Alyawarr Verb Morphology

David Campbell Moore, Bachelor of Arts (Honours).

This thesis is presented for the degree Master of Arts of the University of .

School: Social Sciences.

Discipline: Linguistics.

Thesis submitted in 2012

1

2 Abstract

The Alyawarr Language Region is located in the Sandover and Barkly areas of the . Alyawarr is a language with around 2000 first language speakers. Although the Alyawarr language has been studied since the 1930s, only recently has long-term detailed fieldwork been undertaken in Alyawarr.

Alyawarr verbs are agglutinative, suffixing and compounding. Alyawarr is in rich inflectional and derivational morphology. One significant category is Motion.

Chapter One is an introduction to the Alyawarr language and reviews the literature on the language. Verb roots and types of word formation are explored in Chapter 2. Subsequent chapters explore different categories which follow the verb root. Chapter 3 covers Derivation and Number. Chapter 4 explores the categories of Motion and Aspect. Chapter 5 is dedicated to the inflectional categories of Tense and Mood and to complex clauses.

This research represents seventeen years of research in the Alyawarr language based upon extensive text collections and data elicitation. Although this thesis builds on that of Stanham (1972), Turtle (n.d) and Yallop (1977), it contains new perspectives on understanding the Alyawarr verb and further refines previous analyses.

This thesis is also distinctive and interesting because of further detailed information about in Alyawarr. Basic Motion verbs are explained and the category of Motion including an account of vertical motion. Another feature of this thesis is an exploration of the aspectual system of Alyawarr and its relationship to the categories of tense and mood along with a number of hitherto undescribed inflections.

3

Table of Contents

Table of Contents ...... 4 Acknowledgements ...... 7 Table of Abbreviations ...... 8 Figures ...... 11 Tables ...... 11 1 Introduction ...... 14 1.1 The Alyawarr language region...... 14 1.2 Classification of the Alyawarr language ...... 14 1.2.1 Typological features of the Alyawarr language ...... 15 1.2.2 Thesis orthography: Sounds and their representation in Alyawarr ...... 15 1.3 Review of Linguistic Research in the Alyawarr language ...... 17 1.3.1 Early Research: Wordlists ...... 18 1.3.2 Anthropological research in the Alyawarr region ...... 19 1.3.3 Professional Linguistic Research from 1930 ...... 19 1.3.4 Verb morphology ...... 22 1.4 Linguistic Variation ...... 22 1.5 Social context of the language ...... 23 1.5.1 Wider contacts with other languages of ...... 23 1.5.2 Mutual intelligibility and relatedness of Alyawarr dialects ...... 25 1.5.3 Speakership and language ownership ...... 27 1.5.4 Areal Features ...... 27 1.6 Language change ...... 28 1.6.1 The profile of Alyawarr in Central Australia ...... 28 1.6.2 Mobility ...... 28 1.6.3 Demographics ...... 29 1.6.4 Alyawarr and English ...... 30 1.6.5 The influence of literacy and education ...... 30 1.6.6 Two way education ...... 31 1.6.7 The influence of media ...... 31 1.7 The data and sources ...... 32 1.7.1 Participant observation ...... 32 1.7.2 Types of data ...... 32 1.7.3 Texts ...... 33 1.7.4 Genre ...... 33 1.7.5 Elicitation ...... 34 1.7.6 Observed speech...... 35 1.7.7 Translations ...... 35 1.7.8 Introspective comments of speakers ...... 35 1.7.9 Learning the Alyawarr language ...... 36 1.7.10 The roles of language consultants ...... 37 1.7.11 Recording techniques ...... 38 1.8 The outline and scope of this work...... 38 2 The Alyawarr Verb...... 39

4 2.1 Introduction ...... 39 2.2 Verb structure ...... 39 2.3 Properties of Verb Roots: Transitivity ...... 40 2.4 Verbal Sub-categories ...... 41 2.4.1 Basic Motion verbs ...... 42 2.4.2 Basic Stance verbs ...... 46 2.5 Compound Verbs ...... 49 2.5.1 Other evidence for compound status ...... 50 2.5.2 Subclassification of compounding types ...... 53 2.5.3 Compounds based upon IV and TV ...... 56 2.6 Verbal Reduplication ...... 60 2.6.1 Semantics of reduplication and iconicity ...... 61 2.6.2 Pre-bases ...... 62 2.6.3 RED.elp Attenuative ...... 64 2.6.4 RED.ep Frequentive ...... 67 3 Derivation and Number ...... 70 3.1 Introduction ...... 70 3.2 Derivation ...... 70 3.3 Verbalising suffixes ...... 70 3.3.1 Intransitive Verbaliser -irr ...... 71 3.3.2 Attribute Intransitive Verbaliser -elh ...... 71 3.4 Mediopassive -elh ...... 71 3.4.1 Reflexive ...... 73 3.4.2 Passive ...... 76 3.4.3 Middle ...... 76 3.4.4 Antipassive ...... 76 3.5 Reciprocal -err ...... 79 3.6 Transitive Verbaliser ...... 81 3.6.1 Nominal bases ...... 81 3.6.2 Verb bases ...... 82 3.6.3 Transitiviser -ern ...... 84 3.7 The Category of number ...... 84 3.7.1 Plural -err and -arr ...... 86 3.7.2 -enherr ...... 86 3.7.3 -erl.ew plural with Basic Motion verbs ...... 87 3.7.4 RED.ew- ...... 87 3.7.5 -elhelerr ...... 88 3.7.6 -arlelherr ...... 88 3.8 Discrete -am ...... 89 4 Motion and Aspect ...... 91 4.1 Introduction ...... 91 4.2 The category of Motion ...... 91 4.3 PATH morphemes ...... 92 4.3.1 -eyn Return ...... 92 4.3.2 -enh Point ...... 95 4.4 Sequenced Motion compounds ...... 98 4.4.1 DO&MOTION -erl...... 99 4.4.2 MOTION&DO -ey ...... 101

5 4.4.3 Reduplication and Motion compounding ...... 104 4.4.4 RUN ...... 105 4.4.5 Compositionality and motion compounds ...... 109 4.4.6 Compounding vs. morphemic complexes ...... 110 4.5 Rapid actions...... 111 4.5.1 -erl.iw SIDE ...... 111 4.5.2 (-erl).iw Small Repetitive Reduplication ...... 112 4.5.3 -erl.ayn AWAY ...... 114 4.6 Action along a path ...... 115 4.6.1 -erl.ap Convey something along ...... 115 4.6.2 -erlenty.akng Distributed motion ...... 117 4.7 Vertical movement and ‘do while Y approaches’ ...... 118 4.7.1 -ey.aynt ‘UP’ ...... 119 4.7.2 -erl.arrern ‘DOWN’ ...... 122 4.8 Cultural explanation for complex motion verbs ...... 124 4.8.1 Example of Motion...... 125 4.9 ASPECT...... 127 4.9.1 Continuous Aspect ...... 128 4.9.2 -erl.an State...... 130 5 Obligatory verb marking ...... 132 5.1 Tense and Aspect ...... 132 5.1.1 -eyel Present ...... 133 5.1.2 -eyenh Future ...... 135 5.1.3 -ek/-ew Past Perfective ...... 136 5.1.4 -enh Past Imperfective IMPP...... 137 5.1.5 Relative Past Tense ...... 139 5.1.6 -eyartingkerr Past Habitual ...... 140 5.1.7 -ey.inkwern Past Habitual2 ...... 140 5.2 Mood ...... 141 5.2.1 Imperative ...... 141 5.2.2 -ey Hortative ...... 142 5.2.3 -etyek, -eyew, -eyek Purposive...... 142 5.2.4 -em Potential...... 143 5.2.5 -ey.angenh, -eyang Negation ...... 148 5.3 Complex Sentences ...... 150 5.3.1 Purposive in complex sentences ...... 151 5.3.2 -ey.alkenh, -ey.akenh, Subsequent ...... 151 5.3.3 -enty Irrealis ...... 152 5.3.4 -emer Subjunctive ...... 157 5.3.5 -el Same Subject ...... 159 5.3.6 -ekerr/-ewerr Apprehensive ...... 160 5.3.7 -ekerr-awaty Before ...... 161 5.3.8 Characteristic -ey.angker and -enh.RED ...... 161 5.4 An alternative classification ...... 162 5.4.1 Aspectual Category Two: Intentional Mood ...... 163 5.4.2 Aspectual Category Three ...... 163 6 Future Research ...... 166 References ...... 167

6 Acknowledgements

Thanks are due to my supervisor John Henderson of the Linguistics Discipline Group in the School of Social Sciences who encouraged this project from the beginning in 2005. I have had many long distance conversations with John while I worked on this project and have benefitted from his wealth of knowledge of Central Australian languages. Thanks also to Alan Dench and the Department of Linguistics at UWA whose support has enabled me to understand the issues involved in analysing languages

