitutes a worthwhile endeavor. For those chapters that constitute grammatical sketches of particular languages, we have devised a basic outline that each author has been asked to adhere to as far as possible. The outline was prepared in the light of experience with materials on a number of Amazonian languages, and provides for focussed discussion of a number of the conjectured common features of the languages of the region discussed in the previous section. This outline is presented as a table of contents to part I of the volume. Furnishing grammatical sketches of this type is the main objective of the Handbook, and projected descriptions for future volumes, following the same outline, include the following languages. Arawakan: Cariban: Panoan: Tupian:

Other:

Paumari Waura Macushi Capanahua Guajajara Kamaiurä Oiampi Nadeb Nambikuara Sanuma Yagua

Part II of this volume consists of two papers reporting facts which should be of special interest to both typologists and others working in specific theories. The two languages concerned, Guajajara and Yagua, are genetically unrelated and are geographically about as far apart as it is possible to be in Amazonia.

INTRODUCTION

23

Yet they share a peculiar constellation of constituent order regularities that is not only previously unattested, but is specifically disallowed by one recently proposed universal (Hawkins 1979:626; 1980:201; 1983), namely: (a) (b) (c) (d)

VSO N-P G-N N-A

(verb-subject-object) (postpositions) (prenominal genitives) (postnominal attributive adjectives)

Part III also contains two papers. They are comparative morphosyntactic studies of the Arawakan stock, based on published and (a considerable amount of) unpublished material. The first paper reports on eight languages spoken in Brazil and the other on eight languages spoken in Peru. The literature does not contain very much on the syntactic characteristics of the languages of this major linguistic stock, especially those spoken in Brazil and Peru. These papers should at least provide some concrete information that may be useful to South American comparativists, although they are far from being complete accounts of any single area of the morphosyntax of any of the languages included in the survey. As noted earlier in this section, we plan to have more complete grammatical sketches of at least two Arawakan languages in a future volume.

NOTES Hemming's previous book was The Conquest of the Incas (Hemming 1970). We have more to say about mistakes in Amazonian demography in the section on language classification below. As mentioned above, the remnant "Carib" populations of the area (including the Black Caribs of Belize and Nicaragua) are speakers not of Carib but of Arawakan languages. The island Caribs normally ate the males of vanquished rival populations in the islands (the practice of cannibalism was also named after them) but spared their Arawak-speaking women. Possible exceptions known to us include work by Lucy Seki on Kamaiurä (we hope to include a description from her in a future volume), Neusa Carson on Macushi, Yonne Leite on Tapirape, Bruna Franchetto on Kuikuru, and Adair Paläcio on Guato.

24

Derbyshire and Pullum

REFERENCES

Abbreviations UAL JL Lg.

International Journal of American Linguistics Journal of Linguistics Language

Baruzzi, R. G., L. F. Marcopito, M. L. C. Serra, F. A. A. Souza, and C. Stabile 1978 "The Kren-Akorore: a recently contacted indigenous tribe," Survival International Review 3.2:22-24. Davis, I. 1966 "Comparative Je phonology," Estudos Lingüisticos 1:10-24. 1968 "Some Macro-Je relationships," UAL 34:42-47. Davis, S. H. 1977 Victims of the miracle: Development and the Indians of Brazil (Cambridge: Cambridge University Press). Derbyshire, D. C. 1961 "Hishkaryana (Carib) syntax structure," UAL 27:125-142 (Part I), 226-236 (Part II). 1979 Hixkaryana, Lingua Descriptive Studies, 1 (Amsterdam: North Holland). 1981 "A diachronic explanation for the origin of OVS in some Carib languages," JL 17:209-220. Derbyshire, D. C., and G. K. Pullum 1981 "Object-initial languages," UAL 47:192-214. Durbin, M. 1977 "The Carib language family," Carib-Speaking Indians, edited by E. Basso (Tucson: University of Arizona Press). Ebeling, C. L. 1960 Linguistic units (The Hague: Mouton). Fromkin, V., and R. Rodman 1983 An introduction to language (Third Edition) (New York: Holt Rinehart and Winston).

INTRODUCTION

25

Galväo, E. 1979 "The encounter of tribal and national societies in the Brazilian Amazon," Brazil: Anthropological perspectives, Essays in honor of Charles Wagley, edited by M. L. Margolis and W. E. Carter (New York: Columbia University Press). Givon, T. 1983 "Topic continuity and word-order pragmatics in Ute," Topic continuity in discourse: A quantitative cross-language study, Typological StudiesinLanguage, 3,edited by T. Givon (Amsterdam: J. Benjamins). 1984 Syntax: A functional-typological introduction (Amsterdam: John Benjamins). Goodland, R. J. A., and H. S. Irwin 1975 Amazon jungle: from green hell to red desert (Amsterdam: Elsevier). Greenberg, J. H. 1960 "The general classification of Central and South American languages," Men and cultures: Selected papers of the Fifth International Congress of Anthropological and Ethnological Sciences, edited by A. F. C. Wallace (Philadelphia: University of Pennsylvania Press). Greenberg, J. H., ed. 1978 Universals of human language, volume 4: Syntax (Stanford: Stanford University Press). Griffiths, G., and C. Griffiths 1976 Aspectos da lingua kadiweu, Serie Lingüistica 6 (Brasilia: Summer Institute of Linguistics). Hames, R. B., and W. T. Vickers, eds. 1983 Adaptive responses of native amazonians (New York: Academic Press). Hamp, E. P. 1969 "On Maxacali, Karaja, and Macro-Je," UAL 35:269-270. Hanbury-Tenison, R. 1973 A question of survival for the Indians of Brazil (New York: Charles Scribner's Sons). Hawkins, J. 1979 "Implicational universale as predictors of word order change," Lg. 55:618-648. 1980 "On implicational and distributional universals of word order," JL 16:193-236. 1983 Word order universals (New York: Academic Press).

26

Derbyshire and Pullum

Heelas, R. 1978 "An historical outline of the Panara (Kreen-Akarore) tribe of Central Brazil," Survival International Review (London) 3.2:25-27. Hemming, J. 1970 The conquest of the Incas (London: Macmillan). 1978 Red gold: The conquest of the Brazilian Indians (London: Macmillan). Kakumasu, J. Y. 1976 "Gramatica gerativa preliminar da lingua Urubu," Serie Lingüistica 5:171-197 (Brasilia: Summer Institute of Linguistics). Keenan, E. L. 1978 "Subject final languages," Syntactic typology: Studies in the phenomenology of language, edited by W. P. Lehmann (Austin: University of Texas Press). Lafer, B. M. 1981 "The Surui," in Maybury-Lewis et al. Lambert, J. W. 1978 "Brazil: the conquest that financed the West," Sunday Times (London), April 16. Review of Hemming 1978. Landin, D. 1980 An outline of the syntactic structure of Karitiana sentences (University of London: M.A. thesis). Lathrap, D. W. 1968 "The hunting economies of the tropical forest zone of South America: an attempt at historical perspective," Man the hunter, edited by R. B. Lee and I de Vore (Chicago: Aldine). 1970 The Upper Amazon, Ancient Peoples and Places, 70 (New York: Praeger). Lewis, N. 1979 "The rape of Amazonia," Observer Magazine (London), April 22, 46-61. Lightfoot, D. 1979 Review of Charles Li, ed., Mechanisms of syntactic change, Lg. 55:381-395. Loukotka, C. 1968 Classification of South American Indian languages (Los Angeles: University of California, Latin American Center). Maybury-Lewis, D., J. W. Clay, D. Price, D. Moore, B. M. Lafer, and C. Junqueira 1981 In the path of Polonoroeste: endangered peoples of western Brazil (Cambridge: Cultural Survival).

INTRODUCTION

27

Meggers, B. J. 1971 Amazonia: man and culture in a counterfeit paradise (Chicago: Aldine-Atherton). 1975 "Application of the biological model of diversification to cultural distributions in tropical lowland South America," Biotropica 7:141-161. Meggers, B. J., and C. Evans 1957 "Archaeological investigations at the mouth of the Amazon," Smithsonian Institution Bureau of American Ethnology Bulletin 167 (Washington D.C.: United States Government Printing Office). 1983 "Lowland South America and the Antilles,' ncient South Americans, edited by J. D. Jennings (San Francisco: W. H. Freeman), 287-385. Moore, D. 1981 "The Gaviäo, Zoro, and Arara Indians," in Maybury-Lewis et al. Moynahan, B. 1978 "Brazil: the last frontier," Sunday Times Magazine (London), June 18. Noble, G. K. 1965 Proto-Arawakan and its descendants, Indiana University Research Center in Anthropology, Folklore, and Linguistics, Publication 38, UAL 31, 3, Part II (Bloomington: Indiana University). Pullum, G. K. 1981 "Languages with object before subject: a comment and a catalogue," Linguistics 19:147-155. Ribeiro, D. 1957 "Culturas e linguas indigenas do Brasil," Educaqäo e Ciencias Socials (Rio de Janeiro) 2.6:1-102. Rodrigues, A. 1967 "Grupos lingüisticos da Amazonia," Atos do Simposio sdbre a Biota Amazonica 2:29-39. (Rio de Janeiro: Conselho Nacional de Pesquisas). 1974 "Linguistic groups of Amazonia," Native South Americans: Ethnology of the least known continent, edited by P. J. Lyon (Boston: Little Brown). Schwartzman, S. 1984 "Linguistic humor and the maintenance of Krenakore identity under contact," UAL 50:232-237. Steward, J. 1949 "South American cultures: an interpretative summary," Handbook of South American Indians (Washington D.C.: U.S. Government Printing Office), Vol. 5, 669-672.

28

Derbyshire and Pullum

Taylor, D. 1958 "The place of Island Carib within the Arawakan family," UAL 24:153-156. Thubron, C. 1978 "Genocide in the making," Sunday Telegraph (London), April 9 Review of Hemming 1978. Villas Boas, O., and C. Villas Boas 1973 Xingu: The Indians, their myths (New York: Farrar, Straus and Giroux). Van Wambeke, A. 1978 "Properties and potentials of soils in the Amazon basin," Interciencia 3:233-244. Voegelin, C. F., and F. M. Voegelin 1977 Classification and index of the world's languages (New York: Elsevier). Williams, E. 1981 "Language acquisition, markedness, and phrase structure," Language acquisition and linguistic theory, edited by S. Tavakolian (Cambridge: ΜΓΓ Press).

PART I Grammatical Sketches

Outline of Contents for each chapter in Part I INTRODUCTION SYNTAX OF THE SENTENCE OR CLAUSE 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Word order Parataxis Ellipsis Reflexives, reciprocals, unspecified arguments Passives Causatives Comparatives, equatives Coordination Pragmatic and discourse characteristics Interrogatives Imperatives Negation Anaphora Subordinate clauses SYNTAX OF PHRASE TYPES

15

16 17 18

19 20

Noun phrase structure 15.1 Marking for case 15.2 Genitives 15.3 Modifiers 15.4 Nominalizations Pronoun system Adpositional phrase structure Verb or verb phrase structure 18.1 Tense 18.2 Aspect 18.3 Mood/Modality 18.4 Person, number, gender 18.5 Voice 18.6 Other categories 18.7 Incorporation 18.8 Auxiliary verb system Adjective phrase structure Adverb phrase structure

32

OUTLINE OF CONTENTS PART I

21

PARTICLES 21.1 Sentence particles 21.2 Discourse particles 21.3 Verification particles

22 23 24

PHONOLOGY MORPHOLOGY IDEOPHONES REFERENCES

APALAI Edward Koehn and Sally Koehn INTRODUCTION Apalai is a language of the Carib family spoken in the Northern regions of the Maicuru, Paru and Jari Rivers of Para, Brazil. In the last three decades the tribe has been decimated by various diseases contracted through their contact with civilization. At present there are around 350 speakers of Apalai. Partly as a result of their depleted ranks and because of their proximity, the Apalai have integrated with the Wayana Indians (also Carib language family) to a considerable degree. There are very few "pure" Apalai. Although the material cultures are indistinguishable the linguistic culture is intact. The first language of the offspring of an integrated marriage is generally that of the mother, although a high degree of bilingualism is not uncommon. The Carib family is scattered throughout northern and central South America. Apalai is not mutually intelligible with its nearest family neighbors: Wayana and Hixkaryana of Brazil and Tirio and Carib (Galibi) of Surinam. Apalai-Portuguese bilingualism is limited to the use of bartering terms or "small talk" with transient rubber or jaguar hunters and government representatives. Apart from the Apalai-Wayana bilingualism in "mixed" families, referred to above, Apalai is spoken exclusively within the culture. SYNTAX OF THE SENTENCE OR CLAUSE 1 Word order The basic clause types of Apalai include: Intransitive, Transitive, Expressive, Copular, and Equative. The nuclear constituents of these clauses are described below in sects. 1.1 - 1.5, and peripheral constituents, which apply in general to all clauses, are treated in sect. 1.6. 1.1 Transitive clauses. The transitive clause has an object followed by a verb, and a subject which follows the verb or precedes the object: OVS or

sov. (1)

kaikuxi etapa-V toto, papa jaguar kill- IP 3PL father 'They killed a jaguar, father's group.'

(2)

mame tamy matary epekaty-ase then tobacco 1 +buy- RP Then I bought tobacco.'

tomo 3PL

34

Edward and Sally Koehn

(3)

opi t- yri- se karapareta medicine NF-make-CMPL Karapareta 'The medicine was made by Karapareta.'

(4)

mame n-epinopyase mose then 3-treat medically- RP this one Then this person treated him medically.'

(5)

papa kyn-uo- ne father 3- kill-DP 'Father killed it.'

(6)

papa father

ruka Luke

kana- ry canoe-POSSN

a by rahkene really

apiro- ase open up-RP

mä THEME INTROD 'Father opened up Luke's canoe.'

The basic order of the constituents of the transitive clause is shown as OVS even though very few clauses include all three constituents. Those that do are usually discourse initial sentences. One of the reasons for the lack of clause level constituents in Apalai discourse is the use of devices for marking thematic participants. Once a participant is identified he is referred to by pronominal affixes bound to the verbs of the main or subordinate clauses. In a sample of one hundred Transitive clauses in running text, forty of the clauses have no free subject or object. Fifty of them have a free object occurring before the verb but no free subject. Five of them are SV with the subject occurring before the verb and no free object. Only three are OVS. Another one is SOV. There is also an clause with the subject and verb understood. This may not be a representative sample of all the possibilities since VS can also occur (see (4)), but its occurrence is rare. When a free object occurs in a clause it comes before the verb. When a free subject also occurs it can appear before or following the OV with a 3:1 preference for the position following OV. Additional evidence suggesting that this is the basic pattern is found in the order of constituents in subordinate clauses. (See sect. 14.) 1.2 Intransitive clauses. The intransitive clause has a subject and verb which may occur in either order: SV or VS. (7)

papa n-otuh-no father 3-eat- IP 'Father ate.'

APALAI

(8)

moroto ynän-oturuäse there l +3-converse-RP There we conversed.'

(9)

mame t- önuh- se papa then NF-go up-CMPL father Then father went up above.'

(10)

bere pona Belem to

t- oyto-se NF-go- CMPL

35

kakoxi above konohno, brother in law

poeto child

me typy- ty a- htao DENOM 3RFLX-wife-POSSN be-when 'My brother-in-law went to Belem when his wife was a child.' 1.3 Expressive clauses. The expressive clause has an expression followed by the verb ka 'say' or 'do' and the subject of that verb: (11)

y, oehno ropa, ty- ka- se mokyro i- py- ty yes, 1 +come-IP back NF-say-CMPL that one 3-wife-POSSN ' "Yes, I came back," said his wife.'

There is a variant form a, which occurs for certain tense-aspect categories (see sect. 18.1 and (12) below). The expression may be an ideophone or series of ideophones representing an action, the noise of an action, the noise of an animal, the words of a song, or a direct speech quote. The following examples illustrate a song (12), direct speech (13), and an ideophonic expression (14). (12)

sakura, sakura, sakura, the sweet manioc drink

a- Vko, say-CONT

poet child

ome DENOM

pitiko j- a- htao small 1-be-when ' "Sakura, sakura, sakura," I sang when I was a little child.' (13)

saih saih ka- ry moro ey-a go away go away say-NOMLZR that 3- by ' "Go away! go away," she kept saying.'

(14)

koih kaxi-ko, a- Vko aisore hm-a paddle do- IMP say-CONT Aisore 3- be+PRES ' "Paddle!," Aisore was saying to me.'

y-a 1-to

Ex. (14) has an embedded ka clause within another ka clause. The ideophone is part of the embedded clause.

36

Edward and Sally Koehn

1.4 Copular clauses. The copular clause has a subject, copular verb exi 'be,' and complement. The complement is expressed by a descriptive (adjectival-adverbial type) word or postpositional phrase. The verb may be deleted when it is recoverable from the preceding discourse. The normal order of constituents is: COMPL S V. (15)

poeto me exi- ne kamakuao child DENOM 1+be-DP Kama creek-at Ί was a child at Kama creek.'

(16)

morara oximome toh kyn-exi-ne like that together 3PL 3- be- DP '(The trees) were close together.'

(17)

moroto yna tapyi- V there 1+3 house-POSSN 'There was our house.'

(18)

aimo pyno mana boy caring for 3+be+PRES 'He cares for the boy.'

kyn-exi-ne 3- be- DP

1.5 Equative clauses. The equative is a nonverbal clause whose nucleus consists of two nominals, one of which may be a demonstrative pronoun. This is the preferred construction when a nominalized verb phrase is used. One nominal functions as subject (usually the first constituent) and the other as predicate complement: Subj PC. (19)

mokyro ituakyry that one jungle person That man (is) a jungle person.'

(20)

a-yto-ry 3-go-NOMLZR 'He was going.'

(21)

Ikaponato a-oryxi-ry the next one 3-sister-POSSN 'His sister was the next one (born).'

moro that

1.6 Peripheral constituents. Optional constituents of clauses are Indirect Object and various oblique objects which I will refer to here as Adjunct. Indirect object is a postpositional phrase with the relator α 'to, by'. (22)

y-pyrc ekaro- ase 1-gun+POSSN 1+give-RP Ί gave my gun to him.'

ey-a 3- to

APALA1

(23)

37

moky ene- kose aporo, n-yka-V toto that one look-IMF briefly 3-say-IP 3PL ' "There he is, look at him," they said to father.'

papa a father to

Adjuncts include all constituents which express semantic roles of location, time, manner, cause, and purpose. They may be adverbials or postpositional phrases. (See Derbyshire 1979:39). Usually not more than two adjuncts precede a clause nucleus, one being an introducer and the other a time or location word (28), whereas up to 3 or 4 adjuncts can follow the nucleus: (24)

mame oepyase ituh- taka then 1 +come-RP forest-to "Then I came to the woods.'

(25)

mame t- oyto- se ropa tumy then NF-go- CMPL again 3REFLX-father+POSSN 'Then he went back to his father.'

(26)

jeneh- ne toto 3SlO-bring- DP 3PL 'They brought me here.'

(27)

pina barbed arrow

ke with

a to

taro-na here-to

tuo-py GEN. PREF-kill-NEG+NOMLZR

rohxo t- yka-se INTNSF NF-say-CMPL ' "That's not what you kill with barbed arrows," he said.' (28)

mame kokoro yto-Vko ase then tomorrow go-CONT 1+be+PRES "Then tomorrow I'm going to catch fish.'

kana fish

any- se lift up-PURP

2 Parataxis The juxtaposition of phrases within the clause and of clauses within the sentence is of frequent occurrence. The main clause constituents are usually expressed first, followed by an expanded or modified form of one or more of those constituents. Examples of juxtaposition of noun phrases, subject and object, are the following: (29)

mame yrokokoro ituimano t- oyto-se, then next day Ituimano NF-go-CMPL 'Then the next day Ituimano went, Takyimano.'

takyimano Takyimano

38

Edward and Sally Koehn

The two names represent the same person in a (discontinuous) appositional relationship. (30)

irukaira Irukaira

t- aro- se NF-take-CMPL

tymeretamu- ru, 3REFLX-father in law-POSSN

irukaira Irukaira 'He took Irukaira, his father-in-law.' (31)

paruru enepy-ase jake tomo, i-py- ty banana bring-RP Jake 3PL 3-wife-POSSN 'Jake's group including his wife brought bananas.'

maro with

Examples of juxtaposition of adverbial phrases are the following: (32)

t- oyto-se NF-go- CMPL

nyh- se sleep-PURP

iporiry creek

kuaka at,

apu Assai

eukuru kuaka drink at 'He went to sleep at the creek, at the Assai drink creek.' (33)

moto-na ke onuh- to- ko tapyi there-to POLITE go up-PL-IMP house 'You all go up over there on to the house ridge.'

mypatarä-naka ridgeon to

The phrases of the foregoing examples are dislocated phonologically from the preceding constituents, each phrase having its own intonational pattern and pause between it and other phrases. The dislocation is rightward. Much more infrequent in Apalai speech are examples of leftward dislocation as in (34) and (35). (34)

te se- ino PAUSE there-from

inikahpo-ino up river-from

aorihGEN. PREF-deceased-

tyä tapyi- V kyn-exi-ne mäka myhto NOMLZR house-POSSN 3- be- DP mango at foot of Over there, up river, was the deceased one's house beneath the mango tree.'

APALAI

(35)

oty what

ropa, again

monohne that kind

39

tökehko, PL, etc.

oty keh what else

moro, that

y pape morohne poko atamorepa-Vko oh yes paper that kind about REFLX-teach- CONT 'What next, all those things, what do you call them, studying paper and all those things.' The sequence of phrases is often discontinuous, as in ((29), (30), and (33)), with only the first of the list appearing in the pre-verbal· position. Other examples of discontinuous sequences show that the expansion phrase need not be sentence-final (36) and that constituents within a direct speech (complement) clause may be expanded by a phrase that occurs after the main verb of the sentence (37): (36)

tawapaira Tawapaira

t- aro- se NF-take-CMPL

tokono 3REFLX-brother in law

urakana-se hunt- PURP 'He took Tawapaira his brother-in-law to hunt.' (37)

arimi monkey

wo-hpo kill-PAST +NOMLZR

ase t- yka-se l +be +PRES NF-say-CMPL

y-meretamu- ru zopa-hpono 1-father in law-POSSN feed-PAST+NOMLZR ' "I am a monkey killer" he said, "one who provides for my father-in-law." ' Phrase parataxis may function as a coordination mechanism (sect. 8), as a method of clarification ((30), (32), (33), and (36)), or as a means of modification or emphasis (31). No examples of parataxis occurring within subordinate clauses have been noted. As for juxtaposition of finite verbs, each verb is considered in our analysis to be a complete clause rather than a sequence of serial verbs within a clause. (38)

mame tokare then openly

pyra NEG

tosa- ry 3REFLX-place-POSSN

t- akoh-se. NF-cut- CMPL

akoh-ko opomoko paruru arika-ko cut- HIS plant(manioc)-HIS banana plant-HIS 'Then secretly he cut himself a field. He cleared it, planted manioc and he planted bananas.'

40

Edward and Sally Koehn

Another kind of series of verb phrases is the repetition of a verb phrase or ideophone indicating the passing of time or repeated action. (39)

tyse 3REFLX-mother+POSSN rahkene. really

apori- ry feather-POSSN

enahka-se use up-CONT.PAST

enahka-se use up-CONT. PAST

enahka-se use up-CONT.PAST

enahka-se t- onahka-se moro use up-CONT. PAST NF-use up-CMPL that 'He really used up his mother's feathers. He used them and used them and used them till they were used up.' (40)

mokyro that one

i-nio 3-husband+POSSN

pekä flap

pekä flap

pekä flap

pekä tiwi flap catch on 'Her husband flapped (his wings, clumsily flying) then caught on (to the landing place with his beak).' There is ample evidence of clause juxtaposition in Apalai narrative and dialogue. Both main clauses (41) and subordinate clauses (42) occur juxtaposed, the function being coordination, modification, clarification or emphasis. (41)

(1)

oehno 1+come-IP

papa, father

apori- ry feathers-POSSN (2)

u-muku-ru 1-son- POSSN ry POSSN

(3)

poko occ.with

j-uhpoty 1-feathers+POSSN

u-muku-ru pyre 1-son- POSSN arrow +POSSN poko. occupied with typyre 3REFLX-arrow+POSSN n-ase. 3-be+IP enahka-Vko. use up-CONT

aporifeathers-

APALAI

(4)

41

naeroro oy-a oehno so 2- to l -t-come-IP ' "I've come, father, to get feathers for my son's arrows. My son has been occupied with making arrows, and he is using up all my feathers. That's why I've come to you." '

In (41), the main clauses (2), (3), and (4) clarify the first clause. (42)

zumo pyra ro big NBG yet

ta- htao 3REFLX-be-when

kyn-etyne 3- become drunk-DP

pitiko ro ta- htao small still 3REFLX-be-when 'When he was not yet big he got drunk, when he was still small.' See sect. 22.5 for basic intonation patterns relating to phonologically dislocated clauses and phrases. 3 Ellipsis In Apalai clauses any or all nonverbal constituents can be omitted leaving only the verb and its bound pronominal markers. The nominal subject can be omitted once it has been expressed. As long as it remains the topic or main participant it is only referred to by pronouns or indirect references. (43)

papa father

to group

kV-yto-ne 3- go- DP

parahta rubber

zupise repe. look for- PURP FRUST

taro-ino kV-yto-ne toto kapo here-from 3- go- DP 3PL airstrip 'Father and group went to look for rubber. From here they went to the airstrip.' (44)

pona to

öko wo- ne toto. papa kyn-uo- ne wild turkey kill-DP 3PL father 3- kill- DP 'They killed a wild turkey. Father killed it.'

When the topic needs to be expanded or limited a full nominal may occur as subject (44). The person involved as subject of a preceding clause may be the indirect object of a following clause without any overt reference (45):

42

(45)

Edward and Sally Koehn

zuarohko 3+know-quite

pyra NEG

ase. 1+be+PRES

enaroro 3PL

zuaro 3+know

na. ekaro-Vko roke toh mon-exi-ano 3+be+PRES tell- CONT only 3PL 3- be-RP didn't witness it. They know it (firsthand). They were telling it (to me).' If the verb form is completive, the verb affixes refer to third person subject for intransitives, expressives and equatives, and to third person objects for transitives. When the action is thematic over the span of several clauses the verb can be deleted (46). Often the ideophonic representation substitutes as a proform for its corresponding verb (47): (46)

te PAUSE ro yet

öko turkey

yna- a- htao. 1+3-be-when

to 3PL

ynan-uopito- ne 1+3- shoot-begin-DP

moroto-ino that- after

zumo pyra big NEG

arimi to monkey 3PL

rahkene really

zumo-hxo yna- a- htao rahkene big- more 1 +3-be-when really 'We began to kill wild turkeys when we were not yet big. After that (we killed) monkeys when we were bigger.' (47)

mame pokö poko poko poko kae- hxo then flap flap flap flap high-more

repe FRUST

wewe po toto tree on 3PL 'Then away they flew quite high up in a tree.' There is a high occurrence of partial or incomplete sentences in Apalai discourse. One reason is the heavy use of direct speech quotes in narrative. Much of the dynamic of a story is carried along in the form of direct quotes, which include partial sentence responses. This extends to most cognitive processes, which are stated in the form of speech quotes. When the context, whether behavior or direct speech, provides part of the information load, the quote may be less than a complete sentence and still be semantically complete. Another reason for incomplete sentences is the tendency to reduce redundancy by omitting overt markers for the participants once they have been established as "thematic." A "thematic" participant is one who has been introduced by name or descriptive statement and is central in that part of the

APALAI

43

discourse. Clues such as kinship terms may be used to refer to related participants, while the main participant is not overtly expressed or may be indicated by means of a pronominal prefix. (48)

mokyro that one

kurumu vulture

rahkene. certainly

morarame then

t- yka-se NF-say-CMPL kyrykyry-Vko get into-CONT tose meat

poko occ.with

ke with

typy- e ADJVZR-wife- ADJVZR

a-oryxi-ry 3-sister-POSSN

O-ekyhxo 2-pet+POSSN-INTNSF hxo, INTNSF hxo. INTNSF

tosa- ry SREFLX-place-POSSN

otyro veg.food

morara thus hna otyro 3+be+PRES veg.food kutikutima-Vko messCONT

mame tokare then ADJVZR-tell t- akoh-se. NF-cut-CMPL

pyra NEC

akoh-ko cut- HIS

opomoko paruru aryka-ko plant(manioc)-HIS bananas plant-HIS 'This man married a vulture woman. It started with a complaint from his sister, "Your pet bird keeps messing in the food I'm fixing." Then without telling (anyone) (he) made a clearing. (He) cut it down, planted manioc and planted bananas/ Note that the story begins by introducing the hero in a descriptive statement in title or preview form: "This man married a vulture woman.' The next participant is introduced as his sister who, by speaking, creates a situation calling for a response. The only linguistic clue then that the series of actions to follow is attributed to the hero is the third person pronoun marker on the verbs. The cultural expectation provides additional support, the role of cutting a field being that of a man and not of a sister or bird-wife. The sister's speech constitutes an indirect speech act, the form being declarative whereas the intent is imperative: "Do something about the situation." The response to such forms is often behavioral instead of verbal. (See E. Koehn 1976 for description of behavioral dyads.) 4 Reflexives and reciprocals Reflexivity and reciprocity are expressed by a verbal prefix: os-l ot-l at-1 e-. Reduplication of the first syllable of the verb indicates repeated action and always results in reciprocal meaning, as in (51):

44

Edward and Sally Koehn

(49)

n-osereh- no 3-REFLX-scare-IP 'He scared himself.'

(50)

ynan-oturu- ase 1 +3- REFLX-tell-RP 'We really conversed.'

(51)

atatapoi-Vko toto REFLX-REDUP-grab-CONT 3PL "They (were) grabbing each other (fighting).'

(52)

ekurika-Vko toh mon-exi-ano REFLX-wash- CONT 3PL 3be-RP 'They were washing themselves.'

(53)

oturu-keh xipo nyh- se ynan-yto-ase ropa REFLX-tell-CESS first sleep-PURP 1 +3-go- RP again 'Having stopped our conversing we went to sleep again.'

(54)

t- osetapa-se toh NF-REFLX-hit- CMPL 3PL 'The monkey group stumbled.'

rahkene really

meku tomo monkey group

This prefix (os-/ ot-/ at-1 e-) functions as a detransitivizer, changing transitives into intransitive constructions. Another way of detransitivizing transitive verb stems is to add the discontinuous sequence of prefix and suffix: n-...-ta 'detransitivizer'. The person-marking prefixes and tense-aspect suffixes are those used with intransitive verb stems: (55)

oty ka-toh o-nepekah-tano what do-PURP 2-DETRANSVZR-buy- DETRANSVZR-IP 'Why did you shop?'

Reciprocals take the same form as reflexives except for the form which occurs with the postpositional pyno. In this case, two forms are used: ose- and oxi-. There is preference among some speakers to consider osepyno the reflexive 'care for oneself, and oxipyno the reciprocal 'care for each other'. There is a possessive reflexive prefix t-, which occurs with nouns and certain postpositions. The phrase containing this prefix may function as object, indirect object, adjunct, or genitive of a possessed NP. It may occur with any constituent that can be inflected for person of the possessor other than when it is the subject of the main clause. It applies only to third person.

APALAI

45

(56)

töxiry z-uru-Vko 3REFLX-daughter-POSSN 3-tell-CONT 'He was talking to his own daughter.'

(57)

tytapyi-V 3REFLX-house-POSSN

taka to

kyn-ako 3- be+IP

t- oyto-se NF-go- CMPL

ropa again

ynororo, 3PRO

nyh- se sleep-PURP 'He went back to his own house to sleep.' (58)

turui 3REFLX-elder brother+POSSN

tupi field +POSSN

poko occ.with

eroh- nöko mokyro work-CONT that one 'He is working at his elder brother's field.' (59)

t- yka-se tyya NF-say-CMPL 3REFLX-to 'He said (it) to himself.'

rokene only

(60)

tope a tV- oturu- po- se 3REFLX-friend+POSSN to NF-REFLX-tell-CAUS-CMPL 'He questioned his own friend.'

(61)

töxiry maro 3REFLX-daughter-POSSN with 'He went with his own daughter.'

t- oyto-se NF-go- CMPL

The possessor reflexive is controlled by the subject of the main clause and may occur on the subject of a subordinate clause or other constituents in a subordinate clause: (62)

bere Belem

pona to

t- oyto-se NF-go- CMPL

konohno, brother in law

poeto me typy- ty a- htao child DENOM 3REFLX-wife-POSSN be-when 'My brother-in-law went to Belem when his wife was a child.' (63)

tykana-ry 3REFLX-canoe-POSSN

nery-topoVpyry sink-NOMLZR(place)-PAST +POSSN

pona yto-Vko mana to go- CONT 3 +be +PRES 'He is going to the place where his canoe sank.'

46

Edward and Sally Koehn

The reflexive t- can occur in nonfinite clauses such as the purposive and certain adverbial type clauses. It may be controlled by the subject of the superordinate clause ((64)-(66) and (69)-(72)) or by the (underlying) subject of the nonfinite clause in which it occurs ((67) and (68)). (64)

yto-Vko go- CONT

mokyro that one

mana 3+be+PRES

töxiry SREFLX-daughter-POSSN

ene-se see-PURP 'That one is going to see his (own) daughter.' (65)

typy- ty 3REFLX-wife-POSSN

zuru-keh tell- CESS

oeh- nöko come-CONT

first

ropa mana back 3 +be +PRES 'When he finishes talking to his wife he will come back.' (66)

toky 3REFLX-pet+POSSN

zupi-ryme look-NOMLZR-DENOM

oeh- nöko come-CONT

ropa mana back 3 +be +PRES 'He will be coming back looking for his pet.' (67)

irukaira t- aro- se tymeretamu- m Irukaira NF-take-CMPL 3REFLX-father in law-POSSN 'His father-in-law, Irukaira, was taken (by him).'

(68)

eropa, ömy 1+2+go+IMP 2+father+POSSN

aro- ko, take-IMP

t- yka-se NF-say-CMPL

typy- ty a 3REFLX-wife-POSSN to ' "Let's go! And take your father," he said to his wife.'

(69)

toh 3REFLX-meat+POSSN

wo- toise shoot-FRUST

y-a 1-by

y-meretamu1-father in law-

ru y-nomo- no POSSN 1-abandon-IP 'My father-in-law abandoned me for my almost having shot his meat.'

APALAI

(70)

seromaroro now

47

atamorepa-vko REFLX-teach- CONT

ropa again

mose this one

pape paper

mero- tohme tyya write-NOMLZR-DENOM 3REFLX-by 'Now this person is studying again to write on paper himself.' (71)

poeto-me child-DENOM

ta- htao 3REFLX-be-when

tupa Tüpa

kyn-epuru- ne 3- REFLX-sting-DP 'When she was a child Tüpa was initiated (stung with ants).' (72)

mame kajama then manioc meal

epekah-se buy- PURP

n-oeh- no 3-come-IP

omise hungry

toexi- ryke 3REFLEX-be- NOMLZR-REAS Then he came to buy manioc meal because he was hungry.' 5 Passives There are personal passives in Apalai. The bound person marking prefixes of the active verb are replaced by the prefix t- 'NONFINTTE', and the tense and aspect marking suffixes of active verbs are replaced by -se 'COMPLETIVE'. (73a)

kuany-no 3Sl+2O-lift- IP 'He lifted us.'

(73b)

t- any-se kymoro NF-lift-CMPL 1+2 'We were lifted up by him.'

ey-a 3- by

Comparing the syntactic functions of the nominal constituents of an active clause with those of a passive one we find the following: (i) The subject of the active intransitive remains the subject of the passive construction. It can precede or follow the verb. (74a)

Ituimano n-yto-no Ituimano 3-go- IP 'Ituimano has gone.'

(74b)

t- oyto-se Ituimano NF-go- CMPL Ituimano 'Ituimano has gone.'

(ii) The object of the active transitive becomes the subject of the corresponding passive clause while the subject of the active transitive becomes an agentive phrase adjunct marked by the postposition a 'to, by' (which is also the marker of indirect object). Any indirect object of the active construction remains an indirect object of the passive construction.

48

(75)

Edward and Sally Koehn

mame yna t- aomika-se then 1+3 NF-greet- CMPL

toto 3PL

a, by

nohpo tomo woman group

a roropa by also Then they all greeted us, the women folk also.' The agentive phrase of the passive clause may occur before the verb, or following it, or both. (76)

aimo a t- aro- se, u-muku-ru boy by NF-take-CMPL 1-son- POSSN 'It was taken by the boy, by my son.'

a by

The clause constituent order in the passive construction is a matter of prominence. The constituent which occurs first, whether a nominal or a verb, is the more prominent. If a nominal is repeated or expanded, the first occurrence of the constituent (generally the more generic form) is considered nuclear and subsequent mention is its expansion. As for tense and aspect possibilities, the passive form carries the meaning of completed action. When accompanied by an auxiliary verb the passive functions as an adjectival (79): (77)

moroto-ino t- yarakah- se that- after NF-cageCMPL 'After that (he) caged his pet.'

(78)

mame tyaramari then 3REFLX-feather suit+POSSN Then (she) removed her feather suit.'

toky 3REFLX-pet+POSSN t- ouse NF-remove-CMPL

In the above two examples the t- '3REFLX' is the same third person as the underlying subject, which does not show up on the surface. (79)

moroto-ino t- önuh- se that- after NF-go up-CMPL 'After that they are gone up.'

toh 3PL

mana 3+be+PRES

The primary function of the t-V-se construction is to take the underlying subject/agent out of focus and give prominence to the object or to the activity or property that describes the object. The Apalai t-V-se form resembles the Hixkaryana form of 'derived adverbial' described in Derbyshire (1979:176), t-V-so, t-V-xe, but it seems to have a

APALAI

49

different function. The Apalai form which corresponds to the function described for the 'derived adverbial' in Hixkaryana is t-V-semy as in: tonahsemy

'edible vegetable to be eaten'

toserny

'meat to be eaten'

tyrohsemy

'pestilence, disease to be caught'

tarikasemy

'seedlings to be planted, seeds'

takohsemy

'jungle to be cut down'

The differences are: (1) the Apalai t-V-se form occurs only with past completive meaning; (2) it functions as a verbal predicate of a clause or sentence; and (3) it can cooccur with the underlying subject of the transitive as in: kaikuxi tonese eya 'The jaguar (was) seen by him.' The Apalai t-V-se form sometimes occurs with the auxiliary (copular) verb and functions as an adjectival complement (see (79)). Not all verbs, however, cooccur with the copula in their t-V-se form; examples of those which do are: toorihse 'it died'

toorihse kynexine 'it was dead'

tose 'he ate it (meat)'

tose kynexine 'it was eaten'

Examples of those which do not cooccur with the copula are: tooehse 'he came'

* tooehse kynexine

tonese 'seen'

* tonese kynexine

tuenikehse 'forgotten'

* tuenikehse kynexine

tanoryse 'dried'

* tanoryse kynexine

6 Causatives The intransitive verb in Apalai is made causative by the derivational affixes -ma, -nohpo. When the intransitive construction is causativized the intransitive subject becomes the object of the causativized verb and a new subject, the causer, is introduced:

SO

Edward and Sally Koehn

(80a)

poeto otuh-nöko child eat- CONT 'The child is eating.'

mana 3+be+PRES

(80b)

poeto otuh-ma- Vko child eat-CAUS-CONT 'He is feeding the child.'

(8la)

nyh- nöko mana sleep-CONT 3 +be +PRES 'He is sleeping.'

(81b)

poeto nyh- ma- Vko mana child sleep-CAUS-CONT 3+be+PRES 'He is putting the child to sleep.'

(82a)

tapyi tae nutüta- no house out of 3S3O-leave-IP 'He left the house.'

(82b)

tapyi tae nututa-nohpo- no house out of 3S3O-leave-CAUS-IP 'He caused (him) to leave.'

(83a)

paru tu- kuma-se water NF-rise- CMPL 'The river rose.'

(83b)

paru tu- kuma-nohpo-se river NF-rise- CAUS-CMPL 'He caused the river to rise.'

mana 3+be+PRES

ey-a 3- by

When a transitive construction is causativized the causer again becomes the subject, the subject of the basic transitive clause has the form of an oblique object, and the direct object remains the direct object (see (84b)). The transitive verb in Apalai is made causative by the suffix -po. (84a)

pape aroase paper 3 +take-RP '(He) took the paper (book, letter).'

(84b)

mame then "Then Then

pape aropoase ropa ey-a paper 3 +take-CAUS-RP back 3- to or by he sent the paper back to him again' or he caused him to take the paper back again.'

rahkene really

APALA1

(85)

y-mety 1-loincloth+POSSN

eneh- po- ko bring-CAUS-IMP

51

aja 1 + mother+POSSN

a, t- yka-se papa by NF-say-CMPL 1+father+POSSN 4 "Get my mother to bring me a loincloth," my father said.' Note the ambiguity in (84b), where eya may refer either to the recipient or to the causee. Some intransitive stems which have been causativized by the suffix -ma may be causativized a second time with either the intransitive causativiser -nohpo or the transitive one -po. (86a)

aimo otuh-ma- Vko boy eat-CAUS-CONT 'He is feeding the boy.'

mana 3+be+PRES

(86b)

aimo otuh-ma- poVko mana boy eat-CAUS-CAUS-CONT 3+be+PRES 'He is getting (someone) to feed the boy.'

(86c)

aimo otuh-manohpo-Vko mana boy eat- CAUS- CAUS- CONT 3+be+PRES 'He is overseeing the child's eating.'

The verb ka I a 'say, do' and the copula exi I eh la 'be' both have the causativized form kamexipo 'cause to be, say, or do'. (87a)

koih a- Vko paddle do-CONT 'He is paddling.'

mana 3+be+PRES

(87b)

koih ny-kamexipo- no paddle 3- do+CAUS-IP 'He caused (him) to paddle.'

(88)

tehme ny-kamexipo- no quiet 3- be+CAUS-IP 'He caused (him) to be still.'