7

Table of Abbreviations

A Transitive Subject ABL Ablative -they AED Alyawarr to English Dictionary (Green 1992) AED2 Alyawarr to English Dictionary revision ALL Allative -warl APP Apprehensive -ekerr 5.3.6 ADV Adverbialiser -el ASSOC Associated with -areny ATT Attenuative 2.6.3 AVERS Aversive -kety, -wety AWAY Away from -erl.ayn 4.5.3 BASE&DO Move to base and do V -ey.alp 4.4.2.2 BEFORE Before -ekerr-awaty 5.3.7 CHAR Characteristic -ey.angker 5.3.8 CHAR1 Characteristic 1 -angker 5.3.8 CHAR2 Characteristic 2 -enh 5.3.8 COM Comitative -akert CONJ Conjunction -ap CONT1 Continuous 1 -aynt 4.9.1.1 CONT2 Continuous 2 -erl.aynt 4.1.9.2 CONV Convey -erl.ap 4.6.1 DAT Dative -ek/-ew DISC Discrete -am 3.8 DISTR Distributive -erlenty.akng 4.6.2 DAT Dative case -ew, -ek DO&BASE Do V and then move to base -erl.alp 4.4.1.2 DO&GO Do action and go -erl.alh 4.4.2.1 DO&RUN Do action and run -erl.elp 4.4.4.2 DOWN Do verb action while -erl.arrern 4.7.2 ECAED Easternmoving anddown Central Arrernte to English Dictionary (Henderson and Dobson 1994) EMPH Emphasis -ey ERG Ergative -el EX Exclusive marking on -en pronoun FIRST Occurs first in sequence -tangkwel FOC Focus -an FREQ Frequentive -ep 2.6.4 FUT Future -eyenh 5.1.2 GO&DO Go and then do action -ey.alh 4.4.2.1 HAB Past Habitual -eyartingkerr 5.1.6

8 HAB2 Past Habitual -ey.inkwern 5.1.7 HORT Hortative -ey 5.2.2 IMP.S Imperative -Ø 5.2.1 IMP.DU Dual -enh(err)atherr 5.2.1 IMP.PL Plural -enh(err)arey 5.2.1

IMPP Past Imperfective -enh 5.1.4 INS Instrumental -el IRR Irrealis -enty 5.3.3 IV Intransitive verbaliser -irr 3.6 LOC Locative -el, -ew, -ek, -itwew, - itwek, -itwel LOTS Lots of, numerous -angketyarr MED Mediopassive, middle voice -elh 3.4 NEG Negative, negation -eyang, ey.angenh 5.2.5 NOM Nominative NP Noun phrase NPL Nominal Plural -rnem O Object, grammatical object PERL Perlative, through -angkwarr PP Past Perfective -ek/-ew 5.1.3 PL Plural POINT Action occurs at point on -enh 4.3.2 path POSS Possessive -kenh/-wenh POT Potential -em 5.2.4 PRES Present tense -eyel 5.1.1 PRIV Privative -weny PURP Purposive -etyek, eyew 5.2.3 QUOT Quote -wenh REAS Reason -wety, -kety REC Reciprocal -err 3.5 RED Reduplication REP Small repetitive movement RET Return -eyn 4.3.1 ROUND around -anyem RUN&DO Run and do -ey.elp 4.4.4.1 S Intransitive Subject SAE Standard SBJV Subjunctive -emer 5.3.4 SEMB Semblative -ilkwer SIDE Move to the side -erl.iw 4.5.1 SRC Source -penh SS Same subject -el 5.3.5 STAT State -erl.an 4.9.2 STILL Still occuring -antey SUB Subordinate clause marker -arl SUBS Subsequent action -ey.alkenh 5.3.2 THEN Temporal succession -anem TV Transitive Verbaliser -il, -elhil 3.6 UP Do action while moving -ey.aynt 4.7.1 upwards

9 1S First person singular 1D First person dual 1PL First person plural 2S Second person singular 2D Second person dual 3S Third person singular 3D Third person dual 3PL Third person plural

10

Figures

Figure 1: Map of the Alyawarr language region and adjacent languages...... 13 Figure 2: Example of motion suffixes ...... 126

Tables

Table 1: Alyawarr Consonants ...... 16 Table 2: Alyawarr ...... 16 Table 3: Percentages of cognates shared (Hale 1962:181) ...... 20 Table 4: Alyawarr Simple Verb Structure ...... 39 Table 5: Verbal Subcategories ...... 42 Table 6: Alyawarr Basic Stance verbs ...... 47 Table 7: Types of Alyawarr compounds ...... 49 Table 8: Verbs occurring in lexical compounds ...... 53 Table 9: Types of verbal compound ...... 57 Table 10: Alyawarr Basic Stance compounds ...... 58 Table 11: Deictic Motion compounds ...... 59 Table 12: Verbal reduplication patterns in Alyawarr ...... 61 Table 13: Occurrence of verbs with FREQ and ATT ...... 63 Table 14: Derivational Suffixes ...... 70 Table 15: Verb roots which combine with Mediopassive (including compounds) ...... 72 Table 16: Mediopassive functions ...... 73 Table 17: Some verbs with antipassive interpretation ...... 77 Table 18: Properties of Antipassives ...... 78 Table 19: Number marking on Alyawarr roots and derived stems ...... 85 Table 20: POINT-marked verbs ...... 98 Table 21: Sequenced motion compounds ...... 98 Table 22: RUN&DO verbs and their interpretations ...... 106 Table 23: DO&RUN verbs and their interpretations ...... 107 Table 24: Rapid complex markers ...... 111 Table 25: Small repetitive Reduplication ...... 113 Table 26: Habitual action along a path complex markers ...... 115 Table 27: vertical movement markers ...... 119 Table 28 State, Adopt posture and Agentive forms ...... 119 Table 29: Downward motion forms in the C. Strehlow wordlist ...... 123

11 Table 30: Imperfective Aspect markers ...... 128 Table 31 Continuous Aspect with modal markers...... 128 Table 32 Alyawarr Obligatory Verb Markers ...... 132 Table 33: imperative markers ...... 141 Table 34: Aspectual classes of Alyawarr verbs ...... 163 Table 35: Intentional Mood markers ...... 163 Table 36: Alternative State Markers ...... 164

12 Figure 1: Map of the Alyawarr language region and adjacent languages.

13 1 Introduction

1.1 The Alyawarr language region

Alyawarr [aʎawar] is a language of Central Australia, in the Northern Territory of Australia. There are around 2000 first language speakers of Alyawarr. See §1.6.3 for discussion. In pre- contact times, the traditional Alyawarr-speaking region was located in the area of the Sandover and Bundey rivers. According to Yallop (1969), Alyawarr is spoken in an area ‘roughly bounded by Hatches Creek, Utopia, MacDonald Downs and Ooratippra’.

The Alyawarr region has expanded beyond the traditional boundaries. See Figure 1. The Alyawarr region is unusual in comparison with other language regions in Central Australia because the Alyawarr-speaking population is decentralized. Today most Alyawarr speakers live in small settlements and outstations. Although they have a high level of mobility, their movements are usually within the Alyawarr region. The area covered by the Alyawarr region is shown by the map of the Alyawarr language region and adjacent languages in Figure 1.

1.2 Classification of the Alyawarr language

The Alyawarr and their neighbours use the name ‘Alyawarr’ for themselves and are identified by a common language. Alyawarr is a member of the Arandic group, a subfamily of the Pama- Nyungan family of Australian languages. Other are spoken to the south and west of the Alyawarr region while non-Arandic languages are spoken to the north and east of the region. Spencer and Gillen (1899) referred to Alyawarr as part of the ‘Arunta nation’. Hale (1962:183) claimed that, ‘dialects referred to as Alyawarr appear, in the Plenty River area, to grade into dialects referred to as Aranda’. Dixon (2002:670) regards all of the Arandic languages with the exception of Kaytetye as one language and claims that their separation into distinct languages by Wilkins (1989) is done ‘partly on sociopolitical criteria’. For further discussion of this view refer to §1.5.3.

The spelling in the official orthography for the language is ‘Alyawarr’. The language has also been variously spelled as Il(l)aura, Iljauara, Ilyauarra, Ilyowra, Ilowra, Illyowra, Illura, Alyawara, Ilawara (Tindale 1974:226). Some of the spellings such as Aliwara, Aliawara and Aljawara were Kaytetye pronunciations according to Tindale. He claims that other spellings such as Ilawara and Jaljuwara were Ngalia (Warlpiri) pronunciations and ‘western tribal’ pronunciation. Since that time there have also been additional spellings: Aljawara (Yallop

14 1969), Alyawarra (Yallop 1977, Turtle 1977) and (Lyon and Parsons 1989). The current spelling was adopted for the Alyawarr to English Dictionary (Green 1992).

1.2.1 Typological features of the Alyawarr language

Among the languages of the world Alyawarr has a relatively high number of place contrasts for consonants. Six points of articulation are distinguished for stops and nasals as in Table 1: bilabial, lamino-dental, alveolar, retroflex, palatal and velar. The three Alyawarr vowels are shown in Table 2.