(89)

epo hnae ny-kamexipo- no enough 3+be+PRES 3- say+CAUS-IP 'He made him say, "It's enough." '

52

Edward and Sally Koehn

7 Comparatives There are three ways in which comparison is expressed: (i) Juxtaposition of negative-positive clauses: (90)

mopo Mopo

zumo pyra big NEG

kyn-exi-ne, 3- be-DP

akono 3+brother in law+POSSN

zumo big 'Mopo was not big like his brother-in-law.' (91)

mokyro that one

zumo n-exi-ase big 3-be- RP

apito-rymano begin-NOMLZR-DENOM-NOMLZR

okomino exi-kety zumo-hko pyra after be- NOMLZR big- quite NEG 'The first one was big. The next not quite so big.' (ii) Successive positive clauses in which certain particles mark the degree of the quality: (92)

kure akuri good coati

n-ase, 3-be+PRES

poinoko wild pig

kure kuhse, good superlative

tomepore delicious 'Coati is good, wild pig (is) really good.' (i.e. "The wild pig is better than the coati.') (iii) Within the clause by use of a postpositional relator motye: (93)

men motye akuri zumo mana squirrel greater coati big 3+be-l-PRES coati is bigger than a squirrel.'

8 Coordination There are no exact equivalents for "and", "but", and "or" for expressing coordination on the clause or phrase level. Coordination is accomplished by juxtaposition of two or more clauses with nonfinal intonation on all but the final clause of the series, or by the use of postpositional particles, (e.g. tökehko, kehko 'etc.' (Note: 'etc' is the meaning of the two Apalai words)). Some additive type examples from texts are:

APALAI

(94)

mame tokare then ADJ-openly t- akoh-se. NF-cut-CMPL

pyra NEG

53

tosa- ry 3REFLX-place-POSSN

akoh-ko. cut- HIS

paruru banana

opomoko. plant(manioc)-HIS

aryka-ko plant-HIS Then (he) secretly cut himself a field. (He) cut it, planted manioc and bananas.1

(95)

senohne this +PL

ϊ- kyry-ry 3-food-POSSN

to- kehko PL-etc

t- yese. paruru NF-cook-CMPL banana

t- ye- se kehko. t- upurihma-se. NF-cook-CMPL etc NF-mashCMPL 'She cooked his food. She cooked bananas, etc., mashed and strained them.'

t- ahpi- se NF-strain-CMPL

Both active and passive verbs can be coordinated as clauses but not an active and passive in the one construction ((94) and (95)). Adversative coordination is accomplished by juxtaposition of negative and positive clauses with or without the presence of the conjunction yrome 'but':

(96)

morara thus

pyra NEG

exi-ko. be- IMP

i-rumeka- ko. 3-abandon-IMP

a-ramariny 2-feather suit-POSSN

senohne this +PL

kyry-neme do- NOMLZR-DENOM

exi-ko be- IMP

rahkene really 'Don't be that way, but abandon your feather suit, and be my cook (and wife).'

(97)

morara like that

pyra NEG

ke POLITE

exi-ko be- IMP

kure well

atamorepa-ko REFLX-teach- IMP 'Don't be like that but try to do it well.'

(98)

ytoyto pyra hxo go+REDUP NEG INTNSF

ywy repe. I FRUST

yrome maixi flu but

eneh- n ko hxo toto bring-CONT INTNSF 3PL Ί don't go anywhere but they bring the flu here.'

54

Edward and Sally Koehn

The alternative coordination is expressed as a proposition and "if not then y," or marked by a postpositional particle, such as hkotano ((99) and (100)), where there is a semantic component of "or": (99)

küre well

hkotä rather

mase 2+be+PRES

j-epe. 1-friend+POSSN paxinepo arrow

morara like that

potyry tip+POSSN

a- htao be-if

emuhka-Vko hitCONY

mase 2+be+PRES 'You will have to do better my friend, or if you keep on that way (don't improve), you will be hit by an arrow tip.' (100)

eropa hkotano 1+2+go+IMP rather

u-muku-ru 1-son- POSSN

ky- nomo- Vko 1+2-leave-CONT

mana 3+be+PRES 'Let's go faster (and not the way we are) or my son will leave us behind.' In coordinated clauses all constituents except the verb can be omitted, with the subject and object being expressed only as verb affixes. The normal means used for coordinating phrasal constituents of the clause is juxtaposition in the expansion part of the clause ((101) and (102)). (101)

mame käpo then game

upiko, look for- HIS

itamuru much

käpo, game

maixipuri tapir

to- kehko kapau to- kehko emero PL-etc deer PL-etc all kinds 'Then he hunted game, a lot of it, tapir and deer, etc.' (102)

mame oyhto- to-ko then descend-PL-HIS to- kehko, PL-etc

mokä those

ratoratori vulture species

moromorori vulture species to- kehko, PL-etc

kurumu to- kehko ayra to- kehko king vulture PL-etc vulture species PL-etc 'Then they descended, the magot eaters, the knife bills, the king vulture group, and the blackheaded group.'

APALAI

55

(103)

j-ene j-ene a- Vko eremia-Vko ywy 1-see 1-see say-CONT sing- CONT I ' "Someone sees me," I would say singing.' The phrases being coordinated can be comprised of a single Sometimes the comitative maro can mark coordination: (104)

konohno brother in law

to group

n-oepy-ase 3-come-RP

jakiku Jakiku

word.

to group

maro etiino with Pedrinho 'My brother-in-law Pedrinho came with Jakiku and group' or 'My brother-in-law Pedrinho and Jakiku came.' More complex nominalized constructions may also be coordinated juxtaposition: (105)

aja 1 +mother+POSSN

to group

papa l+father+POSSN

orih-topoVpyry die-NOMLZR-PAST+POSSN

by

orih-topoVpyry die- NOMLZR-PAST+POSSN roropa also

enara sero poko nase these this occ.with 3+be+PRES 'This is about the death of my mother and father.' 9 Pragmatic and discourse characteristics Certain types of discourse-conditioned phenomena are referred to in other sections of this paper: ellipsis (sect. 3), anaphora (sect. 13), and discourse particles (sect. 21.2). Other pragmatic and discourse characteristics are being studied, but there is nothing definitive to report as yet. 10 Interrogatives The yes-no questions are marked by question intonation: high pitch with stress on the penultimate syllable, low pitch on the final syllable.

(106) m-oeh- no? 2-come-IP 'Hello!' or 'You have come?'

56

Edward and Sally Koehn

(107) m-oeh- no ropa? 2-come-IP again 'Hello!' or 'You returned?'

(108) tutuko se pyra Brazil nut want NEC 'Don't you want Brazil nuts?'

hm-a? 2- be+PRES

(109) omise hm-a? hungry 2- be +PRES 'Are you hungry?' When the yes-no question is used as a greeting the expected response is affirmative unless the one making the response makes an attempt at humor. Other yes-no questions are neutral.

(110) m-oeh- no aimo? 2-come-IP boy 'Hello, my boy' or 'You have come, my boy?'

(Π1)

_

y,

oehno kene! yes, 1 +come- IP just now 'Hello!' or 'Yes, I have just come.'

Question word questions are marked by the question word as well as by special word order, the question word generally coming at the front or near the front of the sentence. Any of the main clause constituents can be questioned: (112)

onoky n-oeh- no? who 3-come-IP 'Who arrived?'

(113)

onoky u-muku-ru who 1-son- POSSN 'Who ate my son?'

5- no? eat-IP

APALAI

(114)

onoky keh sekere who POLITE talk 'Who is it that's talking?'

(115)

onoky hm-a? who 2- be +PRES 'Who are you?'

(116)

onoky m-uo- no? who/what 2- kill-IP 'What did you kill?'

(117)

onoky a m-ekaro-no? whom to 2- give- IP To whom did you give (it)?'

(118)

oty ekaro-Vko m-a what give- CONT 2- be+PRES 'What will you give me?'

57

a-Vko? do-CONT

y-a? 1-to

Peripheral constituents (adjuncts) may be specific question words or may be general question words followed by a postposition or nonfmite verb form, depending on the precise semantic distinction being made. (119)

otoko- ino m-eneh- no? where-from 2- bring-IP 'Where did you get it?'

(120)

oze n-ae? which way 3-be+PRES 'Which direction is' it (from here)?'

(121)

otoko m-exi- ne? where 3- be- DP 'Where were you?'

(122)

otoko nohpo n-a where woman 3-be+PRES 'Where is my wife, mother?'

(123)

otara a- htao oeh- nöko what be-when come-CONT 'When is he coming back?'

(124)

tätahtao m-uo- no? what time 2- kill-IP 'When did you kill it?'

aja? mother +POSSN ropa again

n-ae? 3-be+PRES

58

Edward and Sally Koehn

(125)

otäto hxo emypo- Vko how INTNSF revenge-CONT 'How am I going to get revenge?'

(126)

oty what

ka-toh do-PURP

ahno? 1+be+PRES

o-netapa-ta, 2-DETRANSVZR-kill- DETRANSVZR +IP

mopo? Mopo 'Why did you murder, Mopo?' (127)

otäto t- yri- se how NF-do- CMPL 'How did he do it?'

ey-a? 3- by

(128)

oty ka-se yto-Vko m-a? what do-PURP go- CONT 2- be+PRES 'For what purpose are you going?'

The (verbal) action can be questioned in a nonspecific way by using either a postpositional question phrase with the copular verb or by using the verb ri 'to make': (129)

oty poko m-a? what occ.with 2- be+PRES 'What are you doing?'

(130)

oty hxo ri- Vko m-a what INTNSF make-CONT 2- be+PRES 'What are you making, mother?'

aja? 1+mother+POSSN

The expression or quotation clause type can also be questioned: (131)

otara n-ykä? what 3-say/do+DP 'What did he say (or do)?'

(132)

otara aVko n-a? toh what say/do-CONT 3-be+PRES surprise ' "What is it saying?" 'Oh! I was startled." '

ereh- no 1+startle-IP

59

APALAI

Certain constituents of subordinate clauses can be questioned as in the following sentences: (133)

otoko where

xixi sun

a- htao be-when

erama- Vko turn back-CONT

ropa again

syt- a- tose? 1+2-be-PL+PRES 'When will we all turn back?' (134)

oty what 'What 'After

ka χΐρο apoto t- uka- se ey-a? do first fire NF-light-CMPL 3- by did he do before he lit the fire?' (literally, doing what did he light the fire?')

(135)

onoky any- se n-yt what lift- PURP 3-go+IP 'What did they go to catch?'

toto? 3PL

The possessor constituent of a noun phrase can be questioned but not the possession: (136)

onoky tapyi-V who house-POSSN 'Whose house (is) this?'

sero? this

(137)

onoky rato- V m-ematonanoh-no? who knife-POSSN 2-stealIP 'Whose knife did you steal?'

The nominal constituents of a postpositional phrase can be questioned but not the postposition itself (129). Only one constituent of a sentence can be questioned. The question word usually occurs in initial position in a question but can be preceded by an introducer (see sect. 1.6) or by information of greater prominence than the question. When the questioned constituent moves to the front of the sentence, any phrase level modifiers go with it. When other constituents precede the question word in a sentence they are usually phonologically dislocated. (138)

mame oty poko m-a? then what occ.with 2- be+PRES 'Then what are you doing?'

60

Edward and Sally Koehn

(139)

xihxi, kasuru se, otäto rahkene? sister beads want how many really 'Sister, (you) want beads, how many?'

(140)

j-uru, otoko 1-bread+POSSN where 'Where is my bread?'

(141)

konohno eky, otoko brother in law pet+POSSN where 'My brother-in-law's pet, where is it?'

n-a? 3-be+PRES n-ae? 3-be+PRES

Question words are also used in rhetorical questions. The intent of the rhetorical question is not to ask for information but to give it or to express surprise, frustration or annoyance. (142)

oty ropa rehta moro? axikaru what again FRUST there sugar cane 'What else (was) there? Sugar cane.'

(143)

onoky who

j-ezuh- nöko 1-answer-CONT

hxo INTNSF

n-a 3-be+PRES

roropa also

ywy rokuh osezuh- nöko äse I DEDUCT REFLX-answer-CONT 1+be+PRES 'And who will respond (as I speak)? It looks like I will respond to myself.' Answers to questions can be in the form of incomplete sentences, or simply the response words y 'AFFIRMATIVE' or arypyra 'NEGATIVE'. (144)

onoky hm-a? ywy who 2- be+PRES I 'Who are you?' '(It is) I.'

The response to a question can be a counter question: (145)

xihxi, sister otoko where 'Sister, 'Where

kasuru beads

se, want

otäto rahkene? how many really

hn-a? otyh se? 3- be+PRES what this you want beads, how many?' are they?' 'What (are) these?'

APALAI

61

The question intonation (see (106-111)) is obligatory only with yes-no type questions.

11 Imperatives Imperative sentences are signalled by the morphology of the verb, the imperative intonation pattern, and the absence of a subject. Often they are accompanied by a vocative or other evidence of direct speech. The verb suffix which marks tense, aspect and number is affected and the particles kety 'EMPHATIC' or pahne 'POLITE' may also occur. There are special imperative forms for all person-number combinations. The complete paradigms for both motion and non-motion imperatives are given in sect. 18.3. Examples of non-motion imperative sentences follow: (146)

i-kaparu nymyry apoi-ko 3-war club+POSSN genuine grab-2IMP 'Grab his war club, brother-in-law.'

(147)

ikuh-xi GEN.PREF-try- 1IMP 'Let me just try it.'

(148)

äpo n-oeh- no IMP 3-come-IP 'Let him come.'

(149)

s- otuh-ne ropa y-pary 1 +2-eat- IMP again 1-grandchild-POSSN 'Let's eat again, grandchild.'

rokene only

Examples of motion imperative sentences are: (150)

nohpo eneh- ta aja woman bring-MOT. IMP mother 'Go fetch me a wife, mother.'

(151)

ene-tamy see-MOT. IMP 'Go and see.'

(152)

sepy- tase. eropa 1 +2-bathe-MOT. IMP 1 +2 +go +IMP. DL 'Let's go for a swim. Let's go.'

(153)

sepy- tatose. ehmaropa 1 +2-bathe-MOT. IMP .PL 1 +2 +go +IMP. PL 'Let's all go for a swim. Let's go.'

kono brother in law

62

Edward and Sally Koehn

Negative imperatives require the use of the copula and do not have a motion category. The person, number, and imperative are signalled in the affixes of the copula. (154)

yto-pyra eh-to-ko go- NEG be-PL-IMP 'Do not go (you all).'

(155)

y-pyre an-aro- pyra 1 -bow, arrows +POSSN 3-take-NEG '(You all) do not take my bow and arrows.'

eh-to- ko be-PL-IMP

There is also a negative imperative prefix os- cooccurring with the suffix -no, although this is rarely used: (156)

tupito epery os- enah-no, n-ase field fruit NEG-eat- IMP 3-say+PRES ' "Don't eat the fruit of the field," he said.'

As for degrees of imperative, the continuum from polite invitation through moderate to strong order is communicated by postpositional particles: se, ke, pahne, papa, kety, ty. Polite: (157)

otuh-ko se .

(158)

ene-ko ke .

. .

. 'You may eat' or 'Do eat'. . 'Look, ok?'

Moderate: (159)

ene-ko pahne

. Take a look at that, please.'

(160)

enah-ko papa

. 'Eat it and I mean that.'

Strong: (161)

eroh-ko kety .

. 'Get busy and work!'

(162)

t-yri-ko ty .

'Do it whether you want to or not!'

APALAI

63

The usual response to an imperative is behavioral, whether reported or implied. Other responses are: affirmative or compliant, negative or noncompliant, and hesitant or requesting clarification.

(163)

eropa ke 1+2+go+IMP POLITE

j-epe. 1 -friend +POSSN

mame then

oseh- to- ko rahkene come-PL-HIS really ' "Let's go, friend." Then they came.' (164)

piu son in law

a j-oh by 1-meat+POSSN

tuopo-ko, GEN. PREF-kill-CAUS-IMP

kuto. frog

reh pina, EXPECTANT arrow

t- yka-se. NF-say-CMPL

se here

pina pona ka-ra poko m-a- tou kuni arrow onto do-NEG occ.with 2- be-PRES grandmother ' "Get son-in-law to shoot me some game frog. Here is an arrow," she said. "That is not the kind of game you get with arrows, grandmother." ' (165)

tana over here

sä like

enehbring-

ko tarn! IMP grandfather

syryryry gliding

mya rahkene away really ' "Bring (the canoe) over here grandfather." He glided farther away.' (166)

papa father

z-oty 3-food+POSSN

i- tonka- ko 3-gather-IMP

okomo. larva

nary, a- Vko pakero oserehnöko doubt say-CONT already REFLX-be afraid-CONT ' "Gather father's food, the larva." "I doubt," he (was) saying already, being frightened.'

64

(167)

Edward and Sally Koehn

y-mety 1-cloth+POSSN a, by oty what

t- yka-se NF-say-CMPL ka-toh, be-NOMLZR

eneh- po- ko bring-CAUS-IMP

aja l+mother+POSSN

papa. 1+father+POSSN y-mety 1-cloth+POSSN

eneh- po- ko, bring-CAUS-IMP

a- Vko n-ae? t- yka-se j-eny say-CONT 3-be+PRES NF-say-CMPL 3-mother+POSSN ' "Get my mother to bring me a loincloth," my father said. "Why is he asking for a loincloth?" said his mother.' 12 Negation 12.1 Sentence negation. The negation of a sentence is accomplished by a derivational suffix -pyra / -ra 'NEGATIVE' on the verb transforming it into a negative adverbial complement of a copular verb. The distribution of the allomorphs is phonologically defined (-ra occurring with all verb stems ending with a, and -pyra occurring with all others. The form -pyny is a negative nominalizing suffix One who does not do the action'). The person-marking subject prefix and the suffixes marking tense, aspect, and number appear on the copula. In the case of transitive verb stems the person object marker is prefixed to the negative adverbial. The inflected copula is sometimes omitted, as in (169b): (168a)

isapokara ene- no jakuruaru lizard 1 +see-EP Ί saw a jakuruaru lizard.'

(168b)

isapokara on-ene-pyra jakuruaru lizard 3- see-NEG Ί did not see a jakuruaru lizard.'

(169a)

morarame n-otuh-no then 3-eat- IP 'Then his grandchild ate.'

(169b)

morarame otuh-pyra i-paniry then eat- NEG 3-son in law-POSSN 'Then his son-in-law did not eat.'

(170a)

ywy exi-ase I be- RP 'It was I.'

a-ken 1-be+IP

i-paniry 3-son in law-POSSN

APALAI

(170b)

ywy ka-ra n-exi-ase I be-NEG 3-be- RP 'But it was not I.'

65

repe but

12.2 Constituent negation. Adverbiale and postpositional relators are negativized by the same form described above for sentence negation: pyra. Wheq, it occurs with constituents other than verb stems, it is written as a separate (postpositional particle) form, rather man a suffix. (171)

mame yto- ko then 3+go-HIS

nakua-taka, water-to

nakua-taka water-to

pyra. NEG

aza hkotoko t- oyto-se kyn-exi-ne? where perhaps NF-go- CMPL 3- be- DP 'Then he went to the river, not to the river. Where was it he went?' (172a)

(173)

axi pyra oehno fast NEG 1+come-IP came slowly.' ynan-exi-ase 1 +3- be-RP

(172b)

axi oehno quickly 1+come-IP came quickly.'

ameke pyra rahkene far NEG really

owotoh stand up- NOMLZR

pona to 'We were not far from the destination.' (174)

i-pytyme pyra mosero 3-wife+POSSN-DENOM NEG that one 'She is not his wife but his sister.'

(175)

mokyro that one

i-nio 3-husband+POSSN

pekä flap

a-oryxi- ry 3-sister-POSSN pekä flap

kuatiwi grab on

typoty-ry ae rokene kae pyra 3REFLX-beak- POSSN by only high NEG 'The husband flew clumsily and reached a low perch with his beak.' For most simple adverbs the negative expresses the antonym: mosa 'tall', mosa pyra 'short', tuhke 'many', tuhke pyra 'few', kure 'good' or 'well', kure pyra 'unwell' or 'bad'. When the constituent to be negativized is a subordinate clause, the verb stem is inflected by the negative derivational suffix -pyra / -ra.

66

(176)

Edward and Sally Koehn

yto-pyra go- NEG

j-a- htao 1-be-if

kaeae high +NOMLZR-by

äkorehma-Vko 2+help- CONT

ase sero poko 1+be+PRES this occ.with 'If I don't go on the airplane, I'll help you with this.' Nominal constituents can also be negativized with the resulting meaning that there were none of that item around or that the participant in focus did not own one, or that the identity of the nominal is in question: e.g. 'not a man but a spirit.' (177)

orutua pyra man NEG 'There were no men around.'

(178a)

tonu-remy ADJVZR-eye-ADJVZR-NOMLZR 'seeing person'

(178b)

tonu-re ADJVZR-eye-ADJVZR+NOMLZR 'That one is a blind person.'

(179)

ka-ra be-NEG

mokyro that one

rato pyra knife NEG 'There was no knife.'

(180a)

wo- tono shoot-NOMLZR 'sharp shooter'

(180b)

wo- to ka-ra shoot-NOMLZR be-NEG 'That one is no hunter.'

mokyro that one

The underlying form of \votd is wotono (see sect. 22.6 for rules). (18la)

onokyro wo- ne animal shoot-NOMLZR One who shoots animals'

(181b)

onokyro on-uo- pyny animal 3- shoot-NEG+NOMLZR One who does not shoot animals'

APALAI

67

(182a)

onokyro wo- ry waro exi-kety animal shoot- NOMLZR know be- NOMLZR One who is knowing how to shoot animals'

(182b)

onokyro wo- ry waro-V ka-ra animal shoot- NOMLZR know-NOMLZR be-NEG 'He is not one who knows how to shoot animals.'

Descriptives can be negativized by pyra: (183a)

typy- e (183b) ADJVZR-wife-ADJVZR 'married'

typy- e pyra ADJVZR-wife-ADJVZR NEG 'single'

(184a)

tonu-re ADJVZR-eye-ADJVZR 'seeing'

(184b)

tonu-re pyra ADJVZR-eye-ADJVZR NEG 'bund'

(185a)

maixi-hpe cold- infected 'sick with a cold'

(185b)

maixi-hpe pyra cold- infected NEG 'not sick with a cold'

Some nominal negations follow: (186)

enah-pyny eat- NEG+NOMLZR 'fruit-type food not to be eaten' (see sect. 15.4 for discussion of -pyny).

(187)

tonah-se GEN.PREF-eat- NOMLZR 'It is not edible food.'

ka-ra be-NEG

There can be not more than one negation element in a clause, but there can be one in each main and subordinate clause in each sentence. (188)

omise pyra j-a- htao otuh-pyra hungry NEG 1-be-when eat-NEG 'When I am not hungry, I do not eat.'

ase 1+be+PRES

Edward and Sally Koehn

68

13 Anaphora Anaphora is expressed by deletion under the following conditions: (i) deletion of the quotation margin in direct speech sentences: (189)

to! oh!

ohpa fish

ere- ry liver-POSSN

otoko where

hn-a? 3- be+PRES

seny, here

ohpa fish

ere- ry liver-POSSN

j-epe. 1 -friend +POSSN y, yes

J-epe. 1-friend+POSSN

ke with

j-epe. 1-friend+POSSN

tomepore delicious

tomepore? delicious

t5- ne GEN.PREF-eat-NOMLZR

ere- ry liver-POSSN

m-a 3-be+PRES

5- ne. eat-NOMLZR y. OK.

reh expect i-kuh-ko 3-try-IMP

ahse, ohpa 1+be+PRES fish panne AMELIORATIVE

mame t- 5- se then NF-eat-CMPL

ey-a 3- by

' "Oh! the fish liver, my friend." "Where is it?" "Here, the fish liver is delicious, friend." "(Is it) good?" "Yes, I'm an eater of it; I eat fish liver. Try it my friend." "OK." Then he ate it.' The frequency of this deletion seems to vary with different narrators. In a legend text of 139 quotes, 61 are marked with quote tags and 78 are not marked. In another legend of 33 quotes (narrated by another Apalai) only 9 are marked with the quote margin. In a text of made-up dialogue between a father and son, there is not a single quote tag used.

APALAI

69

(ii)

deletion of subject and verb in interrogative sentences:

(190)

eropa! j-epe 1+2+go+MP 1-friend+POSSN yto-Vko, go- CONT

m a

~ 3-be+PRES

i-ty- se 3-poison-PURP

t- yka-se. NF-say-CMPL

otoko na? ikuhpo kuaka? where to lake to ' "Let's go! My friend is going to poison (fish)," he said. "Where (are we going)? (Are we going) to the lake?" '

(iii)

deletion of subject and verb in answers to questions:

(191)

azah where

yto-Vko go- CONT

syt- ah? 1+2-be+PRES

t- yka-se. NF-say-CMPL

ikuhpo kuaka ... t- yka-se NF-say-CMPL to lake ' "Where are we going?" he said. "To the lake ...," he said.' Deletion of subject and verb is very frequent in dialogue. Anaphora is also expressed by personal pronouns in the case of third person referents. Both deictic and nondeictic pronouns are used anaphorically. The free pronouns are also anaphorically related within the clause to the person marking prefix of the verb when they occur as subject or direct object. (192)

sytytase 1 +2-poison fish-MOT. IMP

kokoro tomorrow

j-epe, 1 -friend +POSSN

t- yka-se. i-pyty se ynororo NF-say-CMPL 3-wife- POSSN want 3 ' "Let us poison fish tomorrow, my friend," he said, but he was wanting his wife (the other man's).' (193)

ohpahko sucker fish-EVAL moseh this one

ynoro. 3

s5- tase, 1 +2-eat-MOT. IMP otuh-kose eat- IMP

ke POLITE

t- yka-se NF-say-CMPL j-epe. 1-friend+POSSN

tukasere potu fat nice ' "Let's eat sucker fish," he said. "Here it is. Eat some, my friend. It seems nice and fat." '

70

Edward and Sally Koehn

(194)

päna n-ae UNCERT 3-be+PRES 'This is the one.'

sero this

(195)

mokyro kurumu-ke typy- e that one vulture-INST ADJVZR-wife-ADJVZR 'That one (was) married to a vulture (woman).'

(196)

na BSTROD

y-rato 1-knife+POSSN

yro 3INAN

mkuh really

moro. that

rahkene really

kura good+NOMLZR

rukuku-hxo moro really that "That is my knife, yes. And it certainly is a good one.* There is a third person reflexive prefix /- which occurs with nominale, nominalized verbs and postpositional relators to signal the same referent as the subject of the clause or of a superordinate clause. (See sect. 4 for details and examples.) Anaphora is also expressed by certain discourse particles, including ropa 'back again' and roropa 'additive' which relate the item they modify to something that occurred earlier in the discourse. Also the particle ro 'still' occurs with nominate, personal pronouns, actions and situations to identify them as the same or similar to the antecedent. (197)

m-oeh- no 2-come-IP

ropa? back

t- yka-se. NF-say-CMPL

oehno 1+come-IP

ropa back

ke papa t- yka-se POLITE father NF-say-CMPL ' "Hello, you came back," (he) said. "I came back, Dad," (he) said.' The particle ropa 'again' relates the return of the participant to the location or setting from which he left to begin his travels. (198)

akoroka-ase-Vtyky. 1+clear-RP- finish

mame tapyi then house

zomye around

roropa also

akoroka-ase pitiko atarika-hme exi-ryke 1 +clear-RP little weedy-DENOM be- NOMLZR-REAS finished clearing weeds. Then I cleared a little around the house as well because it was weedy.'

APALAI

71

Other means of anaphoric reference include the particle hro 'in response to your question' and ta 'reported information'. (199)

m-epy- no? y, 2-swim-IP? yes, ' "Did you swim?" "Yes I swam." '

(200)

i-meretamuruh 3-father in law-POSSN

epy 1+swim+IP

ta REPORT

hro RESP

n-ynomoka-no, 3-rejectIP

t- yka-se z-oh wo- to-ise NF-say-CONT 3-meat+POSSN shoot-PL-FRUST 4 "He says his father-in-law rejected him," he said, because he missed his game.' Within the clause there can be a sequence of anaphorically related items. (201)

tamuxi mokyro i-meretamuru old man that one 3-father in law-POSSN 'There was the old man, his father-in-law.'

The deictic personal pronoun mokyro is a cataphoric reference to 'father-inlaw' while the prefix i- 'his' is anaphoric, referring to Ituimano mentioned eight lines earlier in the opening sentence of the story. (202)

m-oeh- no 2-come-IP

ropa? back

y yes,

oehno 1 +come-IP

ropa back

t- yka-se NF-say-CMPL

mokyro i-py- ty that one 3-wife-POSSN ' "Hello, you're back." "Yes, I came back," she said. She was his wife.' In this quotation margin of a direct speech sentence there are two types of anaphora which overlap: (i) the anaphora within the clause, the antecedent being ipyty 'his wife' following the anaphor mokyro, and (ii) anaphora across sentence boundaries, (- 'his' being the anaphor relating to an antecedent orutua 'man' occurring earlier in the discourse. Juxtaposition is the principal form of coordination, and the general rules relating to anaphora across sentence, clause, and phrase boundaries apply; however, the reflexive prefix t- cannot relate anaphorically across sentence boundaries.

72

(203)

Edward and Sally Koehn

arimi wowo repe. epuka-ra monkey 1+shoot+REDUP+IP FRUST fall- MEG Ί shot the monkey repeatedly but it didn't fall.'

kyn-ako 3- be+IP

All subordinate clauses are nonfmite. Anaphora can occur between main and subordinate clauses with the antecedent occurring in the main clause. The subordinate clause may precede or follow die main clause and there may be anaphoric or cataphoric (i.e. "backward anaphora") reference. Where the subject of the main clause is the antecedent, the reflexive prefix t- is the anaphor in the subordinate clause (204). (204)

poeto-me ta- htao t pa kyn-epuru-ne child-DENOM 3REFLX-be-when T pa 3- sting-DP 'When she was a child T pa was initiated (stung with ants).'

(205)

mame oepyase ropa kajama aro- χϊρο then 1 +come-RP again manioc meal take-first 'Then I returned having first delivered the manioc meal.'

In (205) the subject of the subordinate clause is coreferential with the first person subject of the main clause and is not overtly expressed. In example (206) the third person experiencer of the subordinate clause is the anaphor and its antecedent is m- 'third person subject' of the main clause. (206)

orih-py die-NEG+NOMLZR

m-a 3-be+PRES

ro yet

repe FRUST

t- ynahpa-se exi- ryke NF-poison-CMPL 3 +be-NOMLZR-REAS 'He should not have died but (he did) because of his being poisoned.' A sequence of subordinate clauses can have anaphors that refer to the same antecedent in the main clause: (207)

atamorepa-po- Vko REFLX-teach- CAUS-CONT

mo that

makina-ponaka machine-on

tatamorepa- tohme moro poko 3REFLX-REFLX- teachNOMLZR-DENOM that occ.with pape mero- tohme tyya paper write-NOMLZR-DENOM 3REFLX-by '(He is) causing himself to be taught on the machine in order that he learn, in order that he write on paper (a letter).'

APALAI

73

14 Subordinate clauses 14.1 Subordinate constructions in general. Only nonfmite verb forms occur in subordinate clauses. These clauses consist of nominalizations, often embedded in postpositional phrases, or some other construction, derived from a verb by the addition of affixes, and functioning as an adverbial constituent of the main clause. The nominalization or other derived form can be inflected for person of possessor and can contain clause elements such as adverbs, locatives, etc. The subordinate clauses usually follow but can precede the main clause. Fronting of the subordinate clause, like other elements in a main clause, is used for emphasis or highlighting. If the main clause verb has a direct object or subject or some other element preceding it, then die adverbial clause will follow the main clause. Nominalized constructions that occur as the complement of a copula normally precede the copula, which is syntactically the main verb. The following examples illustrate: (i) nominalization embedded in a postpositional phrase: (208)

tooriky-ry 3REFLX-dieNOMLZR

se pyra j-epe want NEG 1-friend+POSSN

kyn-exi-ne repe 3- be-DP FRUST 'My friend did not want to die (but he died anyway).' (209)

ywy papa I 1 +father+POSSN

ooriky-ry die- NOMLZR

a concerning

tuaro pyra GENPREF- know NEG did not understand my father's death' (210)

sero n-ase j-eroh- topovpyry this 3-be+PRES 1-work-NOMLZR-PAST+POSSN "This is about the work which I have done/

(ii)

other derived form:

(211)

moroto-ino there- after

kV-yto-ne 3-go-DP

poko occ.with

je ekepyry 3 +mother +POSSN body

eseka- hpöko eahma- se bite- PAST+NOMLZR-GRP invite- PURP 'After that he went to invite the ones who killed his mother.'

74

Edward and Sally Koehn

(iii)

subordinate clause following and preceding the main clause:

(212)

ko to wo- ne zumo-hxo wild turkey GRP 1 +kill-DP big- DSTNSF Ί killed wild turkey when I was bigger.'

(213)

poeto-me ro j-a- htao kana anypito- ne child-DENOM still 1-be-when fish 1+catch-begin-DP 'When I was still a child I began to catch fish.'

j- a- htao 1-be-when

The main clause with the finite verb has the primary tense marking. Derived nominals may also have a more limited tense marking when they are functioning as a subordinate clause. The basic word order of finite clauses is OVS. In nominal constructions the NP (i.e. possessor), which is the underlying S of an intransitive construction and the underlying Ο of a transitive construction, always occurs preceding the derived nominal. The underlying S of a transitive construction is most frequently found occurring before the derived nominal and before the possessor (object) NP, although it can also occur after the nominal. It is marked in the nonfinite subordinate clause by the postposition α 'to, by', as in (69, 70) which are repeated here as ((214) and (215)). In both these examples, this a phrase follows the derived form. (214)

toh 3REFLX-meat+POSSN

wo- toise shoot-FRUST

y-a I-by

y-meretamu1-father in law-

ru y-nomo- no POSSN 1-abandon-IP 'My father-in-law abandoned me for my almost having shot his meat.' (215)

seromaroro now

atamorepa-Vko REFLX-teach- CONT

ropa again

mose this one

pape paper

mero- tohme tyya write-NOMLZR-DENOM 3REFLX-by. 'Now this person is studying again to write on paper himself.' 14.2 Relative clauses. There is no regular finite form of relative clause. There are various means used to obtain the same effect as such a clause: simple nominalization; placing NPs together in a paratactic relationship, with intonational break; descriptive sentences usually involving an equative clause; or some combination of these means.

APALAI

75

(216)

mame kanawa aro- ko repe zakare konöto then canoe take-HIS but alligator large 'Then he took a canoe which was only a large alligator but (it functioned as a canoe).'

(217)

ku2S1O-

akuohko take across-IMP

rokene only

tarn, t- yka-se grandfather NF-say-CMPL

mokyro, kuto n-akuotyhpyry that one frog 3-take across-NOMLZR +PAST ' "Take me across the water, grandfather," said the one whom the frog had taken across.' (218)

arimi monkey

s- tatose 1 +2-eat-MOT.PL.IMP

to ka-ra NOMLZR be-NEG

onokyro game

t- yka-se NF-say-CMPL

woshoot-

on-uo- pyny 3- shoot-NEG+NOMLZR

onokyro wo- ry waro-V ka-ra game shoot-NOMLZR know-NOMLZR be-NEG ' "Let's go eat monkey," said the one who was not a shooter, who never shot game, who didn't know how to shoot game.' (219)

mame imepy then another

t- ooeh- se NF-come-CMPL

makarita Margaret

samo, like

imepy another

ropa pone samo again Pone like 'Then another one came, one like Margaret, and (there was) another one like Pone.' In example (217) the nominalized form kuto nakuotyhpyry 'the one whom the frog had taken across' follows the pronoun mokyro 'that one' in a paratactic relationship, but it could stand alone as subject of the clause. 14.3 Time and conditional clauses. In the adverbial time clause there is no marking of tense distinctions. These are shown only in the finite verb forms of the main clause. There are four forms of derived time adverbial distinguished by their suffixes: -htao 'simultaneous action,' -xipo 'after, first' tohto 'before,' -tane 'until'. All take the inflectional prefixes that mark the person of the possessor.

76

(220)

Edward and Sally Koehn

seraranoVpo like this- NOMLZR-PAST

ka-ra be-NEG

ase 1+be+PRES

nuaseme j- a- htao young person-DENOM 1-be-when "That's not the kind I was when I was young.' (221)

moroto there

kamakuao j-uane Kama(river)-at 1-dance-DP

poeto-me child-DENOM

ro j- a- htao still 1-be-when Ί danced there at the Kama river when I was still a child.' (222)

araiba t oeh- n ko m-a nonoaty Araiba GRP come-CONT 3-be+PRES tree bark 'After cutting the bark, Araiba and family are coming.'

(223)

papa ruka kana- ry apiro-ase father Ruka canoe-POSSN open-RP 'Father opened Luke's canoe before going on.'

(224)

senohne this+PL

akoroka-Vko clear- CONT

toto 3PL

akoh-χϊρο cut- first

toyto-tohto 3REFLX-go- before

tumyko 3REFLX-father +POSSN-PL

oeh- tane ropa come- until again 'They are clearing this until their father returns.' The foregoing examples show that tense is marked only in the verb of the main clause, the time adverbial always having the same form. There is no distinction between time and conditional adverbials. The form -htao is glossed 'if or 'when' depending on the context (225). (225)

omise aw-a- htao otuh-ko hungry 2- be-if/when eat- IMP 'If you are hungry, eat' or 'When you are hungry, eat.'

14.4 Contrary to fact conditionals. Contrary to fact conditionals also have the form -htao (sect. 14.3) but in this case the main clause has a special nominalized verb form:

APALAI

(226)

opotu fast water

nymyry genuine

a- htao be-if

77

a- rypyra be-NOMLZR-NEG

kahpyry ugly

exi-ry toh asakororo be- NOMLZR 3PL both 'If the water had been fast there would be nothing left of the two of them.' 14.5 Manner subordinate clauses. Manner clauses take the form of derived nominals followed by certain postpositions, including me 'DENOMINALEER' (sometimes written as a suffix -me). The manner clause answers the question ot to 'how?' (227)

oepyase kana 1 +come- RP fish

anymy-ry liftNOMLZR

me axi DENOM fast

pyra NEG

rahkene really Ί came slowly as I caught fish.' (228)

kana fish

anymy-ase 1+lift-RP

j-oepy- ry 1-come-NOMLZR

me kaeDENOM high-

tokohxo NOMLZR-INTNSF Ί caught a great pile of fish as I came.' 14.6 Purpose clauses. There are two basic types of purpose clauses: motion purpose and nonmotion purpose. Motion purpose clauses occur only as the complement of verbs of motion: intransitive verbs oep 'come', and yto 'go', and transitive verbs enep 'bring', aro 'take', and enyok 'send'. The nucleus of the complement clause is a verb stem plus the suffix -se which marks it as 'MOTION PURPOSE'. Transitive forms also have a prefix marking the person of the possessor (object). (229)

seroae today

karaiwa nonindian

to GRP

n-yto-no 3-go- IP

parahta rubber

wo-se cut-PURP

anatukuaka Anatu(creek)- at 'Today a group of men went to cut rubber along the Anatum.' (230)

moroto-ino ynan-yto-ne nyh- se anaxikarahpi-taka there- from 1 +3- go-DP sleep-PURP Anaxikarahpa-to 'From there we went to Anaxikarahpano to sleep.'

78

Edward and Sally Koehn

(231)

pape eneh- se oehne paper bring-PURP 1 +come-DP came to bring a letter.'

(232)

ky- pipoh-se oeh- nöko l +2-beat- PURP come-CONT 'They are coming to beat us up.'

toto 3PL

The underlying subject of motion purpose clauses is always the same as the subject of the main clause, which accounts for the fact that there are no person-marking prefixes in nontransitive forms. Prefixes on transitive forms follow the usual rule in nonfmite derived forms of being possessor prefixes that refer to the underlying object. Nonmotion purpose clauses occur as the complement of verbs other than those involving motion. The nucleus of the complement clause takes the form of a verb stem plus -toh-me 'PURPOSE' (-nominalizer-denominalizer). (233)

mame amihto- Vko then 3+string-CONT

tysarah snap

ka-tohme, kana do-NOMLZR-DENOM fish

wo- tohme enara shoot- NOMLZR-DENOM that's all 'Then I will string it so that it snaps, so that it shoots fish.' (234)

morara thus

exi-ryke be- NOMLZR-REAS

ynan-urumeka-ne 1 +3- leave- DP

rahkene really

xiaro yna- yhtotohme here 1 +3-descend-NOMLZR-DENOM Therefore we left the place really to descend to here.' 14.7 Cause clauses. Cause clauses have as their nucleus a derived nominal with the suffix -ry 'ACTION NOMINALEATION, POSSESSION', followed by the postposition ke 'because of (REASON), by means of (INSTRUMENTAL)' (sometimes written as a suffix -Jte).This postposition also occurs with simple nouns, usually with the instrumental meaning. (235)

mame moro ynan-urumeka-ne then that 1 +3- leave- DP

ropa again

maikuato-hpe antinfested

exi- ryke be- NOMLZR-REAS Then we abandoned that place again because of its being infested with ants.'

APALAI

(236)

morarame thus like

otuh-pyra eat- NEG

ene-ryke see-NOMLZR- REAS

79

i-paniry 3-son in law-POSSN tyya 3REFLX-by/to

moro that

nuriame filthy

sä like

exi- ryke be- NOMLZR-REAS "Then her son-in-law wouldn't eat, because he had seen it, because it was repulsive.' (237)

morara thus

ka- ryke say-NOMLZR- REAS

ey-a 3- by

yto- ase l +go-RP

toto 3PL

eneh- se ropa bring-PURP again 'Because he said that, I went to bring them back.' (238)

müpo-jana rat- person

on-ene-pyra 3- see-NEG

toto 3PL

yto-ryke go- NOMLZR-REAS

imepy pona another +NOMLZR to 'The rat persons were not seen because they went to another place.' 14.8 Result clauses. Result statements are handled either by means of a cause clause, where the main clause constitutes the statement of the result, or by a sequence of two sentences juxtaposed. In the latter case the result may be implied without any formal marking, or the second sentence may contain the postpositional phrase morara exiryke 'therefore, because of that' or the conjunction naeroro 'therefore'. This usually occurs at the beginning of a sentence and can refer to either the immediately preceding sentence or a longer stretch of the preceding discourse. (239)

anno- hpe people-full of exi-ryke be- NOMLZR-REAS

a- Vko say-CONT myaro over there

kyn-exi-ne. 3- be- DP papa my father

morara that yto-pyra go- NEG

eh-se be-CONT. PAST "They said it was crowded. So Father would never go over there.'