Alyawarr is agglutinative and suffixing. Noun cases and verb tenses are marked by suffixes. Grammatical functions are expressed by nominal case inflections and is comparatively free. There are three core cases, Ergative, Nominative and Accusative. Non-core cases include Dative, Possessive, Locative and Instrumental. Alyawarr has a ‘split case’ system with ergative-absolutive alignment for nominals and nominative-accusative alignment for pronouns with the exception of 1sg and 2sg which have tripartite alignment.

Unlike many other Australian languages Alyawarr has no conjugation classes. It has a rich inflectional and derivational morphology. Verbs are agglutinative, suffixing and compounding. The structure of the Alyawarr verb is elaborated in detail in §2.1. One significant category of verbal morphology conveys the motion of the Subject and whether the motion ‘was before, concurrent with or after the action referred to by the verb’ (Koch 1984). Obligatory elements include the verb root and an inflection which denotes tense-aspect, mood or clause-joining.

1.2.2 Thesis orthography: Sounds and their representation in Alyawarr

This thesis accepts the current phonological analysis of Alyawarr and Arrernte (Breen 1990). It utilises the orthographical conventions of the common Arandic orthography, the development of which is outlined in Breen (2005: 95-99) and shown in Table 1 and Table 2. The aim of this section is to briefly explain the current phonological analysis of Alyawarr and the orthography which has developed from that analysis and which is used throughout this thesis.

15 Table 1: Alyawarr Consonants

Bilabial Interdental Apico- Retroflex Prepalatal Palatal Velar Alveolar stop p th t rt yt ty k

Pre- pm tnh tny kng stopped

Nasal m nh n rn yn ny ng

Lateral lh l rl yl ly

Flap/trill rr

Approxi w r y h mant

Table 2: Alyawarr Vowels

FRONT MID BACK i (we)

e

a

Notes:

When /e/ follows a rounded consonant (written with a w), it is realized as a high back rounded , [u] if there is a following non-. When /e / follows a palatal consonant (written with a y), it is realized as a mid to high , “i” if there is a following consonant. Otherwise (i.e. at the end of a word) it will sound like [ə] (schwa), or not be pronounced (especially in running speech).

Linguists had long been aware that the majority of words in northern Arandic languages including Alyawarr begin with a vowel, (Yallop 1977:28, Strehlow 1944:46). Yallop (1977:29) claimed that morphemes were vowel-initial in Alyawarr and that this included suffixes (Breen 2005: footnote 8). The set of consonants for Arandic languages was established by the 1970s and represented in the orthography using digraphs rather than diacritics. The same symbols could be used to represent all of the Arandic languages with some minor variations. However

16 the vowels were more problematic. Yallop (1977) and Turtle (1977) both used the triangular system of three vowels which had been found for other Australian languages, represented by the letters , and . The first orthography for Alyawarr (Yallop 1977, Turtle 1977) was similar to that of the Pfitzner orthography which was used for Western Arrarnta (Yallop 1977:7).

Breen (1977) argued that Antekerrepenh had two vowels. The most noticeable change in the orthography proposed by the School of Australian Linguistics (SAL) was the use of to represent the centralized vowel or schwa. A common orthography was developed for the Central Australian Arandic languages.

In November 1987 another meeting was held to fix the orthography for Alyawarr. Major decisions which were taken at the meeting:

1. To drop 2. To retain 3. Hyphens to be used, at least for pedagogical purposes, but details not decided. 4. Final to be written on words only when it is stressed.

As Breen (2005:100) explains, another orthography meeting in was held in 1992 to finalise the orthography for Alyawarr prior to the publication of the Alyawarr to English dictionary (Green 1992). Apart from some minor changes in hyphenation conventions (Breen and Green 1995) and some dialectal spelling variations, the orthography of the Alyawarr to English Dictionary (AED) has been used for this thesis.

1.3 Review of Linguistic Research in the Alyawarr language

Early researchers focused their efforts on the language of the regions which had been occupied by European settlers, particularly Arrernte (Austin-Broos 2009:19). The first observations of Alyawarr speakers were made by explorers and ethnographers who travelled the periphery of the Alyawarr region in the nineteenth century and had brief contact with speakers. The explorer David Lindsay (1889:664) wrote ‘the language is also the same as that of the Macumba and Finke districts’. In his chapter on language ‘Die Sprache’, it is evident that the German ethnographer Erhard Eylmann (1860-1926) had some familiarity with Arandic languages (Eylmann 1966 (1908)). During his expedition of the 1890s, he visited the Frew River and would have been in contact with Alyawarr speakers. However, it was the mid-twentieth century before the Alyawarr language came to the attention of scholarly researchers. The analysis of Arrernte has influenced that of Alyawarr, which is particularly evident in the linguistic description of Yallop (1977). More specifically, the analysis of particular morphemes has been influenced by the Arrernte analysis. Often there has been a pseudo-similarity between forms and

17 an assumption that the Alyawarr marker has the same function as the homophonous Arrernte marker. The difficulties of the analysis of such ‘false friends’ will be referred to in the relevant sections.

The purpose of the following sections is to survey the previous literature on the Alyawarr language and related Arandic varieties, documenting the sources which were available for the purpose of writing this grammar.

1.3.1 Early Research: Wordlists

The first phase of research in the late nineteenth century resulted in collections of words from Arandic languages.1 Some early wordlists were compiled by F.J. Gillen (1855-1912) and W.B Spencer (1860-1929). One of Gillen’s wordlists of 114 words recorded in his journal of 1875 is the earliest wordlist of Lower and Southern Arrernte words. Spencer and Gillen (1899:70) mention the ‘Ilaura’. Walter Roth (1861-1933) worked to the east of Alyawarr country for four years as a surgeon in far Western . During this time he documented Arandic languages closely related to Alyawarr, namely: Ayerrerenge and Antekerrepenh (Roth 1897). Roth’s works contain comparative wordlists of these languages containing a limited number of items drawn from common semantic domains.

Hermannsburg Mission was established by Lutheran missionaries in the Western Arrarnta- speaking area in 1877. There was a long tradition of language research in the Lutheran mission tradition beginning with Kempe’s grammar which was the first systematic attempt at compiling a grammar of Aranda (Kempe 1891). Carl Strehlow (1871-1922) made a wordlist of 7124 Aranda words while living at Hermannsburg Mission (1894-1922). These words appear to have been taken from several Arandic dialects. Strehlow refers to the closely related dialect ‘Northern Aranda’, referring to Arandic varieties to the north of Hermannsburg in his ethnographic work Die Aranda (Strehlow 1909) although apparently there are no specific mentions of Alyawarr.

1 A variety of orthographies have been used for writing Arandic languages since the late nineteenth century. The orthography adopted by the Lutheran Mission (FRM) was ‘Aranda’ which was the foundation for ‘Arandic’ which followed the orthography of the German missionaries at Hermannsburg Mission which had been established in 1877. Later FRM used ‘Arrarnta’ for Western Arrarnta. In the common orthography developed by the School of Australian Linguistics (SAL) and used by the Institute for Aboriginal Development (IAD) ‘Arrernte’ is used for Eastern and Central Arrernte andother Arandic languages and dialects of Central Australia.

18 1.3.2 Anthropological research in the Alyawarr region

In August and September of 1930 anthropological research was carried out amongst Alyawarr speakers at MacDonald Downs Station by a team from the South Australian Museum and the Board for Anthropological Research at the University of Adelaide (Ford 1966: 142). The MacDonald Downs site was chosen because of the traditional nature of Alyawarr society and because the children of the pastoralist Chalmers family could speak Alyawarr and interpret the language for the researchers. Norman B Tindale (1900-1993) collected wordlists, personal names and items of ethnographic interest. He learned from Jessie Chalmers and Don Chalmers who had lived at the station and who acted as guides and interpreters. Tindale recorded words with an orthography of the Royal Geographical Society (known as RGS II) which was designed for the recording of place names (Walter 1988). Later he transcribed Alyawarr words with a modified version of the International Phonetic Alphabet which has been called the Adelaide University Phonetic System (AUPS). (See Monaghan 2008).

Macdonald Downs was the site of research by archaeologists and anthropologists throughout the early 1970s. In 1971 Woodrow Denham conducted research amongst Alyawarr people at MacDonald Downs station (Denham 1975, 1978). James O’Connell and Lewis Binford studied people in the same area and also gathered data on the Alyawarr population (O’Connell 1979, Binford 1984, 1986). Richard Moyle concentrated upon the musical aspects of songs during fieldwork trips from 1977 to 1980 (Moyle 1986:1). His research focused upon Alyawarr speakers on the Aharreng estate at Ammaroo station and included lexical items. Jeannie Devitt researched food-gathering and use of natural resources of Alyawarr and closely related language groups (Devitt 1988, 1994).