80

(240)

Edward and Sally Koehn

orih-py die- NEG

m-a 3-be+PRES

exi- ryke. be- NOMLZR-REAS

ro yet

morara thus

repe FRUST

t- ynahpo-se NF-poison-CMPL

exi-ryke be- NOMLZR-REAS

ynan-urumeka-ne rahkene 1+3- leave- DP really 'He would not have died but (he did) because he was poisoned. Therefore we left.' (241)

onoky a- Vko who say-CONT kaikuxi-jana jaguar-people

puhko o-uru-Vko, EYEWITNESS 2-tell-CONT asa 2+mother+POSSN

o- ne? eat-DP

naeroro therefore

tehme exi-ko a- Vko ase settled be-IMP say-CONT 1+be+PRES 'Who is it that is telling you the jaguars ate your mother? That's why I'm saying settle down.' (242)

toh 3REFLX-meat+POSSN

wo- toise shoot-FRUST

y-a 1-by

y-meretamu- ru y-nomo-no 1-father in law-POSSN 1-leave-IP 'My father-in-law abandoned me for my almost having shot his meat.'

14.9 Comparative and equative clauses. Subordinate clauses expressing degree such as comparative and equative take the form of derived nominale followed by certain postpositional: saaro 'like', motye 'greater, above', zokonaka 'fraction of (literally 'into its body'), and the manner postposition me 'denominalizer'. (243)

yto-Vko go-CONT

ase 1+be+PRES

bere Belem

pona FAB kana- ryae to FAB plane-POSSN-by

mose yto-topoVpyry saaro that go- NOMLZR-PAST+POSSN like 'I'm going to Belem on a FAB plane like that one did.' (FAB: 'Forga Aerea Brasileira')

APALAI

(244)

tururume running water

bufaro buffalo

81

yto-ry go- NOMLZR

motye more

kawaru horse

yto-Vko axi- hxo go- CONT fast-INTNSF Ά horse goes faster than a running water-buffalo.' (245)

itamurume- hxo moreINTNSF

tose meat

t- apoi-se NF-take-CMPL

oy-a 2- by

ywy I

zokonaka pixo roke t- apoi-se y-a fraction small only NF-take-CMPL 1-by Ά large amount of meat was taken by you. I got only a fraction as much.' 14.10 Desire clauses. Statements of desire are expressed as the complement of the copula. A noun or derived nominal is followed by the postpositional relator se 'desirous of. This relator is the only way to express the verbal concepts 'love', 'like', or 'want' (in the same way that waro 'knowing, aware of is a relator). It occurs with simple nouns and with bound possessor/object pronouns.

(246)

o-se ase 2-like 1+be+PRES Ί like you.'

(247)

y-py- ty se 1-wife-POSSN like 'He wanted my wife.'

(248)

yna aro- ry se eh-se repe 1+3 take-NOMLZR want be-CONT.PAST FRUST 'They were wanting to take us but (we won't let them).'

(249)

yna wory se pyra ase 1+3 shoot- NOMLZR want NEC 1+be+PRES Ί don't want us to get shot (by those people)' or Ί don't want those people to shoot us.'

(250)

j-yto-ry se ase 1-go-NOMLZR want 1+be+PRES Ί want to go to Belem.'

(251)

poponi oepy- ry se ase Poponi come-NOMLZR want 1+be+PRES Ί want Poponi to come to Belem.'

kyn-exi-ne 3- be-DP

bere Belem

(mokaro a) (those by)

pona to bere pona Belem to

82

(252)

Edward and Sally Koehn

j- oh 1-meat+POSSN

wo- ry shoot-NOMLZR

se want

ase 1+be+PRES

poponi Poponi

a by 4 1 want Poponi to shoot game for me.'

14.11 Complex subordinate constructions. There are two types of complex constructions involving subordinate clauses: 1) paratactic and 2) recursive. The paratactic, which is common at all levels of the syntax, consists of a series of two or more clauses which have the same syntactic relationship to the rest of the sentence. They function to: a) enable additional peripheral elements of the clause to be added without overloading any one clause; b) provide increased identification of an item, thus avoiding ambiguity. (253)

mokyro that one

on-ene-pyra 3- see-NEG

j-exi-ryke 1-be-NOMLZR-REAS

exi- ase 1 +be-RP

pata village

pona yto-pyra to go- NEG

typoko-ke ADJVZR-busy-ADJVZR

j-exi-ryke 1 -be-NOMLZR-REAS didn't see him because I didn't go to the village because I was busy.' (254)

otuh-pyra eat- NEG

i-paniry 3-son in law-POSSN

t- one-se NF-see-CMPL

exi- ryke nuriame sä exi-ryke be- NOMLZR-REAS repulsive like be- NOMLZR-REAS 'Her son-in-law refused to eat because he had seen it, because it was repulsive.' The recursive type of complex construction occurs where one subordinate clause is embedded in another, as in (255), where töxiry arory se is the embedded complement of toexiryke.

(255)

töxiry 3REFLX-daughter-POSSN

aro- ry take-NOMLZR

se twant 3REFLX-

oexi- ryke t- aro- se kyn-ako be- NOMLZR- REAS NF-take-CMPL 3- be+IP 'Because he wanted to take his daughter he took her.'

APALA1

83

Nonfinite clauses can also be embedded as the object (256), subject ((257)-(259)), or adjunct ((260)-(262)) of the main clause: (256)

j-oepy- ry eraxima-ko 1-come-NOMLZR wait- IMP 'Wait for my coming.'

(257)

any-to kyn-oeh- ne lift-NOMLZR 3- come-DP 'The lifter (formal inviter to feast) came.'

(258)

moroto kae poroh- toh there high one landing-NOMLZR "There the airstrip is (located).'

(259)

j-umy, 3-father+POSSN

kuto frog

mana 3+be+PRES

n-akuotyhpyry 3-take across-NOMLZR +PAST

z-umy irapare 3-father+POSSN Irapare 'His father, the father of the one the frog took across (is) Irapare.' (260)

mame ynan-yto-ase then 1+3- go-RP

nyh- se sleep-PURP

mokä those

nyh- topoVpyry pona sleep- NOMLZR-PAST+POSSN to 'Then we went to sleep at those people's former sleeping place.' (261)

oepyase kana 1+come-RP fish came fishing.'

anymy-ryme liftNOMLZR-DENOM

(262)

kana anymy-ase j-oepy-ryme fish liftRP 1-come-NOMLZR-DENOM caught fish as I came.'

14.12 Indirect discourse. Indirect statements, assertions, or commands are usually expressed as direct quotes. Rhetorical questions are also used to express certain kinds of indirect questions. Example (263) illustrates one way in which a rhetorical question is used to express an indirect form of question: (263)

otäto mon-exi-ano tarame? yrome zuaro how many 3be-IP perhaps but know '(I) don't know how many there were.'

pyra MEG

84

Edward and Sally Koehn

The use of direct quotation to express commands and statements which would often be translated into English by indirect forms is illustrated in (264M266): (264)

kuto frog

hko etapa-poko EVALUATION kill- CAUS-IMP

n-yka-hko 3-say-EVALUATION

papa father 'Father said, "Go kill frog meat" ' or 'Father had ordered frog meat.' (265)

mose this one

ene-ko see-IMP

panne POLITE

t- yka-se NF-say-CMPL

to3REFLX-

oryxi-ry a sister-POSSN to ' "Look at this one" (he) said to his sister' or 'He told his sister to look at it.' (266)

o-wo-Vko ase t- yka-se 2-kill-CONT 1+be+PRES NF-say-CMPL ' "I am (going to) kill you," he said' or 'He said he would kill me.'

14.13 Direct quotation. Direct quotation is used extensively. It may be marked with a quote tag, vocative, use of first and second person referents in the quote, or a response which is either a speech act or a behavioral response. The quotation can occur before or after the quote tag. There are many examples scattered throughout this paper. SYNTAX OF PHRASE TYPES 15 Noun phrase structure 15.0 Introduction. The noun phrase is defined as the constituent that can occur as the subject or direct object of a sentence or as the head of a postpositional phrase. The internal structure consists of a nucleus and optional peripheral elements. The nucleus is: (i) a single noun or pronoun (simple noun phrase); or (ii) a sequence of nouns, each constituent pair of which is in a possessorpossessed relationship (possessed noun phrase). The peripheral elements are pre- or post-posed modifiers and/or post-posed particles. 15.1 Marking for case. There is no case marking of subject and direct object relations. Word order normally distinguishes the functions of noun phrases as subject and object. See sect. 1.

APALA1

85

Bound prefixes mark subject of intransitive and copula verbs object of transitive verbs. See sect. 18.4. Any other syntactic relationship in which a noun phrase is than as subject or object) requires the use of a postpositional noun phrase. Indirect object is marked with the relator α expresses the recipient role. (267)

and subject and involved (other relator with the "to, by" and

taky m-ekaro-ase ey-a bow 2-give- RP 3- to 'You gave the bow to him.'

The relator a also marks agent in derived constructions such as nominalizations and causatives. (268)

kanawa axika-ry aimo a canoe carve-NOMLZR boy by 'the carving of the canoe by the boy'

(269)

aimo nyh- mapono boy sleep-TRNSTVZR-CAUS-IP

j-eny 3-mother+POSSN

typaxiry a 3REFLX-sisterPOSSN by 'The mother caused the sister to put the boy to sleep.' 15.2 Genitives. In possessed noun phrases the possessor (genitive) precedes the possessed (noun head). The possessed object is the morphologically marked item. It has the possessor prefix (which is 0- when there is a free form possessor) as well as the possession suffix -ny, -ry, or other less common variants: (270a)

nohpo 0-kyry- ry woman 3-thing-POSSN 'the woman's possession'

(270b)

i-kyry- ry 3-thing-POSSN 'her/his possession'

See sect. 16 for the set of pronominal affixes. There is no syntactic difference between alienable and inalienable possession. However, there is a distinction between the two types of possession since there

86

Edward and Sally Koehn

are some items—principally kinship terms, generic terms for foods, and some body parts-which are obligatorily possessed and are always marked by the possessor prefix and possession suffix. (271)

typaxi-ry 3REFLX-sister-POSSN 'his own sister'

Other items are optionally possessed: the unmarked form is the nonpossessed one, and the addition of the possessor prefix and possession suffix marks it as possessed. (272)

rato; 'knife'

a-rato- ny 2-knife-POSSN 'your knife'

There are also obligatorily non-possessed items. These never occur with the affixes that mark possession. They include the names of plants, animals, persons, and natural phenomena. (273)

paruru 'banana'; nuno 'moon'; orutua 'man'; konopo 'rain'

There is no formal difference between temporary and permanent possession. Present and past possession are marked by noun suffixes. The suffix marking present possession is: -ny 'POSSESSION'. Its variants are -ry, -ty, -0, and stem reduction. The suffix marking past possession is: -hpyry or -Vpyry 'POSSESSED ΓΓΕΜ PAST'. (In general -hpyry occurs replacing the -ry form of the present and with nominalizing suffixes; -Vpyry occurs elsewhere. However, if the stem ends in -ry, as in some obligatorily possessed words, then the suffix -Vpyry is used, e.g. ykyryry 'my possession', ykyryrypyry 'my former possession'.) There is no marking for length or duration of possession (remote, historical or recent). (274)

j-omi-ry 'my speech'; j-omi-hpyry 'my ancient words'

(275)

y-nekaro-ry 'what I am giving'; y-nekaro-hpyry 'what I gave'

(276)

y-tapyi-ny 'my house'; y-tapyi-ny-Vpyry

'my former house'

Certain obligatorily possessed items such as generic terms for food can be specified by placing the specific word in apposition with the generic word: (277)

a-napyry; 2-fruit/veg.- POSSN 'your fruit/vegetable'

a-napy-ry paruru 2-fruit-POSSN banana 'your fruit, the banana'

APALAI

87

15.3 Modifiers. Modifiers in noun phrases occur following the head of the phrase or occasionally preceding it. There are adjective/adverbial type words such as: konöto 'large'; zumo 'big'; pitiko 'small'; kure 'good'; popyra 'bad'; nymyry 'genuine'; tuhke 'many'; mosa 'tall'; axiny 'quickly'. Some modifiers are formed by simply adding to any noun the postposition (sometimes written as a suffix) me 'DENOMINAUZER': poeto me'child-like, as a child'; anusasa me 'sweet'; itamuru me 'much'. (278)

y-kaparu-nu nymyry apoi-ko 1-club- POSSN genuine grab-IMF 'Grab my genuine club, friend.'

j-epe 1 -friend +POSSN

(279)

moroto kana pisarara roke ynananymy-ase there fish small only 1 +3S3O-liftRP 'There we only caught some small fish.'

(280)

yna tapyi pitiko ynanamory 1+3 house+POSSN small 1 +3S3O-construct-NOMLZR Our small house, the one we constructed.'

Other modifiers are verb or noun stems with derivational affixes t(y)-...-rele 'adjectivizer/adverbializer': t-onor-e 'dry' (anory 'be dry'); t-onu-re 'seeing' (onu 'eye'); ty-py-e 'with a wife' (py 'wife'). The etymology of some forms is impossible to reconstruct: tüpore 'hard'. Other modifiers are noun stems with the derivational suffix -hpe 'infested, infected, plagued, negative value attributed': maxi-hpe 'infected with flu'; ahno-hpe 'full of people, crowded'. An alternative way is for the adjective and adverb type words to be nominalized and right dislocated as an expansion or addition to the main clause. (281)

moroto there

kana fish

ynananymy-ase, l+3S3O-liftRP

pisararäkö small+NOMLZR+PL

rokene only 'There we caught some fish, just small ones.' (282)

yna tapyi 1+3 house+POSSN

ynanamory 1+3S3O-construct-NOMLZR

zumo-no big- NOMLZR Our house, the one we constructed, a big one.'

moro, that

88

Edward and Sally Koehn

There is no specific form of relative clause construction. See sect. 15.4 for discussion of various forms of nominalization that can function as relative clauses. Articles do not exist in Apalai. Demonstratives (see sect. 16) function as complete noun phrases rather than as a single constituent of a noun phrase. Numerals can occur before or after the word they modify. The preferred position for numerals is in the expansion of the sentence (284): (283)

morarame asakoro toto, nohpo tomo and then two 3PL woman COLL 'And then (there were) two of them, women.'

(284)

aruma apuka-V aporo, asakoro rokene reed pick- IP first two only 'First he cut reeds (weaving fibers), only two.'

(285)

omametone paihxina ynananymy- ase five species of fish 1 +3S3O-liftRP 'We caught five paihxina (a type of fish).'

Quantifiers are not distinguished syntactically from other adjective type words. (286)

kana 0anymy- ase fish 3S3O-liftRP 'He caught a lot of fish.'

kana fish

tuhke many

hxo INTNSF

(287)

orino to- kehko t- apuh- se clay PL- etc NF-mold-CMPL 'Various kinds of clay objects were molded.'

Tökehko indicates a group of more than one, whereas tuhke indicates a large number, (tökehko is made up of the following particles: tomo 'COLLECTIVE', ke 'AMELIORATIVE' and hko 'EVALUATION'.) The plural markers in noun phrases function as group markers rather than plural number. These groups normally share some semantic similarity such as edible plants, reptiles, jungle trees, etc. The markers belong to a small closed set of modifying particles and occur postpositionally in the periphery of the noun phrase. The particle tomo (shortened to to in phrase medial position) most often occurs with non-possessed nouns. (288)

aimo to t- oyto-se boy PL NF-go- CMPL The boy and his family went' or "The boys went.'

APALAI

89

The particle komo occurs with possessed nouns and nominalized verbs with second and third person involvement. (289)

o-muku-ru 2- son- POSSN 'your (pi) son'

komo COLL

Other modifying particles are: -Vpo 'DEVALUED' (290)

pata-Vpo 'the abandoned village' -into 'augmentative' or 'variety of

(291)

paruru- imo 'large cooking banana' poinoko-imo 'variety of jungle pig' or 'domesticated pig' okoi-imo 'anaconda, boa constrictor, big snake' -hne 'homogeneous group* (with demonstrative pronouns)

(292)

mokaro-hne kae-na those- GRP high 'Those ones are high.'

toh 3PL

m-a 3- be+PRES

See sect. 21.1 for a more complete list. 15.4 Nominalizatkms 15.4.1 Types of nominalizations. The nominalization of verb stems includes: (i) -ry 'possessed nominalization of action'. This suffix is added to transitive or intransitive verb stems and cooccurs with person marking possessor prefix or possessor noun phrase. For the forms marking past possession, which also apply to nominalizations (-hpyry, -Vpyry), see sect. 15.2. (293)

j-oepy- ry eraxima-ko 1-come-NOMLZR wait- IMP 'Wait for my coming.'

(294)

karao apoi- ry bird catch-NOMLZR 'my catching the bird...'

y-a 1-by

(ii) -pyny 'negative nominalization of action referring to the actor or potential actor'. This occurs with both intransitive and transitive verb stems.

90

Edward and Sally Koehn

The transitive stem has an obligatory possessor prefix signalling underlying object of the verb. (295)

kana an-any-pyny fish 3- lift-NEG+NOMLZR One who doesn't fish'

(296)

nyh- pyny sleep-NEG +NOMLZR One who doesn't sleep'

(iii) -topo 'thing, time, or place associated with the action.' This cooccurs with person-marking possessor prefixes, (-topo becomes -toh when followed by another word.) (297)

j-epy- topo 1-bathe-NOMLZR 'my bath place'

(298)

xixi kuhtopo sun measure- NOMLZR 'measurer of the sun (clock)'

(299)

moroto kae poroh- toh there high+NOMLZR landing-NOMLZR 'There is the landing strip.'

mana 3+be+PRES

(See sect. 15.2 for past possession markers that may cooccur with this nominalizer.) (iv) -tamity 'payment for the action'. This occurs with person-marking possessor prefixes. (300)

o-eroh- tamity 2-work-pay +NOMLZR 'your wages for work'

(301)

upo kurika-tamity clothes wash- pay+NOMLZR 'payment for washing clothes'

(v)

-ne 'agent of present action'. This occurs with transitive stems only.

(302)

parata wo-ne rubber cut-NOMLZR One who cuts rubber'

APALAI

(303a)

j-ypipoh-ne mokyro 1-beat- NOMLZR that one 'That is the one who hits me.'

(303b)

o-pipoh-ne mokyro 2-beat- NOMLZR that one 'That is the one who hits you.'

91

The plural of this form is the suffix -nanomo 'plural agent of present action'. (304)

parata wo-nanomo rubber cut-NOMLZR +PL Ones who cut rubber' -hpono 'agent of past action'. This occurs with transitive stems.

(305)

j-eky wo- hpono 1 -pet +POSSN shoot-NOMLZR +PAST 'the one who shot my pet'

The plural form is: (306)

j-eky wo- hpo 1 -pet +POSSN shoot-NOMLZR +PAST 'the ones who shot my pet'

komo PL

(vi) -kety 'actor of present action'. This occurs with intransitive stems only. The plural form is -tyamo. (307)

wa- kety dance-NOMLZR One who dances' -hpyry 'actor of past action' (see sect. 15.2). There are no possession markers, but a 'general prefix' occurs (see sect. 18.4).

(308)

aoriky-hpyry GEN .PREF-die- NOMLZR +PAST 'the dead one' or 'the one who died'

(vii) -ny-...-ry object resulting from an action. This occurs with transitive stems only. The -ny- is preceded by possessor person markers. (309)

y-nymero-ry 1 -NOMLZR-write-NOMLZR 'the thing I am writing'

92

Edward and Sally Koehn

Object resulting from a past action' is expressed by substituting the -ry with -hpyry. (310)

o-nymero-hpyry 2-NOMLZR-write-NOMLZR +PAST 'the thing you have written'

(311)

wewe netapa-hpyry wood NOMLZR-hit- NOMLZR+PAST 'the one hit (killed) by the tree'

(312)

wewe netapa-tyamo wood NOMLZR-hit-PL+NOMLZR+PAST 'the ones hit (killed) by the tree'

(viii) -senano 'a brand new item, subject of intransitive verb stem, or object of transitive verb stem, the result of a recently performed action.' The resultant form is obligatorily non-possessed. (313)

enuru-senano born- NOMLZR 'new-born one' (pi: enuru-senä komo)

(314)

tyri-senano GEN.PREF -do- NOMLZR 'freshly made object'

(ix) -semy 'product of action either future or present'. This is also a nonpossessed nominalization, lacking any possessor prefixes or noun phrases. (315)

paratu orino apuh-semy plate clay mold-NOMLZR 'material which is to be molded into a plate'

(316)

mokyro etapa- semy that one kill-NOMLZR 'that one who is to be killed'

(317)

tyriiku risemy wheat flour make-NOMLZR 'something that is to be made from flour'

APALAI

93

(x) -tozo 'companion in the action'. This form cooccurs with person marking possessive prefixes. Its occurrence is rare in the language. (318)

eroh-tozo work-NOMLZR 'his work partner'

(xi) -tono 'general nominalization'. This is a non-possessed form which occurs only with transitive stems and does not have any prefixes. (319)

any- tono kyn-oeh- ne lift- NOMLZR 3- come-DP "The lifter (inviter) came.'

(xii) -no, -to, and -my adjective/ad verb and postposition nominalization. If the stem ends with a vowel preceded by a consonant, the vowel changes to a before adding -no or -to: (320)

kure 'good' —> kurano 'the good one'

(321)

tonure 'seeing' —> tonuremy One who sees'

(322)

kae 'high' —» kaeno 'the high one' e.g. airplane

(323)

maxihpe 'with cold/flu' —> maxihpano One with cold/flu'

(324)

wetukene 'sleepy' —» wetukenato One who is sleepy'

(325)

w^i se mana 'he likes manioc' —» wyi sato One who likes manioc'

(326)

kanawao 'in a canoe' —> kanawaono One in the canoe'

15.4.2 Grammatical relations in nominalized constructions. Where the nominalizations resulting from the above processes are possessed forms, the intransitive subject and transitive object become the possessor noun phrase or prefix of the nominalized verb, reflecting an ergatively organized system of derivational morphology. (327a)

j-oepy- ry 1-come-NOMLZR 'my coming'

(327b)

u-muku-ru 1-son- POSSN

(328a)

j-etapa-ry 1-kill- NOMLZR 'my being killed'

oepy- ry come-NOMLZR 'my son's coming'

94

(328b)

Edward and Sally Koehn

u-muku-ru 1-son- POSSN

etapa-ry kill- NOMLZR 'my son's being killed'

The transitive subject becomes a possessor noun phrase or prefix attached to the postposition a 'to, by'. (329a)

karau apoi- topoVpyry bird catch-NOMLZR-PAST 'my having caught the bird'

(329b)

karau apoi- topoVpyry bird catch-NOMLZR-PAST 'the boy's having caught the bird'

y-a 1-by aimo a boy by

15.4.3 Syntactic functions of nominalized constructions. The nominalized expressions function as nuclear constituents of a sentence and also occur in equative sentences and as head of postpositional phrases. Equative sentence: (330)

a-yto-ry ropa moro isawä 3-go-NOMLZR again that sand 'He is going to the beach.'

pona to

Subject in copular sentence: (331)

moroto kae poroh- toh there high+NOMLZR landing-NOMLZR 'There is the high thing's place to land.'

Object: (332)

j-oepy- ry eraxima-ko 1-come-NOMLZR wait- IMP 'Wait for my coming.'

Subject of intransitive sentence: (333)

kae n-oeh- no high+NOMLZR 3-come-IP 'The plane arrived.'

mana 3+be+PRES

APALAI

95

In postpositional phrases: (334)

epy- toh pona to- yto-se bathe-NOMLZR to NF-go-CMPL 'He went to the bath place.'

(335)

töxiry 3REFLX-daughter-POSSN

ekaro-ry give- NOMLZR

se towant 3REFLX-

exi-ryke be- NOMLZR-REAS 'because he wanted to give his own daughter (in marriage)' 16 Pronoun system 16.1 Free pronoun forms. Within the free personal pronoun set there is a division between animate and inanimate pronouns. Third person pronouns, both animate and inanimate, distinguish between deictic and nondeictic, with three deictic forms. Both animate and inanimate pronouns distinguish between individual and collective. Animate Nondeictic

non-coil

coll

1

ywy

1 +2 1 +3 2 3

kymoro yna omoro ynoro

kymarokomo yna amarokomo ynaro

mose mokyro moky

moxiamo mokaro mokamo

Deictic 3 near 3 medial 3 remote

Inanimate non-coil coll

sero yro

serohne

seny moro many

senohne morohne monohne

The third person deictic pronouns are the equivalent of demonstrative pronouns. There are general question words: onoky 'who', which substitutes for animate nouns, and oty 'what', which substitutes for inanimate nouns. (336)

onoky mokyro; who that one 'Who is that?'

(337)

oty moro; what that 'What is that?'

umukuru mose my son this one 'This is my son.' arimi z-ery monkey 3-tooth +POSSN "This is a monkey tooth.'

seny this

96

Edward and Sally Koehn

There are no indefinite pronouns but the category can be expressed by the use of the first person inclusive forms (cf. Derbyshire 1979:130). 16.2 Bound pronoun forms. The form of personal pronoun used for the indirect object is bound. The full set is: 1 2 3 3REFLX 1+2 1+2PL 1+3 2PL 3PL

y-a oy-a ey-a tyy-a kyy-a kyy-a xine yna-a oy-a xine ey-a xine

me-to you-to him-to himself-to us-to us ail-to us-to you ail-to them-to

There are no gender or class distinctions. The above forms also serve to represent goal of motion. (338a)

oseh- ko y-a come-IMP 1 -to 'Come to me.'

(338b)

tumy 3REFLX-father +POSSN

a ropa to again

to- yto-se NF-go-CMPL

ropa ynororo again 3 'He went back again to his father.' There are bound person pronouns that occur as prefixes on verb stems. These are listed in sect. 18.4. Possession in Apalai is marked by a possession suffix cooccurring with a person-marking prefix on the possessed noun. Free pronouns are not normally used to mark possession, except for yna Ί+3'; occasionally other person forms are used to express emphasis, e.g. ywy tapyiny 'my house'. First person possessor pronoun: jj- becomes u- in the presence of -Cu initial stems; j- becomes y- in the presence of CV initial stems where V is not u.

APALAI

(339)

jepity umükuru ykyryry

'my medicine' 'my son' 'my thing'

97

(j-epi-ty) (u-muku-ru) (y-kyry-ry)

Second person possessive pronoun: o0- becomes a- in the presence of -Qz or -a initial stems; in the latter case a plus a becomes a. (340)

oepity akanary äpitunu

'your medicine' 'your canoe' 'your basket'

(o-epi-ty) (a-kana-ry) (a-apitu-nu)

Third person possessor pronoun: i1- becomes j- replacing stem-initial z', z- before stem-initial w; and 0- before a vowel other than u. (341)

imükuru jery epity zupony

'his 'his 'his 'his

son' tooth' medicine' clothes'

(i-muku-ru) (i-ze-ry) (O-epi-ty) (z-upo-ny)

Third person reflexive possessor pronoun: tt- becomes tu- before stem-initial w or CM; ty before stem-initial consonants other than a CM sequence; and to- before stem-initial e. (342)

tapony tykyryry tumükuru toepity

'his 'his 'his 'his

own own own own

bench' thing' son' medicine'

(t-apo-ny) (ty-kyry-ry) (tu-muku-ru) (to-epi-ty)

First person inclusive possessor pronoun: kyky- becomes ku- before stem-initial vowel or CM sequence. (343)

kuepe kykyryry kumükuru

Our friend' Our thing' Our son'

(ku-epe) (ky-kyry-ry) (ku-muku-ru)

98

Edward and Sally Koehn

There are at least 20 idiosyncratic forms whose prefixes for first and third persons are not predictable; e.g.:

(344)

(345)

(346)

(347)

joko

'my body'

(j-oko)

zoko

'his body'

(z-oko)

jokuru

'my drink'

(j-oku-ru)

aokuru

'his drink'

(a-oku-ru)

woryxiry

'my sister'

(w-oryxi-ry)

aoryxiry

'his sister'

(a-oryxi-ry)

waxtry

'my leg'

(w-axi-ry)

jaxiry

'his leg'

(j-axi-ry)

Four kinship terms are idiosyncratic:

(348)

papa omy jumy /zumy tumy kumy

'my father, father' 'your father' 'his father' 'his own father' Our (incl.) father'

(349)

aja asa jeny /eny tyse kyse

'mother, my mother' 'your mother* 'his mother' 'his own mother' Our (incl.) mother'

(350)

pipi oohpyry aohpyry tohpyry kuaohpyry

'aunt, my aunt' 'your aunt' 'his aunt' 'his own aunt' Our (incl.) aunt'

(351)

eo eory aory tory kuory

'uncle, my uncle' 'your uncle' 'his uncle' 'his own uncle' Our (incl.) uncle'

There are no reflexive pronouns or reciprocal pronouns other than the reflexive possessor prefix t- (see above and sect. 4). There are no relative pronouns.

APALAI

99

17 Adpositkmal phrase structure Only postpositions occur. These are frequently referred to as relators in this description to distinguish them from other kinds of postpositions, e.g. particles. The postpositional phrase is defined in terms of its constituents: an obligatory nucleus consisting of a noun phrase followed by a relator postposition, or a possessed form of the relator without any preceding noun phrase; and optional particles that follow the nucleus. It is further defined by its syntactic function as an adverbial that occurs as either the complement of a copular sentence or the adjunct of any type of sentence. (352)

oturu-Vko ton n-exi-ase aimo poko/ talk- CONT 3PL 3-be- RP boy occ.with "They were talking about the boy/ about me.'

(353)

pata pona ropa to- yto-se village to again NF-go-CMPL 'They went back to the village/ to it.'

toto/ 3PL

y-poko 1-occ.with

i-pona 3-to

Both relators and particles are postpositions but are distinguished on both syntactic and morphological grounds. Syntactically, relators only occur following nouns or as free forms (with possessor prefixes), whereas particles can follow any category of word. Morphologically, only relators can be inflected (for possession) and be the subject of derivational processes (negation and nominalization). Postpositional relators cannot occur without arguments, nor with more than one argument. Their argument can only be a noun phrase or a person-marking prefix attached to the relator itself; the prefix is one of the set of prefixes that occur with nouns to mark the person of the possessor. In this way the relator can occur without a noun phrase head, but it cannot be stranded in the sense of being separated from its noun phrase head in the sentence. The only elements that can modify postpositional relators are particles, and usually these seem to modify the postpositional phrase as a whole rather than just the relator. (354)

u-tupi pona ropa 1-field+POSSN to again 'I'm going back to my field.'

yto-Vko go-CONT

ase 1+be+PRES

The relator is pona 'to'; ropa 'again' is a discourse particle that relates to the whole nucleus of the phrase.

100

Edward and Sally Koehn

The relator used in many cases depends on the nature/shape of its object:

liquid fire small container large place flat place pole shape river hammock

in/on kuao htao ao tao po poko nao tapo

into/onto kuaka htaka oka taka pona pokona naka tapona

via/from kuae htae ae tae poe pokoino nae tapoe

Other common relators are: püto 'close'; samo 'equal/similar'; se 'want'; pyno 'care for'; waro 'know'; poko Occupied with'; a 'to/by/about'; wino 'from (a person)'; akoxi 'toward'; kuroko 'through' or 'around'; etonie 'across from'; myhto 'at the foot of; ino 'from'; winoino 'from (a person).' 18 Verb and verb phrase structure 18.1 Tense. Tense is one of four categories marked by verb suffixes; the others are aspect, number, and mood. Mood is only marked for imperative and does not involve tense and aspect in the forms it takes. The unmarked mood is indicative and the suffixes in this set are composite forms involving tense, aspect, and number. The basic forms for the indicative are as follows: Finite Immediate Past Recent Past Completive Recent Past Continuative Distant Past Completive Distant Past Continuative Historic

Noncollective -no -ase -asene -ne -sene -ko

Nonfinite Nonpast Continuative

-Vko

Past Completive Past Continuative

t-...-se -se

Collective -tou, Vtoto -atose, -ase toto -atosene -tone, -ne toto -sene toto -ko toto

-toryko toh (mono) -Vko toh (nae) t-...-se toto -se toto

Note: The -Vko form includes the tilde (~) as part of the suffix, but the V which is nasalized is the final vowel of the stem. (See sect. 15.4 for

101

APALAI

nominalizations, and sect. 14 for subordination forms -se 'purposive' and -htao 'temporal'.) 'Completive' and 'Continuative' are used as general terms each of which includes more than one specific aspect. 'Collective' is related to person. The nonpast forms occur as complements of the copular verb. The construction is: verb stem plus continuative aspect suffix followed by copular verb. It is the copular verb that has the person-marking prefixes and nonpast tense suffix. Forms for ka la 'say, do' are:

1

+

2

1

3

2

Finite Immed. Past noncoll coll

sykano sykatou

ykano ynanykano

nykano nyka toto

mykano mykatou

Recent Past Cmpl noncoll coll

sase satose

ase ynanase

nase nase toto

mase matose

Dist. Past Cmpl noncoll coll

sykane sykatone

ykane ynanykane

kynako kynako toto

mykane mykatone

Nonfinite Nonpast Cont Past Cont Cmpl

äko kase tykase

'doing, saying' 'was doing, was saying' 'done, said'

Copula Forms Finite Nonpast noncoll coll

1+2

1

sytase sytatose

ase ynanase

mana ma toto

mase matose

Nonpast Uncert noncoll coll

sytah sytatohu

ha(no) ynanah

hnae toh nae

hma(no) matohu

102

Edward and Sally Koehn

1+2

Immed. Past noncoll coll

sakene satokene

akene ynanakene

kynako toh kynako

makene matokene

Recent Past Cont noncoll

sexiase

exiase

mexiase

sexiatose

ynanexiase

nexiase mon-exi-ano toh nexiase toh mon- -ano nexiase toto

Recent Past Cmpl noncoll coll

sexino sexitou

exino ynanexino

nexmo toh nexino nexi toto

mexmo mexitou

Dist. Past Cont noncoll coll

sexiasene sexiatosene

exiasene ynanexiasene

nexiasene nexiasene toto

mexiasene mexiatosene

Cmpl noncoll coll

sexine sexitone

exine ynanexine

kynexine toh kynexine kynexine toto

mexme mexitone

coll

Historic noncoll coll Nonflnite Nonpast Cont Past Cont Cmpl

exiko ehtoko

exiko ehse toehse

'being, becomin 'was being, livi 'became'

mexiatose

APALAI

103

The copula or auxiliary verb has basic stem forms exi and a. Whereas all transitive and intransitive verbs are dependent on the auxiliary verb in the nonpast and nonpast uncertain tenses, the auxiliary or copula is independent in these tenses. When dependent forms of the other verbs occur, it is the auxiliary that carries the subject person-marking prefixes and the tense suffixes. The form mon-...-αηο 'RECENT PAST CONT' is a discontinuous affix used to indicate action performed out of sight of the speaker. The nonpast forms are used with reference to present, future, and universal, without any formal distinctions between the three types of tense /aspect. (355)

urakana-Vko hunt- CONT

mana 3+be+PRES

'He is hunting now' or 'He will be hunting soon' or 'He hunts.'

The nonpast uncertainty forms of the copula have a modal value. They are used to express interrogatives, and also noninterrogatives, in which case they usually cooccur with certain particles which mark degree of certainty or authority for assertions. In all these uses they can refer to present, future, or universal. (356)

wa- Vko dance-CONT

(357)

eremia-Vko sing- CONT

mah 2+be+PRES oko perhaps

'Do you dance?' or 'Are you dancing?' or

hnae 3+be+PRES

'Maybe he sings' or 'Maybe he is singing' or 'Maybe he will sing.'

Past is divided into three degrees of remoteness: (i) Immediate past refers to actions done the same day or the previous night. (358)

nu-tuta-no 3- exit-IP

'He came out.'

(ii) Recent past refers to actions done on the previous day or any time earlier up to a period of a few months (this is relative to the total situation). Events of only a few weeks past may be expressed with the distant past suffix. (359)

n-eroku-ase 3-work- RP

'He worked.'

104

Edward and Sally Koehn

For repeated action ('continuative') in the recent past, the suffix -ne is added. (360)

eroku- ase-ne 1+work-RP-CONT

Ί worked many times.'

(iii) Distant past refers to actions done any time previous to the recent past. (361)

akohne lS3O+chop-DP

Ί felled it (tree).'

Distant past continuative: (362)

pake long ago

a- htao be-when

arimi wo-se monkey kill-CONTPAST

pyrou arrow

ke toto INST 3PL 'Long ago they used to kill monkeys with arrows.' The historic tense is expressed by the form -ko and is limited to occurrence with third person subject and object referents. It is found only in legends, where it expresses main event line action. (See E. Koehn 1976 for the function of -ko in Apalai narrative.) (363)

apoi-ko grab-HIS

'He grabbed it.'

Future tense is expressed by nonpast forms. Tense-aspect distinctions apply not only to finite verb forms but also to certain nonfinite forms as well. (See sects. 14.1,15.2, and 15.4.). The imperative mood forms have number, but not tense /aspect distinctions. 18.2 Aspect. The aspectual markers in the language are all related to ways of viewing duration of a situation. There are two different kinds of markers in the verb: (i) the completive/continuative dichotomy expressed in the tenseaspect-number suffix system; and (ii) ingressive and terminative aspects which are expressed by derivational suffixes added to verb stems.

APALAI

105

Habitual, continuous, and progressive aspects are expressed by the continual!ve component of the verb suffixes: (364)

poeto-me child-DENOM

j-a- htao 1-be-when

itu jungle

zuno exi- ase-ne / fear 1 +be-RP-HAB

ehse ywy be-CONTPAST I 'When I was a child, I was often afraid/lived in fear of the jungle' (365)

mokyro oepy-ryhtao that one come-NOMLZR-when 'When he came, I was eating.'

otuh-n ko eat-CONT

(366)

wyi ekei-Vko toh n-exi-ase manioc bake-CONT 3PL 3-be- RP 'They were making a lot of manioc bread.'

akene 1+be+IP

pitiko little

roke just

pyra NEC

Ingressive aspect is expressed by adding the suffix -pito 'begin' to any verb stem. The focus of the meaning can be either on the beginning of the action or on the primacy of the participant in relation to other participants in the ensuing action. (367)

kana any-pito- ne poeto-me ro fish lift-begin-DP child- DENOM still Ί began to catch fish when I was still a child.'

j- a- htao 1-be-when

Terminative aspect is expressed by adding -keh 'CESSATIVE' to any verb stem: (368)

otuh- keh- no 1 +eat-CESS-IP

Ί finished eating.'

The terminative aspect is also expressed TERMINATrVE'. (369)

ynan- ua- ase-Vtyky 1 +3- dance-RP- TERM

by the clitic particle

Vtyky

'We finished dancing.'

The completive component of the verb suffixes expresses semelfactive and punctual aspects. (370)

n-yto-ase anatu pona 3-go- RP Anatum to 'He went to Anatum.' (one action)

106

Edward and Sally Koehn

(371)

n-yto-ase ropa anatu pona 3-go- RP again Anatum to 'He went back to Anatum again.'

ropa again

18.3 Mood/Modality. The only mood distinction in the verb inflection system is that between indicative and imperative. The indicative mood is unmarked and contrasts with the imperative. The imperative mood is expressed primarily by a set of verb suffixes that show components of number and motion and which differ in form according to the person of the subject. Optative and hortative values are included in this mood. It is reinforced by: (i) The cooccurrence of the particle ah or apo 'IMPERATIVE' which is obligatory for the third person imperative, there being no distinctive imperative suffix for that person. The suffixes used are the tense/aspect forms -no and -Vtoto 'IMMEDIATE PAST'. (372)

ah IMP

n-yto-V 3-go- IP

toto 3PL

'Let them go.'

(ii) The optional occurrences of the imperative form of the verb 'to go' eropa. The shortened form pa is used when the command involves motion away from the location of the speaker and first person is included in the action. (373)

pa sepytase

'Let's go, let's go swim.'

The paradigm of nonmotion imperatives is as follows:

2

Noncoltective -ko

Collective -toko

1 -xi 1 +3

ynan- -xi

1+2 s- -ne

s- -tone

3

ahjapo n- -V toto

ah /apo n- -no

any-ko any-toko eneh-xi ynan-eneh-xi synyhne synyhtone ah nytono ah nytö toto

'Lift it.' 'You all lift it.' 'Let me bring it.' 'Let us (excl) bring it.' 'Let's sleep.' 'Let's all sleep.' 'Let him go.' 'Let them go.'

The set of motion imperative suffixes is used when the action to be performed involves distance from the location of the speech act.

APALAI

Noncoltective -ta

2

1 -taxi, -tamy I +2 s- -lose

Collective -tatoko

s- -tatose

eneta äporo enetatoko enetaxi senetase senetatose

107

'Go see it first.' 'You (pi) go see it.' 'Let me go and see it.' 'Let's go see it.' 'Let's all go see it.'

There is no special motion form for third person imperatives. See also sect. II for imperative constructions, including some discussion of ways of expressing different degrees of imperative. The second person imperative forms of the verb ka 'say, do' are irregular: nonmotion motion

noncollective collective noncollective collective

kaxiko kaxitoko kaxita kaxitatoko

'Say/do it.' 'You all say/do it.' 'Go say it, go do it.' 'You all go say/do it.'

There is no special form for expressing the idea of obligatoriness (e.g. 'must do anything'), other than the imperative forms. Potential mood, the equivalent of 'can do', is expressed by zuaro, 'to know, know how to'. (374)

kanawa axika-ry waro mana canoe carve-NOMLZR know 3+be+PRES 'He knows how to carve out a canoe.'