1.3.3 Professional Linguistic Research from 1930

As noted in §1.3.2, the Alyawarr language had been heard by explorers in the late nineteenth century and some words were recorded at Macdonald Downs in 1930. Significant in Australian linguistics 1930-1960 was the Adelaide School, which included the linguistically-trained TGH (Ted) Strehlow (1908-1978) (Moore 2008) and Tindale (Monaghan 2008). Methods of recording speech improved markedly during the period 1930-1960 through the use of recording devices and phonetic transcription. Tindale and Strehlow used the symbols of a version of the International Phonetic Alphabet (I.P.A) which had been developed at Adelaide University. An Aranda dictionary was being compiled by Strehlow in the late 1950s with around 3160 entries as an extension of his father Carl Strehlow’s wordlist. The techniques of phonemics and morphemics were not widely adopted in Australian linguistics until the 1950s and 1960s (Moore 2008), after the time in which Strehlow did his linguistic field research. Some grammatical

19 analysis was previously undertaken of Eastern Arrernte dialects by T.G.H Strehlow in his detailed account of the phonetics and grammar of Arrernte dialects (Strehlow 1944). His grammar was the only published grammar of an Arandic language until the 1970s. In a letter to Suzanne Allbright (nee Shepley) concerning a name of Alyawarr origin, Strehlow indicated that he had ‘never done any research work in the Ilaura area’ (correspondence between Strehlow and Allbright 29th November 1958, ). Strehlow had a limited interest in Alyawarr, regarding Alyawarr as a non-traditional group because they had contact with explorers, pastoralists and with miners at Hart’s Range to the south of the Alyawarr Region (Strehlow commenting on Denham application in a letter to Ronald Berndt 24th March 1970, AIAS correspondence file). According to Capell, (1962:101), at the time of writing Alyawarr was ‘not as yet recorded’.

In the late 1950s professional linguists who were trained in descriptive linguistic techniques began working on Australian languages in what McGregor (2008) describes as the third period of Australian linguistics. Strehlow’s focus on the languages to the south of Alyawarr (Green 2001:33) was one of the reasons why the linguist Ken Hale undertook research in the Alyawarr- speaking region. Hale collected extensive recordings and notes on Australian languages including Alyawarr in 1959 (Green 2001:33). He conducted a linguistic survey in 1959-60 and his paper (Hale 1962) showed an early application of the method of lexicostatistics to Australian languages. The survey utilized a 100 word list to elicit words from ten localities including words from two dialects of Alyawarr as shown in Table 3.

Table 3: Percentages of cognates shared (Hale 1962:181)

AAl MDAl Ak PRAr ASAr An HgAr HyAr LoAr Ka 41 38 37 33 31 32 30 28 29 AAl 83 77 70 64 60 60 56 55 MDAl 83 72 67 59 60 58 57 Ak 79 70 60 62 62 64 PRAr 87 75 74 68 65 ASAr 83 86 77 67 An 82 70 54 HgAr 84 66 HyAr 65

20 Key Ka Kaytetye from Murray Downs Kaytetye from Murray Downs AAl Alyawarr from Ammaroo MDAl Alyawarr from MacDonald Downs Ak Akarre, Aketyarre or Akarre-Aketyarre from the Plenty River area. PRAr Arrernte from the Plenty River area west of Ak ASAr Eastern Arrernte from An Anmatyerr from Napperby HgAr Western Arrernte from Hermannsburg HyAr Pirtima or Southern Arrernte, from Henbury LoAr Lower Arrernte from Dalhousie

Colin Yallop researched the Lake Nash dialect of Alyawarr at the end of 1966 and throughout 1967 completing his PhD thesis at Macquarie University which was published as Alyawarra Grammar (Yallop 1977). He checked some points of his work with Alyawarr residents at Ammaroo and Alice Springs. A large number of texts were recorded. Elicitation was recorded on field cassettes 877-1630 which are archived at the AIATSIS. I listened to these texts and transcribed some of them for this thesis.

The Summer Institute of Linguistics (SIL) began working in the Alyawarr region in 1972 with the intention of translating the Bible into Alyawarr. Janet Stanham researched Alyawarr at Lake Nash and Murray Downs, compiling a sketch grammar and small dictionary (Stanham 1972). Nancy Turtle researched phonology at Alekarenge (Warrabri, Ali-Curung) in the 1970s (Turtle 1977). Turtle and Carol Morris compiled an unpublished grammar of Alyawarr in the 1980s, working with speakers of the northern dialect at Alekarenge. A number of fluency drills, mimicry drills and texts were recorded, for the purpose of learning Alyawarr. Discourse markers were analysed by SIL linguists (Bierbaum 1987). An Alyawarr mini-Bible consisting of one third of the New Testament and selections from two books of the Old Testament was published in 2002. An Alyawarr Songbook was published by SIL and an Alyawarr hymnbook by Finke River Mission containing many song translations.

Alyawarr wordlists were compiled by Ken Hale, Colin Yallop, Nancy Turtle and Carol Morris. The Arandic Languages Dictionary Program of the Institute for Aboriginal Development began to collect data in the 1980s. These words were checked at Ampilatwatja and Alpurrululam in 1985. Early wordlists included a Picture Vocabulary (Breen 1986). Breen worked on the wordlist which was compiled in March 1989 and was organized in semantic domains and which became available as the Alyawarr Wordlist (Arandic Languages Dictionary Programme 1990). Jenny Green had worked in the Alyawarr region in 1977-78 on community literacy and art. She

21 began collecting data for the Alyawarr to English dictionary in the early 1990s. The Alyawarr to English Dictionary (henceforth: AED) was published by the Institute for Aboriginal Development Press in 1992. A revision is currently in progress. David Blackman worked at McLaren Creek (1992) and Epenarra (1994-2004) and provided extra words for the dictionary. David Moore worked with Alyawarr speakers to collect additional data from Amperlatwaty from 1995 onwards.

Barbara Sayers of the Summer Institute of Linguistics taught a discourse analysis workshop in Alice Springs in 1998. The main purpose of the workshop was to analyse Alyawarr language data at a level above the clause. Other workshops were continued by linguists from SIL and the Finke River Mission in 2004 and 2005 which involved discourse analysis of Alyawarr texts.

1.3.4 Verb morphology

Grammars of Alyawarr have been compiled since the latter part of the twentieth century. Yallop (1977) admitted that his work was a sketch grammar. The AED (Green 1992) contains a more basic account of the grammatical suffixes which was necessary in a lexicographic work. David Wilkins researched the Mparntwe dialect of Arrernte (Wilkins 1989) in the Australian morphosyntactic tradition. Wilkins brought more recent theory and collaboration with other researchers to his research in Arrernte. David Strickland worked as a Bible translator with SIL at Amengernternenh (Utopia Clinic) from 1998-2000 and completed a basic grammar (Strickland 1998). This work was an extension of the earlier work of Yallop (1977) and the AED and involved the analysis of some morphemes which had not been considered by earlier researchers.

1.4 Linguistic Variation

The Alyawarr language is a large and viable language, having the sixth largest number of speakers of all of the Aboriginal languages in Australia (excluding creole languages), according to the Australian population census of 2001 by the Australian Bureau of Statistics (http://www.abs.gov.au). There is a direct relationship between land and the Alyawarr language, a situation which has been reported for other Australian languages (Rumsey 1993). The relations between people and language are complex and there is no simple correspondence between these two aspects of Alyawarr identity. The Alyawarr language itself forms the primary link between a number of separate social and political units.

There is variation in Alyawarr which can be attributed to age factors and social settings i.e. non- geographical factors. These include avoidance registers, sign language and the language of

22 songs. There is a child-directed speech register which is a simplified code. Otherwise there appear to be few significant age-related differences in the language of the kind which has been reported by O’Shannessy (2005) for . The modalities of song and signs are derived from and dependent upon spoken language. They can be explained by the use of the normal spoken language but not vice versa.

In this thesis I concentrate upon the main vehicle of communication in Alyawarr society which is the ordinary spoken language of adult speakers. In the next two sections I will attempt to account in more detail for the sources of linguistic variation and change in Alyawarr. In the first part I will concentrate on dialect variation as a geographic factor in variation and in the second part I will examine change as a historical process of variation in the language.

1.5 Social context of the language

Geographically-determined language variation is indicative of the patterns of long-term social interaction in the Alyawarr region. Alyawarr speakers have had relatively recent contact with the mainstream Australian culture. In traditional times they lived in exogamous patrilineal groups in the Sandover and Bundey River areas. Their population was probably distributed in a similar way to that which Strehlow (1965:142) describes for Western Arrarnta as reflecting an essential ‘disunity of the tribe’ (Strehlow 1947:1). The size and composition of each group was limited by environmental factors and the availability of resources (O’Connell 1979:100).