Degree of certainty on the part of the speaker is marked by two suffixes and one particle: -my -na rahkene

deduction uncertainty certainty, prediction, warning, consummation, legendary fact

18.4 Person and number. The persons of the subject and the direct object are coded in the verb prefix. There are three sets of forms: transitive, intransitive, and copula. In the transitive set each prefix is a composite form that marks subject and object. Collective number is coded in the verb suffix. The collective forms relate to the person of the subject and/or object. The 1 +3 category is expressed by yna, ynan; the form remains unchanged whether the referent includes only a single third person or a group of third persons. The collective suffix refers to the person of the subject with intransitive stems and the copula, and to the person of either the subject or the object or both with transitive stems, so that in the latter case there can be ambiguity in

108

Edward and Sally Koehn

the first person inclusive, second person and third person forms. The collective marker on the verb is usually in agreement with the collective marker on the noun or pronoun subject or object; occasionally, however, the noun or pronoun may be collective when the verb is noncolleciive. The paradigm of the person-marking prefixes in the verb is shown in the following matrix: OBJ 1 + 2 2

SUB 1+2 2 1 1+3 3

POSS

1

1 + 3 3

s(y)-

«000-lyk(y)-

Jt(y)-

o-

o-

j(y)-

j-fy-

yna-

yna-

INTRANS COPULA/ EXPR S S s(y-) 5o-\mm-

0-ty-

0-

ynan(y)-

ynan(y)-

"00-

"00-

kyn(y)-

JkynOO-

ynannkyn(y)-

t- REFLX ι- NONREFL·*

The subject-object markers can be compared with the nominal possessor markers as shown in the lower row of the chart above. (See sect. 16.2 for further details.) The form kyn- '3' occurs only with distant past tense forms (-ne). The vowel v-[t] occurs with C-initial stems. The v- changes to u- before CM, following a rule of vowel harmony. (375)

y-kurika-no —»· u-kurika-no y-pipoh- no Ί hit it.'

Ί washed it.'

The 3 subject+3 object prefix is n(y)- in the absence of a free object, and 0- when a free object is present. (376a)

aimo n-any-no boy 3-lift- IP "The boy caught it.'

(376b)

aimo kana any-no boy fish lift- IP 'The boy caught the fish.'

APALAI

(i)

109

Intransitive 2S has two morphologically conditioned forms: o- occurs with a subclass of stems including 'sleep' -nyh: o-nyh-no 'you slept' 'cry' -xita: o-xita-no 'you cried' 'climb' -onuh: o-onuh-no 'you climbed up'

(ii) m(y)- occurs with the subclass including 'go' -yto: m-yto-no 'you went' 'eat' -otuh: m-otuh-no 'you ate' 'come' -oeh: m-oeh-no 'you came' The m(y)- form always occurs with reflexive forms of the verb. There are several verb stems which occur with a special form of prefix in first and second person imperative forms and nonpossessed derived nominals. This prefix is called 'GENERAL PREFIX'. With transitive imperative forms the prefix refers to third person object. The form of the prefix is: i- with consonant initial stems (147), and t- or a- with vowel initial stems (27), (162), (164) and (308); cf. Derbyshire 1979:149-151. 18.5 Voice. 4).

See passive (sect. 5), causative (sect. 6), and reflexive (sect.

18.6 Other categories. There are suffixes which indicate nonfinite verbal forms, that is, the derived nominals and pseudo nominals. They include -ry 'NOMINALIZER' and -topol-toh 'thing/place NOMINALEER'. (See sect. 15.4 for the full range of nominalization possibilities.) (377)

j-yto-ry moro 1-go-NOMLZR that 'That was my going.' ( was going.')

(378)

watoh moro dance-NOMLZR that 'That was the placeAime to dance.'

18.7 Incorporation of nominal elements into the verb. One type of incorporation involves noun stems cooccurring with derivational suffixes to form verb stems; e.g. opi 'medicine'; epinohnöko 'giving medicine to' (a transitive verb, the object of which is the recipient of the action).

110

Edward and Sally Koehn

-noh 'administer1 (379)

eo epinohnöko mokyro uncle give medicine to-CONT that one 'He is giving medicine to my uncle.'

mana 3+be+PRES

-pa 'benefactive' (380)

napi 'vegetable (sweet potato)'; inahpäko 'feeding (someone)' -to 'enhance'

(381)

amity

'string'; taky amih-to-Vko

'stringing a bow'.

Certain body parts follow this pattern of verbalization, as with derivational suffixes such as: -po 'cause to sever or leave' (382)

oma 'hand'; ema-Vko 'throwing'; ema-po-Vko 'dropping' or 'losing' -ka, -hka 'extract with force'

(383)

ipihpyry esena-ry

'its skin'; ipi-ka-Vko 'skinning' 'his throat'; esena-hka-Vko 'strangling'

-ta 'produce' (384)

munu

'blood'; mu-ta-Vko

'bleeding'

Incorporation occurs with other possessed nouns: -a 'bear or reproduce' (385)

i-muku-ru 'his/her child (male)'; imükua-Vko 'bearing a child' -hma 'benefactive'

(386)

epe 'his friend' (i.e. originally 'trading partner'); epe-hma-Vko 'paying it' akoro-ny

'younger sibling of same sex' or 'another set of similar things'; akore-hma-Vko 'helping' -kah 'weave'

APALAI

(387)

111

epe 'his friend'; epe-kah-nöko 'buying it'

19 Adjective phrase structure There are no adjectival phrases as such. See sect. 15.3 for possibilities relating to modifiers in noun phrases. 20 Adverb phrase structure An adverbial phrase may consist of an adverb followed by one or more particles. (388)

axi kuhse n-yto-V fast very 3-go- IP They went very fast.'

toto 3PL

(389)

zumo-hxo mokyro pitiko rokene bigger that one little only 'That fellow (is) only a little bigger.' 21 PARTICLES

Particles follow words of any class other than the ideophone, and never occur as free forms or in isolation. They are noninflected, nonderived words and can be divided into three subclasses on the basis of function and position of occurrence in relation to other particles in a sequence: i) modifying particles, which usually occur first in any such sequence and normally modify a noun or adverb; ii) discourse particles, which usually occur between the modifying and verification particles in any sequence and whose normal function is to relate the phrase or sentence in which they occur to some other part of the discourse; iii) verification particles, which generally occur last in any particle sequence and whose function is to express the attitude or relationship of the speaker to what he is saying. 21.1 Modifying particles. Modifying particles are sentence particles in the sense that their function is limited to the sentence in which they occur. For additional treatment of modifying particles see sect. 15.3, where noun phrase particles are described. Certain particles could be included in more than one class: e.g. nymyry 'genuine' freely occurs modifying nouns or adjectives/adverbs: kure nymyry 'genuinely good', umükuru nymyry 'my genuine son', wewe nymyry 'real wood'.

112

Edward and Sally Koehn

Some particles seem to fit in the sentence particle set: rokene Only' and pyra 'negative', but may under certain circumstances also be significant as a discourse particle: masaralupa rokene 'false rubber only'. The focus is on the lack of the object of search — in the particular discourse context, there was only false rubber, not the real thing for which they were searching. The sentence (modifying) particles are listed and illustrated below. hko pyra 'not quite as' (390)

zuaro hko pyra ase 3+know quite MEG 1+be+PRES Ί don't know it quite (as well as someone else does).' hxo 'degree superlative, intensifier'

(391)

mame tyrato- ke then ADJVZR-knife-POSSN. ADJ

pyra hxo MEG degree

hano, 1+be+IP

ako mokyro akono say+CONT that one brother in law ' "I don't have any knife at all," said brother-in-law.' kuhse 'to a superlative degree' (392)

i-muku-ru arypyra kuhse 3-son- POSSN NEC SUPERL 'His son (did not do) anything at all.' nymyry 'genuine'

(393)

opotu fast water

nymyry genuine

a- htao be-if

arypyra nothing

kahpyry ugly

exi-ry toh asakororo be- NOMLZR 3PL both 'If the water had been really fast, there would be nothing left of either of them.' pohto 'somewhat, medium quantifier' (394)

tuhke many

pohto somewhat

roke only

toh 3PL

n-exi-ase, 3-be- RP

watohtao dance-NOMLZR-at There were not too many at the dance place.'

pukuro 'medium quantifier'

APALAI

(395)

113

anusasame pukuro rokene sweet medium only '(It was) just medium sweet.' kokonie night 'evening'

pukuro medium degree

rokene Only' (396)

asakoro roke airiki two only bacaba (fruit) 'There were only two bacaba fruit.' rukukuh 'superlatively, uniquely'

(397)

kure rukukuh nae good SUPERL 3+be+PRES 'It is really good.' tiko 'diminutive'

(398)

t- akoh-se kae NF-cut- CMPL high 'It was cut quite high.'

tiko DIMESf

imo 'augmentative' (399)

okoi imo t- otapa-se ey-a snake big NF-kill- CMPL 3- by big snake was killed by him.' komo 'collective'

(400)

oexiry komo 'your (pi) daughter' Vpo 'devalued'

(401)

upö po 'used domes'

hne 'homogeneous group' (402)

moro hne kure those GRP good 'Those were good.'

kynako 3+be+IP

21.2 Discourse particles. Discourse particles are optional phrase constituents that follow a head word (or head complex like a possession noun phrase) from

114

Edward and Sally Koehn

any of the major word classes. The primary scope of the discourse particle is the head word it governs and it usually relates that word to some other part of the discourse. The discourse particles are listed and illustrated below. They are taken from a concordance of 28,000 morphemes from Apalai texts. (The concordance was prepared at the University of Oklahoma Computer Support for Linguistics Field Research which was partially supported by National Science Foundation Grant GS-1605.) aporo 'first in time', 'temporary duration of activity or state'; the event it modifies precedes some other event. (403)

kana fish

any my- ase 1 +catch-RP

aporo, araku first fish species

kotyxi fish species

pahxina enara rokene fish species those only 'First I caught fish, araku, kotyxi, and pahxina, just those (species).' (404)

oepuase 1 +come-RP

taro-na here-to

aporo, first

pakeimo early

opi medicine

edrink-

se aporo PURP first 'At first I came here early to drink some medicine first.' (405)

motu- ke pyra aporo motor-INST NBG temporarily Temporarily without an outboard motor.'

(406)

mame se se se i-paixi- ry aporo then (approaching) 3-sister-POSSN first 'Then someone approached, (it was) her sister first.' hkotano 'rather (it is...)', i.e. correction of something said or implied.

(407)

ywy kara exi- ase repe I be+NEG 1+be-RP but 'It wasn't I, rather it was they.'

moxiä those

hkotano rather

mä 'title or discourse topic marker' The whole clause is marked as a title or thematic sentence which is to be elaborated in succeeding sentences. A variant form is mya 'title marker mood)'.

APALAI

115

(408)

kokonie osema akoroka-ase yesterday path 1 +clear-RP 'Yesterday I cleared a path.'

(409)

isapokara muku-ru nomono kaire lizard baby-POSSN 3+enter- IP mosq. net baby jakuraru lizard entered the mosquito net.'

(410)

müpo-jana rat- people

toh 3PL

ase 1+be+PRES

repe FRUST

kyn-exi-ne 3- be-DP

mä title

mä. title

yrome mose CNJ that one

taka mä into title

On-ene-pyra 3- see-NBG ekaro-Vko tell- CONT

na 3+be+PRES Once there were some rat people. I didn't see them, but that one is telling it.'

(411)

sakura drink

eno 1+drink-IP

mya. title

ese drink-PURP

ropa again

ke (?)

yto-Vko ropa ase go-CONT again 1+be+PRES drank sakura. I'm going back again to drink again.' rahkene 'consummation, finally completed action, fulfilled expectation' (412)

mame t- oehse ropa yna rahkene then NF-arrive- CMPL again 1 +3 finally 'Then we came back again (at end of set of actions).' reh I refine 'expectation of a response' (ne is dropped sentence medially)

(413)

moro reh öty that expect 2+meat+POSSN 'Here is some meat for you.'

(414)

n-osereh- no hxo rehne 3-REFLX-scare-IP INTNSF expect 'He must have really frightened himself.' rehta 'attention holder'.

(415)

onoky rehta who ATTN 'Who are you?'

mano 2+be+PRES

komo PL

116

(416)

Edward and Sally Koehn

oty ropa rehta ano what again ATTN 1+be+PRES 'What else shall I tell?' repe 'but, contrary to expectation, frustrative'

(417)

kaikuxi-jana jaguar- people

apuru-ru lock up-NOMLZR

se want

kyn-exi-ne repe 3- be-DP FRUST 'He wanted to capture the jaguar people but (was not able to).' ro 'continuity or sameness (of an object or action.)' (418)

tuhke ro kyn-ako many still 3- be+IP "There were still many.'

(419)

xia ro ynan-oepy-ase here still l -l-3-come-RP 'We came back to here again.'

ropa again

rohxo 'even more, INTNSF' (420)

ynan-opiase 1+3- look for-RP

rohxo INTNSF

repe, FRUST

arypyra be+NFJG

masaratupa tree species

rokene only 'We looked even more for (rubber) but there was not any, only the masaratupa tree.' ropa a) 'repetition', 'a second or subsequent item or action in opposition to a previous one' b) 'return to a former state or location', 'reference to previous object, state, or action' (421)

oty ropa exi-Vko nae? what REPET be- CONT there is 'What else is going to happen?' roropa 'a second or subsequent item or action, likewise'

APALAI

(422)

117

mamao roropa kyn-exi-ne repe fruit also 3- be-DP FRUST 'There was also papaya there, but ...' roro 'alone, only just'

(423)

ywy roro koih a- Vko I alone paddle do-CONT (was) the only one paddling.'

21.3 Verification particles. The function of these is to express the speaker's evaluation of, or his relation to, his utterance: i.e. his claim to knowledge or uncertainty, impatience, irritation or honorifics, emphasis or pity. The verification particles are listed and illustrated below. For some of these forms further testing is needed to ascertain their precise range of meaning. hko 'speaker's claim to knowledge (EVALUATION)' (424)

zuaro kuh ase hko, kase repe know CERT l+be+PRES quite sure said FRUST ' "I am quite sure I know it", he said (but it didn't happen that way.)'

(425)

omoro hko you EVALUATION 'You are the one to decide (and I know it).' hkotoko 'possibly'

(426)

serae hkotoko axikäko tarame today perhaps carve+CONT maybe 'Perhaps he is carving (canoe) today.'

mana 3+be+PRES

hro 'response marker to question or situation', Obviously' (427)

opi toto orih-mato medicine 3PL kill-NOMLZR The poison was deadly to them.'

(428)

m-epy- no 2-bathe-IP 'Did you bathe?' y, AFFIRM

epy bathe

hro RESP

hro RESP

'Yes, I sure did.'

ka 'impatience (with imperative)'

118

(429)

Edward and Sally Koehn

eneh- ko ka bring it-IMP IMPATIENCE 'Bring it and hurry up about it.' ke I keh 'ameliorative or polite form' (with imperative, question, etc.)

(430)

kana ke any- ko fish POLITE catch-IMP 'Catch some fish, please.'

(431)

oty keh popyra nae what POLITE bad there is Ί wonder what is wrong.' kene 'at this point in time relative to speaker', 'immediacy'

(432)

oehno kene I +come-IP just Ί have just arrived.'

(433)

eropa kene let's go now 'Let's go now.' kety 'irritation', occurs with imperatives or questions

(434)

oty kety what IRRITATION 'What is it?'

(435)

eropa kety let's go IRRITATION 'Come on!' kuh 'certainty'

(436)

ajana kuh kyn-exi-ne Wayana CERT 3- be-DP 'He was definitely Wayana.' myhene 'pity for person in trouble'

(437)

mame yto- ko then 3+go-HIS

japyry tree

poko- na tiko side of-to DIMIN

hxo INTNSF Then he went up a topless tree, poor guy.'

myhene pity

APALAI

119

pahne 'ameliorative (with imperative)' (438)

ene- xi pahne 1 +see-IMP AMELIORATIVE 'Let me see please.'

(439)

n-ekuh-no pahne 3-try- IMP AMELIORATIVE 'Let him try it.' päna 'uncertainty, doubt'

(440)

küre päna ma good UNCERT 2+be+PRES trust you are fine.'

repe FRUST

papa 'moderately emphatic, with slight irritation' (occurs with imperative forms) (441)

tyri-ko papa GEN.PREF-put-IMP EMPH 'Put it there, will you!' puh I puhko 'eye witness evaluation' (in certain contexts)

(442)

moro puh t- onah- se that EYEWITNESS NF-finish-CMPL could tell it was all gone.'

(443)

otäto puhko ke how many EYEWITNESS POLITE 'How many did you see working?'

rohke only

eroh-nöko work-CONT

nae 3+be+PRES

sätä 'promise or oath, truly' (spoken lightly or with humor) (444)

typyre-ke ADJVZR-arrow-POSSN.ADJ certainly do own an arrow.'

sätä truly

tano 'deduction' or 'assumption' (445)

m-otuh-no tano 2-eat- IP DEDUCT 'You must have eaten.'

(446)

i-se pyra tä m-ase 3-want NEG DEDUCT 2- be+PRES 'You apparently do not want any.'

äse 1+be+PRES

rahkene really

120

Edward and Sally Koehn

toko 'evidently' (447)

t- oyto-se toko NF-go-CMPL EVIDENTLY '(He) evidently left.' ty 'emphatic and without regard for the opinion of the recipient of the order'

(448)

osehko ty 2+come-IMP EMPH 'Come here! (whether you want to or not)' 22 PHONOLOGY

22.1 Syllable structure. Syllables consist of a nuclear vowel, V, with or without initial and/or final consonantal margins, C. Syllable patterns are thus V, VC, CV, and CVC. The predominant pattern is CV. This fact, together with considerations of stress placement, leads to the interpretation of nonsyllabic high vocoids as consonants w [u] and y [i]. 22.2 Segmental phonology Consonants. There are twelve consonants represented in the orthography as (with corresponding phonetic values in parentheses): p, t, k, h [9], m, n, z, s, χ [s], r [f], w, and j [y]. p, t, k, and h are normally voiceless, unaspirated stops at bilabial, alveolar, velar and glottal points of articulation respectively: pone ['pone] 'piranha'; taky [•taki] 'bow'; ihmo ['Pmo] 'egg.' p shows no significant allophonic variation. t has a palatalized form preceding and following i. The palatalization of t results in an alveopalatal affricate [c]: pltlkO [pi'ciko] 'small', k has a morphophonemically conditioned voiced variant [g] which occurs at morpheme boundaries after elision of a final unstressed vowel and preceding morpheme initial stressed a: yt ko ase [i. 'to. 'ga.se] Ί am going.' m and n are nasal continuants normally produced at bilabial and alveolar points of articulation respectively: moe ['moe] 'far'; nuno ['nuno] 'moon', m exhibits no significant variation; n has a palatalized form preceding and following i: kokonie [koko'nie] 'yesterday'; inapyry [ina'ptri] 'his vegetables or fruit' z is a voiced alveolar fricative: zery ['zeri] 'tooth.'

APALA1

121

s and are voiceless fricatives produced at alveolar and alveopalatal points of articulation respectively: seny ['seni] 'here'; kurehxo [ku're9so] 'very good.' r is an alveolar flap: rato ['fato] 'knife.' w and j are nonsyllabic vocoids produced at bilabial and alveolar points of articulation respectively: wäko ['wäko] 'dancing'; jeny ['yent] 'his mother.' w is the nonsyllabic counterpart of u being a voiced high close back rounded vocoid. j is the nonsyllabic counterpart of i being a voiced high close front unrounded vocoid.

Vowels. There are six oral and six nasal vowels: i, i, y (i), y (t), u, u, e, e, a, ä, o, ö. i and its nasal counterpart are regularly high close front unrounded vocoids: ipyty [i'pitt] 'his wife'; ipary [i'pafi] 'his back*. y and its nasal counterpart y are normally high close central unrounded vocoids: epyko [e'piko] 'bathing'; epyko [e'piko] 'bath'. In stressed syllables preceding a, y sometimes has a [g] (voiced fricative quality: oya [oya] or [oga] 'to you'. (It is a 3-vowel sequence [o.'i.a]). u and its nasal counterpart u are regularly high close back rounded vocoids: upupuru [upu'puru] 'my foot'; usehpo [Q'se9po] 'hair'. e and its nasal counterpart e are normally mid-front unrounded vocoids ranging from close to open: eneko [e'neko] 'look'; eneko [e'neko] 'looking'; e usually occurs in open form [ ] in nonstressed syllables before p; epe [epe] 'friend', a and its nasal counterpart ä are regularly low open central unrounded vocoids: etako [e'tako] 'listen'; etäko [e'tako] 'listening'. and its nasal counterpart ö are normally low back rounded vocoids ranging from close to open: aroko [a'foko] 'take'; aröko [a'foko] 'taking'; generally occurs open [3] in final syllables and closed in reduplicated syllables; toto ['toto] 'they'; kuto [kuto] 'frog'.

22.3 Borrowed segments from Portuguese: f [p] bilabial voiceless fricative, especially used in names borrowed from Portuguese culture: firipe [pi'fipi] 'Felipe'; b [b] bilabial voiced stop: here [be'fe] 'Belem' city.

22.4 Phonotactics. Distribution of phonemes within the syllables is as follows. All consonants occur in the initial C slot of CV and CVC syllables.

122

Edward and Sally Koehn

Only h and are found in the syllable final non-word final C position; otuhkose [otu9'kose] 'eat'; köxtapänöko ['köstapä'nöko] or ['kösitapä'nöko] 'becoming chilly'. All vowels may occur in the V slot of V and CV syllables, and all oral vowels and the nasalized low central vowel occur in closed syllables: tähpome [täh'pome] Old-man-like'. All consonant-vowel sequences occur with the exception of the following: w plus u; j plus i; s or z plus i; x plus y. Consonant clusters result from the juxtaposition of syllable-final with syllable-initial consonants. Glottal stop has been found to precede all syllable-initial consonants except j. A wide variety of two and three vowel sequences have been recorded (in which each vowel occurs in a separate syllable): nae 'have'; oeary 'your cord'; oeonary 'your nose'. There are apparently no restrictions as to syllable combinations within words, except that VC and CVC syllables have not been found in word final position, except for one case of a CVC syllable in this position: tarn 'grandfather'. 22.5 Suprasegmentale. Apalai utterances may be segmented into words which are significant both as grammatical and as phonological units. The phonological unit may consist of one or more grammatical words bounded by pause. (449)

se pona yto-Vko ase there to go-CONT 1+be+PRES 'I'm going there.' [/se.po.na/*.to.ga.se/] (syllables are separated by periods and phonological units are marked by slashes)

(450)

m-oeh- no ropa aja 2-come-IP again mother 'You've come back, Mother.' [/moe9.no.ro.pa/a.ya/]

Phonologically, words are sequences of syllables in which the penultimate syllable is stressed: paruru [pa'fufu] 'banana'. A few words exist which have only one syllable: , po. These words may carry stress because of the intrinsic stress which accompanies nasalization or because of elision: kanawa ae ytöko [kana'wae i'töko] 'going from the canoe'. Secondary stress falls on closed and on nasal syllables bordered by at least one unstressed syllable: köxtapänöko ['köstapä'nöko] 'becoming chilly.' Secondary stress may fall on the initial syllable of words of four or more syllables when it is followed by an unstressed syllable: xirikozuru ['sifiko'zufu] 'grass.' The stress pattern is somewhat flexible when several of the stated factors occur. Length occurs with nasalization. Nasalized vowels are long in contrast with nonnasalized vowels: aräta [a'rä:ta] 'howler monkey.'

APALAI

123

In discourse, nasalization and length occur on the final syllable of sentence initial connectives and at the end of utterances which constitute a major transition between thoughts: mame [mame:] 'and then'; ynanoturuase. te. ma/ne...[inanoturuase: te mame:...]. 'We talked. PAUSE. And then...'

(451)

oepyase 1 +come-RP ese drink-PURP

taro-na here-to aporo. briefly

aporo briefly opi medicine

pakeimo early

opi medicine

enyase. 1 +drink-RP

te. PAUSE

moroto-ino that- after Ί came here first early to take some medicine. I took it. (pause) And then...' In (451), the final e of eny-ase is nasalized and lengthened. The basic intonation patterns are: falling, which occurs sentence final; and level or rising, which occurs at the end of clauses and dislocated phrases within a clause or sentence, when these are not sentence final. 22.6 Morphophonology. A number of generalizations can be made about the system of pronunciation of the language, including certain morphophonological processes: (i) Words never begin with a consonant cluster. (ii) A sequence of vowel plus nasal consonant in morpheme final position, and preceding another consonant, becomes a nasalized vowel: anym-se -» anyse 'fished.' (iii) Any stop in morpheme final position becomes glottal stop preceding a consonant: nyk-se —> nyhse 'slept.' The following rule is complementary to rules (ii) and (iii). (iv) The morphophonemic change in consonant-final verb stems is a transition vowel y [i] , central unrounded, occurring epenthetically between the stem final consonant and the vowel initial suffix. Generally, if the final stem vowel is u, the transition vowel harmonizes to u: otuk-ase —*· otukuase Ί ate'; jy-nyk-ase —» jynykyase Ί slept.' However, there are a few idiosyncratic forms such as: enetuput-ase —» enetuputyase Ί recognized it'; oep-ase —> oepuase Ί came.' (v) The morpheme final syllables no, ny, and my, when following a nonconsonantal segment, are lost before stops, and the nonconsonantal segment is nasalized: nytono toto —» nyt toto 'they went.'

124

Edward and Sally Koehn

(vi) Elision of vowels at morpheme junctures is fairly common, with the second vowel becoming dominant if the two are different: airiki-amuru —» airik-amuru 'palm leaf.' (vii) Any semivowel becomes consonantal when it is bounded on both sides by a vowel or when it is preceded by a k or silence and followed by a vowel. The rule operates from left to right kujujuari [kuiuiu'ah] —» [kuyuyu'ari] 'bird species'; wäko [u'ako] 'dancing'; kwirikwiri [ku'iriku'iri] —» ['kwifi'kwiri] 'bird species'. (viii) HI tends to become nonvoiced between nonvoiced segments or in morpheme final position following a nonvoiced segment: pitiko [pltlkO] 'sma11

23 MORPHOLOGY

This is treated in earlier sections. 24 IDEOPHONES The ideophone is a noninflected onomatopoeic word that denotes an action that is normally expressed by a finite verb form. It is the only class of word that cannot be followed by a postpositional particle. It functions normally as a distinct sentence constituent, carrying the same meaning as that contained in the finite verb of that sentence (i.e. the basic part of the meaning, not including person, tense etc., which are not part of the meaning of the ideophone) (452): (452)

apoi-ko repe kyry grab-HIS CONTRAEXP grabbing (IDEOPH) 'But it (the frog) grabbed him instead.'

It may also function as: (i) a kind of direct object of the verb ka 'say, do', in which case it normally precedes that verb, as in (14); (ii)

a separate sentence, occurring in isolation (453); and

(iii) a substitute for a verb, cooccurring with nominals and/or adverbials to constitute a separate sentence, as in ((40) and (47)). (453)

kuto jakuohno. pyhseky frog 3SlO-take across- IP jumping(IDEOPH) ' "The frog took me across," (he said). Act of jumping.'

The ideophone may be a single morpheme (e.g. kyry 'action of taking hold of) or a sequence of reduplicated forms (e.g. koe koe koe koe 'action of

APALA1

125

paddling')· In the latter case the number of repeats of the form may be from two to ten or more but it is usually not more than six. Several different ideophones may occur in sequence. Ideophones also show some variations from the normal phonology of the language. One or two additional phonological units occur: [6], voiced implosive bilabial stop, as in [6e] 'dying', [g] voiced velar fricative, and [ae] open central low unrounded vowel, as in [pe.gae pe.gae] 'flapping (wings) clumsily'. In the phonotactics, there is a word final consonant h [9] that does not occur elsewhere: sysyh [sist9] 'climbing'. Ideophones are used frequently in everyday conversation and in story telling. Normally they consist of one or two syllable morphemes. The following list is not exhaustive, but presents ideophones which occur in three legends.

be

'dying'

koe koe koe koe

'paddle'

peei

'shoot'

kute kute kute

'(frog) croak'

pyhseky

'jump'

pyh tere

'jump into a canoe'

ti

'arrival of moon'

syryryry

'gliding movement of canoe'

sororo

'moving of canoe in shallow water'

syry tope topö

'falling into the water'

sysyh

'climbing a tree'

kui kui

'screaming'

soko soko

'munching'

seky seky

'creep up'

tä

'hit hard object against another hard object'

126

Edward and Sally Koehn

wezo wezo

'squeeze through'

popo popo

'flapping of wings'

ty ty ty

'person walking'

pyre pyre

'coming out, appearing'

woe

'lifted up'

toh toh toh toh toh toh

'beat to death'

kuh soro

'finish beating and throw out'

te rokene

'short time passing'

pore

'arrived'

wyyy

'clearing field'

soko soko soko soko

'digging of debris'

somo

'pick up'

puto

'put dirt on top of, pour out'

tu tu tu

'stamping'

tototo tototo

'bird falling, struggling to flap wings'

wywywywy

'hammock swinging'

uroruro

'trees falling'

tororira

'flute music'

tyry tyry tyry tyry tyry

'gourds rattling'

soposopo

'stepping into water'

tututututu

'fast approach'

APALAI

kyryh

'take hold of

pupupupu-tyryry

'sound of wind'

127

REFERENCES Derbyshire, D. C. 1979 Hixkaryana, Lingua Descriptive Studies 1 (Amsterdam: North-Holland). Koehn, E. 1976 "The historical tense in Apalai narrative," UAL 42:243-52. Koehn, S. 1976 Semantic clause analysis (ms). To appear in a Portuguese version (Brasilia: Summer Institute of Linguistics).

CANELA-KRAHÖ* Jack Popjes and Jo Popjes 0 INTRODUCTION The Ramkokamekra Canela are part of the Je linguistic family, spoken on the great central plateau of Brazil. Other Je-speaking people include the Timbira, Apinaje, Xerente, Xavante and Kajapo, located mosdy south west of the Canela (Nimuendaju 1946, Crocker 1961). The Canela language is spoken with minimal dialect variations by three separate groups. The Ramkokamekra Canela live in a village of about 800 people about 50 miles south of Barra do Corda, Maranhäo. About 350 Apanjekra Canela live 30 miles west of the Ramkokamekra Canela village. Some social interaction occurs, and there have been a few cases of intermarriage. The language is also spoken by the 900 Kraho of Northern Goias, over 200 miles south west of the two Canela villages. There has been no social interaction to speak of in recent years between the Kraho and the Canela villages. The total number of speakers is thus around 2000. Bilingualism and acculturation vary from village to village as well as between age groups and between sexes. Generally die most bilingual are the young men in their teens and early twenties. In the Ramkokamekra village at least 70 people can read and write, at least to some extent, in Portuguese and in Canela. There has been extensive contact for over 50 years with the surrounding Portuguese speaking Brazilian population. The Canela-Kraho have, however, been able to preserve their culture and life style, customs and language, to a remarkable degree. Although both bilingualism and acculturation are increasing in all locations, thus far the changes are relatively small. Both groups appear to be growing economically dependent on the government. SYNTAX OF THE SENTENCE OR CLAUSE 1 Word order There are 10 clause types: six verbal clause types and four non-verbal. They are distinguished on the basis of their obligatory constituents. In the listing that follows, the constituents appear in their normal, unmarked order. Following the discussion of the ten clause types (sects. 1.1 - 1.10), there is a summary of basic word order patterns (sect. 1.11) and a description of peripheral clause constituents (sect. 1.12). 1.1 Transitive. Differences between this clause type and the other types are most clearly seen in the past tense forms. The clause nucleus consists of a

CANELA-KRAHO

129

subject, a past tense marker, a direct object, and a verb. The subject may be either a pronoun, a noun phrase or a prefix attached to the past indicative tense marker te. (See sect. 16 for free and bound pronouns). If the subject is a noun phrase, the tense marker te occurs immediately following it, without a prefix. The object may be a noun phrase or a person prefix bound to the verb. There is one subgroup of transitive verbs where the object in the third person is not overtly marked by a prefix. The long form of the verb occurs with the past tense. (See sect. 23). The examples that follow include some with present and future tenses also. (1)

wapo te i-xec knife PAST 1-cut 'The knife cut me.'

(2)

co he! que ha wapo a-xec warning 3 PUT knife 2-cut 'Careful, the knife will cut you!'

The following examples show the object unmarked. (3)

i-te h hk r 1-PAST 3+buy Ί bought it.'

(4)

wa ha ρϊχό j hk 1 PUT fruit buy Ί will buy fruit.'

(5)

humre te cakwin man PAST 3 +beat 'The man beat it.'

(6)

humre apu a-cakwi man CONT 2-beat The man is beating you.'

(7)

cu-te hitep 3- PAST 3 +cut 'He cut it.'

(8)

ca pi jitep 2 wood cut 'You cut wood.'

The following examples show the object as a noun phrase.

130

Jack and Jo Popjes

(9)

i-te ρϊχό 1-PAST fruit Ί bought fruit.'

j hk r buy

(10)

humre te rop cakwin man PAST dog beat The man beat the dog.'

(11)

cu-te pi jitep 3- PAST wood cut 'He cut the wood.'

See sect. 23 for the morphophonemic rules which explain the initial j/h and x/h alternations. The following examples show the object as a prefix on the verb. (12)

i-te a-pupun 1-PAST 2-see Ί saw you.'

(13)

a-te ih-kre 2-PAST 3- plant 'You planted it.'

The quotative is another type of transitive sentence (see sect. 14). 1.2 Stative transitive. This is a special form of transitive, directly related to the Stative clause (see sect. 1.5). It has most of the formal characteristics described above for transitive clauses, including the same word order, object noun phrase or verb prefix, and subject noun phrase or prefix. In this case, however, the postposition (to which the subject prefix may attach) can be either te or m . When te occurs it carries the meaning 'habitual state' not 'past tense' as in the non-stative transitive clauses. When m occurs it means 'temporary state'. Both te and m occur with the same meanings in simple stative clauses (sect. 1.5). Free pronoun subjects do not cooccur with the prefixed te and ma.

(14)

i-te h pa 1-HAB 3 +fear Ί live afraid of it.'

(15)

a-te amji kin 2-HAB self like 'You always like yourself or 'You are a happy person.'

CANELA-KRAHÖ

(16)

cu-te rop jape 3- HAB dog pity 'He always has pity on the dog' or 'He wants the dog near him.'

(17)

hümre ata te rop kin man that HAB dog like 'That man (habitually) likes dogs.'

(18)

i-mä 1-TEMPRY like you.'

(19)

i-mä rop kin 1-TEMPRY dog like like the dog.'

131

a-kin 2-like

1.3 Pseudo-transitive. This clause has some characteristics of a transitive construction and some of an intransitive. The subject and past tense marker te occur exactly as described above for transitive clauses, and the long form of the verb occurs. The person prefix of the verb, however, agrees with the subject, and what is semantically the direct object is expressed as an oblique object, by means of a noun phrase or person prefix followed by one of the relators: to, ma, kam. The verbs which govern ma are those where a human object would be expected, e.g. 'divorce', 'teach', 'play with', etc. The verbs which govern to appear to be those where a non-human object would be expected, e.g. 'steal', 'trade', 'rip', 'scratch', etc. These two postpositions ma and to occur in peripheral clause constituents to signal 'benefactive' and 'instrument' respectively. (20)

a-te po kam a-catöc 2-PAST deer at 2-shoot 'You shot (at) the deer.'

(21)

ca ha po kam a-catöc 2 PUT deer at 2-shoot 'You will shoot (at) the deer.'

(22)

i-te a-mä i-cator 1-PAST 2-to 1-arrive found you (arrived to you).'

(23)

a-te carä cahäcre 2-PAST goats 'You raised goats.'

to INST

a-pijapar 2-raise

132

Jack and Jo Popjes

1.4 Intransitive. The clause nucleus consists of a subject and a verb. The subject may be a free subject pronoun, a noun phrase, or a person marker prefixed to the verb. With the majority of intransitive verbs, the long form of the verb (see sect. 23) occurs in the past tense, and the subject is prefixed to the verb, unless it is a noun phrase. With a small sub-class of irregular intransitive verbs, the long form of the verb does not occur in the past tense unless the verb is non-final in the verb phrase. When the short form of the verb occurs, the subject may be a free pronoun or noun phrase. (24)

i-rit 1-see •I saw.'

(25)

capi jäpir Capi climb 'Capi climbed.'

(26)

pe capi api DP Capi climb 'Capi climbed (long ago).'

(27)

a-jöt 2-sleep 'You slept.'

(28)

a-mör nare 2-go NEC 'You did not go.'

(29)

ihnö kam ca mo yesterday 2 go 'You went yesterday.'

(30)

i-crer 1-sing sang.'

(31)

pe ca ere DP 2 sing 'You sang (long ago).'

1.5 Stative. The simple Stative clause contains a verb that has a single, uninflected form (contrast the stative transitive, described above, where the verb is inflected for person of the object). The subject and postpositions which cooccur with it are the same as for the stative transitive, and the same set of

CANELA-KRAHO

133

verb roots occurs in both constructions. Each member of the set can cooccur with both te 'HAB' and mä 'TEMPRY' (cf. (33) and (34)). (32)

cu-te kry 3- HAB cold 'He is always cold/chilly/feverish.'

(33)

il-

(34)

i-mä pa 1-TEMPRY afraid am afraid (right now).'

(35)

a-mä präm 2-TEMPRY hungry 'You are hungry.'

(36)

i-mä 1-TEMPRY 'I'm cold.'

te pa HAB afraid am always afraid.'

kry cold

1.6 Adjectival. The distinguishing characteristic of the adjectival clause is the class of verb: all members of the class occur also as adjectives in noun phrases and as adverbial modifiers in verb phrases (see sects. 15.3, 18.6 and 19). As verbs in adjectival clauses they are always inflected, by means of a prefix, for person of the subject. A free form subject may also occur. (37)

in-tyt 3- narrow 'It/he/she is narrow.'

(38)

rop im-pej dog 3- good 'The dog is good.'

(39)

i-paham 1-ashamed am ashamed.'

The four non-verbal clause types are distinguished by the presence or absence of a subject and, when it occurs, the means by which it is related to the predicate complement.

134

Jack and Jo Popjes

1.7 Equational. The equational clause consists of a subject and a complement. The subject may be a noun phrase or a demonstrative. The complement is a noun phrase.

(40)

ita ken DEM stone "This is a stone.'

(41)

ata-je ankrare DEM-PL children 'Those are children.'

(42)

pur itekje field 1possession "The field is mine.'

(43)

ita i-pur DEM 1-field "This is my field.'

1.8 Identifications!!. The identificational clause also consists of subject and complement, but in this case there is a relator postposition pe, which functions as a copula. It follows the subject, which may be a noun phrase or a prefix attached to pe. The complement is a noun phrase. (44)

capi pe mehi Capi COP Indian 'Capi is an Indian.'

(45)

i-pe capi 1-COP Capi Ί am Capi.'

(46)

a-pe hakryj cate 2P rich NOMLZR 'You are a rich person.'

When the noun hapryljapry 'name' is the subject, the form of the relator copula is te rather than pe, and the complement is a proper name. (47)

hapry te capi name COP Capi 'His name is Capi.'

(48)

humre japry te capi man name COP Capi "The man's name is Capi.'

CANELA-KRAHÖ

(49)

135

i-japry te capi l-name COP Capi 'My name is Capi.'

1.9 Existential. The existential clause consists of a complement and a subject, occurring in that order. The complement may be a person prefix, a noun phrase or a locative phrase, followed by one of the relators ma, ri or kam. The relator ma is used for persons, and ri and kam indicate location. The subject is a noun phrase. (50)

pur kam pöhy field RELTR corn 'There is corn in the field.'

(51)

anket kam patti bushes RELTR snake 'There is a snake in the bushes' or "There are snakes in the bush.'

(52)

curi co there- RELTR water 'There is water.'

(53)

i-mä pöhy 1-RELTR corn have com.'

(54)

capi mä catöc Capi RELTR gun 'Capi has a gun.'

(55)

cu-mä pi 3- RELTR wood 'He has wood.'

(56)

curi pi there- RELTR wood 'There is wood.'

The English gloss 'has/have' may be misleading-the emphasis in Canela is not on possession but on existence or availability. 1.10 Temporal.

(57)

The temporal clause consists of a single uninflected word.

awa night 'It is night.'

136

Jack and Jo Popjes

(58)

cacro hot 'It is hot.'

(59)

awcapätko dark 'It is dark.'

(60)

hamre finished 'It is finished.'

1.11 Basic word order. The basic word order for both main and subordinate clauses may be summarized as follows: -

transitive and stative transitive: subject-object-verb pseudo-transitive: subject-oblique object-verb intransitive, stative and adjectival: subject-verb equational and identificational: subject-complement existential: complement-subject

For the purpose of focus or emphasis the object in transitive clauses may be fronted, and phonologically dislocated by pause, and the verb prefix then occurs. Since there are several other ways to mark emphasis (see sect. 16) this fronting of the object is rare, being found most often in the context of correcting a misunderstanding in the minds of the hearers. (61)

i-te po curan 1-PAST deer kill killed a deer.'

(62)

po, wa i-te ih-curan deer, 1 1-PAST 3- kill 'It was a deer I killed.'

1.12 Peripheral clause level constituents. There are eight peripheral clause level constituents. All but two, aspect and manner, are marked by postpositions. Time commonly occurs clause initial. The other seven elements, location, benefactive, instrument, directional, comparison, manner and aspect, commonly occur between the subject and the object in transitive clauses, between the subject and the verb in other verbal clauses, and before the first nuclear element in non-verbal clauses. These peripheral elements may occur, subject to certain semantic constraints, in all clause types. Directional words and phrases are usually found in intransitive clauses with intransitive verbs of movement. They usually precede the verb.

137

CANELA-KRAH

The normal maximum number of peripheral constituents in a clause is four. This number is further reduced to two if they take the form of long phrases or a series of phrases. The number of constituents occurring between the subject and the object or verb does not normally exceed two. When the number is more than two, the preferred ordering is as follows: time precedes the subject; location and instrument follow die verb; benefactive, comparison, aspect and manner are not normally found either before the subject or after the verb. In transitive clauses the object is normally followed by the verb. (63)

wa ha pur kam cu-m pi jakep 1 PUT field in 3- for wood cut Ί will cut wood for him in the field.'

(64)

ape na wa ha pur kam cu-m morning in 1 PUT field in 3- for Ί will cut wood for him in his field tomorrow.'

(65)

ape morning

na wa ha in 1 PUT

pur field

kam cu-m in 3- for

ρϊ wood

jakep cut

ρϊ jakep, wood cut

waky to axe with Ί will cut wood for him in his field tomorrow with an axe.' (66)

caxwa night

ri at

wa ha 1 PUT

in-to 3- eye

kaj na away from

aracri quietly

cu-pe 3- MAL

ρϊ jakep, pur kam wood cut field in Ί will quietly cut wood from his field tonight without him seeing it (taking it away from him, for his negative benefit).' 2 Parataxis Both phrases and clauses may be juxtaposed. 2.1 Phrases. There is juxtaposition of both noun phrases and adverbial phrases. This juxtaposition occurs both before and after the verb. (67)

i-te 1-PAST

amji self

kam awke to, at left with

hakati snake

imput neck

na around

i-pyr 3-grab Ί grabbed the snake around the neck with my left hand.'

138

Jack and Jo Popjes

(68)

wa te, i-japen xä wyr, 1 go 1-woik NOMLZR to 4 1 went to work, to my cane weeding.'

(69)

ha, hey,

me PL

i-kra, 1-child

me PL

can cane

i-wawy, 1-son in law

japrar weed

ca me 2 PL

wyr to amji self

mä hare,... to say 'Hey, my children and my sons-in-law, you say to yourselves,...' Phonological dislocation occurs when the phrases are long or when they occur following the verb. Such dislocation is to the right, toward the end of the clause, following the verb. Normally, discontinuous juxtaposition is limited to one phrase after the verb. (70)

harkwa mä xa, ken kre kam mouth to stand hill hole in '(He) stood in the opening of the cave.'

Discontinuous sequences are common, especially sequences involving locative or time phrases. The function of juxtaposed phrases is commonly one of modification, that is, each phrase narrows the spectrum of meaning of the previous phrase. (71)

ken kam hikwa, ken kre kam hill in live hill hole in '(He) lived in a cave in the hills.'

(72)

caxwa ri hyrmä me mo, awcahti xä wyr night at to him PL go midnight NOMLZR toward "They went to him at night, before midnight.'

In (72), hyrmä is a frozen form meaning literally 'there where he is.' Phrase parataxis can occur in subordinate clauses but is not as common there as in main clauses. (73)

i-te 1-PAST

a-te 2-PAST

ihmutri, there

capi Capi

jürkwa house

cahhyr na a-pupun beat SUBORD 2-see 4 1 saw you beat your child there at Capi's house.'

ri, at

a-kra 2-child

CANELA-KRAHO

139

2.2 Clauses. There is also juxtaposition of clauses, the function being either coordination or subordination. Juxtaposed clauses are coordinate in function in a listing of activities, all going on more or less at the same time, or within the same time period, though not necessarily in the order stated. (74)

wa apu amra, 1 CONT cry

apu i-toj, CONT 1-jump

apu ajhu, CONT stumble

apu CONT

amji jam!, ajco apu amrä self bury HAB CONT cry continually cried and jumped and stumbled and threw earth on myself, (I used to be) crying all the while.' Juxtaposed clauses may also involve a subordinate relationship, e.g. result or reason: (75)

pe wa i-pym, pe inxe PAST 1 1-fall PAST mother 'My mother died when I was bom.'

ty die

(76)

wa ha ma ajcahu, i-mä 1 PUT away run 1-TEMPRY will run away because I'm afraid of it.'

hüpati 3+fear

3 Ellipsis Any clause constituent other than the verb phrase and the indirect object with its postpositions can be omitted when it is recoverable from the general context. The deleted element is marked as a prefix on the verb if it is the object, or the subject of an intransitive clause, and on the tense marker if it is the subject of a transitive clause: (77)

hümre te po curan man PAST deer kill 'The man killed the deer.'

(78)

cu-te po curan 3- PAST deer kill 'He killed the deer.'

(79)

cu-te ih-curan 3- PAST 3- kill 'He killed it.'

140

Jack and Jo Popjes

Certain elements, such as instrument, may also be omitted under identity in coordination. (80)

cu-te wapo pyr ne to po curan 3- PAST knife take and INST deer kill 'He took the knife and killed the deer with it.'

The verb phrase can be replaced by a dummy verb, which consists of the (main) verb to 'do, make' and hajyr/hane 'thus'. (81)

a-te pur poc, wa i-te ita to hajyr 2-PAST field burn 1 1-PAST this do thus 'You burned off your field and I did likewise.'

(82)

que ha 3 PUT

capi Capi

pur field

to REL

cuto, set fire

wa ha 1 PUT

hanea thus

ita to hane this do thus 'Capi will set fire to his field and I will do the same.' When the context is very specific as in the case of responses to questions, even the verb phrase can be omitted. (83)

jü kam ca te? pur wyr where 2 go? field to 'Where are you going? To the field.'

The verb may also be omitted from the subordinate clause in causative sentences or indirect commands. (84)

i-te jaco jö na i-prö to ihtyj 1-PAST Jaco food about 1-wife SUBORD assign assigned my wife to (prepare) Jaco's food.'

(85)

wa ha capi pi na cu-me 1 PUT Capi wood about 3- throw (order) ordered Capi to (cut) the wood.'

The verb of the quotation formula that introduces direct speech can also be omitted (see sect. 14). 4 Reflexives and reciprocals Reflexivity is expressed by an invariable form amji 'REFLX (= self)'.

CANELA-KRAH

(86)

wa amji cakwin l REFLX hit Ί hit myself.'

(87)

a(pu) me amji cakwin CONT 3PL REFLX hit 'They are hitting themselves.'

141

The following examples show the reflexive element in various coordinate and subordinate clauses. (88)

wa ha ikre wyr te ne amji 1 PUT house to go and REFLX Ί will go home and scratch myself.'

(89)

jaco Jaco

te PAST

pi jakot, wood pile up

cu-te 3- PAST

cakre scratch amji REFLX

m for

hakep cut

ataje DEM 'Jaco piled up the wood he cut for himself.' The reflexive is always controlled by the subject and can function as either a direct object or an oblique object (benefactive, addressee, locative): (90)

i-te amji pytar 1 -PAST REFLX defend Ί defended myself.'

(91)

cu-te amji m ρϊ jakep 3- PAST REFLX for wood cut 'He cut wood for himself.'

(92)

capi amji m ih-cakoc Capi REFLX to 3- speak 'Capi speaks to himself.'

(93)

i-te ih-curan ne amji cacec 1-PAST 3- kill and REFLX back Ί killed it and put it behind myself.'

Reflexives may also occur in nominalizations.

rum toward

hir put

142

Jack and Jo Popjes

(94)

me amji pupun cateje PL REFLX see NOMLZR- PL 'acquaintances, neighbors, people who know themselves* (i.e. people who have gone through hard times, have suffered troubles)

(95)

amji kam hapac xä REFLX at hear NOMLZR 'self hearing thing (thought, idea, custom, lifestyle)'

Reciprocity is expressed by an invariable form ajpen commonly in conjunction with the plural me, meaning each other, and functions much as the reflexive amji. It is controlled by the subject and functions as the direct object or oblique constituent. (96)

jaco Jaco

me capi and Capi

te PAST

pi here wood twig

jakep cut

ne me and PL

to EMST

ajpen caprec REC1P beat 'Jaco and Capi cut twigs and beat each other with (them).' (97)

jaco me capi ajpen mä pi jakep Jaco and Capi RECIP for wood cut 'Jaco and Capi cut wood for each other.'

(98)

jaco Jaco

me capi and Capi

te PAST

pi here wood twig

jakep cut

ame to 3PL 1NST

ajpen cahhyr pram te RECIP beat want because 'Jaco and Capi cut twigs because they wanted to beat each other with (them).'

5 Passives There are no passive constructions.

6 Causatives Any verb can be made causative by the use of the transitive verb to I ton 'make/do'. The verb to be made causative is placed in a subordinate clause, which is marked as such by the postposition na, and which immediately precedes the causative verb. The direct object of the causative verb agrees with the subject of the verb in the subordinate clause. In the case of the third person, the direct object marker

CANELA KRAHO

143

on the causative verb is zero (see (100) and sect. 23). When the verb in the subordinate clause is transitive, the subject is omitted. (99)

capi te i-jöt Capi PAST 1-sleep 'Capi made me sleep.'

(100)

i-te i-prö jäpen 1-PAST 1-wife work made my wife work.'

(101)

pahhi amji kin na me pato chief REFLX like SUBORD PL 1INCL- make The chief makes us have a festival/enjoy ourselves.'

(102)

que 3

ha pahhi PUT chief

na i-to SUBORD 1-make

me PL

na ton SUBORD 3+make

pamä 1INCL- for

amji REFLX

kin- ti like-much

na me pato SUBORD PL lINCL-make 'The chief will make us very happy/have a festival.'

(103)

capi te i-pej Capi PAST 1-good 'Capi made me good.'

na i-ton SUBORD 1-make

7 Comparatives and equatives Comparatives are expressed by means of postpositional phrases. Equatives and similitives are expressed by verbal predications. Comparison is expressed by two postpositions which follow the standard of comparison. They are hirö pe 'more than', and hapyre mä 'less than'. The subject or item to be compared occurs clause initial and it is followed by the standard of comparison, which is followed by the postposition, followed by the rest of the clause. (104)

jaco capi jiro pe Jaco Capi more than 'Jaco is bigger than Capi.'

cati big

(105)

capi kryt jiro pe hi Capi Kryt more than body 'Capi jumps better than Kryt.'

(106)

kryt jaco japyre mä Kryt Jaco less than 'Kryt is smaller than Jaco.'

to DMST

in-crire 3- small

in-toj 3- jump

144

Jack and Jo Popjes

Tense markers or relators occur following the subject and before the standard of comparison. (107)

capi te kryt jiro pe Capi PAST Kryt more than 'Capi cut more wood than Kryt.'

pi jakep wood cut

(108)

kryt mä capi japyre ma Kryt TEMPRY Capi less than 'Kryt likes the dog less than Capi.'

(109)

que ha kryt jiro pe pi jakep 3 PUT Kryt more than wood cut 'He will cut more wood than Kryt.'

rop kin dog like

Comparison is also expressed by the postposition kin te 'unfavorable comparison'. The standard of comparison occurs clause initial followed by the postposition, followed by the subject or item to be compared, followed by the rest of the clause. (110)

capi kinte kryt crire Capi COMPAR Kryt small 'Kryt is small in (unfavorable) comparison to Capi.'

(111)

capi kinte kryt te pi crere ne hakep Capi COMPAR Kryt PAST wood few and cut 'Kryt cut few trees in (unfavorable) comparison to Capi.'

(Numerals and quantifiers like crere 'few' act like verbs and always require to be separated from another verb by ne 'and'). Equation is treated by linking the two nouns in a coordinate noun phrase which is both the subject of a main clause in which the verb is ipipen 'be equal' and of a subordinate clause which contains the attribute or action being compared. In the case of an active verb, the subordinator is to. Within context the attribute can be omitted. (112)

capi me kryt cati pipen Capi and Kryt big equal 'Capi and Kryt are equally big.'

(113)

capi me kryt te pi jakep Capi and Kryt PAST wood cut 'Capi and Kryt cut wood equally.'

to me SUBORD PL

i-pipen 3-equal

CANELA KRAHO

(114)

capi me kryt me in- toj Capi and Kryt PL 3- jump 'Capi and Kryt jump equally.'

145

to me SUBORD PL

i-pipen 3-equal

The coordinate noun phrase can occur not only as a subject as in the examples above but also as a direct object and as a benefactive adjunct. (115)

jaco te rop me ropo curan to i-pipen Jaco PAST dog and cat kill SUBORD 3-equal 'Jaco killed equal numbers of dogs and cats.'

(116)

jaco Jaco

te PAST

rop me dog and

ropo cat

curan kill

xä ita to NOMLZR DEM SUBORD

i-pipen 3-equal 'Jaco killed the dog and cat in the same way (equally).' (117)

jaco te pa me capi mä hämjör Jaco PAST 1EXCL and Capi to pay 'Jaco paid me and Capi the same (equally).'

to i-pipen SUBORD 3-equal

Similarity between two items, where the similarity is expressed by a stative verb, is treated as follows: the item being compared for similarity occurs clause initial, and is followed by the copula pe, which is followed by the standard of similarity, which is followed by the verb hurac pyrac 'be similar'. (118)

capi pe kryt pyrac Capi COP Kryt similar 'Capi is like Kryt.'

(119)

a-pe hüräc 2-COP 3+similar 'You are like him.'

(120)

capi pe kryt cati Capi COP Kryt big 'Capi is as big as Kryt.'

pyrac similar

A stative or adjectival verb may follow the standard of similarity and is then followed by the verb hürac Ipyräc (120). Similarity between two items, where the similarity is expressed by an active verb, is treated as follows: the item being compared for similarity occurs clause initial, and is followed by the postposition te, which is followed by a subordinate clause containing an active verb and the subordinator to, which is

146

Jack and Jo Popjes

followed by the standard of similarity (either a nominal or verb prefix) and the main verb hürädpyräc 'be similar'. (121)

i-te i-picahur to a-pyrac 1-HAB 1-run SUBORD 2-similar am like you in running.'

(122)

a-te a-toj to i-pyrac 2-HAB 2-jump SUBORD 1-similar 'You jump like me, the same as I.'

(123)

capi te pi jakep to kryt Capi HAB wood cut SUBORD Kryt 'Capi cuts wood like Kryt does.'

pyrac similar

Similarity can also be expressed by a single clause, using the inflected postposition cuxa 'like'. The subject or item being compared for similarity occurs clause initial, followed by the standard of similarity and the postposition, followed by the verb. (124)

wa ha a-cuxa ape 1 PUT 2-like work will work like you.'

The standard of similarity can be displaced rightward, to follow the verb. (125)

i-te to hajyr ne i-japen, a-cuxa 1-PAST do thus and 1-work 2-like worked in the same way as you did.'

There is also a clause level particle quet which denotes superiority of the subject, leaving the standard of comparison implied and unstated. (126)

wa ha quet po cura 1 PUT COMPAR deer kill however (in positive comparison to others) will kill the deer.'

8 Coordination 8.1 Clause coordination. Clauses may only be conjoined in chronological order. They are conjoined in three ways: with the conjunctions ne and mä and by juxtaposition. (For functions of juxtaposition other than coordination, see sect. 2).

CANELA-KRAHO

147

The choice of conjunction depends on the tense of the verbs in the two clauses and the person of the subject. When the subject of two or more consecutive clauses is the same, they are joined by the conjunction ne. The conjunction can be omitted when the first clause is in the past tense and the second in the future, as in (139). (127)

capi te po curan ne que ha Capi PAST deer kill and 3 PUT 'Capi killed a deer and will eat it.'

cuku 3+eat

When (1) the subject of the second clause is third person and not coreferential with the subject of the first clause, (2) the second clause is not in the future tense, and (3) there is no time phrase initially in the second clause, then the clauses are joined by the conjunction mä. (128)

a-te po curan mä capi apu cuku 2-PAST deer kill and Capi CONT 3+eat 'You killed a deer and Capi is eating it.'

(129)

capi Capi

apu ajcahu CONT run

mä hixi and his wife

apu CONT

no lie down

ne gor and sleep 'Capi is running and his wife is lying down and sleeping.'

(130)

i-te a-pupun mä capi te 1-PAST 2-see and Capi PAST saw you and Capi also saw you.'

hanea also

ne a-pupun and 2-see

This is the basic pattern, although there are certain additional constraints having to do with the combination of tenses and the presence or absence of time phrases second clause initial. The conjunction ne is also used to join two clauses when they have different tenses, though in chronological order ((131) and (133)), and when a time word occurs second clause initial (135). If the person of the subject in the second clause is either first or second person and the tense is non-future (first clause is past and second is present) the conjunction ne freely fluctuates with juxtaposition ((131)-(132) and (133)-(134)). (131)

capi te po curan ne wa apu cuku Capi PAST deer kill and 1 CONT 3+eat 'Capi killed the deer, and I'm eating it.'

148

Jack and Jo Popjes

(132)

capi te po curan, wa apu cuku Capi PAST deer kill 1 CONT 3+eat 'Capi killed the deer, I'm eating it.'

(133)

i-te pi jakep ne ca apu 1-PAST wood cut and 2 CONT Ί cut the wood and you are piling it up.'

(134)

i-te pi jakep, ca apu hakot 1-PAST wood cut 2 CONT 3+pile up Ί cut the wood, you are piling it up.'

(135)

i-te po curan ne ape na que 1-PAST deer kill and tomorrow on 3 4 1 killed a deer and tomorrow he will eat it.'

hakot 3+pile up

ha cuku PUT 3+eat

The clauses are also joined by juxtaposition when both clauses are in the past tense: (136)

i-te po curan, a-te 1-PAST deer kill 2-PAST Ί killed the deer, you ate it.'

ih-krer 3- eat

Clauses of which both verbs are in the future tense are joined through juxtaposition. (137)

wa ha po cura, que ha 1 PUT deer kill 3 PUT 4 1 will kill a deer, he will eat it.'

cuku 3+eat

(138)

que ha ta iwryc jicu, wa ha 3 PUT rain fall stop 1 PUT 'When it stops raining, I will hunt.'

awjahe hunt

Also where there is no time word second clause initial, and the second clause is in the future tense, the clauses are joined through juxtaposition. Compare (135) where there is a time word with (139): (139)

cu-te po curan, que ha ih-kre 3- PAST deer kill 3 PUT 3- eat 'He killed the deer and he will eat it.'

The conjunction cakro carries the idea 'but, although, in spite of, contrary to expectation'. It may join any verbal clauses, irrespective of tense.

CANELA KRAHO

(140)

i-jäpen 1-work

cati big

cakro, but

cu-te 3- PAST

149

nee MEG

i-mä 1-to

hämjör pay

n are MEG 'Although I worked hard, he didn't even pay me.'

(141)

ca ha ajcahu cakro, ca ha 2 PUT run but 2 PUT 'You will run but you will not arrive.'

a-cator 2-arrive

nare MEG

The negative construction nee...nare carries the idea of 'but, not even'. This construction is used only in the second clause (see sect. 12, also (140)), and may follow the conjunctions ne and ma 'and': (142)

i-picahur ne nee i-cator 1-ran and NEC 1-arrive ran but didn't (even) arrive.'

(143)

capi Capi

te PAST

pahhi chief

mä for

häpen work

nare NEG mä nee cu-te and NEG 3-PAST

cu-mä hämjör nare 3- to pay NEG 'Capi worked for the chief but he didn't even pay him.' The negator nare functions as a conjunction carrying the meaning Or if not, then'. In this usage, nare occurs between the two clauses, and the intonation pattern usually associated with negation does not occur (see sect. 12 for normal use of nare with special intonation and the long form of the verb): (144)

ca ha 2 PUT

pur field

wyr te, to go

nare ca ha NEG 2 PUT

kri village

wyr te to go 'You will go to the field or if not you will go to the village.'

(145)

ca ha 2 PUT

pur field

wyr a-tem to 2-go

nare, NEG

ca ha 2 PUT

kri wyr te village to go 'If you do not go to the field, you will go to the village.'

150

(146)

Jack and Jo Popjes

wa ha 1 PUT

ita today

kam i-tem on 1-go

nare na, wa ha NEG SUBORD 1 PUT

ape na te, nare, wa ha amcro nöa tomorrow on go NBG 1 PUT day other 'If I don't go today, I will go tomorrow, or if not, then I'll go another day.'

na te on go

More than two clauses may be coordinated by ne, mä and juxtaposition, and these then carry no semantic load other than simple conjunction and subject/tense contrast. The conjunction is repeated each time, except that lists of actions sometimes occur as juxtaposed clauses (see sect. 2). Syntactic coordination which expresses a semantic function of subordination does not extend over more than two clauses. (147)

i-te po curan, ne ih-krer, 1-PAST deer kill and 3- eat 'After killing the deer, I ate it and sang.'

ne i-crer and 1-sang

In the second or following clauses the subject and object free forms can be omitted if they are the same in both clauses. In transitive clauses the subjecttense relator te is omitted. The verb and the verb phrase level subject marker apu 'CONTINUATIVE', cannot be omitted. Verb prefixes may not be omitted. 8.2 Phrase coordination. Nouns may be coordinated into noun phrases with two or more heads linked by the conjunction me, which is phonologically linked to the head preceding it. (148)

wa apu pöhy 1 CONT corn

tepti red

me, kwyr and manioc

jakare white

me, aryjhy and rice

kre plant 'I'm planting red corn, white manioc and rice.'

(149)

capi me kryt ma te Capi and Kryt away go 'Capi and Kryt go away.'

When the idea of accompaniment is in focus, the conjunction me follows the second head. (150)

capi kryt me kri wyr te Capi Kryt and village to go 'Capi went to the village with Kryt.'

CANELA-KRAHO

151

I do not know of any constraints on the types of noun phrases that may be coordinated. Thus, a simple noun and a complex nominalized clause can be linked by the conjunction me 'and': (151)

jaco Jaco

me pahhi and chief

kot after

me PL

ipa follow

cate- je te NOMLZR-PL PAST

po pupun deer see 'Jaco and the chiefs followers saw a deer.' 9 Pragmatic and discourse characteristks In other parts of this paper there is discussion of a number of discourseconditioned phenomena: ellipsis and dummy verb (sect. 3), various types of anaphora (sect. 13), free form pronouns used for emphasis (sect. 16), certain anaphoric uses of demonstratives and cooccurring postpositions (sect. 16), and the fronting and left-dislocation of the direct object, with cooccurrence of a pronominal copy prefixed to the verb, for purpose of focus and emphasis (sect. 1). Here I discuss ways in which foregrounding and backgrounding is distinguished in Canela narrative discourse. These terms, for the two principal types of discourse information, are used here in much the same way as by Hopper and Thompson (1980) and many others: foregrounding consists of main events that are important to the narrator's purpose, whereas backgrounding includes both events and nonevents that generally support the foregrounded material but are not themselves part of the main thread of the discourse, e.g. setting, evaluation, description, reference to custom, flashback, recapitulation. The two types of information are distinguished primarily by contrasting aspect particles and past tense markers, and the (irregular) use of the long form of verbs (foregrounded) and the short form of verbs (backgrounded). Customary/habitual aspect is marked by hot pe (foregrounded) and ajco (backgrounded). The regular past tense form te ('recent' in sect. 18.1) is used in foregrounded, and a special form pe 'DISTANT PAST' in backgrounded, sections of the discourse. The long form of the verb is used in foregrounded clauses even when it occurs clause final, thus over-riding the usual syntactic conditioning (see sect. 23). The short form of the verb is found in backgrounded clauses, except when the continuative aspect particles apu and ame occur. Two examples are given below to show contrasting foregrounded and backgrounded sequences from two texts. Single parentheses signal foregrounding, and square brackets backgrounding (the specific devices are thus marked in the Canela, and the whole sections in the English translation):

152

Jack and Jo Popjes

(152)

meh-te PL- 3+shin

DP

[ajhu] run (short form)

[ ] DP

ne takamry and slowly

to INST

jitep, [pe] cut DP (invariable form)

[ajco] CUST

me PL

ma [pra] ne [ajco] away run and CUST (short form) me PL

(cu-te) 3- PAST

mo. ne me and PL go (invariable form)

hahto many

[api] climb (short form)

to INST

ih-rwen 3- few (invariable form)

ne and

me PL

mo ne (cu-te) me kam me (ih-tem) go and 3- PAST PL LOG PL 3-begin (invariable (long form) form) [It cut their shins and they used to stumble. It used to run away and climb up.] (Slowly they became fewer. And many of them went and began to attack it.) (153)

ne and

pea then

(cu-te) 3- PAST

cama ah-na only it-to

me PL

(iren) put in (long form)

ih-cucac, ne nee me 3- singe and NEG PL (invariable form)

to INST

ih-kä 3- skin

nare yhy, NEG yes

mam first

cama me only PL

[pe] DP

[ajco] CUST

ne me and PL

caxär take off (long form) pryre animal

CANELA KRAHO

ih-cucäc. pea 3- singe then (invariable form)

(cu-te) 3- PAST

kwyr manioc

153

kam me in PL

cu-mä 3- to

ih-cupu 3- wrap (invariable form) '(Then they just put it into the fire and singed off the hair. They didn't even take the skin off.)' '[Yes, long ago they used to only singe the skin.] (And then they wrapped it in manioc.)' 10 Interrogatives Interrogative sentences are marked by an overall interrogative intonation pattern (sect. 22) and a question word which occurs clause initial. Any constituent of the sentence may be questioned. Polar (yes-no) questions are marked by the interrogative marker xä which occurs sentence initial. The sentence constituents remain in their normal order and a polar question intonation pattern occurs. (154)

a-te po curan 2-PAST deer kill 'You killed a deer.'

(155)

xä a-te po curan Q 2-PAST deer kill 'Did you kill a deer?'

(156)

capi pejti Capi good 'Capi is well.'

(157)

xä capi pejti Q Capi good 'Is Capi well?'

Most constituents have a corresponding interrogative constituent which then occurs initially in the clause. If the constituent in question does not have a corresponding interrogative morpheme, as for example, aspect morphemes, the element is moved into initial position in the clause and is preceded by the interrogative marker xä.

154

Jack and Jo Popjes

(158)

cu-te ramä jakep 3-PAST already wood cut 'He already cut the wood.'

(159)

xa ramä cu-te pi jakep Q already 3- PAST wood cut 'Did he already cut the wood?'

The interrogative markers for the subject are: jurn (ma) - 'who, human, singular' jianje - 'who, human, plural' ampo (ma) - 'what, non-human' In the past tense in transitive clauses mä follows the tense marker te.

(160)

jüm mä ata who DEM 'Who is that?'

(161)

ampo mä ata what DEM 'What is that?'

(162)

jüm te mä ita ton who PAST DEM do/make 'Who made/did this?'

(163)

ampo te mä ita ton what PAST DEM do/make 'What did this?'

The interrogative markers for the object are: jüm - 'whom, human, singular' jümje - 'whom, human, plural' ampo - 'what, non-human' (164)

jüm ca a-te hömpun what 2 2-PAST see 'Whom did you see?'

(165)

ampo ca ha kre what 2 PUT eat 'What will you eat?'

CANELA-KRAHÖ

155

The interrogative markers for destination are: jü jü jit jü

ri - 'where to, specific destination' kam mä - 'where to, general destination' pin - 'where from' cajpuw - 'what destination' or 'what course'

(166)

jü ri ri capi mo where to Capi go 'Where is Capi going?'

(167)

jü kam mä capi mö where to Capi go 'Where is Capi going?'

(168)

jü pin ca mö where from 2 go 'Where are you coming from?'

The interrogative marker for time is jü caxuw. (169)

jü caxuw ca ha te when 2 PUT go 'When are you going?'

(170)

jü caxuw ca a-cator when 2 2-arrive 'When did you arrive?'

The interrogative marker for location in verbal clauses is jü ri ri and in equative clauses is jü ri mä. (171)

jü ri ri capi xa where Capi stand 'Where is Capi standing?'

(172)

jü ri mä capi where Capi 'Where is Capi?'

Interrogative markers for reason are: ampo na - 'why' ampo na mä - 'emphatic why' ampo na ri - 'emphatic why'

156

Jack and Jo Popjes

ampo cwyrjape ri - 'intensified why, for what reason' ampo caxuw ma - 'intensified why, for what purpose' (173)

ampo na cu-te to hajyr why 3- PAST do thus 'Why did he do that?'

(174)

ampo caxuw mä cu-te pi ita jakep why (purpose) 3- PAST wood DEM cut 'Why on earth did he cut this wood?'

The interrogative marker for manner or state is jü mä mä...te, where ju ma mä occurs clause initial and te occurs clause final. (175)

jü mä mä cati te how big 'How big is it?'

(176)

jü mä mä a- te how 2'How are you?'

(177)

jü mä mä ahna hö pore te how much 3 POSSN money 'How much does it cost? (how much is its money?)'

(178)

jü mä mä hömpun xä te what saw NOMLZR 'What manner of thing was seen?'

Only main clause constituents may be questioned. The constituents of a noun phrase may be questioned with question markers jüm...mä and ampo...ma, where jüm or ampo occur clause initial preceding the noun phrase and mä following the noun phrase. (179)

jüm jö rop te mä who POSSN dog PAST 'Whose dog bit you?'

a-xar 2-bite

(180)

jüm to mä cu-mä ha who eye 3- to hurt/sick 'Whose eye is sore/hurting?'

(181)

ampo pi mä prin jö pi ita what wood piqui POSSN wood DEM 'What kind of wood is "piqui" wood?'

CANELA KRAHO

157

The noun phrase of postpositional phrases may be questioned by the markers jüm mä ri 'to whom' and jüm na ri 'about whom'. (182)

capi te prejaka Capi PAST Prejaka 'Capi spoke to Prejaka.'

mä ih-cakoc to 3- speak

(183)

jüm mä ri capi cakoc to whom Capi speak 'To whom did Capi speak?'

(184)

jüm na ri cu-te about whom 3- PAST 'Whom did he ask about?'

(185)

jüm mä ri capi cakoc about what Capi speak 'What did Capi speak about?'

ih-cukij 3- ask xä te NOMLZR

The construction jüm mä ri...te in (185) has not been completely analyzed as yet, but it is probably related to the common idiomatic question jü ma ri cute 'what happened?'. Answers to polar questions take no special form. Commonly, the clause is repeated, leaving out the interrogative marker and the question intonation pattern. An affirmative yhy or negative que usually precedes the answer to polar questions. (186)

xä capi te po Q Capi PAST deer 'Did Capi kill a deer?'

curan kill

(187)

yhy, cu-te ih-curan yes 3- PAST 3- kill 'Yes, he killed it.'

Answers are often minimal, being only: yhy - 'yes' que - 'no' pypyn - 'don't know' hapa - 'no', which is used most often within a discourse where the narrator asks a question for rhetorical effect and then answers it himself negatively.

158

Jack and Jo Popjes

Answers to non-polar questions normally occur without any special intonation pattern. The element on which the question focusses tends to be fronted in the answer. (188)

ju pin ca te where from 2 go 'Where are you coming from?'

(189)

pur pin wa te field from 1 go 'I'm coming from the field.'

Compare the unmarked clause: (190)

wa pur pin te 1 field from go 'I'm coming from the field.'

11 Imperatives Imperative sentences are marked by omission of the subject and by the imperative intonation pattern (sect. 22). Imperatives occur only in the second person, singular and plural. The positive and negative forms of the imperative sentence are basically the same. The negator nare is added after the verb to negate the sentence, and the normal negation intonation pattern occurs (see sect. 12). The negator nare causes the verb to be non-final in the clause, so that the long form of the verb occurs (ex. (194) and sect. 23). (191)

po cura deer kill 'Kill the deer!'

(192)

pahhi kin chief like 'Like the chief!'

(193)

ere sing 'Sing!'

(194)

rop to a-japet dog OBL.OM 2-startle 'Don't startle the dog!'

(195)

a-tyj 2-strong 'Be strong!'

nare MEG

CANELA-KRAHO

159

With certain verbs the person-marking prefix is obligatory, whatever the mood ((194) and (195)). Other verbs have two forms, one with the prefix and the other without, and the imperative form for these verbs is always the one without the prefix. Common negative responses to imperatives are:

(196)

wa kra 1 NEC 'No, I won't!'

(197)

que, wa ha ton NBG 1 PUT do 'No, I won't do it.'

nare NEG

For use of another form of negative response to imperative, wyr, see sect. 12. Hortatives differ from the imperatives by the presence of the obligatory first person inclusive cu (dual), or cu me (plural). There is a special form of persuasive or cajoling hortative, marked by ne (200): (198)

cu pa-htyj 1INCL 1- strong 'Let's be strong.'

(199)

ha cu me po ATTN 1INCL PL deer 'Hey, let's kill a deer.'

(200)

ha cu ne ATTN 1INCL PERSUASIVE 'Hey, let's go, okay?'

cura kill

The negative form of the hortative follows the same pattern as the negative imperative form, i.e. the negator nare follows the verb. (201)

cu rop to pajäpet 1INCL dog OBL.OM lINCL-startle 'Let's not startle the dog.'

(202)

cu pancrer 1BSICL 1INCL- sing 'Let's not sing.'

nare NEG

nare NEG

160

Jack and Jo Popjes

A positive response to the hortative takes the form:

(203)

ha cu ATTN 1INCL 'Let's go then.'

mer then

12 Negation Sentence negation is expressed by means of the negator nare. Negation sentences expressing imperative or hortative mood and negative responses to imperatives have distinct intonation patterns (see sect. 22). The negator occurs at the end of the sentence, that is, following the verb. (204)

cu-te po pupun nare 3- PAST deer see NEG 'He did not see a deer.'

The negator affects the form of the verb in that it may be preceded only by the long form (see sect. 23). Any constituent may be negated, subject, object, indirect object, verb or any of the peripheral constituents, and the negator usually follows the constituent to which it refers. Any word may be negated: verbs, nouns, including nominalized verbs, pronouns, aspect or manner markers, etc. When constituents which are marked by a postposition are negated, the negator most commonly follows the postposition: (205)

hyrmä nare to him NEG 'not to him, not in his direction'

(206)

kri kam nare village in NEG 'not in the village'

(207)

mehi nare Indian NEG 'not an Indian'

(208)

ihkanhoc to book INST 'not a teacher'

(209)

ahtüm nare later NEG 'short time (not a long time)'

hahkre cate nare teach NOMLZR NEG

CANELA KRAHO

161

Occasionally the negator precedes the postposition (214), when it appears to reflect a subtle distinction in meaning. In a sentence the non-verbal constituents are commonly negated by placing them clause initial and then negating the verb. Compare the normal order of the positive statement (210) with the corresponding negative statement in which the focus of the negation is on the instrument phrase (211): (210)

wa ha catöc to po 1 PUT gun INST deer will kill the deer with a gun.'

cura kill

(211)

catöc to wa ha po curan gun INST 1 PUT deer kill won't kill the deer with a gun.'

nare MEG

An element of the clause may be negated more than once with the result being positive. (212)

cu-mä amji kin nare nare 3- TEMPRY happy NEC NEG 'He is not unhappy (he is happy).'

(213)

ca ha güpar nare nare 2 PUT listen NEG NEG 'You will not not listen (you will listen).'

Strings of two or three phrases or nouns can be negated. (214)

i-te 1-PAST

catöc gun

me cuhe and bow

me ko and club

me wapo and knife

nare NEG

kam po curan OBL.OM deer kill didn't kill the deer with a gun, bow, club or knife (but I killed it).' To negate only the instrument but not the verb, the clause retains its normal order. The negator follows the instrument and precedes the postposition kam, which has the meaning 'instrumental' in (214) - (216). (215)

wa ha catöc nare kam po 1 PUT gun NEG OBL.OM deer will kill a deer without a gun.'

cura kill

162

Jack and Jo Popjes

This can also be negated entirely by negating the verb. (216)

wa ha catöc nare kam po 1 PUT gun NEC OBL.OM deer won't kill a deer without a gun.'

curan kill

nare NEG

There are several other more specific negators. There is a way of using nare in combination with the negator nee which negates the verb and carries the idea 'not even'. The nee follows the subject (see also sect. 8). (217)

capi nee pur wyr ih-tem Capi NEG field to 3- go 'Capi did not even go to the field.'

nare NEG

The negative response to an imperative or hortative is expressed by the negator wyr. It can be used to refer to any person.

(218)

wyr wa ha NEG 1 PUT won't do it.'

(219)

wyr que ha hane NEG 3 PUT thus 'He won't do that (like that).'

to do

To a suggestion that someone go out and kill a deer without using a gun, the response would be: (220)

wyr wa ha catöc nare kam po NEG 1 PUT gun NEG OBL.OM deer can't kill a deer without a gun.'

cura kill

To playfully contradict a statement of fact, the expression que, hoto pe is used. (221)

-I saw you in town yesterday, -que, höto pe! -not so!

To contradict a statement of fact by lying, or to contradict a wrong statement, the negator que occurs as the first element in the response. (222)

que, wa nee curi no 1 NEG there 'No, I didn't go there.'

i-tem 1-go

nare NEG

163

CANELA KRAHO

A class of verbs takes the particle no, which both negates and nominalizes the verb. Most frequently the augmentative ti or the diminutive re is suffixed to the particle no. (223)

hüpar ipicahur increr

- Obey' - 'run' - 'sing'

hüpar no ipicahur nöti increr nöre

- 'a disobedient person' - 'non-runner' - 'non-singer'

13 Anaphora Deletion is one means of anaphoric reference in Canela. Certain items may be omitted if the context carries the meaning. Third person morphemes are often omitted in this way (see sects. 3 and 23): (224)

cu-te wapo pyr ne 0-to 3- PAST knife take and 3-DSfST 'He took the knife and killed it with (it).'

ih-curan 3- kill

(225)

capi pur wyr te ne 0-kam häpen Capi field to go and 3-in work 'Capi went to the field and worked in (it).'

A system of personal pronouns is also used as a means of anaphoric reference. Complete lists of free and bound form are given in sect. 16. The reflexive and reciprocal pronouns amji and ajpen (described in sect. 4) are also a means of anaphoric reference. (226)

capi te amji Capi PAST self 'Capi hit himself.'

cakwin hit

(227)

jaco te po pupun ne ih-curan Jaco PAST deer see and 3- kill 'Jaco saw the deer and killed it.'

(228)

jaco te po pupun ne amji ma Jaco PAST deer see and self for 'Jaco saw the deer and killed it for himself.'

(229)

capi te po curan, que ha Capi PAST deer kill 3 PUT 'Capi killed the deer and will eat it.'

(230)

capi te po curan, ih-krer Capi PAST deer kill 3- eat 'Capi killed the deer for food.'

ih-curan 3 -kill

cuku 3+eat

xä caxuw NOMLZR PURP

Jack and Jo Popjes

A verb may be replaced by a dummy (pro-verb) or be deleted altogether, under conditions described in sect. 3. Location pronouns commonly occur with the postposition ri or mä indicating place. The free form he which frequently follows location pronouns indicates that the location is in sight of both the speaker and the hearer, and the speaker may even be pointing to it. curi - 'there, at previously stated location' hyrma - 'there, to a previously identified person' atari - 'there, further away' itari - 'here, closer by' ihmutri - Over there' ihmutxa he - Over there, look' All of the above means are used to express anaphora with the antecedent preceding. They may be used in any syntactically appropriate situation. (231)

capi Capi

te PAST

haren 3(J)+tell

jaco Jaco que so that

pupun see pahhi chief

ne cute and 3(C)-PAST te PAST

ih3(J)-

pahhi chief ken bad

mä to

mä OBL.OM

men pram te harkwa kot throw want because 3(C)+word after 'Because Capi wanted the chief to make trouble for him (Jaco) he spied on Jaco and told the chief.' Capi as the subject is referred to by cu in the second clause and zero in the displaced adjunct at the end of the sentence. Jaco as object is referred to by zero in the second clause and ih in the third. 14 Subordinate clauses Subordinate clauses are marked by postpositions and one free form word, and by a subordinate clause intonation pattern (sect. 22). The postpositions na and to are commonly used as general subordinators; te 'because' occurs mainly with the verb pram 'want'. (232)

wa i-xä na apu i-jujaher 1 1-sick SUBORD CONT 1-hunt went hunting when (while) I was sick.'

to mo SUBORD go

CANELA KRAHO

(233)

jaco Jaco

me capi and Capi

te PAST

165

pi here wood twig

jakep cut

ame 3PL

to INST

ajpen cahhyr pram te RECIP beat want because 'Jaco and Capi cut twigs because they wanted to beat each other with them.' The word incwyrjape 'reason', can be used with two distinct intonation patterns (see sect. 22): in one, it is phonologically a part of the preceding clause and functions like a subordinator; and in the other it is phonologically bound to the clause which follows, functioning more like a sentence connective. (234)

i-mä hüpa incwyrjape 1-TEMPRY 3+fear reason feared it, therefore I ran.'

i-picahur 1-run

Skewing occurs between syntax and semantics. What is syntactically the main verb is often semantically the subordinate verb and vice versa. This occurs most often when a syntactically main verb is a verb of motion. (235)

cu-te pi jakep to mo 3- PAST wood cut SUBORD go 'He went along cutting wood.'

The syntactically subordinate verb jakep 'cut' is semantically the main verb, whereas the syntactically main verb mo 'go' is semantically the subordinate verb. In indirect speech quotation the postposition na occurs most commonly. It occurs as the final element of the indirect quotation. Indirect statement: (236)

cu-te 3 -PAST

i-mä 1-TEMPRY

amji self

na SUBORD 'He told me that he likes you.'

jaren, told

cu-mä 3- TEMP

a-kin 2-like

166

Jack and Jo Popjes

Indirect question: (237)

a-mä 2-TEMPRY

jüri where

i-tem 1-go

xä ita NOMLZR this

na i-jahkre SUBORD 1 -learn

well 'You know where I am going.' Indirect commands have the same syntactic form as causatives (see sect. 6). The postposition na occurs most frequently; the purpose word caxuw also occurs. (238)

cu-te a-mör na a-mä 3- PAST 2-go SUBORD 2-OBL.OM 'He ordered you to go.'

harkwa order

(239)

cu-te a-mör xä caxuw a-rnan 3- PAST 2-go NOMLZR PURP 2-command 'He commanded you to go.'

(240)

cu-te a-mör xä na a-män 3- PAST 2-go NOMLZR SUBORD 2-command 'He commanded you to go.'

Complement clauses occur with the verb of the subordinate clause in the long form and marked with na. (241)

wa jum pupu i-picahur 1 someone see 3 -run see someone running.'

na SUBORD

Direct quote complements are marked by the word hajyr/hane 'thus', which occurs following the quote, from which it is marked off intonationally. The quote is preceded by the introductory quotation formula, which is a complete transitive clause in itself. (242)

cu-te 3- PAST

i-mä 1-to

harkwa word

ton, i-mä make 1-TEMPRY

hane thus 'He said to me, "I like you." ' The verb of the quotation formula can be omitted.

a-kin, 2-like

CANELA-KRAHO

(243)

cu-te i-mä, i-mä 3- PAST 1-to 1-TEMPRY 'He said to me, "I like you." '

a-kin, 2-like

167

hane thus

Direct quotes are much more common than indirect quotes. SYNTAX OF PHRASE TYPES 15 Noun phrase structure 15.1 Marking for case. There are several morphological ways for distinguishing the syntactic and semantic functions of noun phrases. Peripheral clause constituents with noun phrases have one of the set of postpositions described in sect. 17, representing various semantic distinctions. Certain occurrences of subject also require specific postpositions. The subject of verbs denoting feelings such as fear, hunger, attraction, lust, cold, etc. is marked by one of two postpositions: mä indicates 'temporary state' and te 'habitual state'.