1.5.1 Wider contacts with other languages of Central Australia

Yallop (1969) documents the most significant ceremonial connections between the Alyawarr and other language groups in Central Australia. Mythological Dreaming tracks or paths link Alyawarr speakers with their neighbours (Yallop 1969:194). Ceremonial activities have traditionally been the meeting point of speakers of different social groups. People maintain contact with their kin and associate with speakers of other, often unrelated language groups at ceremonial gatherings. Different language groups shared altyerr ‘dreamings’. Some Alyawarr speakers claim kinship relationship with Arabana speakers, approximately 1500 kilometres away at Port Augusta through possession of common altyerr affiliations (D. Thompson, p.c). According to early sources (Roth 1897) there were speakers of Arandic dialects living near the Queensland border in the late nineteenth century. The Ayerrerenge and Antekerrepenh languages were Arandic languages which appear to have been mutually intelligible with Alyawarr.

23 In the early twentieth century Alyawarr speakers began to settle in the areas of other language groups whose numbers were reduced drastically because of white settlement (Lyons and Parsons 1989). Some Alyawarr speakers migrated to Lake Nash, which had previously been inhabited by speakers of Pwelany and other languages.

In the mid-twentieth century there was a general move towards larger settlements, for example the settlement of Warrabri (Alekarenge) which was established in 1956. Yallop (1977:1) claimed that “Arandic peoples now live in concentrations determined by the pattern of white settlement and administration rather than by traditional Aboriginal life”. This was the situation that prevailed throughout the Northern Territory at the time that Yallop researched the Lake Nash dialect. Yet his claim was less applicable to Alyawarr than to other groups in Central Australia. With the exception of Warrabri, there were no large welfare settlements established close to Alyawarr land. Many Alyawarr worked on stations in close proximity to their traditional land, particularly at the pastoral properties of Derry Downs and Macdonald Downs. Denham (1978:14) highlights the difference between the three groups of Alyawarr then resident (1971-72) at Macdonald Downs, Lake Nash and Warrabri- ‘it seems likely that divergence from pre-contact conditions was greatest at Warrabri settlement and least in the Macdonald Downs- Derry Downs area’. Recent movement patterns tend to suggest that cultural priorities determine Alyawarr patterns of movement to a large degree. Autonomy is highly valued by Alyawarr speakers and has meant a lack of conformity to patterns of settlement determined by ‘white settlement and administration’.

Due to a combination of cultural and economic factors, opportunities for work in the pastoral industry declined at the same time that welfare payments became more available to Aboriginal people in the Northern Territory. Alyawarr control of land was facilitated by the Aboriginal Land Rights (Northern Territory) Act of 1976 (Hagan and Rowell 1978, 1979). In the late twentieth century there was a revival of interest in gaining land tenure in the region. Outstations were established on traditional land and were controlled by family groups. The return to traditional lands has allowed continuity between contemporary and traditional patterns of settlement. O’Connell (1979:117) identifies this recent development as ‘the reassertion of faith in traditional culture and the corresponding rejection of much that is European which is inherent in the land rights controversy, the outstation movement and other recent developments in Northern and Central Australia’.

As a consequence of the decline of language groups adjacent to Alyawarr, such as Wakay and Pwelany, there was a diaspora of Alyawarr beyond their traditional country. In far western Queensland and the southern Barkly region Alyawarr speakers are now numerically dominant. The diaspora has probably led to a strengthening of Alyawarr as a ‘discrete social entity’ (Yallop 1969:189). Alyawarr is the language of a socially and culturally intact people. Although

24 a large proportion of Alyawarr speakers live beyond the boundaries of their traditional land, they continue to visit their ‘country’ and to participate in initiation ceremonies which are held in the region. Most speakers at Lake Nash, for example, know what land they are affiliated to in the Sandover region.

1.5.2 Mutual intelligibility and relatedness of Alyawarr dialects

In this thesis, three Alyawarr dialects are recognised. The dialects of Alyawarr form a ‘network’ which is continuous with other Arandic languages to the south, north and west. The other Arandic languages appear to be genetically related to Alyawarr, that is, they are historical continuations of an ancestral language (Hale 1962). Kenneth Hale classified Alyawarr as part of ‘Upper Aranda’ and introduced the notion of dialect chain, a dialect being defined as ‘forms of speech which differ in certain respects but which are mutually intelligible or are members of a chain in which adjacent links are mutually intelligible’ (Hale 1962:183). Yallop (1969:187) put forward four reasons why the Alyawarr should be recognised as having a separate identity: their possession of a separate dialect (language), a tribal name, a distinctive culture and their own territory.

Alyawarr appears to lack the distinct internal dialect labels which are reported for other languages (Wilkins 1989, Meggitt 1962). One dialect of Alyawarr is spoken in the north of the Alyawarr area at Epenarra, Murray Downs and Antarrengeny. (See Figure 1, map of the Alyawarr language region and adjacent languages in §1.1). Another dialect is mainly spoken in outstations which have been settled by those speakers who were formerly resident at Macdonald Downs. Lake Nash was originally populated from the southern regions of the Alyawarr area (Yallop 1977:3) but some speakers from the north of the area also settled there. The Lake Nash dialect contains a few unique vocabulary items but is otherwise similar to the southern dialect of Alyawarr.

In this section I will attempt to account in more detail for the sources of linguistic variation and change in Alyawarr. I will concentrate on dialect variation as a geographic factor in variation and examine change as a historical process of variation in the language.

Hale (1962) spoke of a northern dialect around Ammaroo and a southern dialect based at Macdonald Downs which share 83% common vocabulary. I claim on the basis of more recent lexicostatistics that there are probably 150 words which differ for each dialect within the corpus of the AED, which contains 4500 words. This represents around 3% of the total number of lexical entries that have been recorded, meaning that these dialects have 97% of lexical items in common.

25 Yallop (1969:191) comments on the difficulty of using dialect labels for Arandic languages. The Sandover and Barkly regions are multilingual and multi-dialectal. There are many people living in areas adjacent to the Alyawarr region who speak Alyawarr as a second or third language. Alyawarr speakers often marry speakers of the adjacent Arandic languages as Yallop (1977:3) noted at Lake Nash, but this is a traditional pattern and not something adopted only because of contemporary living arrangements. Northern Alyawarr speakers often have kinship ties with Kaytetye speakers and southern Alyawarr speakers to Arrernte speakers. Borrowings into Alyawarr from adjacent languages tend to accentuate the differences between the Alyawarr dialects.

While most words are uncontroversially accepted by all speakers, there is no clear agreement among speakers about what vocabulary items are included within the domain of each dialect and language. The inclusion of a particular word will vary according to the opinions of speakers from different dialects who have different social and linguistic affiliations. Speakers are likely to say that a word used by speakers of another Alyawarr dialect is not Alyawarr at all but is used by speakers of another language (Yallop 1969:192). Green reported that ‘people in communities such as Antarrengeny are inclined to say that some of the Lake Nash words are actually Arrernte’ (Green 1992: xii). This comment reflects the degree of overlap between languages as speakers of the southern and Lake Nash dialects share particular lexical items with the adjacent Arrernte dialects. At the launch of the Alyawarr Picture Dictionary in October 2004 there was considerable debate about the inclusion of some of the items in the dictionary. Although efforts were made to include all dialects in the dictionary, some speakers were unhappy that words from another dialect had been included. One speaker of the northern dialect expressed regret that Lake Nash language had been included in the Alyawarr Picture dictionary because ‘they are not Alyawarr’. A Lake Nash speaker was adamant that their words are ‘proper Alyawarr’.

Nor are there a large number of regular isoglosses for the language. Although some emblematic items in the vocabulary appear to clearly mark the speakers of a particular dialect, these forms only represent a small percentage of items in the lexicon. Nevertheless a few common lexical and grammatical items can be noticeable in speech. Speakers from different dialects regard particular words as diagnostic of ‘essential’ Alyawarr. For example, whether speakers use apey- alheyel or apetyeyel ‘coming’ is thought to separate speakers of Alyawarr from speakers of other languages. Yallop found it disconcerting that many Alyawarr people themselves used apetyem in preference to apey-alhem ‘coming’(apety- is a shared lexical item used by both Alyawarr and Arrernte speakers, apey-alh- is restricted to Alyawarr) . At Lake Nash Arrernte speakers were called ‘petyeme people’ from their use of apetyem ‘coming’ in contrast with apey-alh- which is commonly used by Alyawarr speakers. More conspicuous in speech are

26 grammatical forms which are differentiated for dialect e.g Past Perfective -ew/-ek and the Purposive -eyew/-etyek which have a high frequency of use.

1.5.3 Speakership and language ownership

A distinction can be made between linguistic and political uses of ‘language’ similar to that made by Dixon (1976:214). This distinction corresponds to the difference between communicative competence in a language and being an owner of the language, a situation also reported for Mparntwe Arrernte (Wilkins 1989). The situation for Alyawarr in terms of ‘speaker’s identification’ is similar to that which holds for Arrernte. Language ideology plays a part in intelligibility (mutual intelligibility) and what speakers are able to understand (Wilkins 1989:10) and not only linguistic features of the languages.