(244)

i-mä 1-TEMPRY 'I'm cold.'

(245)

rop ita mä hüpa dog this TEMPRY 3+fear 'This dog is afraid of it.'

(246)

cu-mä a-kin 3- TEMPRY 2-like 'He likes you.'

(247)

cu-te pram 3- HAB hungry 'He is habitually hungry.'

(248)

capi te ho ita xen Capi HAB food this lust 'Capi lusts for this food habitually.'

(249)

i-te hüpa 1-HAB 3+fear live afraid of it.'

kry cold

The subject in identificational clauses is marked with the postposition pe, which seems to function as a copula.

168

Jack and Jo Popjes

(250)

i-pe mehi 1-COP Indian 'I'm an Indian.'

(251)

humre ita pe hah-kre man this COP ear-hole "This man is ignorant.'

keatre badly

The subject of transitive verbs is marked by te 'PAST' when the action is in the past (see sect. 1). (252)

i-te cu-mä hör 1-PAST 3- to 3+give 4 1 gave it to him.'

(253)

a-te catöc to po curan 2-PAST gun INST deer kill 'You killed the deer with a gun.'

(254)

cu-te i-pe pöhy jähkij 3- PAST 1-from corn steal 'He stole the corn from me.'

Elsewhere, the subject is not morphologically marked, and the direct object is also not marked. The syntactic function of noun phrases is distinguished primarily by word order (sect. 1). 15.2 Genitives. Nouns are divided into two classes: a) Inalienably possessed nouns occur with obligatory possessor person prefixes or free form nominal possessors. Most of these nouns denote body parts or kinship relationships. For possessor prefix forms, see sect. 16. In inalienably possessed noun phrases, the possessor is followed directly by the possessed.

(255)

i-to 1-eye 'my eye'

(256)

capi to Capi eye 'Capi's eye'

(257)

me panPL 1INCLOur forefathers'

quetje forefathers

CANELA-KRAHO

(258)

a-quetti 2-uncle 'your uncle'

(259)

pahkra 1INCL- child Our child'

(260)

capi kra Capi child 'Capi's child'

(261)

in-to 3- eye 'his eye'

169

b) Alienably possessed nouns can occur without a possessor. (262)

pi ton pry

'wood' 'armadillo' 'path'

In alienably possessed noun phrases, the possessor is followed by the possession marker ho I jo, which is followed by the possessed item. (263)

i-jö 1-POSSN 'my knife'

wapo knife

(264)

capi jo Capi POSSN 'Capi's field'

(265)

hö pur 3+POSSN field 'his field'

pur field

Non-specific possession of inalienably possessed nouns is not common, but may be shown by the use of the first person inclusive prefix pa, or the plural form me pa, or with the plural form alone: me, meh, mem, men. (266)

pan-to Our eyes, eyes, one's eyes' me pan-to jaxajre 'eye diseases' men to jaxajre 'eye diseases'

170

Jack and Jo Popjes

Non-specific possession of a few inalienably possessed nouns is shown by the non-specific indefinite third person prefix ah: ahkrare 'children', cf. ihkrare 'his children' 15.3 Modifiers. Nouns may be modified as follows: a) with an adjective, which is an adjectival verb stem.

(267)

rop pej dog good 'good dog'

(268)

rop no dog another 'a dog (one of a number of dogs)'

b) with demonstratives ita 'this', itaje 'these', ata 'that', ataje 'those'.

(269)

rop ita dog this 'this dog'

(270)

hümre ata-je man that-PL 'those men'

c) with specifiers, where both the specifier and the specified are nouns. The specifier is followed by the specified item, which is a more generic term. (271)

prin par piqui tree ' "piqui" tree'

(272)

crow kd burity grove ' "burity" grove'

There is a subtype of the specifier noun phrase in which the specifier is linked to the specified item by the postposition te. The relationship to the specified item is to specify the material out of which the specified item is made. (273)

carec te clay SPECFR 'clay animals'

pry re animals

CANELA-KRAHO

171

d) with the particle xwy 'deceased'. (274)

a-quet xwy 2-uncle DECEASED 'your late uncle'

e) with the particle kwyj 'female name'. (275)

pyt kwyj sun FEM 'Sun girl (i.e. girl named Sun)'

f) with a relative clause, in which the head noun either occurs within the relative clause (276) or precedes it (277). The relative clause may be extraposed, as in (277). The relative clause is signalled by an obligatory demonstrative (ita, ata, etc.) occurring after it. This demonstrative functions as a relative pronoun corresponding to the head noun. There are no restrictions on the syntactic function of the head noun. (276)

i-te humre te rop curan 1-PAST man PAST dog kill Ί saw the man who killed the dog.'

(277)

wa i-te rop pupun, 1 1-PAST dog see Ί saw the dog Capi killed.'

(278)

i-te humre pe rop curan 1-PAST man MAL dog kill Ί saw the man whose dog I killed.'

(279)

i-te h mre m rop curan ata pupun 1-PAST man BEN dog kill DEM see Ί saw the man for whom I killed the dog.'

capi Capi

ita pupun DEM see

te PAST

ih-curan 3- kill

ata pupun DEM see

The preferred ordering of modifiers is as follows: 1st order (these could cooccur): xwy kwyj

'deceased marker' 'female name'

2nd order - adjective (pej, tyj, etc.), or relative clause 3rd order - demonstrative (ita, ata, etc.)

ata DEM

172

Jack and Jo Popjes

Other than first order modifiers, not more than one of each type can cooccur. A noun or coordinate noun may be followed by up to three modifiers of different types. The usual limit is two modifiers. (280)

pe a-quet xwy pejti ata gare DP 2-uncle DECEASED good- very DEM 2+tell 'Your late, very good uncle, that one, told about you.' (gare is a contraction of a-jare '2-tell'.)

(281)

rop tyc ataje dog black those 'those black dogs'

(282)

i-te h mre xwy tyj 1-PAST man DECEASED strong Ί saw this late strong man.'

(283)

i-te h mre xwy pe rop curan 1-PAST man DECEASED DP dog kill Ί saw the late man whose dog I killed.'

ita pupun DEM see ita pupun DEM see

15.4 Nominalizations. Verbs are nominalized in several ways: a) with the nominalizer χα, which denotes a thing, place or an event. (284)

i-cator 1-arrived Ί arrived.'

(285)

i-cator 1-arrived 'my arrival'

(286)

wa apu tep pro 1 CONT fish catch 'I'm catching fish.'

(287)

tep pro x fish catch NOMLZR 'fish net, or fishing place'

x NOMLZR

b) with the nominalizer cate (plural cateje), which denotes the agent/doer of the action. This is usually human, but occasionally refers to an animal when such is in the agent role.

CANELA-KRAHO

(288)

i-te pryti jamar 1-PAST cows look after looked after cows.'

(289)

pryti jamar cate cows look after NOMLZR One who looks after cows (cowboy)'

(290)

i-picahur 3-run 'runner'

173

cate NOMLZR

c) with the 'size' suffixes re 'diminutive' and ti 'augmentative'. These may nominalize any verb phrase which denotes a person; re occurs more frequently than ti. (291)

im-prar tyj 3- run strong/well 'He runs well.'

(292)

im-prar tyjre 3- run strongDIMES 'He's a good runner.'

(293)

im-prar tyjre ita ma te ne 3- run strong- DIMIN DEM away go and This good runner goes away and kills a deer.'

(294)

in-to cara 3- eye widen (in fear) 'His eye widens (in fear).'

(295)

in-to carare 3- eye widen- DIMIN 'a person who is afraid'

po deer

cura kill

d) with the negative nominalizer no which nominalizes a certain class of verbs denoting a person while at the same time negating the meaning of the verb. (296)

hüpar ipicahur

(See also sect. 12).

Obey' 'run'

hüpar no ipicahur no

'disobedient person' 'non-runner'

174

Jack and Jo Popjes

e) with the postposition xi which denotes a person who is an expert or one who habitually practices the action of the verb. (297)

apu a-hej CONT 2-lie 'You are lying.'

cf.

(298)

a-hej xi 2-lie NOMLZR 'You are a liar.'

(299)

a-ken 2-bad 'You are bad.'

cf.

(300)

a-ken xi 2-bad NOMLZR 'You are a bad person'

Two classes of verb have an agentive nominalized form which does not correspond to any of the above ways of nominalizing. The verb in its long form and with a third person prefix may have a nominalized meaning, as well as its normal active verb meaning. It then carries the meaning of an act done by the agent. (301)

capi te häpen Capi PAST work 'Capi worked.'

(302)

capi ma häpen wyr te Capi away work to go 'Capi goes to his work.'

(303)

cu-te hahkre 3- PAST plant 'He planted.'

(304)

ipicxar apicahur

cf.

hahkre

'his planting'

'He laughed' or 'his laughing' 'You ran' or 'your running'

Two nominalizations may occur on the same verb. The meaning is carried by the last one to occur. (305)

ihhej ihhej xi xa ihhej xi

'He lies.' 'his lie' 'He is a liar.'

16 Pronoun system The pronoun system has both bound and free definite personal pronouns. There are also indefinite personal pronouns which differentiate between specific, non-specific, negative indefinite and plural. There are also demonstrative, reflexive and reciprocal pronouns.

CANELA-KRAHO

175

16.1 Personal pronouns. The same bound prefix forms occur for verbal object and subject markers, and for nominal possessors. There are no gender or class distinctions, but there is a distinction between first person inclusive and exclusive. The system is as follows: ιpa-, pah-, pam-, pan-

Ί exclusive' Ί inclusive'

a-

'2'

ih-, in-, i-, cu-, 0-

'3'

Any prefix on verbs or relators can be pluralized by the plural free form me or one of its variants (sect. 18.4). (306)

i-te capi 1-PAST Capi Ί saw Capi.'

pupun see

(307)

capi te i-pupun Capi PAST 1-see 'Capi saw me.'

(308)

capi te i-m h ren Capi PAST 1-to 3+told 'Capi told it to me.'

(309)

capi te me i-pupun Capi PAST PL 1-see 'Capi saw us (exclusive).'

There is also a set of free form pronouns which may occur as subject: wa pa cu ca que ta je (310)

Ί exclusive, unmarked' Ί exclusive, emphatic' Ί inclusive' '2' '3' '3 emphatic' '2 or 3, relative'

ha cu je ne po no cura hey 1INCL relative and deer a kill 'Hey, relative, let's go and kill a deer.'

176

(311)

Jack and Jo Popjes

i-te je me po 1-PAST relative and deer 'My relative and I killed a deer.'

no curan a kill

These free forms are not obligatory, but occur in certain discourse situations which have not been fully studied yet. It appears they may be omitted when the subject is clear from the immediate context, and marked by a prefix on the tense postposition, verb or relator. (312)

(wa) i-te 1 1-PAST Ί saw a deer.'

po deer

pupun (optional) see

(313)

wa po pupu (obligatory) 1 deer see Ί see a deer.'

The emphatic forms of the subject pronoun cooccur with the unmarked forms. The emphasis is commonly contrastive. (314)

wa ha po cura 1 PUT deer kill Ί will kill a deer.'

(315)

pa, wa ha po cura 1 1 PUT deer kill 7 will kill a deer (emphatic).'

(316)

wa ha pa po cura 1 PUT 1 deer kill 7 will kill a deer (emphatic).'

The following types of indefinite pronoun occur: - specific indefinites: the free forms jurn 'someone' and ampo 'something'. These forms are also used as interrogative pronouns, with the meanings 'who' and 'what' respectively (sect. 10).

(317)

j m j pen someone work 'someone's work'

(318)

i-te ampo 1-PAST something Ί saw something.'

x NOMLZR

pupun see

CANELA-KRAHO

177

- non-specific indefinite: the 1st person inclusive pa-: (319)

pa-nto pa-xä pan-to xa

Our eyes, one's eyes, eyes' Our sickness/disease' 'eye diseases'

- negative indefinite: the negator nee jum...nare: (320)

nee jurn te NEC someone PAST 'No one killed a deer.'

po deer

curan nare kill NEC

The plural free form me with the third person automatically refers to animate beings in general, to human beings more specifically, and most specifically to Indians: ihcunea 'all of it' ken cunea 'all the stones' mehcunea (me ihcunea) 'all the Indians/people/animate beings' 16.2 Demonstrative pronouns: ita itaje

'this' 'these'

ata ataje

'that' 'those'

They may be used both independently and adjectivally: (321)

ita pi DEM wood This is wood.'

(322)

pi ita wood DEM 'this wood'

(323)

ataje ken DEM stone 'Those are stones.'

(324)

ken ataje stone DEM 'those stones'

These demonstrative pronouns are also used to indicate previous reference in discourse. In this case the demonstrative pronoun is used in conjunction with some sort of relator. The following examples refer to previous statements in the discourse: ita na ita cwyrjape

'in reference to this' 'because of this'

178

(325)

Jack and Jo Popjes

capi te häpen cati, ne cute amji mä pur catia nor. ita cwyrjape que ha cumä höa to ho. 'Capi worked hard and put in a huge field for himself. Because of this he will have plenty of food.'

16.3 Reflexive and reciprocal pronouns

(see sect. 4):

amji and ajpen respectively 16.4 Interrogative pronouns jam jümje ampo

(see sect. 10):

'who, singular' 'who, plural' 'what'

16.5 Relative pronouns. described in sect. 15.3.

These take the form of demonstratives and are

17 Adpositional phrase structure The language is postpositional. Postpositions are not normally separated from the noun phrase although some rare irregular occurrences have been noted. They cannot occur as independent forms. Postpositions relate to people, things, actions and locations. Below they are listed in the third person inflected form, with a hyphen separating the third person prefix from the stem (0 indicates a zero third person marker): cu-ma cu-pe cu-ri cu-rum cu-mam cu-na cu-te ih-kot ihah-na 0-to 0-kam 0-wyr

'for, to, for his benefit' (benefactive) 'from, to his loss, to his negative benefit' (malefactive) 'there' (locative) 'toward there' (directional) 'before it' 'in front of it' 'he' (transitive past tense, or stative habitual) 'after it' 'from there' (locational source) 'about it' 'with it' 'in, into, at it' 'toward it'

CANEIA-KRAHÖ

179

Postpositional phrases can be modified by another postpositional phrase ((327 and 328)). (326)

wa ha capi mä ampo 1 PUT Capi to something will give something to Capi.'

go give

(327)

wa ha capi mä pit mä ampo 1 PUT Capi to only OBL.OM something will give something to Capi only.'

(328)

a-te wapo to pit mä 2-PAST knife INST only OBL.OM 'You killed the deer with only a knife.'

(329)

wa ha pur kam ape 1 PUT field in work will work in the field.'

(330)

wa ha curi ape 1 PUT there work will work there.'

(331)

jü pin ca te where from 2 go 'Where are you coming from?'

carä deer

go give curan kill

18 Verb and verb phrase structure See sect. 23 for some general morphological characteristics of verbs. 18.1 Tense.

Tense in Canela is marked as follows:

(1) Future tense is expressed by the morpheme subject (ha fluctuates between dialects with kra). (332)

ca ha a-mä 2 PUT 2-TEMPRY 'You will like him.'

ih-kin 3- like

(333)

que ha i-mä catöc 3 PUT 1-to gun will have a gun.' (literally, 'It, a gun will be to me.')

(334)

capi que ha rop cakwi Capi 3 PUT dog beat 'Capi will beat the dog.'

, which occurs following the

180

Jack and Jo Popjes

(2) Recent past is expressed in transitive clauses by the postposition te 'PAST', which follows a free form subject or has the subject person prefix attached to it, and which always has the long form of the verb cooccurring with it. In intransitive clauses recent past is only partially marked: when the verb is clause final, the long form of the verb occurs (except for a small sub-class of intransitive verbs, see sect. 1.4); when the verb is not clause final, the long form always occurs anyway, whatever tense is to be understood, so intransitive verbs then are not morphologically marked for tense (see sect. 23): (335)

i-te rop cakwin 1-PAST dog beat (long form) Ί beat the dog.'

(336)

capi j pir Capi climb (long form) 'Capi climbed.'

(3) The distant past tense is indicated by pe 'DP' and the short form of the verb (see sects. 9 and 23). (337)

pe wa rop cakwi DP 1 dog beat 'Long ago I beat the dog.'

In this paper, 'PAST' always refers to recent past and 'DP' to distant past. (4) Present tense is not morphologically marked. 18.2 Aspect. Aspect in a clause is expressed by certain verb forms and/or particles or postpositions. The aspectual distinctions in the language are as follows: (1) Continuous action: - apu 'continuing', a particle. (338)

apu i-cakoc CONT 1-speak Ί am speaking.'

- to mo 'to go along doing for the duration of a process'. The subordinative postposition to is preceded by an active verb and followed by the uninflected verb of motion mo.

CANELA-KRAHO

(339)

181

wa ρϊ jakep to mo 1 wood cut SUBORD go Ί go along cutting wood (as part of the process of preparing a field).'

Other ways of expressing continuous action, involving an active verb and postposition, followed by another verb, are: - to ipa 'to live doing the action'. (340)

wa kwyr kur to ipa 1 manioc eat SUBORD live Ί live eating manioc/always eat manioc.' - to incr n and kam incra 'to continue the action'.

(341)

a-cakoc to a-cra 2-speak SUBORD 2-continue 'You continue to speak.'

(342)

capi te rop kin kam in-cr Capi HAB dog like SUBORD 3- continue 'Capi continues to like the dog.'

(2) Completed action (all particles): - rarna 'already completed'. (343)

i-te rarna pi jakep 1-PAST already wood cut Ί already cut the wood.' - curm

(344)

i-te curm pi jakep 1-PAST now wood cut Ί just now cut the wood.' - corm

(345)

'just now completed'.

'still to be completed'.

wa ha corma pi jakep 1 PUT still wood cut Ί will still cut the wood.'

182

Jack and Jo Popjes

-par 'all'. (346)

i-te i-jö krer 1-PAST 1-food eat ate all my food.'

par all

- partu 'altogether, completely' (limited to transitive verbs). (347)

i-te i-jö krer partu 1-PAST 1-food eat completely ate all my food completely.'

(3) Habitual (particles or postpositions): - hän ne 'customary', and - hot pe 'customary'. (348)

xä hot pe a-mä waky Q CUST 2-to axe 'Do you usually have axes?' - te 'habitual, permanent' (with stative verbs only, see sect. 1).

(349)

cu-te rop kin 3- HAB dog like 'He is a dog lover.' - ajco 'used to, customary' (with distant past only).

(350)

pe wa ajco apu to hane DP 1 CUST CONT do thus always used to do that.'

(4) Iterative: - hiper 'repeated'. (351)

hiper i-mä hare REPET 1-to say Tell me again' or 'Say it again for me.'

CANELA-KRAHÖ

183

There are also verbs which have special forms for inherently repetitive aspect: single action ihcakwm har ihcuran har men

repeated action ihcahhyr incjej hipej incjen iren

'hit it' 'enter' 'kill it' 'put it in' 'throw it down'

Another class of verbs marks inherently repetitive aspect by duplication of the stem: ihhähäc ihpecpec capreprec ihpypym

'hiccup' 'drip' 'whip' 'fall repeatedly, stumble, several items one after another'

The other two aspects are expressed by inflected verbs. (5) Ingressive: - kam ihtem, kam te 'begin' (literally, kam 'SUBORD' and te 'go'). (352)

que ha ih-cakoc kam te 3 PUT 3- speak SUBORD go 'He will begin to speak.'

(353)

capi mä rop kin Capi TEMPRY dog like 'Capi began to like the dog.'

(6) Terminative: - hicu 'stop'. (354)

cu-te ih-cakoc 3- PAST 3- speak 'He stopped speaking.'

jicu stop

- hipej 'finish'. (355)

cu-te ih-cakoc 3- PAST 3- speak 'He finished speaking.'

jipej finish

te PAST

kam ihtem SUBORD 3+go

184

Jack and Jo Popjes

- ihhi 'stop'. (356)

i-te to ih-kin 1-PAST OBL.OM 3- like stopped liking him.'

hi stop

- krä cura(n) 'finish'. (357)

i-te i-jäpen krä curan 1-PAST 1-work finish finished my work.'

18.3 Mood. See also sect. 10 for interrogative and sect. 11 for imperative. Further study is needed for the understanding of some modal distinctions: conditional involves the use of the future tense marker, and debitive has to do with the tense marker te, the descriptive ihtyj and the norninalization marker xa, but these constructions have not yet been adequately analyzed, and are not discussed further here. Inflected verb forms express some modal types: - optative: pram 'want', caca 'not want, hate, dislike'. (358)

i-mä pöhy kre pram 1-TEMPRY corn plant want want to plant corn.'

(359)

wa pöhy kre caca 1 corn plant hate/dislike don't want/I hate to plant com.' - degree of certainty: tyj 'be certain'.

(360)

wa ha i-tyj ape 1 PUT 1-certain work will definitely work.'

Particles are more often used to express modal values: ma apu ahnaa awki märmä

'emphatic certainty' (commonly used with caca 'hate') 'confirmation' (usually with ita or as tahnaa, see below) 'doubt, possibility' 'maybe'

CANELA-KRAHÖ

xwymä ajki pijamä

185

'possibly might' 'possibly' 'possibly'

Some examples are: (361)

wa ma apu pöhy kre caca 1 EMPH.CERT com plant hate really hate to plant corn.'

(362)

märmä que ha ape maybe 3 PUT work 'Maybe he will work.'

(363)

xä ca ha pijamä ape Q 2 PUT possibly work 'Are you possibly going to work?'

Exclamative expressions which reflect modal values are: co co he ajco ampeaj ham tahnaa

'Careful' Take care! Look out!' 'Remember! (what I told you and be careful)' "That's right! Truly!' "That's right!'

18.4 Person and number. See sect. 16 for a listing of the person-marking prefix forms. Person prefixes on the verb agree with the subject (pseudotransitive, intransitive and adjectival verbs) or the object (transitive and stative transitive verbs). (See sect. 1 for other examples). (364)

capi apu ih-cakoc Capi CONT 3- speak 'Capi is speaking.'

(365)

capi te a-pupun Capi PAST 2-see 'Capi saw you.'

Number is sometimes expressed by the particle me (meh, mem, men) 'plural', usually where the referent is human and, more specifically, Indian (see sect. 16.1). The particle precedes the subject prefix in intransitive clauses and the object prefix or free object in transitive clauses. The context alone indicates whether the subject or the object is being pluralized.

186

Jack and Jo Popjes

(366)

rop kwy te cukoj cämxar dog GRP PAST monkey bite 'The dog pack bit the monkey.'

(367)

hümre te rop kwy man PAST dog GRP "The man beat the dogs.'

(368)

hümre te me rop cahhyr man PAST PL dog beat The men beat the dog.'

cahhyr beat

Number is also expressed by numerals, quantifiers and demonstratives. These more clearly indicate to which constituent the plural form refers. pyxit inkre no

One' 'three, many' One of

ipijakrut kwy

'two' 'some, group'

The suffix je pluralizes certain nominalized nouns and the demonstratives: ataje 'those', itaje 'these'. ken itaje 'these stones' ihcaköc cateje Orators'

ken ita 'this stone' ihcaköc cate Orator'

18.5 Voice, valency (1) An adjectival or intransitive verb may be formed into a transitive or pseudo-transitive verb respectively, by means of the postposition to 'INST': - haka (adjectival) 'it is white'. (369)

i-te to haka 1 -PAST 3 + INST 3 + white whitened it with (it).'

(370)

i-te to a-jaka 1-PAST 3+INST 2-white whitened you with (it).'

(371)

i-te haka xä to ikre jaka 1-PAST white NOMLZR INST house white whitened the house with whitener (paint).'

CANELA-KRAHO

187

The above three examples indicate that the use of to as a transitivizer with adjectival verbs results in a fully transitive clause (te 'PAST' cooccurs with object-marking prefixes, see sect. 1.1). - irit (intransitive) 'he looked'. (372)

cu-te to 3-PAST 3+INST 'He looked with it.'

i-rit 3-look

- atem (intransitive) 'you went'. (373)

a-te penhoc cucwyr- ti to a-tem 2-PAST rubber wheeled-big INST 2-go 'You went in the rubber wheeled thing/truck.'

Examples (372) and (373) show that to with intransitive verbs results in a pseudo-transitive clause (te 'PAST' cooccurs with subject agreement prefixes, see sect. 1.3). (2) There is a sub-class of verbs which have stems that are inherently neutral and which may be formed into either transitive or intransitive stems through morphophonological processes. This is explained in sect. 23. 18.6 Modifier. Adjectival verbs occur in their uninflected forms as modifiers of other (inflected) verbs (see sect. 15.3 for a similar modifier function in noun phrases): ihcuran pej intoj pej ihkin hi ihkrer jicu

'(someone) killed it well' 'he jumped well' 'he stopped liking' 'he stopped eating'

19 Adjective phrase structure The possible combinations of modifier sequences in noun phrases are discussed in sect. 15.3. Both adjectives (in noun phrases) and adjectival verbs can be followed by the suffixes // 'augmentative' and re 'diminutive'. The suffixes primarily modify the adjective/verb but in the case of re the noun is often modified as well. (374)

rop ita im-pej- ti dog DEM 3- good-AUG 'This dog is very good.'

188

Jack and Jo Popjes

(375)

wapo ita im-pej- re knife DEM 3- good-DIMIN "This knife is a good little one.'

(376)

rop pej- ti ih-tyc dog good-AUG 3- dead The very good dog is dead.'

20 Adverb phrase structure There are no phrasal constructions with adverb heads. 21 PARTICLES There are no particles other than conjunctions (sect. 8) interrogative markers and response words (sect. 10), and negation words (sect. 12). 22 PHONOLOGY The phonemes of Canela are described and given in the standard phonetic representations of Pike (1947). The practical orthographic symbols used in this paper are given in parentheses following the phonetic symbol. For a fuller description, see Popjes and Popjes (1971). Consonants There are three voiceless unaspirated stops at the following positions: labial - p (p) alveolar - t (t) velar - k (c,qu) These all have voiced allophones occurring syllable initial in unstressed syllables following voicing, and syllable final preceding voicing. There are two voiceless aspirated stops at the following positions: alveolar - ts (x) (affricate) velar - k" (k) There are three voiced nasal occlusives at the following positions: labial - m (m) alveolar - n (n) velar - η (g)

CANELA-KRAH

189

There are four non-occlusives at the following positions: labial - v (w) alveolar - 1 (r) palatal - y (j) velar - h (h) l (r) is a voiced alveolar lateral which has a flap allophone which occurs intervocalic, utterance initial and following consonants; y (j) is a voiced frictionless palatal continuant which has a fronted, alveolar grooved fricative allophone [z] which occurs final in consonant clusters and initial in stressed syllables; and h (h) is a voiceless frictionless continuant which has two allophones, a front velar fricative [x] which occurs before high oral vowels and a glottal stop [9] which occurs syllable final preceding consonants. Vowels All vowels are voiced. There are five front unrounded vowels (three oral and two nasal) at the following positions: i (i) high close e (e) mid close ε (e) low open ϊ (ϊ) high close έ (e) low open

oral oral oral nasalized nasalized

There are five back unrounded vowels at the following positions: ϊ e έ ϊ έ

(y) high close oral (y) mid close oral ( ) low open oral (y) high close nasalized ( ) low open nasalized

There are five back rounded vowels at the following positions: u (u) high close oral ο (ό) mid close oral ο (o) low open oral ( ) high close nasalized 5 (o) low open nasalized

190

Jack and Jo Popjes

There are two low open central rounded vowels: a (a) oral ä (ä) nasalized

Syllable patterns Syllable patterns are as follows: V, VC, CW, CVC, CCV, CCW, CCVC. Stress falls on the final syllable of nouns and verbs (and most other words) in isolation.

Vocalic release of consonants All consonants, except kft(k), ts(x), rj(g) and h(h) have a vocalic release utterance final, and in some dialects also utterance medial. The occurrence and strength of this release varies and fluctuates freely between dialects. The quality and occurrence of the release are predictable: v(w) and y(j) syllable final always occur with release, the quality of which is as follows: v(w) has a low central vowel quality [a]; y(j) has a high front vowel quality [i], except when the y(j) follows i in which case its release has a low central vowel quality [a]. All other consonants have a release which has the quality of the preceding vowel, except when the vowel is a, in which case it releases into a high front vowel, [i]. None of these releases are written in the practical orthography. Long vowels contrast with short vowels, but their occurrence is fairly rare, and in the practical orthography vowel length is written only when it is the only distinguishing feature between two otherwise identical words: ma 'benefactive' and mää 'rhea' caxwa 'night' and caaxwa 'salt' pe 'DP' and pee 'until' te 'PAST' and tee 'in vain'

Intonation patterns Some common intonation patterns are illustrated below.

Declarative: (377)

a-jte po cuifan 'You killed the deer.'

Polar interrogative: (378)

xä a-te po curanV'Did you kill the deer?'

CANELA-KRAHO

WH-Interrogative: (379)

jüm mäitaVWho is that?'

(380)

ampo ca ha kre\^What will you eat?'

(381)

jüm jö rop te mä a-xar\'Whose dog bit you?' ^—'l

Imperative: (382)

po_afa 'Kill the deer!'

Negative imperative/hortative: (383)

rop to a-japetjnare 'Don't startle the dog!'

Cajoling hortative: (384)

ha[cu(ne 'Hey, let's go, okay?'

Response to imperative (negative): (385)

wafkra 'No, I won't.'

(386)

que Iwahatonjnare 'No, I won't do it.'

Hortative: (387)

hajcu me po cura 'Hey, let's kill a deer.'

Subordinate: (388)

wa ijxä na ap[i i-jujaher|to mo 'When I was sick, I went hunting.'

Reason sentence (subordinate): (389)

i-mä hijpa incwyrjape i-picapur 'Because I feared it, I ran.'

192

Jack and Jo Popjes

Reason sentence (coordinate): (390)

i-mähöBä incwyrjafce i-picafrur feared it, therefore I ran.' 23 MORPHOLOGY

This section discusses some general morphological characteristics of verbs (see sect. 18 for treatment of particular verb and verb phrase categories). Nearly all verb stems have two morphemic shapes, a long form and a short form. A few verbs have only a single form, e.g. hitep 'cut', cucäc 'singe'. For those with two forms, the long occurs whenever the tense is (recent) past, and elsewhere whenever the verb stem is non-final in the verb phrase. (There is a small sub-class of intransitive verbs which have the long form in the past tense only when the stem is non-final, see sect. 1.4). The short form of the verb occurs elsewhere. Examples: long form

short form

ihcuran ton ihcahhyr cator

ihcura to ihcahhy cato

Examples of the long form are: (391)

ca a-te ton 2 2-PAST make/do 'You made/did it.'

(392)

a-mä ton pram 2-TEMPRY make want 'You want to make it.'

(393)

que ha ton pyrentu 3 PUT make immediately 'He will make it immediately.'

Examples of the short form are: (394)

pe ca to DP 2 make 'You made it long ago.'

'(someone) kills it' '(someone) makes it' '(someone) whips it' 'he arrives'

CANEIA-KRAHÖ

(395)

193

que ha to 3 PUT make 'He will make it.'

One set of transitive verbs has four forms with changes which occur stem initial. When the object is not specified in the clause, the normal long and short forms occur; with a specified object, a second set of long and short forms occurs. Unspecified Object long form short form hömpun hompu hüpa hüpa hüxwyr hüxwy hünar hüna

Specified Object long form short form pupun pupu cupa cupa pyxwyr pyxwy pynar pyna

'see it' 'fear it' 'set it up' 'follow it'

The following morphophonological rule operates throughout a major part of the language: j and occurring phrase medial become h phrase initial. (3%)

po xümre pjen xom

'male deer' 'grains of sand'

(397)

cu-te ampo japror 3-PAST something buy 'He bought something.'

(398)

ijö rop ijäpen

'my dog' worked'

humre horn

'male, man' 'kernels'

cu-te haprör 3-PAST 3+buy 'He bought it.' 'his dog' 'he worked'

ho rop häpen

There are three verb classes in Canela: (1) Verbs which are intrinsically transitive and cannot be detransitivized: ihcura(n) ihcakwi(n)

'kill it' 'hit it'

hompu(n) cake(n)

'see it' 'scratch it'

(2) Verbs which are intrinsically intransitive and cannot be transitivized: irit

'see'

ijyr

'sit'

194

Jack and Jo Popjes

(3) Verbs which are intrinsically neutral and have both transitive and intransitive forms: Neutral Root

japry jaher jaxwy japje japu

Intransitive

'name' 'chase' 'spill' 'track' Tight'

long form

short form

ipijapry ipijaher ipijaxwyr ipijapjer ipijapu

ajpry ajhe ajxwy Φ& ajpu

Transitive

hapry haher haxwy hapje hapu

The generalized rules for forming the actual verbs from the root are: add prefix ipi- for intransitive long form apply metathesis to initial syllable for intransitive short form apply the j -» h rule (see earlier in this section) for transitive form. Transitive verb roots are divided into five sub-classes on the basis of the form of the third person prefix: (i)

(zero) '3'

(399)

hihe 3 +close 'Close it.'

(400)

cakwi 3+beat 'Beat it.'

(u)

ih- '3'

(401)

ih-cura 3- kill 'Kill it.'

(402)

ih-cahhy 3- whip 'Whip it.'

CANELA-KRAHÖ

195

(iii) in-lim- '3' (phonologically predictable within the sub-class, impreceding p, in- elsewhere) (403)

cu-te in-xer 3- PAST 3-pinch 'He pinched her.'

(404)

im-pyn 3- carry 'He carried it.'

(iv) (405)

i- '3'

cu-te i- pir 3- PAST 3-choke 'He choked it.'

(v) CM- '3' - this occurs with one sub-set of roots only with the non-past tenses; (zero), im-, in- or ih- occur with past tense (406)

ca ha cu-xi 2 PUT 3- put down 'You will put it down.'

(407)

a- te hir 2- PAST 3 +put down 'You put it down.'

(408)

ca ha ken xi 2 PUT stone put down 'You will put the stone down.'

(409)

ca ha cu-py 2 PUT 3- carry 'You will carry it.'

(410)

a- te im-pyn 2-PAST 3- carry 'You carried it.'

Intransitive verb roots are divided into four sub-classes on the basis of the form of the third person prefix: (i)

(zero) '3' nor

'He lay down.'

1%

Jack and Jo Popjes

Some members of this class have different stem-initial syllables for the long and short forms: ha I a hapen 'He worked.' ape 'Work.' hah I ah

hahtep ahtep hü law hüjaher awjahe

'He approached.' 'Approach.' 'He hunted.' 'Hunt.'

The ha, hah, and hü forms of the stem occur in the long form of the verb.

(ii)

IÄ- '3' ih-tor

'It flew.'

Some members of this class do not have the third person prefix in the past tense when the verb is final in the clause: (411)

ihnö kam capi yesterday on Capi 'Capi went yesterday.'

(412)

ih-tem nare 3- go MEG 'He did not go.'

(iii)

(iv)

te go

in-lim- '3' in-crer 'He sang.' im-pat 'It broke.' - '3' i-rlt

'He saw.'

Some members of this class have different stem-initial syllables for the long and short forms: (413)

i- pintuw 3- change 'He changed.'

CANELA KRAHO

(414)

197

que ha antuw 3 PUT change 'He is going to change.'

Adjectival verb roots occur in adjectival clauses and some transitive clauses, and also as modifiers in noun phrases and verb phrases (sects. 15.3 and 18.6): (415)

rop ita ih-ken dog DEM 3- bad 'This dog is bad.'

(416)

i- te rop ita ken 1-PAST dog DEM bad did bad to the dog (hurt it).'

Adjectival verb roots are divided into four sub-classes on the basis of the occurrence of the third person prefix: (i)

(ii)

(zero) '3' catac häpre

'It is split.' 'It is wild.'

ih- '3' ihtyj

'He is strong.'

(iii)

in-lim- '3' in·crye 'He is angry.' im-pej 'He is good.'

(iv)

- '3' i-rä

'He is dirty.'

Stative verb roots occur in stative and stative transitive clauses. They have a single uninflected form (no person prefixes) in stative clauses, but are inflected with the object marking prefix in the transitive clause:

(417)

i-mä pa 1-TEMPRY afraid am afraid.'

(418)

i-mä hüpa 1-TEMPRY 3+fear am afraid of it.'

198

Jack and Jo Popjes

24 IDEOPHONES Ideophones are not common, occurring at a ratio of only one to every 100 clauses. They are of two types: (1) a small class, the members of which act as verbs semantically, some occurring in place of the verb, some in addition to the verb. (2) a larger class, the members of which have other more specific meanings, mainly of an exclamatory nature. There are no variations from the normal phonology, other than intonation. Examples of (1): cyt 'shoot with bow and arrow, the arrow glancing off the target' (419)

i-te cyt, cyt, cyt 1-PAST shoot shoot shoot Ί shot and shot (but missed).'

cru 'enter quickly' (420)

wa ikre m cator pe i-te cru 1 house to arrive DP 1-PAST quick Ί arrived at the house and entered quickly.' uu 'last a long time'

(421)

que ha uu ne hamre nare 3 PUT long and finish MEG 'It will be a long time before it is finished (it will last a long time).'

Examples of (2): hatititil Ouch!' (when hot only) an! 'wow!' (at bad happening) paw! 'wow!' (at large size or quantity) akyl Ouch!' (pain of cut, blow etc.)

ne i-xar and 1-enter

CANELA KRAHO

199

NOTES * The data for this description were gathered during field work in the years 1968-1977 under the auspices of the Summer Institute of Linguistics in accordance with its contracts with the Museu Nacional of Rio de Janeiro, the Universidade de Brasilia and the Fundac.ao Nacional do Indio of Brazil. We are indebted to Eunice Burgess of SIL for help in the analysis and presentation of this material at a workshop held in Belem, Para during October and November, 1981. Special thanks go also to our friend and colleague, Dr. William H. Crocker, with whom we have had many stimulating and helpful discussions about the Canela and their language, and to two excellent and patient Canela language informants, who developed over the years into teachers and friends, Raimundo Roberto Capertyc and his son-in-law, Luiz Carlos Jaco Hömpryxy. REFERENCES Crocker, W. H. 1961 "The Canela since Nimuendaju: A preliminary report on cultural change," Anthropological Quarterly 34:69-84. Hopper, P. J., and S. A. Thompson 1980 "Transitivity in grammar and discourse," Lg. 56:251-299. Nimuendaju, C. 1946 The Eastern Timbira, University of California Publications in American Archaeology and Ethnology, vol. 41 (Berkeley: University of California Press). Pike, K. L. 1947 Phonemics: A technique for reducing languages to writing (Ann Arbor: University of Michigan Press). Popjes, Jack and Jo Popjes 1971 Phonemic statement of Canela (Brasilia: Summer Institute of Linguistics ms).

PIRAHÄ Daniel Everett INTRODUCTION Pirahä is a member of the Mura language family, which also included the now (probably) extinct dialects of Bohura, Yahähi, Mura, and, possibly, Tora. The language has generally been referred to in Portuguese and English literature as Mura-Piraha. I have avoided use of this larger term here, since it tends to obscure the distinction between the language family and the language itself. This decision also reflects the preferences of most Brazilian linguists. Although tentative classifications have placed the Mura family in the Macro-Chibcha phylum, these seem to be poorly motivated; and I would, therefore prefer to reserve judgments on genetic affiliation, considering the Mura family to be still unclassified. Pirahä is spoken by approximately one hundred and ten individuals along the Maici River, in the state of Amazonas, Brazil. The people are almost completely monolingual; and though some have had frequent contact with outsiders, for the most part they are unacculturated into Brazilian society. Technologically, the Pirahä maintain an extremely primitive existence, using few tools in their day-to-day living, aside from the bow and arrow, crudely woven baskets, and aluminum pots purchased from outsiders. There are two primary settlements of Pirahä. These are separated by approximately one hundred and fifty kilometers (by river). The village where linguistic work has been done is slightly acculturated, being located close to the mouth of the river and thus seeing traders, rubber workers, hunters, etc. Unlike this group, the upriver village rejects most artefacts of outside culture and is generally hostile to foreigners. Both groups are able to make many artefacts and frequently do. However, the downriver settlement seems to prefer to get by on the barest necessities, begging from passing boats whenever possible. The research on which this description is based was conducted from January to March of 1979 and from April through December of 1980. Further work was carried on at various times (a total of about four months) with Pirahä informants outside the village. The average workday in the village was two hours of desk-elicitation, three hours filing and analysis, and three hours memorizing vocabulary and conversing with villagers in a number of settings— "perambulatory elicitation." In total, the present author has had approximately fourteen months of intensive contact with the Pirahä. However, various other researchers have contributed significantly to the description presented here. I want to thank Arlo and Vi Heinrichs, Steve and

PIRAHÄ

201

Linda Sheldon, and my wife, Keren, for their help on various sections of this paper. Keren Everett (henceforth KE) has contributed greatly to the analyses of verbal morphology and tone. Steve Sheldon (henceforth SS) has also helped a lot with the analysis of verbs and has contributed in one way or another in almost every section. None of these people are to be held responsible, however, for any errors which may have crept in despite their help. SYNTAX OF THE SENTENCE OR CLAUSE 1 Word order 1.0 Introduction. The basic clause types of Pirahä are transitive, intransitive, copular, and equative. These are distinguished from one another according to the potential occurrence of certain constituents in particular clause types. These constituents are in turn marked morphologically and/or through syntactic positioning within the clause. The unmarked, basic constituent order of Pirahä clauses is SOV (subject-direct object-verb). Peripheral constituents, such as indirect or oblique objects may be inserted between subject and direct object slots (cf. sects. 1.5 and 1.6 below). The primary criterion for distinguishing between the clause types mentioned (other than copular and equative, cf. sects. 1.3 and 1.4) is the appearance or potential appearance of the object constituent. 1.1 Transitive Clauses. All transitive clauses are distinguished by the (optional) occurrence of a direct object constituent, this being obligatorily absent in intransitive clauses. The unmarked order for transitive clause constituents is SOV. Evidently, classification by verb type is not a sufficient criterion for distinguishing among clause types. Compare, for example, 'kill' and 'die' in (7) below. (For a discussion of the orthographic conventions employed in this paper and their phonetic value in Pirahä, the reader is referred to the phonological summary in sect. 22.) (1)

ti

xibogi tibai 1 milk drink-INTNSF really drink milk.'

(2)

hi xäpiso xaho-aiihai 3 bark eat- ATELIC-PROX-RELATIVE CERT 'He will eat bark.'

(3)

hi käixihi xoab-aha 3 paca kill- REMOTE-COMPLETE CERT 'He killed a paca.'

202

Daniel Everett

(4)

xisäähai xaita xoho-aopa grasshopper leaf eat- TELIC-IMPERF-REMOTE The grasshopper ate (the) leaves.'

(5)

toipii Parintintin

hi xaooi 3 foreigner

xib-aobihit-TELIC-PERF- PROX-

hi COMPLETE CERT The Parintintin, he arrowed the foreigner.'

This order is considered basic for various reasons. First, given only the basic constituents of (l)-(5), any syntagmatic variations would produce drastic changes in the meaning or focus of the clauses. For example, in (6)

xibogi milk

ti ti- bai 1 drink-INTNSF

without a pause between xibogi 'milk' and ti T the meaning would be 'Milk drinks me.' With a pause between these words, xibogi would be interpreted as topic, resulting in a translation such as 'Milk, I drink a lot of.' (7)

kaixihi hi xoab-aha paca 3 kill- REMOTE-COMPLETE CERT (i) The paca killed it/him' or (ii) The paca died.'

In (7) (cf. (3) above) either (i) or (ii) is a possible translation. Let us remember that the verb xoab is translatable as 'kill' or 'die' depending on the number of arguments in the sentence (cf. above) and the context. Another argument in favor of the claim that SOV is the basic order is its high frequency of occurrence in relation to other possible configurations. A count based on transcribed texts reveals that SOV clauses compose approximately 90% of the total. Finally, supporting the above observations, any other constituent ordering is best understood as representing a marked form such as topicalization, emphasis, clarification, or other type of comprehension-aiding device (e.g., "Heavy Shift"— field want NOMLZR 'wanting a field'

xogai

sogisai

If both C/ and C2 are from the set b, g, h, k, or x, then V is a, as in: xi kob + hoaga —* it see + INCEPT '(He) started to see it.' 22.3.3.2.2 Deletion.

xi kob a hoaga

These observations are again taken from SS (1976):

s Deletion: Desiderative sog becomes og when preceded by a morpheme ending in 'a' as in: hi oa + sogabagai —» hi oa-og-abagai 3 delay DESID-FRUST. ΙΝΓΓ 'He almost wants to delay.' I believe these observations by SS to be essentially correct and have no further comments on them.

320

Daniel Everett

22.3.3.3 Syllable modifications in morphemic combinations. The remarks in this section are further exemplified and argued for in D. Everett (1981) and SS (1974). I will merely describe the grosser aspects of the phonological consequences of morphemic combinations in this section. In D. Everett (1981), it is suggested that syllable structure and stress placement operate syntagmatically without respect for morpheme boundaries, while tone is assigned paradigmatically, i.e., as a melody (or feature) of the word or morpheme. Its assignment to individual vowels is affected by changes in syllable structure and stress placement. The following observations would appear to support the first claim. These observations are limited, however, to noun 4- adjective or noun + verb sequences. For other sequences, see earlier sections (cf., for example, sects. 18 and 19). In sequences such as (471)-(475), the (a) examples show original syllable, stress, and tone configurations, and the (b) examples illustrate the modified forms. (Suffix divisions have been omitted to eliminate confusion with syllable structure—see sect. 18.) (47 la)

'soi CW skin

+

'baa.gi GW.GV sell

(471b)

so.'baa.gi CV .GW.GV

(472a)

si.'toi CV.CW egg

(472b)

si.to.'hoi CV .CV.CW

(473a)

si.po.'ai CV.CV.W feather

(473b)

si.po.'ai.gä CV.CV.W.GV

(474a)

si.'toi CV.CW egg

(474b)

si.'too.ga.ba.'gai CV .CW .GV .GV .G W

+

'sell skin' 'hoi CW two 'two eggs' + 'xi.ga CV.GV hard 'hard feather'

+ xo.ga.ba.'gai CV.GV.GV.GW want 'want an egg'

PIRAHÄ

321

(ö = one vowel manifesting two tones, low + high, simultaneously) (475a)

hoa.'xai CW.CW smoke

+

(475b)

'hoa.xi.oa.ba CW.CV.W.GV

xoa.ba CW.GV high 'high smoke'

The significant observation to be made with regard to the above examples is that high tone in word final accented syllables is moved in morphemic combinations. Moreover, this high tone actually appears to "follow" the accent (cf. below) (although, naturally, to support this hypothesis it would be necessary to (i) record many more such examples; and (ii) show through more detailed argumentation the relevance of this observation — is it relevant?; does it help capture significant generalizations?). When followed by a "perturbable" morpheme whose initial syllable is accented after restructuring, high tone is dislocated to the right (471) and (472). When the final syllable of the first morpheme is accented both prior to and after restructuring, any high tone originating there will continue in that syllable, regardless of its segmental composition (474). When, as in (475), word final accent moves leftward, any high tone originating in that syllable will also undergo left-dislocation, replacing the tone of the adjacent vowel. (476) below represents the class of words mentioned by SS (1974), namely, words which end in

In this class, the high or low tone on the penultimate vowel (a or o) extends rightward in the same syllable. To use a more theoretical terminology, the tone of the "strong" vowel replaces that of the "weak" vowel within the "rhyme" (cf. McCarthy (1979) and D. Everett (1981)). (476a)

sa.'hai CV.CW fat

+

ho.'äo.bä CV.W.GV give

(476b)

sa.'hai.ho.ao.ba 'give fat' CV.CW .CV.W.GV

322

Daniel Everett

23 MORPHOLOGY The basic elements of inflectional morphology are found in sects. 15 and 18. 23.1 Compound words. The criterion used to classify the examples to follow as compound words rather than merely phrasal constructions is semantic. For example, in (477) below, the syntagmeme xabagi soixaoxoisai may be understood as 'toucan beak' or 'saw1, according to the context. However, the majority of speakers who, for example, ask me for a saw (or other instrument with a compound name) find it very amusing and surprising when I make some sort of remark relating 'saws' and 'toucan beaks'. In my opinion, they are not even aware of the relationship unless they stop to reflect for a moment. Of course, this criterion is not a strong one. Therefore, the conclusions presented here are to be taken as tentative; i.e., it may be that some of these examples are simply phrases. 23.1.1 Nouns 23.1.1.1 Noun + noun: (477)

xabagi toucan

+

(478)

xapai foot

+ toii —> xapaitoii 'ladder' handle

(479)

hoii bow

(480)

xapai foot

+

soixaoxoisai —> xabagisoixaoxoisai beak

hoi —> hoiihoi vine

+

'saw'

'bowstring'

soi —> xapaisoi 'shoe' learner

23.1.1.2 Noun + verb. Compare sect. 15.4, where examples of new words formed by subject/agent nominalizations are given. 23.1.1.3 Noun + adjective: (481)

pi + gaia -*· pigaia 'scissors' thorn crooked

(482)

kao + ogiai —>· kaogiai [type of bass (fish)] mouth big

23.1.2 Verbs. Compare sect. 18.8 for a discussion of verb root incorporation. It is difficult to determine whether this process is synchronically productive. In other words, rather than being based exclusively on simple

PIRAHÄ

323

combinations of the sense of each root to produce a sense for the entire stem, the resultant forms often seem somewhat arbitrary in meaning and thus seem to be diachronically formed rather than resulting from synchronic rules. 23.2 Basic word classes. In sects. 19 and 20, we attempt to justify a collapsing of adjectives and adverbs into one class of modifiers. To sum up, from sects. 15-21, we conclude that the basic, morphosyntactically justifiable word classes in Pirahä are: nouns, pronouns, postpositions, particles, modifiers, and verbs. 24 IDEOPHONES I have not observed any onomatopoeic forms or other types of noninflected words similar to ideophones.

REFERENCES Chomsky, N. 1982 Some concepts and consequences of the theory of government and binding (Cambridge: Press). Comrie, B. 1976 Aspect (London: Cambridge University Press). 1981 Language universals and linguistic typology (Chicago: University of Chicago Press). Dana, H.E., and J.R. Mantey 1927 A manual grammar of the Greek New Testament (Toronto: The Macmillan Co.). Derbyshire, D. 1979 Hixkaryana (Amsterdam: North-Holland). Dooley, R. 1982 "Options in the pragmatic structuring of Guarani sentences," Lg. 58.307-331. Everett, D. 1979 Aspectos da fonologia do Pirahä (Universidade Estadual de Campinas: M.A. thesis). 1981 "Tom, acento e silabacäo," Anais V do G.E.L. Araraquara (Sao Paulo). 1982a "Some remarks on minimal pairs," Notes on Linguistics 22.24-30. (Dallas: Summer Institute of Linguistics). 1982b "Phonetic rarities in Pirahä," Journal of the International Phonetics Association 12:94-96

324

Daniel Everett

1983 A lingua Pirahä e a teoria da sintaxe (Universidade Estadual de Campinas: Ph.D. dissertation). 1984a "Clitic doubling and M-chains in Pirahä," Work Papers of the Summer Institute of Linguistics, University of North Dakota session 28:51-89. 1984b "Sociophonetic restrictions on subphonemic elements in Pirahä," Proceedings of the X International Congress of Phonetic Sciences, edited by A. Cohen and M.P.R. Van den Broecke (Amsterdam: Foris Books). To appear a "Referenda no Pirahä e a teoria de 'binding'," Anais do VII Encontro Nacional de Lingüistica (Rio do Janeiro: PUC-RJ). To appear b "Dialogue and the selection of data for a grammar," Dialogue: an interdisciplinary approach, edited by Marcelo Dascal (Amsterdam: John Benjamins). Everett, D. and K. Everett 1984 "On the relevance of syllable onsets to stress placement," LI 15:705-711. Everett, K. 1978 Phonological prerequisites in Pirahä (ms.). 1981 The semantics of Pirahä verbal affixes (ms.). Goldsmith, J. 1976 Autosegmental phonology (Bloomington: IULC). Greenberg, J.H. 1966 "Some universale of grammar with particular reference to the order of meaningful elements," Universals of language, edited by J. H. Greenberg (Cambridge M.I.T. Press). Grimes, J. 1981 Synthesis and feedback in field linguistics (101st meeting: Acoustical Society of America). Harbert, W. 1977 "Clause union and German accusative plus infinitive constructions", Syntax and Semantics 8, edited by P. Cole and J. Sadock (New York: Academic Press). Heinrichs, A.L. 1964 "Os fonemas do Mura-Piraha," Boletim do Museu Paraense Emttio Goeldi 21:1-9. Keenan, E. and B. Comrie 1977 "Noun phrase accessibility and universal grammar," LI 8.333-351. Lefebvre, C. 1980 "Cases of lexical complementizers in Cuzco Quechua and the theory of Comp", Journal of Linguistic Research l(no.2): 91-112. McCarthy, J. 1979 "On stress and syllabification," LI 10:443-465.

PIRAHÄ

325

Muysken, P. 1981 "Quechua word structure," Binding and filtering, edited by F. Heny (Cambridge: Press). Nimuendaju, K. 1948 "The Mura and Piraha," Bulletin 143, Handbook of South American Indians 3:255-269 (Washington: USA Government Printing Office). Pike, K. 1948 Tone languages (Ann Arbor: University of Michigan Press). 1949 Phonemics (Ann Arbor: University of Michigan Press). Searle, J. 1979 Expression and meaning (London: Cambridge University Press). Sheldon, L. 1976 Pedagogical grammar of Mura-Pirahä (ms.). Sheldon, S. 1973 Pirahä relational, a beginning attempt (ms.). 1974 "Some morphophonemic and tone perturbation rules in Mura-Pirahä," UAL 40:279-282. 1976 Pirahä verbal suffixes (ms.). 1977 Mura-Pirahä verbal suffixes (ms.).

URUBU-KAAPOR James Kakumasu INTRODUCTION Urubu-Kaapor belongs to the Tupi-Guarani language family and is spoken by approximately 500 Indians in the northeast corner of Brazil. They live on the tributaries of streams that flow into the rivers Gurupi, Maracac.ume, Paraua and Turiacu in the state of Maranhäo. There are approximately ten villages scattered from the river Gurupi to the river Turiac.u, an area covering some 2800 square miles. Urubu is the name by which the tribe is known to Brazilians. It is the Portuguese word meaning 'buzzard'. In anthropological journals, the tribe is often referred to as Urubu-Kaapor. The Indians refer to themselves as Aa'apor, which means 'forest-dwellers'. In this paper they will be called Urubu-Kaapor. Until the early 1970's, the Urubu-Kaapor Indians were isolated from the outside world. But in the early seventies, a highway was completed from Belem to Säo Luis which passed close to villages near the Paraua river. Because of the highway, the whole tribe has more contact than ever before with the outside world and, as a result, bilingualism is increasing, especially among the younger men. Before the highway, less than five per cent of the people were bilingual.

SYNTAX OF THE SENTENCE OR CLAUSE 1 Word order I shall refer to three clause types in Urubu-Kaapor, distinguished by the type of predicate they contain: declarative, stative, and equational. There is a subset of verbs I shall call statives, differing from declarative verbs in that they do not take the first and second person pronominal prefixes or the imperative prefix. A clause having a stative verb as its predicate is a stative clause. A clause with any other verbal predicate is a declarative clause. A clause with a nominal or adverbial predicate is an equational clause. Declarative clauses may be transitive or intransitive, these differing only in the obligatory absence of an object in the intransitive clause. Table 1 shows the differing features among the clause types.

URUBU-KAAPOR

327

Table 1: Differing features of clause types Declarative 1.

Stative

First and second person prefixes

+

-

2.

Imperative prefix

+

-

3.

Third person prefix

+

+

4.

Subject-verb agreement

+

+

5.

Predicate

declarative verb

stative verb

Equational

nominal or adverbial

1.1 Transitive clauses. I now consider SOV to be the basic word order in the transitive clause, not OSV as in Kakumasu (1976). In the least marked clauses which have both subject and object nominals, the most frequent order is SOV. Clauses in which either the S or Ο nominal is a pronoun, or is followed by ke OM^OC' or riki 'EMPH/FOC', are considered marked types, but some of these are also SOV (see sect. 9, and exs. (4) and (5)). In a corpus of over 3000 clauses, only 20 were unmarked transitive clauses, including (1) - (3): (1)

oropo hu'y pyhyk Oropo 3 +fever 3 +get 'Oropo came down with fever.'

(2)

ajame'e ke urupe je'e- ha after that Lupercio speak-NOMLZR 'After that Lupercio sent word also.'

(3)

pe kamarar jarusu ra'yr raho and friends canoe small 3+take 'And friends took over a small canoe.'

mondo ti 3 +send also

Examples of SOV clauses with marked object (4) or marked subject (5) are: (4)

a'erehe taru nexi ke jo'ok for that reason Taru Nexi OM 3+take 'For that reason Taru married Nexi.'

328

(5)

lames Kakumasu

karai riki muka mopok non-indian EMPH shotgun 3+shoot The non-indians shot off shotguns' or 'It was the non-indians that shot off shotguns.'

In addition, there were 22 SOV clauses containing first person free pronoun subject, including: (6)

ko ihe paper pinim thus I paper spotted Thus I sent a letter to you.'

amu- wyr ISG-CAUS-come

(7)

ihe nde ke a- raho tar katu I you OM ISO-take want well really want to take you (but can't).'

(8)

pe ihe y ajo'ok tipe i-pe and I water ISO-take FRUST 3-to 'And I got water for him (but he didn't drink).'

nde pe you to

tipe FRUST

The second most prevalent pattern for transitive clauses is OSV. There were only three examples in the corpus with both S and O nominals in their unmarked forms: (9)

kangwaruhu maneru paca Mane's father 'Mane's father killed a paca.'

juka 3+kill

(10)

ta'yr jawar ta su'u tipe his child dog PL 3 bite FRUST The dogs bit a small one (but didn't kill it).'

(11)

kawaru je manepäi we- rur ta horse HSY Mane's father 3bring FUT 'It is said that Mane's father will also bring a horse.'

ti also

In addition, there were 10 OSV clauses with the followed by ke or riki (in (13) there is one of the rare occurrences of ke with both the and S nominals, a phenomenon for which I still do not have any satisfying explanation): (12)

pe ke te'e kairarixä trail FOC freely Kairarixä 'Kairarixä was cleaning the trail.'

musorok 3+CAUS-space

u-'am 3-be

URUBU-KAAPOR

329

(13)

u'i ke wira'i ta ke manioc meal OM Wira'i PL FOC 'Wira'i folks didn't eat any manioc meal.'

(14)

pe and 'And 'And

tajahu riki wild pig EMPH the dogs chased the it was the wild pigs

u-'u 3-eat

ym NBG

jawar mahem dog 3+chase wild pigs' or the dogs chased.'

There were 26 OSV clauses with first person subject: (15)

ne ra'yr ihe a- pyhyk your son I ISO-take hold was holding your son.'

axo ISG-be

(16)

ihe räi ihe amoi ta tipe my teeth I ISO-place PUT FRUST wanted to put in my (false) teeth (but didn't).'

There were 27 clauses containing third person plural subject. The word nga 'people' is often used with the meaning 'they', so this group of clauses was separated from the group containing two nominals: (17)

tapi'ir nga u-sa- sak ta tipe tapir 3PL 3-see-see PUT FRUST They tried and tried to see the tapir (but didn't).'

(18)

pe riki katar ngä muhyk and so flu 3PL 3 +C AUS -arrive 'And so they brought the flu.'

A comprehension experiment was conducted to test basic word order. A sentence such as 'John Bill killed' was supplied, and the native speaker consultant was asked, "Who died?" The experiment was conducted with three men, two in their late twenties and the third in his late thirties. Two men caught on to what was being tested and their answers are given on the right in the examples (19)-(22). The man in his thirties is a literate man who has had experience in questions and answers and was one of the two who gave answers. The other man who answered is an illiterate with no experience in question and answer technique. The following sentences were given: (19)

haimü Raimundo

juä John

juka 3 +kill

(John died)

330

James Kakumasu

(20)

juä John

haimü Raimundo

(21)

jawaru black jaguar

(22)

jangwate jaguar

juka 3 -f-kill

(Raimundo died)

jangwate jaguar

juka 3 -f kill

(Jaguar died)

jawarü black jaguar

juka 3 +kill

(Black jaguar died)

After the experiment, the men's comments were that with the object marker ke the one who was killed is very clear but without ke it is not. (The men used the phrase ukwa ran 'he knows falsely' or One faintly understands'). This experiment seems to substantiate the results of the statistical study that SOV is the basic word order. The basic full order for a transitive clause is as follows: Connective Adjunct Subject Object Verb Closure Subject, object, and verb are considered nuclear constituents, while connective, adjunct, and closure are peripheral constituents. It is only rarely that all these constituents actually occur in a single clause, the usual number being four or five. (23)

ajame'e ke after that

ku'erahä next day

jande we

pupur boiled

ja- 'u IPL-eat

ja- xo ti 1PL- be again 'After that, the next day we were eating boiled (paca) again.'

(24)

pe and

taramö today

kaitä Caetano

ju'ipape turtle

pi'a egg

u-sak 3-see

o-ho 3-go

ti, pytuniwe again, early morning 'And today, early in the morning, Caetano went to look for turtle eggs.' All other logically possible orderings of nuclear constituents occur:

VSO: (25)

a'u ym ihe ISG-eat MEG I didn't eat a thing.'

ma'e something

ke FOC

URUBU-KAAPOR

(26)

jo'ok uhu te juwenda 3+take much truly Juvenal 'Juvenal bought a lot of things.'

331

ma'e thing

ke FOC

VOS: (27)

muhury katu 3+CAUS-be happy well "They cheered me up.'

ihe me

ke FOC

(28)

me'e hu te ma'e ke 3+give much truly thing FOC 'Long ago they really gave a lot.'

ngä 3PL ngä 3PL

kwe long ago

OVS: (29)

pahar pahar je kyse me'e ngä quickly quickly HSY knife 3+give 3PL 'It is said that they very quickly gave out knives.'

(30)

päsykwer riki matyr uhu ngä cloth EMPH 3+get much 3PL 'They got a lot of cloth for him.'

i-pe 3-to

SVO: (31)

karai ta hijar non-indian 3 leave 'Non-indians left knives.'

(32)

nde ere-'u ma'e you 2SG-eat thing 'Did you eat anything?'

1.2 Intransitive clauses. SV:

kyse knife

The intransitive clause has the basic word order

(33)

oropo kanim o-ho Oropo 3+be lost 3-go 'Oropo went to hide.'

(34)

arauxu ahem uhu Araujo 3 +shout much 'Araujo shouted a lot.'

332

(35)

James Kakumasu

ihe ramui u-sak o-ho my grandfather 3-see 3-go 'My forefathers went to see (it).'

Intransitive declarative clauses differ from transitive clauses only in the obligatory absence of the object. Otherwise, the basic ordering of both nuclear and peripheral constituents is the same: (36)

pe pay jywyr and Pay 3+return 'And Pay returned.'

(37)

ihe aho ta I ISG-go PUT Til go see.'

o-ho 3-go a- sak ISO-see

Word order may also vary in the intransitive clause and constituents can be fronted for emphasis. Where only nuclear constituents are involved, this means that the focus is on the action expressed by the verb: (38)

u-pe u-pe tamüi 3-stumble 3-stumble grandfather 'The old man stumbled and stumbled.'

(39)

je'e te'e 3 + speak freely 'Pay lied.'

ke mä FOC EXCLM

pay Pay

The fronted constituent may be further emphasized by the addition of a postposition such as riki 'EMPH' (40). Adjuncts (adverbiale and postpositional phrases) appear to be somewhat deemphasized when they follow the subject or verb (40, 41, 42): (40)

ko here

riki ihe EMPH I

ahyk ISO-arrive

a- jur ISG-come

rehe to arrived here at the port.' (41)

ka'aro ne rehe u-sak Ka'aro you on 3-see 'Ka'aro went to see you.'

o-ho 3-go

jarupa port

URUBU-KAAPOR

(42)

pe u-hyk o-ho kaninde and 3-arrive 3-go Caninde 'And he arrived at Caninde.'

333

rehe to

1.3 Stative clauses. The predicate of the stative clause contains adjectival-type verbs. In combination with a subject they function as verbs, but they can also modify a noun in a noun phrase: (43)

pira riki heta fish EMPH many There are lots of fish.'

(44)

pe so'o-nem pe tui and meat-rotten there 3+lie The rotten meat is lying there.'

(45)

tapi'ir i-nem tapir 3-rotten The tapir is rotten.'

The stative verb differs from the declarative verb in the form of the personal pronoun prefix. One subset of stative verbs occurs only with third person subject, and this is signalled by the prefix /'- (45). See sects. 15.3 and 18.4. Stative verbs of another subset have initial h- when the subject is third person and initial r- when the subject is first or second person. See also sect. 18.4. (46)

heta karai many non-indian There are many non-indians.'

(47)

jande reta ym we many MEG 'We are not many' or 'We are few.'

(48)

tamüi he'5 grandfather 3+be tired The old man is tired.'

(49)

ihe re'o I 1 +be tired am tired.'

334

James Kakumasu

Other stative verbs are unmarked as to subject agreement: (50)

ihe katu I good Ί am good.'

(51)

karai katu non-indian good 'Non-indians are good.'

(52)

mirixo riki ahy woman EMPH hurt The women were sick.'

(53)

u'i nixoi te h a'ep manioc meal (there is) none truly much there There wasn't any manioc meal there at all.'

Connectives, adjuncts and closure all occur in stative clauses: (54)

parana ita pupur rehe heta Gurupi river stone boil at many There are many (non-indians) at the Gurupi river rapids.'

(55)

pe and

y water

rymy'y edge

rehe at

riki kamarar EMPH Indian

ta PL

pirer skin

pira red 'And it was on the shore that the Indians' skin was red.' (56)

nixoi a'e ta ke ri none 3 PL FOC still They weren't (around) yet.'

1.4 Equational clauses. The equational clause nucleus consists of a subject and a predicate. The predicate may be a nominal or an adverb: (57)

ihe rakehar ym I wife MEG Ί don't have a wife.'

(58)

eha ym riki awa eye NEG EMPH people 'People were without sight.'

URUBU-KAAPOR

335

Peripheral constituents are less common with equational clauses, but connectives, adjuncts and closure may occur: (59)

a'erehe ko ihe aja ti for that reason here I thus also Tor that reason I am also in the same situation.'

(60)

oropo pewe ri Oropo there still Oropo is still there.'

1.5 Subordinate clauses. Subordinate clauses have fewer peripheral constituents than main clauses and the closure constituent does not occur at all. The subordinator rahä 'if, when' occurs after the verb. See also sect. 14. Typically, adverbial subordinate clauses precede main clauses, as in these examples: (61)

se here

a- jur ISG-come

we again

rahä, when

ihe I

a- jo'ok ISG-take

ta ky PUT FUT.DEF.PURP 'When I come again, I will get it for sure.'

(62)

upa aman u-kyr rahä, jande ja- ho end rain 3-fall when we IPL-go 'When it stopped raining, we went on in vain.'

tipe FRUST

(63)

i-pai 3-father

ke OM

ximbo poison

tuka 3+hit

rahä, when

ta'yn his child

u-'u 3-eat

raho 3 +take 'When his father was beating the poison, (the anaconda) ate his child.' (64)

kyha upa pukwar hammock end 3 -Hie

rahä, when

i-pai 3-father

pytym cigar

me'e- me'e 3 +give-give 'When he finished tying (his) hammock, his father distributed the cigars.' 2 Parataxis Phrases or nominal!zed clauses may follow the main clause to clarify the adjunct, object, or subject. The phrase/clause may or may not be contiguous

336

James Kakumasu

with the main clause constituent it clarifies. Phonologically, there is a dislocation in that the main clause has a falling intonation signifying closure. Syntactically, many clauses have the closure constituent, which is then followed by a phrase or nominalized clause. The clarifying constituent may be a noun phrase (65), nominalized clause (66), or adverbial phrase (67): (65)

karai riki katar jar, ma'e ahy jar non-indian EMPH flu owner some pain owner 'Non-indians are the possessors of the flu, of all diseases.'

(66)

ame'e ke te'öru that one Te'öru

japijapi 3+shoot- shoot

mä'y, kairarixä EXCLM Kairarixä

jingo me'e ke 3 +shoot NOMLZR Te'öru finally killed that one, the one Kairarixä had shot.' (67)

pe there

koty towards

u-'ar 3-fall

o-ho 3-go

ti, again

kairarixä Kairarixä

rake ti beside again On that side it fell again, beside Kairarixä.' Two or three phrases may be juxtaposed to clarify or make more specific one of the constituents: (68)

pe and

muhyk 3 +CAUS -arrive

raho, 3+take

h-ekoha 3-village

pe, to

i-anam koty 3-relative towards 'And he brought (it) to his village, to his relatives.' (69)

u-'u 3-eat

katu well

aja thus

je HSY

ma'e someone

pandu, 3+say

Xerupäi Xerupäi

pandu, kapitam pandu 3 +say captain 3 +say * "He is eating well," thus it is reported that someone said, it was Xerupäi, the chief, who said it.' 3

Ellipsis

3.1 Omission of nominate. In declarative clauses, ellipsis occurs regularly since either the subject or object may be inferred from the context. The verb

URUBU-KAAPOR

337

has a pronominal prefix which marks the subject, whether or not a free nominal occurs. In intransitive clauses, only the verb constituent is obligatory. (70)

u'y o-mbor uhu arrow 3-throw much They (my forefathers) shot a lot of arrows.'

(71)

kyse raho knife 3 +take 'He (the man) carried away the knife.'

(72)

kaninde rehe ja- hyk Caninde to IPL-arrive 'We arrived at Caninde.'

(73)

arauxu jo'ok Araujo 3 +take 'Araujo took it (a picture).'

(74)

ajame'e ke ihe a- raho we ti after that I ISO-take little again 'After that I took him (your son) again for a little while.'

(75)

pe raho and 3 +take 'And he (the man) carried it (the knife) away.'

(76)

pe je u-'u we te je and HSY 3-eat some really HSY 'And it is said that she (Kupi'ipe, a woman) really ate (some of the bananas).'

ja- ho IPL-go

In Stative clauses, the subject nominal may be omitted: (77)

upa je i-juk o-ho end HSY 3-rot 3-go 'It completely rotted away.'

The context of (77) is that a snake bit a man on the leg, which is the referent of i- '3'. (78)

pe mani'ok tui and manioc be there 'And the manioc lay there.'

338

(79)

James Kakumasu

i-membek o-u 3-soft 3-lie They were soft.'

In equational clauses, both the subject and predicate are obligatory: (80)

ihe ra'yr ym I child MEG Ί am without child.'

Third person goal ehe 'to him, for him' occurs following the main verb, without any nominal or pronominal head. First and second person and nominal goals take the form of postpositional phrases. See sect. 17. (81)

u-hyk ym ehe m 3-arrive NEG to him EXCLM 'He didn't find him' (literally, 'He didn't arrive to him.')

(82)

ere-sak te ehe 2SG-see truly to him 'You really saw him.'

(83)

pako ro rehe ihe a- ho banana leaf for I ISG-go Ί went for banana leaf again.'

(84)

kotete jande rehe aman u-hyk nearby us to rain 3-arrive 'Nearby it first rained on us.'

ti again py first

m EXCLM

3.2 Omission of verb. The verb pandu '3+say' may be omitted when it occurs to close a quotation. The normal closure for quotations is aja pandu 'thus he said', but speakers often close with only aja 'thus', or sometimes with aja riki 'thus EMPH'. See also the discussion of direct quotation complements in sect. 14. (85)

upa mumembek aja ihe pe pandu end 3+CAUS-be weak thus me to 3+say 4 "It is completely weak," thus he said to me.'

(86)

ani, aja riki no thus EMPH ' "No," thus (he said.)'

(87)

epu'am we, aja tipe ihe 2SG.ΕΜΡ-stand little thus FRUST I ' "Stand a bit," thus I (said) to him in vain.'

i-pe 3-to

URUBU-KAAPOR

339

In a text where there is conversation, the main verb may be omitted, but there occur forms or substitutes which give clues to the hearer for clear understanding of what is going on. Some forms that occur are aja 'thus', upa 'it is finished', '(someone) finished (something)', ere 'yes, okay', and am 'no'. (88a)

upa nde ere- mujä tamüi end you 2SG-make grandfather Old man, did you finish making (it)?'

(88b)

upa end 'It's finished.'

The full form corresponding to (88b) would be: (88c)

upa ihe amujä end I ISO-make finished making (it).'

(89a)

ihe-ma'e py amujä ta, my-thing first ISG-make FUT 'Mair said, "I'll make mine first".'

(89b)

ere, sarakur pandu, okay Saracur 3+say ' "Okay," said Saracur.'

mair Mair

pandu 3+say

ere okay

Compare the full form of response, corresponding to (89b): (89c)

ere, okay

ne- ma'e your-thing

emujä 2SG.IMF-make

py, first

aja thus

sarakur Saracur ' "Okay, you make yours first," thus Saracur (said).'

4

Reflexives and reciprocals Reflexivity is expressed by the prefix ju-. The result is an intransitive verb:

(90a)

pe je and HSY

so'oran rabbit

jangwate jaguar

ke pukwar OM 3+tie

mä'y EXCLM- CMPL 'And it is said that the rabbit tied the jaguar at that time.'

340

James Kakumasu

(90b)

ko ihe ajupukwar ATTN I ISG-REFLX-tie 'Look, I'm going to tie myself.'

ta PUT

(9 la)

jande ke a'e mu'e us OM 3SG 3+teach 'He was teaching us also.'

ti also

(91b)

makynda xirur machine pants

i-xo 3-be

mujä ha rehe make NOMLZR on

a'am ISG-be

kujä ta woman PL

jumu'e i-xo ti REFLX-teach 3-be also The women are also learning on the sewing machine.'

The reflexive form can be used with reciprocal meaning when prefixed to a reduplicated verb stem: (92)

jangwate ke rehe jutuka tuka katu jaguar FOC on 3REFLX-hit hit good 'He bumped and bumped himself on the jaguar' or They (he and the jaguar) bumped each other.'

te truly

The reflexive/reciprocal form ju- may cooccur with certain postpositions: jupe 'to himself (-pe 'to, towards, at'): (93)

kaninde Caninde

pe to

aho ta ISG-go PUT

aja thus

jupandu 3REFLX- say

jupe 3RFJLX-to ' "I'll go to Caninde," thus he said to himself or 'He decided to go to Caninde.'

juehe 'to/for himself (-ehe 'to, for'): (94)

i-juehe har upa je 3-REFLX- for NOMLZR end HSY

makak ta monkey PL

pandu 3+say

katu i-pe good 3-to 'It is said that the monkeys told everything about themselves to him.'

URUBU-KAAPOR

341

The construction ijuehe har in (94), meaning literally One for himself/themselves', carries the meaning 'about himself/themselves'; cf. ihe r ehe har 'about myself. Although the form jupandu jupe 'he said to himself occurs, the preferred construction is simply the direct quotation followed by pandu 'he said', without the reflexive. The English verbs 'decide', 'wish', 'want' and 'think' are expressed as direct speech followed by the quotation tag: (95)

tuti Tuti

namö with

nahä perhaps

o-ho 3-go

o-u- p 3-lie-LOC

my doubt

aja thus

nahä xa'e u-kwa my perhaps Xa'e 3-know doubt 'Perhaps she went to sleep with Tuti, perhaps Xa'e thought thus.' See sect. 3.2 for the use of aja 'thus' to signal the end of a direct quotation. 5

Passives Passive constructions do not occur.

6

Causatives

Causatives are formed by adding the prefix mu- to intransitive verbs and to some nouns. Only the semantics restricts causativization of lexical items in these categories. The resulting form is a transitive verb, and the third person pronominal prefixes always have a null realization (sect. 18.4): (96a)

pe jande ja- hyk and we IPL-arrive 'And we arrived.'

ja- jur IPL-come

(96b)

pe kyse muhyk and knife 3 +CAUS-arrive 'And he brought the knife.'

(97a)

pe ihe a- pu'am aho ti and I ISO-stand ISG-go again 'And I stood going again' or 'And I stood up again.'

(97b)

ok pyter amu- pu'am house center 1SG-CAUS -stand put up the house center pole.'

342

James Kakumasu

(98a)

ma'e so'o i-nem pe tui tapi'ir i-nem some game 3-rotten there 3+lie tapir 3-rotten 'Some kind of meat lying there is rotten, a tapir is rotten.'

(98b)

pe irapüimbor munem arapuha rukwer and king vulture 3+C AUS-rotten deer meat 'And the king vulture caused the deer meat to rot.'

(98c)

ne riki ere-munem you EMPH 2SG-C AUS-rotten 'You are the one who caused it to be rotten.'

Some verbs undergo a morphophonemic change: mumba munger mondo

'cause to end' (upa 'end', an auxiliary verb). 'cause to sleep' (uker 'he sleeps'). 'cause to go, send' (oho 'he goes').

(99a)

a'ep jande ja- ker there we IPL-sleep 'We slept there.'

(99b)

ne mimi ihe amu- nger your child I ISG-CAUS-sleep put your child to sleep.'

(lOOa)

y pe o-ho ym water to 3-go NEC They didn't go to the stream.'

(lOOb)

pe xuä pu'yr ra'yr mondo i-pe and John beads small 3+CAUS+go 3-to 'And John sent him small beads.'

I have not so far encountered any causativized transitive verbs. Sentences such as "the mother caused the child to eat" are expressed in other ways: paraphrasing in terms of commanding, as in (101), or using a different verb, e.g. jopoi 'give food to' (used of babies, pets, etc.), seen in (102): (101)

e'u ma'e aja tipe 2SG.IMP-eat something thus FRUST 'His mother said in vain, "Eat something".'

i-mäi 3-mother

(102)

ihe rendyr riki ma'e so'o te'e jopoi my sister EMPH some game freely 3-t-feed 'My sister is feeding some kind of game (animal).'

reko AUX

URUBU-KAAPOR

343

Some nouns may be causativized: mukapitam muher mujawar

'cause to be captain' (kapitam 'captain') 'name, cause to be named' (her 'name') 'elope' (jawar 'dog')

Examples are: (103a)

ma'e ne ere-putar, kapitam aja arauxu what you 2SG-want captain thus Araujo 'Araujo (said) thus, "What do you want, captain?".'

(103b)

arahä i-päi mukapitam ta at that time 3-father 3 +CAUS-captain PUT 'At that time his father will make (him) captain.'

(104a)

her ja- moi ta name IPL-place PUT 'We will give (him) a name.'

(104b)

ta'yn ke upa ja- muher child OM end 1PL- CAUS- name 'We finished naming the child.'

(105a)

tajahu riki jawar mahem wild pig EMPH dog 3+chase The dogs chased the wild pigs.'

(105b)

i-namo mujawar 3-with 3 +CAUS-dog 'He eloped with her.'

o-ho 3-go

The reflexive ju- and the causative mu- may cooccur in the order ju-mu-, resulting in an intransitive verb: (106a)

pe irapüimbor i-nem ke te'e u-'u and king vulture 3-rotten OM freely 3-eat 'And the king vulture ate that which was rotten.'

(106b)

pe irapüimbor munem and king vulture 3 +CAUS-rotten 'And the king vulture caused (it) to be rotten.'

(106c)

ihe a- jumu- nem I ISG-REFLX-CAUS-rotten will cause myself to be rotten.'

ta PUT

344

James Kakumasu

(107a)

oropo katu ym Oropo good MEG Oropo is not good.'

(107b)

karai ihe amu- katu ta non-indian I ISG-CAUS-good PUT will make peace with the non-indian.'

(107c)

aja thus

te'e ngä freely 3PL

jumu- katu 3+REFLX-CAUS-good

ha kwe riki NOMLZR past time EMPH That's the way it was long ago that they caused themselves to be at peace.' 7

Comparatives and equatives

7.1 Comparatives. The simplest form of the comparative is one of the suffixes we 'little' or te 'much', with or without katu 'good'. For example, compare the two forms below: katu we 'somewhat good' katu te 'very good' The antonyms of 'very good' are: katu katu

ym ym te

'not good' 'absolutely not good'

The constructions katu we and katu te may occur with verbs to express degree, and the katu may be omitted: me'e me'e me'e me'e me'e

we katu katu katu te

we te

'he 'he 'he 'he 'he

gave some' gave somewhat good' gave well' really gave well' truly gave'

Another form of the comparative is an adverbial phrase with the basic idea of direction upwards or downwards. For a higher degree the following form is used: (108a)

i-'ar 3-above

koty towards

we little

X X

'greater than it'

URUBU-KAAPOR

345

The "X" in the above phrase is expressed by a nominalized form, as in the examples that follow: (108b) (108c) (108d)

i'ar koty we i'ar koty we i'ar koty we

tamüiha old man-NOMLZR

Older than him'

tiha-ha big-NOMLZR

'bigger than him'

puku-ha long-NOMLZR

'longer than it'

For a lesser degree the following form is used: (109a)

i-wyr 3-below

koty towards

we little

X X

'lesser than it'

(109b)

iwyr koty we puku-ha

'shorter than it'

(109c)

iwyr koty we tiha-ha

'smaller than him'

The standard of comparison may be expressed by a prefix, as in the foregoing examples, or by a pronoun (110), or by a noun: (llOa)

a'e riki ne 'ar he EMPH you above 'He is taller than you.'

koty towards

we little

puku-ha long-NOMLZR

(llOb)

a'e ne 'ar koty we tamui-ha

'He is older than you.'

(llOc)

a'e ne wyr koty we tamui-ha

'He is younger than you.'

While the above ways of expressing the comparison are acceptable and are sometimes used, it is more common to use expressions such as 'he is really tall', or by stating a positive-negative contrast that implies the comparison: (lila)

peme'e ja tiha-ha that one like big-NOMLZR 'Like that size there?'

(lllb)

am, puku te no long truly 'No, he is really tall' or 'No, he is taller.'

346

James Kakumasu

(112)

nema'e katu, xuama'e am yours good John's no 'Your's is better than John's.'

(113)

a'e puku, ne am he long you no 'You are not as tall as he.'

There is no special form of correlative comparison to match the English "the more he eats the fatter he gets." A semantically near equivalent would be used, in the form of a conditional sentence with raha 'if, when' (see sect. 14): (114)

u-'u hü rahä, i-xa hü 3-eat lot if 3-fat lot 'If he eats a lot, he will get fat.'

ta PUT

7.2 Equatives. Equatives are expressed by saka 'like' or aja 'thus, like'. Sometimes both forms are used: aja saka 'like that, thus'. The two forms may also combine with the deictic ame'e 'that one' to form the expressions: ame'eja saka 'like that one' or ame'eja 'like that, just like that'; or they may combine with kome'e 'this one' to form: kome'eja saka 'like this one' or kome'eja 'like this one': (115a)

pe riki mair mujä and so Mair 3+make 'And so Mair made (a feather headdress).'

(115b)

sarakur Saracur

pandu kome'eja 3 +say like this

ihe-ma'e my- thing

emujä aja 2SG.IMF-make thus 'Saracur said, "Make mine just like this one".' (116a)

karai ta u-'u katu non-indian PL 3-eat good 'Non-indians eat it also.'

(Hob)

so'o rukwer game flesh

ja thus

saka like

ti also

ngä 3PL

pe to

je a'erehe a'e ta u-'u katu ti HSY for that reason 3 PL 3-eat good also 'It is said that it is just like game meat to them; for that reason they eat it also.'

URUBU-KAAPOR

(117a)

se amö awa u-hyk here another people 3-arrive 'Some different people came here.'

(117b)

ne saka riki you like EMPH '(They were) just like you.'

347

u-wyr 3-come

8 Coordination The equivalent of 'and' as a coordinator is best treated as a discourse feature, since the form used normally occurs in the sentence-initial position. It is the connective pe 'and, and then', which is used to express closely connected temporal sequences. Sometimes there are causal overtones which are best translated as 'so, then'. One speaker used pe twelve times in succession to narrate a sequence of events. An example from part of a narrative follows; (118a-h) is a sequence of sentences: (118a)

ihe asak I ISO-see Til go see.'

ta PUT

aho ISG-go

(118b)

pe pytun rupi tur, kandei pe and night at 3-fcome pitch with 'And/so he came at night with a resin torch.'