Accommodation and code-switching is common in interaction between the speakers of different languages. Language identification ‘appears to be based more on social networks or political affiliations than linguistic features’ (Henderson 1998:5). In conversation speakers use the phrase angka anyent ‘one language’ to emphasis their relationship with speakers of related dialects and languages. Claims are often made that speakers of other dialects speak angka anyent-antey - the ‘one language’ claim is also extended to speakers of related Arandic dialects, for example north- eastern Arrernte speakers who are related to Alyawarr speakers. Alyawarr speakers typically say that their own language is ‘a bit light’ and that the adjacent dialects of other languages are also ‘light’ or easy to understand. The local English term ‘light’ has a geographical sense rather than a chronological sense as in O’Shannessy (2005). Distant dialects of another language are awerr ‘hard’ and anantherr awey-angenh ‘we can’t hear (that language)’. Eastern Anmatyerr (spoken in Utopia outstations) and northeastern Arrernte dialects (spoken at Alcoota and Harts Range) are thought to be easy to understand and share many lexical and grammatical items with the southern dialect of Alyawarr. Compared with Eastern Anmatyerr, Central Anmatyerr is described by Alyawarr speakers as awerr ‘hard’. The corresponding ‘inner’ dialects of neighbouring languages, Central Anmatyerr and more distant dialects of Arrernte are not thought by speakers to be mutually intelligible with Alyawarr.

1.5.4 Areal Features

Alyawarr appears to share some areal features which may have diffused into Alyawarr from non-Arandic languages. Borrowing is frequently from languages said to be mutually unintelligible with Alyawarr (Yallop 1969:187) e.g Warumungu and Warlpiri. (See Figure 1 map of the Alyawarr language region in 1.1) Many Alyawarr speakers are bilingual or multilingual speakers of other languages and have regular social interaction with speakers of

27 other languages. There is a regular traffic of words from languages adjacent to Alyawarr. Often replacement words are borrowed from adjacent languages. Borrowing occurs through name taboos which occur when a person dies and words which sound similar to their name are replaced with a replacement word (often kwementyay), leading to lexical replacement within the language. Some of the loans are of recent origin; others are assumed to have appeared in Alyawarr before recorded history.

1.6 Language change

In the pre-contact situation, language change occurred through social interaction and intermarriage between different language groups. The most common vector of language change in Alyawarr continues to be social interaction with speakers of other languages.

1.6.1 The profile of Alyawarr in Central Australia

The status of Alyawarr as a language determines the degree of its acceptance as a vehicle of communication. A number of factors have acted as inhibitors of language loss and change. The Alyawarr are a culturally conservative people with relatively limited influences from Western society. Most retain a core patrilineal affiliation to their traditional country in the traditional Alyawarr region. Alyawarr speakers have a low profile in the wider Central Australian community. Most speakers live in small remote and dispersed outstations which tend to have strong social controls. As O’Connell (1979: 100) noted, ‘the consumption of alcohol is low and the amount of overt interpersonal hostility and violence negligible’, a situation which continues at the present time. Alyawarr people tend to appear reserved and avoid speaking out at meetings. Most are involved in the hunting and gathering of bush foods. The lack of attention from governments, both Commonwealth and Territory has meant that Alyawarr are often not visible in the public arena but often ‘go under the radar’. There has been limited penetration of external institutions into the Alyawarr region.

Numbers of non-Indigenous people in Sandover and Barkly communities have been low and have exerted little influence over the Alyawarr-speaking population. The land permits which are administered by the (CLC) have controlled social contact, particularly with those outside influences which could have detrimental effects upon Alyawarr language and culture.

1.6.2 Mobility

Devitt (1988:40) observed that ‘the lifestyle of Aborigines of Sandover River region was characterised by mobility’. A number of contemporary factors increase contact between

28 Alyawarr and other language groups in Central Australia. People travel long distances to attend gatherings where they may remain for several days or weeks. Community issues are often resolved at large gatherings.

Greater access to vehicles has increased the distances which may be travelled. In recent times social interaction has also involved football carnivals which have often occurred on a regional basis but can involve teams from more distant and unrelated locations in Central Australia. Many speakers attend the Show (a kind of fair) at Tennant Creek and Alice Springs, regional rodeos and the Harts Range races. The towns of Alice Springs, Tennant Creek and Mount Isa are resource centres for Alyawarr speakers. The towns have attracted Alyawarr speakers who are seeking medical attention and visiting their relatives.

Many Alyawarr speakers are reluctant to go to visit towns because the land is controlled by other language groups, for example Arrernte in Alice Springs. As a result Alyawarr speakers tend to congregate in town camps or suburban houses with other Alyawarr speakers when they visit towns.

The Finke River Mission (FRM) of the Lutheran Church has promoted contacts between Alyawarr and other language groups through courses, including languages such as which is not closely related to Alyawarr. The Aborigines Inland Mission (A.I.M) has conducted Easter Conventions at Cammooweal, Mount Isa and other centres to which large numbers of Alyawarr speakers congregate with speakers of other languages. Both the FRM and AIM organizations have encouraged Alyawarr people to assume leadership positions in their respective churches and encouraged the use of Alyawarr in worship services.

The Central Land Council has conducted large meetings about land tenure which have brought speakers from many different language groups together.

1.6.3 Demographics

Population growth of Alyawarr speakers occurred throughout the twentieth century from earlier estimates of population of a few hundred (Yallop 1977, Denham 1978:13), to more recent estimates of around 1500 (Hoogenraad and Thornley 2003:2) and 2000 (Australian Bureau of Statistics). There are also large numbers of non-first language speakers which includes speakers of Anmatyerr, Arrernte, Kaytetye and Akarre speakers. The Alyawarr speaking population is relatively young. At Ampilatwatja the average age is in the early twenties (Ampilatwatja Health Service, Paul Quinlivan, p.c). As lifestyle diseases are less prevalent than in other parts of Central Australia, there is also a large population of older people. Language and culture is being transmitted to the younger generations and levels of intergenerational conflict appear to be low.

29 The large number of young speakers indicates that Alyawarr is likely to continue to be viable in the future.

1.6.4 Alyawarr and English

Contact between Alyawarr and speakers of English greatly accelerated throughout the twentieth century. The proximity of Alyawarr to the Barkly region and the pattern of travelling and interaction of many Alyawarr speakers means that there are influences from Kriol which is spoken to the north of the Alyawarr region. Many Alyawarr speak a local variety of English which has been described as ‘Aboriginal English’. Some elements of the English influence have arrived through an earlier English established when non-Aboriginal people first came into the area (Henderson 1998:9). There is a distinctive semantics in local Aboriginal English which reflects the underlying semantics of Alyawarr (Harkins 1994).

A community perception is that middle aged people have some fluency in English (Frank Turner, p.c). English is spoken by older more widely-travelled men who worked in the pastoral industry as youths. Older women and young children are often monolingual speakers of Alyawarr.

Alyawarr language often contains ‘loanwords’ which originate with English. English loans are phonologically assimilated to Alyawarr to varying degrees, often depending upon the age of the speaker.

1.6.5 The influence of literacy and education

Access to education in English-only schools is a possible threat to the viability of the Alyawarr language. Schools were established relatively recently in Alyawarr country. By 1960 Welfare Branch schools had been established at Elkedra and Murray Downs. The first school in the Sandover River region was established at Utopia in 1969. Richardson (2001:117) makes the claim that as late as 1979 two hundred children living on stations in the Sandover River were not receiving any form of education. Even at the Utopia School a high rate attrition of teaching staff frequently led to the ‘truncation of education programs and student learning outcomes, necessitating the constant revision of previous coursework’. From the point of view of the Education Department, staffing the schools was a bigger priority than the content of what was taught in the schools. From the point of view of the Indigenous people, ‘they responded to educators whom they liked and participated in the school’s programs primarily because they wanted to get to know the person teaching, rather than because of the content of what was being taught’ (Richardson 2001:157). In the 1980s, homeland learning centres were built on

30 outstations and schools were built on the larger communities following the policies which Richardson describes as ‘tokenistic’. This means many school-aged children in the ‘homelands’ or ‘outstations’ were unable to access education. High levels of mobility, high staff turnover and official indifference have created a situation where Western education has not been accessed by a large proportion of Alyawarr speakers.

1.6.6 Two way education

The Alyawarr language has recently been incorporated in formal educational programs in the Sandover and Barkly regions. School Bilingual education programs have never been established in these regions (Hoogenraad 2001:131). An adult literacy course was developed by Gavan Breen and taught at Alpurrurulam (Lake Nash) in 1983 through the School of Australian Linguistics (SAL). In April 1985 Gavan Breen (Breen 2001: 171) visited seven Alyawarr communities, commenting that “at least ten percent of the thousand-odd Alyawarr speakers attended at least part of an SAL course”. Vernacular literacy has been taught in some Alyawarr schools since around 1995 (Susan Moore, p.c.)

Batchelor College ran an annual Aboriginal Languages Fortnight in the mid-1990s within its teacher education program involving the recording of stories for use in the classrooms. The Batchelor College branch in Alice Springs has been teaching Alyawarr literacy as part of the Certificate in Own Language course. The Northern Territory Department for Education and Training (DET) have run vernacular literacy workshops in the Sandover region. The Alyawarr Picture Dictionary was launched at Ampilatwatja in 2004 and has been widely used in school language and culture programs.

1.6.7 The influence of media

The availability of radio and television in Alyawarr communities has given speakers a greater exposure to English. Many of the smaller outstations haven’t had full access to television, internet and computing facilities so the effects of these media has been limited. The Central Australian Aboriginal Media Association (CAAMA) has made several programs in Alyawarr with English subtitles. These documentary style films are of twenty minutes duration and have been televised on and the ABC channel as part of the Nganampa Anwernekenhe series. These included the Alyawarr language films Crookhat and Camphoo (2005) and Willaberta Jack (2007).

31 1.7 The data and sources

The current work is based upon over 100 000 words of text including many texts which were recorded and transcribed as part of this grammatical description project. I have included material from all of the dialects of Alyawarr in a variety of genres. The data upon which this thesis is based has been gathered over a fifteen year period of language learning and research, including residence in Ampilatwatja for ten years. Where the treatment of verb morphemes has been brief in the work of previous writers, I offer a more fine-grained analysis. In this section I will outline my data gathering methods and the types of data which I have collected. The approach which I have taken in writing this grammar is descriptive and eclectic, drawing on a variety of theoretical approaches and presented in as straightforward terms as possible. I can lay no claim to complete objectivity as a statement of the language but I aim to make my methodology clear.

1.7.1 Participant observation

Rather than using non-linguistic ‘props’ (Hellwig 2006) to elicit information, participant observation associated with text has enabled me to understand the context in which the text is located (see §1.4.8). Observation and recounting of activities enabled me to understand some of the subtler distinctions of language use by understanding the context of the utterances. As much as possible I have used stimuli in the natural environment to elicit data, which means that the informant’s responses are not limited to their understanding of the researcher’s questions. Maps have been used to elicit compound verbs associated with movement.

1.7.2 Types of data

I have used a range of data which reflects a balance between different types. I have tried to ensure accuracy by confirmation across different types and different sources. All data types have strengths and weaknesses which I will explain in the following paragraphs. I have given the data different ‘weight’ within the text corpus according to its reliability and ranked as follows:

(i) Texts in which the language informant has told a narrative on a particular topic including Dreaming texts, local history and hunting texts.

(ii) Elicited sentences. Language informants have been asked to provide an example sentence for the AED. Language is given as a translation of the type, ‘how do you say X?’

(iii) Freely spoken utterances heard in situations where they could be written but not recorded.

32

(iv). Translations from English and other languages into Alyawarr.

(v). Statements from language consultants about a particular grammatical construction.

The material in (i) and (ii) includes texts and sentences which have been gathered by previous researchers and all published and publicly available sources of Alyawarr.

1.7.3 Texts

In constructing this grammar, I have given priority to natural texts. As Payne (1997:368) claims, ‘the more pragmatic, semantic and subtle parts of language are best analyzed via a large body of text data, supplemented by elicitation where necessary’. Although recent fieldwork practice e.g. Heath (1984:5) often emphasises the exclusive use of texts over other forms of data, it is clear that text alone cannot form the only form of data collected by the researcher. While I agree with McGregor’s (2008:414) comments concerning the limitations of elicitation and the need for the careful interpretation of data, I maintain that the real problem with elicited data is with the lack of time spent in fieldwork and language learning which leads to problems with translation and interpretation.

Myth narratives known as altyerr and historical narratives often contain ellipsis, unfamiliar vocabulary and descriptions of customs which are alien to the researcher’s culture. All of the texts in this collection have required elicitation consisting of follow-up checking and confirmation to ensure a clear interpretation of the data. The low frequency of some forms in text means that there is a need to question speakers about these forms and to attempt to elicit further examples of them. As many of the narratives involve the recounting of events in the past, there may be a bias to the description of the remote past or a past that is remembered nostalgically.

1.7.4 Genre

I have collected in a variety of different genres. There is a correlation between particular grammatical forms and genre, in that some forms are more prevalent in particular genre types (see for example McGregor 1990:31), e.g the Potential -em morpheme occurs frequently in hortatory texts.

33 Figure 2: Proportion of each genre as a proportion of the total number of texts: 50

dreaming hunting historical expository hortatory

I have made an effort to achieve a balance by including texts of a variety of genres. Although the text collection contains a broad range of data, there is a bias towards narratives as this is the most common genre and the easiest to record. There is limited conversational data. Hortatory texts are difficult to collect and are limited in the collection. Often samples of a genre are found embedded within another. Often the speaker’s purpose in using narrative text is not simply recounting events but expository; to inform the hearer about the ways in which things were done in the past. As much as possible I have attempted to distinguish locally-defined genre, for example altyerr, Dreaming story. A genre which is not represented in the data is aharlperr or Morning Discourse (Liberman 1985:4,5,102, Wallace 1979:149).

1.7.5 Elicitation

Elicitation can be a useful technique in gaining data. I have always consulted native speakers when transcribing texts and elicited further examples of forms which occur in texts. I have elicited English glosses as translations of textual material from informants. Elicited sentences through dictionary work in the preparation of the revision of the AED (forthcoming) have been a valuable source of data.

There are limitations to the use of elicitation and ‘leading questions’ discussed by linguists such as Strehlow (1947b:171). Strehlow (1947b:168) criticised the work of Spencer and Gillen who elicited Arrernte through English: ‘practically all their errors could have been avoided had either of the two authors possessed a thorough knowledge of any of the languages spoken by their informants’. Vaux and Cooper (2003:20) highlight the problem of ‘priming effects’. I have made use of naturally occurring language rather than attempting to elicit paradigms. I have also avoided acceptability judgements, safeguarding against the danger of eliciting what Payne (1997:369) labels ‘culturally nonsensical sentences’.

As much as possible I have used the Alyawarr language to gain information about grammatical forms. A problem with using the for elicitation is the ‘pseudo-intelligibility’

34 trap (Harkins 1994:182) where the researcher assumes that a form in a local English variety (Aboriginal English) has the same value as it would have in Standard Australian English. This imposes limits on the use of elicitation as a technique (Sutton and Walsh 1979:5). I have been able to converse with speakers monolingually in Alyawarr without often resorting to English. However, I make no claim to ‘native speaker’ proficiency and my conclusions have been tested extensively and triangulated with native speakers of Alyawarr.

1.7.6 Observed speech

Often I have only heard spontaneous utterances with the unaided ear without being able to record them. Often interesting and unusual forms have been heard which represent natural conversational data and which are not susceptible to the same problems which are associated with elicited utterances. I have usually written down the utterance and later confirmed it with speakers.

1.7.7 Translations

Translations into Alyawarr from other languages such as English have also been regarded as valuable data, often revealing forms which are non-existent or sparsely represented in other data. For example, I have referred to the Angka Mwerr-angker (2010) Bible translation and a range of other texts which have been translated since the 1970s. A limitation of the use of translations is the possible transfer of structures from the source text as the translator sought accurate translation from the source text. As TGH Strehlow (1947b:167) had found earlier, a number of words such as altyerr cannot easily be glossed. I have used texts that were translated freely without the use of ‘front translations’ and ‘key terms’ and therefore the circularity of a more fixed approach has been avoided (Chelliah 2001: 163). I have given less priority to translations because I have attempted to study the utterances which reflect the conversations that are occuring in Alyawarr society. The Alyawarr title of the Northern Territory government’s report Ampe Akelyernemane Meke Mekarle “Little Children are Sacred” should be a cautionary warning against accepting as data the kinds of utterances which Alyawarr-speaking people themselves do not actually use.

1.7.8 Introspective comments of speakers

Often speakers will have useful insights into the structure and function of their language. However, there are problems with using introspective statements. One problem is the impressionistic comment of a language informant that two forms are ‘the same’ which are similar but occur in different places in the data, have different functions and are distinct forms.

35 Informants may confirm or negate the researcher’s statements for a variety of reasons, including non-linguistic reasons, often for reasons of content rather than linguistic form. McGregor (1990: 36) was reluctant to use starred forms in compiling a grammar of Gooniyandi because, ‘speakers were as a rule unwilling to label utterances I produced as unacceptable’.

Gratuitous concurrence is a pervasive phenomenon of intercultural conversation in Central Australia (Liberman 1985:198) and occurs when an informant agrees with the researcher as a way of establishing and maintaining relationships in terms of Myers’ (1986) ‘sustaining relatedness’, rather than of evaluating propositional statements. Disagreement is a mark of individuality in Western societies but is not as valued in Aboriginal societies (Liberman 1985:216). Most important is the informants ‘desire to please their questioners- a desire that vitiates such evidence when given in a court of law- renders suspect some of the information gained in this way’ (Strehlow 1947: 170).