(118c)

pe u-hyk, kyse renda rehe and 3-arrive knife place at 'And he arrived at the location where the knife was.'

(118d)

pe kyse amö i-hi and knife another 3-be 'And there was another knife there.'

(118e)

petei kyse one knife There was one knife.'

(118f)

pe raho so 3 +take 'So he took it.'

(118g)

pytun rupi raho night at 3 +take 'He took it at night.'

348

(118h)

James Kakumasu

pe pay u-hyk, kyse renda rehe then Pay 3-arrive knife place at Then Pay arrived at the location where the knife was.'

Closely related to pe is pe riki 'then, and so, and it was that, but'. This occurs less frequently than pe, but it also marks a sequence of events, being a combination of the forms pe 'and' and riki 'EMPH'. It has slightly stronger cause-result overtones than pe. For example, note the following consecutive sentences: (119a)

tapi'iruhu juka cow 3 +kill 'He killed a cow.'

(119b)

pe riki pupur and then 3 -l-boil 'And then he cooked (some).'

Examples from another text follow, again a sequence of consecutive sentences: (120a)

putu'u ym riki 3 +rest NEG EMPH '(The fever) didnf stop.'

(120b)

pe riki kaninde ngi puhä tur tipe and then Caninde from medicine 3+come FRUST Then medicine came from Caninde, but in vain.'

(120c)

ame'e parä ehe that 3+slip to him That went right through him.'

(120d)

pe riki kutuk kutuk tipe ngä and so 3+pierce 3+pierce FRUST people 'And so they gave him many injections, but in vain.'

Comitative is expressed by namö 'with, also', and sometimes has an aspect of coordination, as in (125): (121)

pe ihe aker i-namo and I ISO-sleep 3-COM1T 'And I was sleeping with him.'

a- ju ISO-lie

(122)

tangara riki tuti ke namö Tangara EMPH Tuti FOC COMIT Tangara was with Tuti.'

tui 3 be

URUBU-KAAPOR

349

(123)

y namö jande karaipe jasosok water COMIT we fibrous leaf IPL-pound 'We pounded the fibrous leaves with water.'

(124)

nasui riki ihe namö i-hon Nasui EMPH me COMIT 3-go 'Nasui went with me.'

(125)

wasai ihe a'u u'i a?ai I ISG-eat farinha ate assai with farina.'

namö COMIT

Quite often the comitative phrase occurs at the end of the sentence, following the main predication, functioning to clarify or specify (cf. sect. 2): (126)

a'erehe for that reason

anguxl Agostino

ko thus

ihe my

rok house

pe in

a'e 3

tui, h-akehar namö 3 +be 3-wife COMIT 'For that reason Agostino is in my house, with his wife.' (127)

ma'e something

ke kaitä OM Kaitä

ta PL

kekar 3+hunt

a'e 3

ta PL

o-ho, 3-go,

te'öru namö Te'öru COMIT 'Kaitä folks went hunting, Te'öru also.' (128)

sepetu ja- 'u we ri, pupur namö roast on stick IPL-eat some still boiled COMIT 'We were still eating roasted (meat), boiled also.'

The adversative 'but' is expressed by am 'no' or

raha 'if not':

(129a)

se ihe mimi ke kanim o-ho here my son FOC 3+be lost 3-go ' "My son got lost here," thus it is said.'

(129b)

ani, o-ho riki aja je tukä pandu i-pe no 3-go EMPH thus HSY Tucano 3+say 3-to ' "But no, he went away," thus it is said Tucano said to him.'

(130a)

howy me'e ke erur blue NOMLZR 2SG .IMP-bring 'Bring (me) a blue one.'

rahä HORT

aja thus

je HSY

350

(130b)

James Kakumasu

ani rahä, i-tawa me'e ke erur no if 3-yellow NOMLZR 2SG.IMF-bring 'But if there isn't (a blue), bring a yellow one.'

Another kind of adversative is expressed by pe riki 'and then, but' (see (119) and the preceding discussion): (13 la)

pe haimü kaitä kyse me'e and Raimundo Caetano knife 3+give 'And Raimundo Caetano gave knives.'

(131b)

upa me'e, tarumba amo all 3+give sickle also 'He gave everything, sickles also.'

(131c)

pe riki amo parahy kyse but another 3+angry knife 'But another was crazy for a knife.'

rehe for

There is no special form for expressing Or' coordination, but the idea can be expressed in various ways, the principal one being by the use of nahä my 'perhaps' (cf. sect. 18.3): (132)

o-ho 3-go

nahä my. pyta perhaps 3+remain

nahä my. perhaps

akwa ISG-know

ym ihe NEG I dont know whether he went or stayed.'

9 Pragmatic and discourse characteristics There are two morphological markers for focus and emphasis: ke 'FOCUS or OBJECT MARKER'; and riki 'EMPHASIS'. Both are postposed to the head word or phrase which they highlight. There is no restriction as to the category with which riki can occur. The morpheme ke normally follows a noun or noun phrase, including pronouns and demonstratives, but there are a few examples in my data where it follows a verb that constitutes a nominalized clause functioning as object (see, for example, 106a). In transitive clauses where two nominals occur, one of them may be marked with ke to indicate that the marked one is the object, that is the one to whom the action is done. In many instances, the context or semantics may make it clear as to who did what to whom. But there are occasions when it is not clear and in such cases ke is used, functioning as an object marker:

URUBU-KAAPOR

351

(133)

tuti ke xa'e juka aja i-pe Tuti OM Xa'e 3+kill thus 3-to ' "Xa'e killed Tuti," thus (she said) to him.'

(134)

ame'e pe mataru xa'e ke nupä that with Mataru Xa'e OM 3+hit 'With that Mataru hit Xa'e again.'

(135)

pe xa'e ke mata pukwar and Xa'e OM Mata 3+tie 'And Mata also tied Xa'e.'

(136)

pe tap! xa'e ke kutuk tl and Tap! Xa'e OM 3+pierce also 'And Tapi also pierced Xa'e (with an arrow).'

ti again

ti also

Object marking is not, however, the only function of ke. It can occur with the subject nominal, of intransitive (137, 138) and transitive (139) clauses, and it also occurs in postpositional phrases between the nominal and the postpositional relator (140). As (139) shows, it may occur in a transitive clause with both the subject and the object (cf. 13). I do not fully understand the conditions under which ke is used to express these different functions, but reactions of native speakers to its use with a single nominal in a transitive clause consistently demonstrate that its function is then to distinguish the direct object constituent. For the present, when it occurs with some other constituent, I interpret its function as that of highlighting, and gloss it as 'FOCUS'. (137)

pe xa'e ke manö, aja ixyr pandu käjä and Xa'e FOC 3+die thus Ixyr 3+say Käjä ' "And so Xa'e died," thus Ixyr said to Käjä.'

(138)

xe ihe ke ajupir katu there I FOC ISG-climb well There I was really climbing well.'

(139)

a'u ym ihe ke ISG-eat NEG I FOC didn't eat a thing.'

te truly

pe to

axo ISO-move

ma'e ke thing OM

The particle riki 'EMPH' is stronger than ke when the latter is used with a similar function. The best English translation for riki is often by means of a cleft sentence, as seen in examples (140) - (143), although it should be noted that riki does not always occur with the first constituent of the clause (see 144, 145):

352

James Kakumasu

(140)

tangara riki tuti ke namö tui Tangara EMPH Tuti FOC COMIT 3+be 'It was Tangara that was with Tuti.'

(141)

ihe riki aka'u 1 EMPH ISG-be drunk 'It was I who was drunk.'

(142)

upa riki raho all EMPH 3+take 'It was everything that they took away.'

(143)

jande rehe riki jee kamarar kamarar kamarar us to EMPH 3+say friends friends friends 'It was to us that they called out, "Friends, friends, friends".'

(144)

upa raho riki all 3 +take EMPH 'It was that they took away everything.'

(145)

ihe jai py ramö te I menstruate first recently truly 'It was when I just first menstruated.'

rahä riki when EMPH

Another way to signal emphasis is simply to front the constituent to be highlighted. The particle riki may or may not accompany such fronted constituents: (146)

nixoi riki tame there was none EMPH crab There wasn't any crab.'

(147)

a- juka ym ihe ISO-kill NEG I didn't kill (him).'

(148)

akwa ym ihe ISO-know NEG I don't know.'

(149)

parahy karai angry non-indian The non-indians were angry.'

(150)

heta riki kamarar a'ep many EMPH friends there There were many friends there.'

URUBU-KAAPOR

(151)

ne reko-ha pe pahar you have-NOMLZR to quickly 'Go quickly to your village.'

353

eho 2SG.IMP-go

10 Interrogatives There are basically three ways of marking sentences as interrogatives: (l) phonologically - an interrogative sentence ends with high pitch accompanied by strong stress on the final syllable of the sentence: (152)

upa nde ere-mujä tamüi end you 2SG-make old man Old man, did you finish making (it)?'

(153)

nde xa'e ke ere- juka you Xa'e OM 2SG-kill 'Did you kill Xa'e?'

kwe time past

(2) by using the particle my, which normally occurs sentence-final and seems to carry with it the meaning 'perhaps'; it indicates that the speaker has no idea as to the answer involved in the question (see also sect. 18.3): (154)

ko here

te'e without cause

ta PUT

nde ere-xo you 2SG-be

jupe REFLX-to

my perhaps 'Will you continue to stay here?'

(155)

nde kurumi ta rehe hü ere- xo you young boy PL for much 2SG-be 'Why do you want young boys?'

my perhaps

(3) by using question words. There are seven question words and they always occur sentence-initial. Sentence-final my and mamy 'perhaps' may cooccur with any of the seven words except the homophonous sentence-initial my 'where?'. (!*my oho my 'Where did he go?' may be possible). The homophonous forms perform two quite different functions as well as occurring in different positions in the sentence:

(156)

my o-ho where 3-go 'Where did he go?' (Here my asks for information with regard to location).

354

(157)

James Kakumasu

o-ho my 3-go perhaps 'Did he go?' or wonder if he went.' (Here my expresses simply ignorance or doubt on the part of the speaker).

The seven question words are: ma*e my marä myja awa ma^erehe ma'ewä

'what?' 'where?' 'how?' or 'when?' 'how?', 'how many?', or 'which one?' 'who?' 'for what purpose or reason?' 'why?', usually signalling a rhetorical question.

Examples in full sentences are: (158)

ma'e her mamy what 3 +name perhaps 'What is his name?' (The speaker may or may not be seeking information; he may simply be trying to think what the name was).

(159)

ma'e her what 3 +name 'What is his name?7 (Here the speaker is seeking information).

(160)

ma'erehe nde ere-jur why you 2SG-come 'Why did you come?'

(161)

myja ne tapi'ir ne ere-juka how you tapir you 2SG-kilI 'How do you kill a tapir?'

(162)

myja ere-juka how many 2SG-kill 'How many did you kill?'

(163)

ma'ewä jangwate nde ke su'u why jaguar you OM 3+bite 'Why did the jaguar bite you?' or 'You should not have let the jaguar bite you.'

katu well

URUBU-KAAPOR

355

(164)

awa ko my who this perhaps 'Who is this one?' or 'Who is this person anyway?'

(165)

my kyse i-hi where knife 3-be 'Where is the knife?'

There does not appear to be any device for indicating whether a "yes" or "no" answer is expected. The closest that comes to this is the rhetorical question form ma'ewä (163) mentioned above. Answers to questions may take the form of "yes" or "no" either alone or with other elements (167), a repetition of the main or auxiliary verb (166), an incomplete sentence (167), a single word or phrase (166), or a full sentence (168): (166a)

upa nde ere- mujä tamui end you ISO-make old man Old man, did you finish making it?'

(166b)

upa end 'It is finished.'

(167a)

nde xa'e ke ere- juka kwe you Xa'e OM 2SG-kill time past 'Did you kill Xa'e?'

(167b)

am nai i-pe ihe no mistakenly 3-to I ' "No, that's not true," I (said) to him.'

(168 a)

nde sawa'e ke ere-juka aja i-pe you man OM 2SG-kill thus 3-to ' "You killed the man," thus (they said) to her.'

(168b)

pe je teri and HSY Teri

mäi mother

je'e 3+speak

te INTNSF

ani no

a- juka ym ihe ISO-kill NEC I 'And it is said that Ten's mother said, "No, I didn't kill him".' A "yes" answer may take one of several forms: a}e ty 'yes, okay, that's good', ha Okay, yes', ere Okay, yes', and a*e 'yes, okay'.

356

11

James Kakumasu

Imperatives

Imperative is marked in the verb by the prefixes: e- '2SG.IMP' and pe- '2PL' The pe- form is used for both statement and imperative sentences. No distinction is made between positive and negative forms of imperative. (169)

e'u katu puhä pep e'am 2SG.IMP-eat well medicine there 2SG.IMP-be 'Stay there and keep taking the medicine.'

(170)

mair ra'yr ke ahem ejan Mair son FOC 3+shout 2SG.IMF-run 'Mair's son shouted, "Don't run".'

(171)

pake pe- pyta pehe over there 2PL- remain you+PL 'You folks stay over there.'

(172)

mä pe- pyhyk ym EXCLM 2PL-grab NEG 'Hey, don't grab (it).'

ym NEG

Hortatory expressions are formed by postposing raha to the verb. This form is homophonous with the clause subordinator raha 'if, when'. It occurs with both second person and first person plural imperatives. The response particles ere 'yes' and 'no' may occur alone with raha to express a mild form of imperative (176, 177). The absence of rahä indicates a strong imperative or command (see, for example, 179): (173)

jaho aja riki 1PL- go thus EMPH ' "Let's go," thus (I said).'

(174)

jaho rahä 1PL- go HORT 'Let's go.'

(175)

esak ejur 2SG.IMP-see 2SG.IMP-come 'Come and see.'

rahä HORT

URUBU-KAAPOR

(176)

ere rahä okay HORT Okay, go ahead.'

(177)

am rahä no HORT 'Not yet' or 'Don't do it yet.'

357

The affirmative response particles listed at the end of sect. 10 occur also as responses to imperatives and to declarative statements that express a mild form of imperative (178): (178 a)

pytun rahä jande ja- ho ta night when we IPL-go PUT 'We'll go at night now.'

(178b)

ere aja ihe i-pe okay thus I 3-to ' "Okay," thus I (said) to him.'

(179a)

ajame'e ke urupe after that Lupercio

pandu 3 +say

ihe me

apo now

pe to

ti again

jaho 'y 1PL- go CMPL 'After that Lupercio said to me again, "Let's go now".'

(179b)

ere aja ihe i-pe okay thus I 3-to ' "Okay," thus I (said) to him.'

(180a)

kol ta aho ky tomorrow PUT ISG-go FUT.DEF.PURP Tomorrow I'll go for sure.'

(180b)

a'e ty aja ihe i-pe okay thus I 3-to ' "Okay," thus I (said) to him.'

12

Negation

Sentence negation is marked by the morpheme ym postposed to the verb. The same form is used to express constituent negation and may be postposed to a nominal or adverbial phrase:

358

James Kakumasu

(181)

sawa'e ym man NEC '(He is) not a man.'

(182)

paite ym far MEG '(It is) not far.'

(183)

u-hyk we ym ri 3-arrive little NEG still 'He hasn't arrived yet.'

(184)

pe riki a'i k ym wapyk u-i and so old woman head NEG 3+sit 3-be 'And so the old woman without a head was sitting.'

A negative answer αηΐ 'no' may cooccur with the negative morpheme ym to reinforce the negation. There is usually a phonological break after αηΐ: (185)

am, eraho ym no 2SG.IMF-take NEG 'No, don't take (him).'

Historically, there is, in Tupi-Guarani, another form of negative, expressed by infixing a verb to nd-...-i (Rodrigues 1953:151). A similar construction occurs in Guajajara, where a preverbal form na- 'NEG' occurs with the suffix -z (Bendor-Samuel 1972:85,94-95). In Urubu-Kaapor, only one (uninflected) form has remained of this construction: nixoi 'there is none' (derived from -ixo- '3 be' and n-...-i 'NEG'). 13

Anaphora

Two types of anaphora are dealt with in some detail in other parts of this description: deleted elements (sect. 3), and pronominal forms (sects. 15.2; 16; and 18.4). Here I will restrict my remarks to the special forms that occur as sentence connectives. There are four basic forms that occur as sentence connectives with anaphoric reference: a'engi 'from there', arah 'at that time', ame'e 'that one', and aja 'thus, like that'. Other elements sometimes occur before these words, but as a general rule they occur sentence-initial. (For another connective, a'erehe, which links two clauses in a reason-result relationship, see sect. 14). The connective a'engi 'from there' marks the continuation of a narrative with a change of setting. It can be regarded as a paragraph boundary marker

URUBU-KAAPOR

359

indicating the beginning of a new sequence of events at a new location. It is composed of two morphemes: a'e 'that one, 3', and -ngi 'from', which can be suffixed to pronouns, nouns and adverbs: (186a)

arar rok renda 'ar pe i-hon o-i macaw house place above to 3-go 3-be 'He was passing by the place of Macaw's house.'

(186b)

ku'e raha a'e- ngi o-ho ti next when there-from 3-go again The next day he travelled again from there.'

(186c)

pe nduwi ru koty u-hyk o-ho and Nduwi father towards 3-arrive 3-go 'And he arrived to Nduwi's father's (place).'

(186d)

a'ep u-ker ti there 3-sleep again There he slept.'

(186e)

ku'e rahä je a'e- ngi o-ho ti next when HSY there-from 3-go again The next day he travelled from there again, it is said.'

The connective arahä 'at that time' has reference to some earlier event whose duration may or may not be ended at the time the second event takes place. When the first event continues, arahä indicates simultaneity: (187a)

pete! xo ru jahy kanim one Jon father moon 3+be lost 'Jon's father was gone for one month.'

(187b)

arahä at that time

ma'e kaninde what Caninde

rehe to

o-ho 3-go

ine I

araho ISO-take

tipe FRUST 'At that time I took (him) to Caninde, but in vain.' (188a)

arahä at that time

je HSY

tuti Tuti

u-hyk 3-arrive

mamy perhaps

ti again

ihe raikwer koty me behind towards 'It is said that at that time perhaps Tuti arrived again, behind me (i.e. after I left).'

360

(188b)

James Kakumasu

kaninde Caninde

rehe to

ihe I

aho wä ISG-go just

ihe I

ke FOC

arahä 'y at that time CMPL had just now gone to Caninde at that time.' (188c)

arahä at that time

je tuti HSY Tuti

ke FOC

rehe for

i-pai 3-father

o-ho 3-go

ti also 'It is said that at that time his father also went for Tuti.'

The word arahä is made up of two morphemes: a- 'that, it' and raha 'when'. When the future tense marker ta cooccurs postposed to arahä, it may refer to some coming event, although the reference is still to a previous statement in the discourse: (189a)

warahy sun

ka'aruk afternoon

koty towards

esak 2SG.IMF-see

eho ky 2SG.IMP-go FUT.DEF.PURP 'Go see (him) in the late afternoon.'

(189b)

arahä ta ky aja ihe i-pe at that time PUT FUT.DEF.PURP thus I 3-to ' "At that time (I will go see him)," thus I (said) to him.'

The connective ame'e 'that one' refers to some event or object mentioned earlier in the narrative. It may refer to something immediate or it may refer to something quite distant. The word ame'e is composed of the two morphemes: a- 'that, aforementioned' and -me'e 'NOMLZR', and it may be followed by a postposition or particle: (190a)

ajame'e ke karai küjä a'ep u-hyk after that non-indian woman there 3-arrive 'After that a non-indian woman arrived there.'

(190b)

a'ep amö u-wyr ti there another 3-come also There another one came also.'

URUBU-KAAPOR

(190c)

ame'e awa rehe that one people for That one is the nurse.'

(191 a)

juka te'e 3+kill without cause

361

u-sak-iha 3-see-NOMLZR ngä people

aja thus

ihe me

pe to

ihe my

rakehar pandu wife 3 +say 'My wife said to me, "They killed him without cause".' (191b)

ame'e jo riki ihe akwa that only EMPH I ISO-know only know that (i.e. what she said).'

(192a)

pe mataru mani'ok sosoka ke pyhyk and Mataru manioc pounder OM 3-(-grab 'And Mataru grabbed a manioc pounder.'

(192b)

ame'e pe mataru xa'e ke nupä that with Mataru Xa'e OM 3+hit 'With that Mataru also hit Xa'e.'

(193a)

myra ke wood FOC

y water

'ar above

rupi through

ti also

myrape planks

ke ngä OM people

mupe hu i-ndo 3+CAUS- flat much 3-send 'People flattened out the planks above the water.' (193b)

ame'e that

'ar above

rupi through

jande we

ja1PL-

sak ja- ho see 1PL- go

ti also On top of that we also went to see.'

In (194) and (195), each from a different narrative text, ame'e follows another connective aja (see below) to refer to the whole story that precedes: (194)

aja ame'e upa ke thus that one end FOC Thus that (story) is ended.'

(195)

aja ame'e upa thus that one end Thus that (story) is ended.'

362

James Kakumasu

The connective aja 'thus, like, like that' has reference to whole statements or thoughts that have been mentioned before, and is usually associated with conclusions and summary statements. It embraces more thana/ ' , which most often refers to single events or objects; aja may occur alone at the end of a text to refer back to the whole story:

(196)

aja riki thus EMPH Thus it was' or That's the way it was.'

See also sect. 3.2 for use of aja in quote tags which follow the direct speech. Other examples of aja are: (197a)

jangwate tiki mujan i-ndo aja ihe jaguar AFF 3 +CAUS- run 3-go thus I * "It is the jaguar that chased it," thus I (thought).'

(197b)

aja ihe ke a- kwa- ha thus I FOC ISG-know-NOMLZR Thus I thought' or 'Like that I was knowing.'

(198a)

ma'e something

te INTNSF

mutu motor

ke FOC

nupä 3+hit

ke FOC

hü much

te ehe INTNSF onto it The outboard motor beat hard on the (waves).' (198b)

aja me'e rupi te'e jande thus NOMLZR through freely we 'We (travelled) through (it) just like that.'

(199a)

arahä u-häi at that time 3-split 'At that time (the clay pan) split.'

(199b)

aja te'e mou like that freely 3+CAUS-lie 'It was laid just like that (i.e. split).'

(200a)

ihe mahapyr apyhyk I three ISO-take hold also caught three (crabs).'

(200b)

xo aja a'e pyhyk Jon like that 3 3+take hold 'Jon also caught that many.'

ti also ti also

ke FOC

te INTNSF

URUBU-KAAPOR

363

Backward anaphora (or, cataphora) is expressed by one form that also usually occurs sentence-initial: kuja 'like this', which refers to something that follows: (201)

kuja like this

ihe I

sepetu roast on stick

rehe on

amo- i, ISG-CAUS-be

tumeme four made this many roast on stick, four.' (202a)

riki kuja pandu käjä ixyr EMPH like this 3 +say Käjä Ixyr 'It was Ixyr who said like this to Käjä.'

(202b)

papa father

riki juka EMPH 3+kill

aja thus

pe to

riki ixyr EMPH Ixyr

pandu 3+say

käjä pe Käjä to ' "It was father who killed him," thus Ixyr said to Käjä.'

14 Subordinate clauses One type of subordinate clause is formed by postposing the clitic rahä 'when, if to the verb. It may be preceded by other suffixes and postpositions, but not usually more than two. Subordinate clause verbs do not differ in structure from main clause verbs except for the addition of rahä, and the absence of the closure constituent. Context determines whether the conditional 'if or the temporal 'when' is meant. When the negative ym and the future ta cooccur, the meaning is unambiguously 'if (204, 205):

(203)

asak ehe rahä apandu ta ISO-see to him when 1SG- say PUT 'When (or, if) I see him, I will tell him.'

(204)

ere-hendu ym rahä apetek 2SG-hear MEG if ISO- hit 'If you don't obey, I will hit (you).'

(205)

aman ym rahä aho rain MEG if ISG-go 'If it doesn't rain, I will go.'

ta PUT

ta PUT

364

James Kakumasu

Further examples of subordinate clauses with rahä follow, including contrary-to-fact conditionals (209, 210) (see also sect. 18.3): (206)

ajame'e ke marajä after that Maranhäo

ngi from

ihe I

ahyk rahä, 1SG-arrive when

mataru rehe ihe asak ti Mataru on I ISG-see again 'After that, when I arrived back from Maranhäo, I saw Mataru again.' (207)

pe and

upa end

apirok ISG-peel

rahä, when

ihe I

peir back carrier

a- mujä ISO-make 'And when I finished peeling (it), I made a back carrier.' (208)

a'erehe for that reason

ka'a 3 +defecate

kupe back

a- hupir 1SG-carry

pe ihe on I

ta PUT

rahä, when

araho 1SG- take

pe then

ihe my

awa people

ka'a ha renda pe 3+defecate NOMLZR place to Tor that reason, when she wants to defecate, then I carry her on my back to the place where people defecate.' (209)

ihe jö aho rahä ihe I only 1SG- go if I 'If I went alone I would suffer.'

(210)

a'e 3

ta PL

pandu 3+speak

ym rahä NEG if

pyrara suffer jande we

ta PUT ja- kwa IPL-know

ym ta apo NEG PUT now 'If they had not spoken, we wouldn't know now.'

The use of rahä is not restricted to clauses. It occurs also with phrases, especially those expressing time when the position of the sun is referred to: (211)

warahy sun

jandar noon

rahä, when

paramarajä Para-Maranhao

ja- hyk ja- ho IPL-arrive IPL-go 'At noon we arrived at Parä-Maranhäo.'

pe to

URUBU-KAAPOR

(212)

warahy sun

ka'aruk tree top

te INTNSF

365

rahä when

urupe Lupercio

pandu 3+say

ihe pe me to 'In the late afternoon, Lupercio said to me ...'

(213)

kue rahä jande ja- ho next day when we 1PL- go The next day we travelled again.'

ti again

A time phrase and a subordinate clause, both with rahä, may cooccur in juxtaposition: (214)

warahy sun

ka'aruk te tree top INTNSF

ajur rahä pije ISG-come when Pije

rahä ihe when I

mäi mother

ihe me

pe to

ahyk ISO-arrive kyha hammock

pukwar 3+tie 'In the very late afternoon, when I arrived, Pije's mother tied a hammock for me.' (215)

upa ja- matyr end 1PL- gather

rahä when

ame'e that

pytun rahä night when

jande we

ja- hapy ta tipe IPL-light fire PUT FRUST 'When we finished gathering, on that night, we wanted to light the fire.' A sequence of subordinate clauses may be juxtaposed: (216)

katu good

rahä u-hyk when 3-arrive

rahä apo a- hupir ta when now ISG-lift up FUT

ky'y FUT.DEF.PURP-CMPL 'When (the father) is well, and when he arrives home, I will then lift up (i.e. give a name to) (the child).'

(217)

ta'yn child

ra'yr little

ma'e something

rahä when

hyru 3+receptacle

rahä when

366

James Kakumasu

i-pai ke ma'e ahy i-xo riki 3-father FOC some pain 3-be EMPH 'When the little child already was, when he was in the uterus, his father was with some kind of sickness.' When rahä follows an aspect auxiliary, the sequence signifies that while one event is taking place another event occurs at the same time (see sect. 18.2 for aspect auxiliaries): (218)

o-ho i-xo rahä a'e u-sak 3-go 3-ASP.AUX when 3 3-see 'While he was going, he saw him.'

ehe to him

Generally, the subordinate clause precedes the main clause, but the main clause can be fronted for emphasis: (219)

u-'ar u-'ar te'e wata i-xo 3-fall 3-fall freely 3+walk 3-be 'He stumbled while he was walking.'

rahä ti when also

There are usually fewer constituents in subordinate clauses than in main clauses. Nominalized verbs constitute the nucleus of another type of subordinate clause. It can be embedded in the main clause as subject (221), object (220), or as part of a postpositional phrase (222). See also sect. 15.4. Examples are: (220)

ajame'e ke after that

pete'e there

je HSY

mä pira EXCLM fish

juka 3 -l-kill

me'e ke pyta mo-I mä'y NOMLZR 3+remain 3+CAUS-lie EXCLM-CMPL 'After that, it is said, there it was, the thing he killed the fish with.' (221)

pe kamyjä tiha me'e ke pyta then truck big NOMLZR 3+remain 'Then that big bus stopped again.'

(222)

katu me'e pe ngä mou 3+good NOMLZR to people 3+CAUS-lie The people placed (him) in a good place.'

ti again

Direct speech complements function as the direct object in quotative sentences. These complements can be any kind of main clause or incomplete

URUBU-KAAPOR

367

sentences that can occur alone as interrogatives or responses (sect. 10) or which satisfy any other conditions relating to ellipsis (sect. 3). In particular, connectives and closure constituents may occur. The normal order in quotative sentences is for the direct speech complement to precede the quote tag. There is no indirect quotation, and verbs such as 'wish', 'desire' and 'want' are expressed by direct quotes. (223)

aho ta ky aja ISG-go PUT FUT.DEP.PURP thus ' "I'm going," thus he said.'

(224)

ere-jywyr ym ta 2SG-return NEC PUT

nde you

ke FOC

pandu 3+say mamy aja perhaps thus

ihe i-pe I 3-to ' "Perhaps you won't return," thus I (said) to him.'

(225)

sapukai je'e rahä manö aja rooster 3+speak when 3+die thus ' "He died before dawn," thus (he said to me).'

(226)

kotete nearby

ta PUT

jande we

kupixa garden

ja- mujä IPL-make

ky ti aja je peri pe FUT.DEF.PURP also thus HSY Peri to ' "We are also going to make a garden nearby," thus, it is said, (he said) to Peri.' Semantic subordination is involved in the reason-result type of sentence, but the two clauses that occur in the sequence are both main clauses. The word a^erehe 'for that reason, as a result, therefore' occurs as a sentence connective in the second clause of the sequence: (227)

jixi'u 3 +cry

te'e without cause

a'erehe for that reason

arur ISG-bring

ihe me

pe with

mä, EXCLM

ahijar ISG-leave

i-mai pe 3-mother to 'He cried for no reason at all with me (and) for that reason I returned (him) to his mother.'

368

James Kakumasu

(228)

xuäxi Joäozinho

atu je juka good HSY 3 kill

ta PUT

je, HSY

a'erehe upa ngä tyryk o-ho for that reason end people 3 +move 3-go 'It is said that Joäozinho himself will kill (him); for that reason all of them moved.'

te INTNSF

The word tferehe consists of two morphemes: a*e '3' or 'that', which usually has reference to the protasis, and rehe 'for, to', a postpositional relator. In the protasis there are no restrictions as to the constituents that may occur, so that the closure constituent may occur before a'erehe: (229a)

hoi ru o-ho ym a'e Hoi father 3-go MEG 3 'Hoi's father also did not go.'

(229b)

aruwi Aruwi

ru father

ani no

a'e 3

ti, also

ti also a'erehe for that reason

ka'i ru ani a'e ti Ka'i father no 3 also 'Aruwi's father also did not go and for that reason Ka'i's father also did not go.' 15

SYNTAX OF PHRASE TYPES Noun phrase structure

15.1 Marking for case. The only element that might be considered a case marker is the suffix ke, which marks the object in a transitive clause when two nominals occur. See sects. 1.1 and 9 for a fuller description. (230)

jakare ke kaitä japi alligator OM Caetano 3+shoot 'Caetano also shot an alligator.'

(231)

pe je and HSY

so'oran rabbit

jangwate jaguar

amö another

ke nupä OM 3+hit

tl also

katu well

te mä 'y INTNSF FJCCLM CMPL 'And it is said that the rabbit really beat the jaguar.'

15.2 Genitives. There are two ways to express possession. One is by using possessive pronouns, which also (except for the third person forms) can function as the free form subject in a clause. The second way is by a nominal phrase:

URUBU-KAAPOR

(232)

myra räkä tree branch

369

'branch of a tree'

Possessive pronouns are: ihende- (ne-) i- I h- I x-

'ISO' '2SG' '3SG'

jandepehengä-

(jane-)

'1PL' '2PL' '3PL'

The items in parentheses are used when a nasalized vowel or consonant follows. They also occur varying freely with their alternate forms before nonnasal vowels or consonants. See sect. 22 for the phonological rules concerning nasal segments and progressive nasalization. (233a) (233b)

nde-po 2SG-hand ne- äpui 2SG-nose

'your hand' 'your nose'

(233c)

katar ahy hu flu hurt much The flu was very painful.'

(233d)

ngä-äpüi huwy 3PL-nose blood Their nose bled.'

(233e)

jande-mani'i upa ngä mumba 1PL- manioc plant end 3PL 3+CAUS- end The people finished off all our manioc plants.'

(233f)

imembyr ym me'e ke te'e 3SG-child MEG NOMLZR freely The one who has not had a child died.'

manö 3+die

Nouns are divided into four sub-classes, based on the form of the third person possessor prefix which occurs with each class: h-, /-, x- and Other'. Some hclass nouns are: hamüi 'his grandfather' hape 'his path' häi 'his teeth' hapo 'its root' hembe 'his tongue' hayty 'its nest' The h- changes to r- when the possessor is first or second person or a free form nominal; it changes to t- when the possessor is unspecified:

370

James Kakumasu

hamüi 'his grandfather', cf. ihe-ramüi 'my grandfather' tajahu ramüi 'wild pig's grandfather', cf. tamüi '(someone's) grandfather' In a few cases, /- has come to function as the third person instead of A-: ta*yr

'his son' (not *Aa'yr), cf. ihe-ra*yr 'my son' nde-ra^yr 'your son'

In a few other cases, the form t- 'unspecified possessor' is not used; for example: ma*e rapo ma*e rayty

'some kind of root' 'some kind of nest'

(never *tapo) (never *tayty)

Some /- class nouns are: iäkä ipy ipy'a

'his head' 'his feet' 'his liver'

ijywa ipusu itangwa

'his arm' 'his stomach' 'his buttock'

The /- is dropped and no prefix occurs when the possessor is first or second person or a free form nominal: ihe-äkä nde-py arapuha py'a

'my head' 'your feet' 'deer's liver'

In the x- class of nouns, x- changes to k- when possessed by a non-third person or when not possessed at all. That is, words in this class beginning with k- may or may not be possessed. Thus, the x- class of nouns is less rigid than the A- or - classes in this respect. See the following examples, where the non-possessed form kyha is used in (234a) and (234b) and the possessed form xyha in (234c): (234a)

pe kyha muhem we-rur then hammock 3+CAUS-come out 3- bring Then he brought away (his) hammock.'

(234b)

kyha je putar tipe je hammock HSY 3+want FRUST HSY 'It is said that he wanted a hammock.'

URUBU-KAAPOR

(234c)

pe pytun and night

rahä pe when then

371

x-yha 3-hammock

we-rur 3- bring

xe there

maneru rok pe Maneru house to 'And when it was night, then he brought his hammock there to Maneru's house.' Examples of the x- class of nouns are: xamby xyha xupe xywyr xamarar

'her breast, milk' 'his hammock' 'his back' 'her brother' 'his friend'

ihe ihe ihe ihe ihe

kamby kyha kiipe kywyr kamarar

'my 'my 'my 'my 'my

breast, milk' hammock' back* brother' friend'

The fourth sub-class of nouns is designated Other'. These are non-possessed forms and constitute the largest number of nouns in the language. Some examples are:

Jtfl'fl awaxi kupixa

'forest' 'corn' 'garden'

jawar akaju myra

'dog' 'cashew' 'wood, tree'

The only way to refer to these items as being possessed by someone is to use a possessed form of the very general word ma'e 'thing', which belongs to the i- class of (possessed) nouns: ihe-ma*e ne-ma*e i-ma'e

'my thing, mine' 'your thing, yours' 'his thing, his'

Free form nominals may precede ma*e to indicate the possessor: kapitä ma*e mane ru ma*e

'captain's thing' 'Manuel's father's thing'

Examples of genitive phrases in sentences are: (235a)

arauxu miasu ke pukai Araujo servant FOC 3+call 'Araujo's servant was calling out.'

i-xo 3-be

372

James Kakumasu

(235b)

y rymy'y rehe ngä u-hyk water edge to 3PL 3-arrive They arrived at the edge of the water.'

Nouns belonging to the h- and i- classes, and some members of the x- class, are obligatorily possessed. 15.3 Modifiers. Modifiers of nouns fall into two groups: those that precede the noun (numerals, quantifiers, and demonstratives); and those that follow the noun (either postposed elements or stative verb roots, which combine with the noun to form a compound noun). Compare the following forms: stative verb (236a), compound noun (236b), and stative verb with a full nominal subject (236c): (236a) (236b) (236c)

i-juk 3-rotten

'It is rotten.'

myra-juk wood-rotten

'rotten wood'

myra i-juk wood 3-rotten

The wood is rotten.'

See sect. 22 for the stress rule which explains why these modifiers are regarded as suffixal or incorporated elements. Some modifiers may be postposed to nouns, verbs and certain adverbs (sect. 20):

k a m (u)hu/(u)hu t e keruhü ai we py wä

'good, well' 'big, much' 'true, much, INTNSF' 'very big, much' 'ugly, useless' 'little, somewhat' 'first' 'just, recent'

Nouns X X X X X X X

Verbs X X X X X X X X

Adverbs X X

X X

Examples are: (237 a) (237b)

jawar-uhu dog- big

'big dog' or 'species of jaguar'

mondok-uhu 3-(-cut- much

'He cut a lot' or 'He worked hard at cutting.'

URUBU-KAAPOR

(237c)

(237d)

(237e)

ka'a- te forest-true u-hyk 3-arrive

373

'virgin jungle' katu well

te truly

me'e- te 3+give-truly

'He arrived very well.'

'He really gave (i.e. he gave generously).'

When te and we occur postposed to verbs, they occur in fourth position after the verb. (See the table of verb affixes at the beginning of sect. 18). In addition to te and we, there is another intensifier, tai 'really, finally', which seems to be stronger than te or katu te: (238a)

puhä ke nixoi tai medicine FOC none DSTNSF There wasn't any medicine at all.'

(238b)

ja- 'u ym tai IPL-eat NEG DMTNSF 'We didn't eat at all.'

jande we

ke FOC

Other modifiers that occur postposed to nouns are:

(239)

(240)

ambyr

'deceased':

nose Nose

ruambyr father-deceased

anga

'substitute':

i-pai3-father

anga substitute

kwer Ingwer (241)

(242)

'bone (removed from body)' 'similar, substitute':

küjänungar woman-substitute par

'his foster father' or 'his godfather'

' former':

kä- ngwer bone-former nungar

'deceased father of Nose'

'mistress'

One who exists/dwells':

374

(243)

James Kakumasu

ka'a- por forest-one who dwells in ran

(244)

(245)

'fox'

'PLURAL':

kaitä- ta Caetano-PL warn

(246)

'false, similar to':

jawa-ran dog- similar to/false ta

'forest dweller'

'Caetano folks'

'intended to be':

kapitä- warn captain-intended to be

'future captain'

Modifiers that precede the noun are numerals, demonstratives:

quantifiers,

(247)

h-akehar 3-wife

pe kuja and like this

pytun night

moköi two

pytun night

pe then

and

rehe o-ho ti for 3-go also 'And this many nights, two nights, then he went for his wife also.' (248)

petei renda aja nde ere- mupinim one place thus you 2SG-CAUS- spot One place, thus you wrote.'

(249)

pe amö kamarar we-rur and another friend 3- bring 'And/then another companion brought it.'

(250)

ame'e jarusu rehe ihe asak py ti that one boat on I ISG-see first also 'Also, I saw that (kind of) boat for the first time.'

(251)

kome'e tamüi pyta ma'e ka'aro this one old man 3 -(-stay what Ka'aro This old man stayed, that is, Ka'aro.'

Special rules apply to sequences of pronoun and numeral. When the numeral follows a pronoun, it gives an 'exclusive' meaning to that pronoun. There are no special forms of pronouns distinguishing 'exclusive' and 'inclusive', except by this use of numerals (see sect. 16).

URUBU-KAAPOR

375

(252)

jande moköi jö raho 'y us two only 3+take CMPL '(The truck) finally took only us two (exclusive).'

(253)

jande moköi ja- ho rahä katu ta aja riki we two IPL-go if good PUT thus EMPH ' "If we two (exclusive) went, it would be good," thus (she said).'

(254)

pehe moköi pe- ho you two 2PL-go 'You two go.'

rahä HORT

When the numeral precedes the pronoun, the sequence constitutes a stative clause: (255)

moköi jande two we 'We are two.'

In narrative texts, the numeral often occurs without a head noun in certain restricted contexts, such as in (256). The deleted element is recoverable from the semantics of the sentence or from the context generally: (256)

a'ep moköi jande jathere two we 1PL'We slept there two (nights).'

Relative clauses occur in the The latter do not have any example, (257)). Both types of in an adverbial phrase. See also

ker sleep

form of nominalizations or descriptive clauses. formal marking of nominalization (see, for construction may occur as subject or object or sect. 15.4.

(257)

tamüi i-äkä tuwyr kyha pe wapyk u-i old man 3-head white hammock in 3 +sit 3-be 'An old man who has white hair was sitting in his hammock.'

(258)

amö ahy me'e ta another pain NOMLZR PL Others who were sick came.'

(259)

wasai

(260)

a'engi from there

kaitä Caetano kneaded the ko here

ke FOC

tur 3 come

we-rur me'e ihe 3- bring NOMLZR I fruit that Caetano brought.' ambor 1SG- throw

arur ISG-bring

akamyryk ISG-knead akaju'y cashew

376

James Kakumasu

kaitä mondok me'e ke pe Caetano 3+cut NOMLZR to 'From there I threw (it) down here, to where Caetano cut down the cashew tree.' 15.4 Nominalizations. There are three basic forms of nominalizer: -ha, -har, and me'e ke. The ke in me'e ke may be omitted, without any meaning difference. The ke appears to be the same form as ke 'FOCUS', but in me'e ke it does not function to mark focus or emphasis, at least in many cases. Neither does it always have the object-marking function of ke. The forms -ha and me'e ke nominalize verbs, while -har nominalizes adverbs and postpositional phrases. There are two other variants of ha: -a occurs after the consonant k, and -iha occurs after other consonants. Some speakers use -iha, rather than -