I have often encountered gratuitous concurrence during fieldwork. For example I recently found in answer to a question of linguistic acceptability, “Yeah, you can say that any time you want”. However the informant could not think of how he would use the form. For this reason I have used native speaker judgements and intuitions cautiously.

1.7.9 Learning the Alyawarr language

I began learning Alyawarr in 1995 when employed as a literacy worker with the Finke River Mission of the Lutheran Church, an organization which had been working in Central Australian Indigenous languages since 1877. Initially I recorded some speech and read some of the available literature in the language such as the AED and some of the church translations. Later I began using a more intentional approach to language learning. My language teacher was Banjo Morton, apmerek-artwey or senior traditional owner of Aherrenge country and a speaker of the Southern dialect of Alyawarr. Banjo had worked with linguists previously and emphasised the careful pronunciation of forms, and the use of the ‘correct’ or ‘really Alyawarr’ forms. I listened to some of the paradigm elicitation speech drills which had earlier been recorded at Alekarenge by SIL linguists although I didn’t make much use of them. I learned the language through ‘language experience’, travelling on a number of hunting trips with Peter Ngwarrey Morton (Aherrenge country), the Ross family at Irrultja and the senior women of Aherrenge. During 1996 I recorded stories about the hunting trips. Participant observation enabled me to learn the language in context. Peter Morton sketched a number of drawings and diagrams of the hunting trips. Photos were taken on these trips and often sentences relating to the photos and diagrams were elicited. During this time I also collected texts of altyerr traditional ‘dreaming’ stories with Peter Morton, who has a wide ranging repertoire of traditional stories and cultural knowledge.

36 The stories were illustrated by Peter Morton and made into books for teaching literacy, language and culture at the Ampilatwatja School. Susan Moore and Agnes Ladd worked with local women to illustrate books for teaching literacy at the school. After eighteen months of language learning I was actively involved in interpreting and translating. Simon Ross (Irrwelty) and Frank Turner (Aherrenge) helped me to understand the meanings of Alyawarr verb morphemes, giving me sentences with additional examples of these morphemes. Both of these men had experience as Bible translators, translating the Alyawarr Bible from Western Arrarnta and English. We began translating the Alyawarr Bible in August 1996. Over time I became more aware of the limitations and advantages of translating English texts into Alyawarr. In 1999 I recorded history stories with Donald Thompson Kemarr as part of ATSILIP (Aboriginal and Torres Strait Islander Language Initiative Program) Language Program funding which was accessed through the Institute for Aboriginal Development (IAD) and auspiced by the Aherrenge Council (Aherrenge Community report 1999-2000). Donald is a senior traditional owner of Antarrengeny country. An authority on Alyawarr traditions and multilingual speaker of many languages and dialects, he has been featured in Alyawarr language films such as Crookhat and Camphoo in 2004 (video, Alice Springs, CAAMA Productions) and Willaberta Jack in 2007 (video, Alice Springs, CAAMA productions). In 1999 I began collecting additional items for the AED. In 2003-04 Donald and other speakers worked on the Alyawarr Picture Dictionary project.

I commenced research for this project on grammar in 2005. In recent times I have been working as an interpreter for medical, educational, political, local government and legal organizations. I worked as a tutor for the Own Language program at Batchelor College. During 2007 and 2008 I researched language for the Cultural Signs project in Alyawarr (Turpin and Alyawarr speakers 2009). From 2006 to 2009 I researched the desert raisin project with the CSIRO (Walsh, Douglas and Alyawarr speakers 2009).

Interpreting has given me access to much spoken language which was often not recorded at the time. Widespread involvement in many domains of Alyawarr life and extensive travel in the Sandover, Southern Barkly and Plenty regions has enabled me to gain additional knowledge of the language in use and to confirm and check data gathered in a variety of contexts.

1.7.10 The roles of language consultants

I have recorded a variety of Alyawarr speakers of different ages, sexes and dialects in order to achieve a balance which reflects the demographic realities of the Alyawarr language. The age range of informants was from the thirties to the seventies. Although some forms are only used

37 by older speakers, the grammatical forms appearing in this grammar are in common use by Alyawarr speakers.

Different Alyawarr speakers have fulfilled different roles in writing this grammar which reflect their different talents and interests. The analysis of Alyawarr verb morphemes depends upon accurate translation. I have depended upon consultants to translate the example sentences from Alyawarr to English. As much as possible I have checked translations and ‘triangulated’ the results with speakers other than the original speaker. Speakers have been engaged in a long iterative process of checking and confirming data and better understanding the structure of the language.

1.7.11 Recording techniques

All texts were recorded using a high quality digital recorder. The first period of fieldwork involved collecting texts. In the second period of fieldwork I began to record elicitation sessions and the comments of speakers (McGregor 2008:424 note 17) in order to provide additional contextual background for the text. I was also able to capture the comments of speakers on the recorded material. The wave files from the digital recorder have been saved on DVDs. Earlier tape recordings of Alyawarr texts and elicitation sessions were digitized in 2006 and the earlier transcriptions of these texts were revised.

1.8 The outline and scope of this work

Throughout this thesis, I have categorized verbal morphemes according to formal structural morphological criteria rather than the notional criteria employed by traditional grammar. In addition to formal criteria, functional or semantic criteria can also be used to further categorize the morphemes which make up formal categories, particularly where different morphemes appear to have the same distribution within the verb. Each chapter covers a structural position in the verb, roughly from left to right. Chapter 2 outlines the basic structure of the Alyawarr verb and three processes of word formation. In Chapter 3, derivational morphology and number are discussed in detail. Chapter 4 concerns motion, aspect and verb compounding. The final obligatory morphology of the verb is treated in Chapter 5.

38 2 The Alyawarr Verb

2.1 Introduction

Verbs were introduced as a primary syntactic category in §1.2.1. In this chapter I discuss the structure of the Alyawarr verb in terms of its grammatical categories. The morphological structure of Alyawarr verbs involves a set of categories, mostly determined upon the basis of the distribution of markers within the verb. The categories are realized by the processes of affixation, compounding and reduplication.

Following chapters each describe a particular grammatical category within the verb. Some details of the Alyawarr verb are covered in Yallop (1977:48-67). However, I have made a further analysis, refining and extending the description.

2.2 Verb structure

The morphological structure of the Alyawarr simple (i.e. non-compound) verb is shown in Table 4. I consider complex verbs such as in §2.6. Two positions in the verb are obligatory. The verb root conveys the lexical meaning of the word. The final position takes tense, aspect, mood (TAM) and clause joining suffixes. These are discussed further in Chapter 5. Various non-obligatory elements occur between the root and the final inflection. In the following verb structure diagrams, obligatory categories are marked {} and non-obligatory categories are marked ( ).

Table 4: Alyawarr Simple Verb Structure

------Non-obligatory------Obligatory {verb root} (Derivation) (PATH) (Number) {TAM, inter-clausal}

The most basic form of the verb is a root with a final inflectional suffix:

(1) atw-ek hit-PP {vroot} {TAM} ‘hit’

39 Derivational suffixes which occur after the root are optional and have a syntactic function, changing the transitivity of the verb in which they occur, for example:

(2) atw-err-eyel hit-REC-PRES {vroot} (DERIV: reciprocal, detransitivising) {TAM} ‘hitting each other, fighting’

Derivational morphology is explored further in Chapter 3. The other category of the simple verb structure involves PATH morphemes which describe motion along a path, as in (3). PATH morphemes are described in Chapters 4.

(3) ar-enh-ek see-POINT-PP {vroot} (PATH) {TAM} ‘saw at a point along a motion path’

Complex verbs may be distinguished from syntactic sequences. In compound verbs as in (4), the ordering of elements is fixed and the initial element must precede the second part of the compound2. This can provide evidence for distinguishing phrases from compounds in some contexts though elements of some phrase types are also strictly ordered. Most non-obligatory categories have both compounding and simple affix markers.

(4) ar-enh-ey-err-alp-enh

see-POINT-BASE&DO1-PL-BASE&DO2-IMPP

{vroot1} (PATH) (MOTION1) (NUMBER) (MOTION2) {TAM} ‘they would see while returning’

2.3 Properties of Verb Roots: Transitivity

The most significant property of the verb root is transitivity. Verb roots are strictly classified as either transitive or intransitive (Yallop 1977:124). The argument structure of the clause is determined by the transitivity of the verb. A transitive verb in Alyawarr is defined as one which has the potential to occur with an ergatively-marked NP as in (5). The term ‘potential’ is used because the ellipsis of case-bearing arguments is common and not every clause will show the morphosyntactic properties that indicate its transitivity. Transitive roots are also distinguished

2 Alyawarr clauses have a relatively free order of the major constituents and all orderings are possible, although AOV is the preferred ordering in transitive clauses.

40 on morphological grounds by being able to occur with Mediopassive and Reciprocal markers. An intransitive verb cannot occur with an accusative argument or an ergative one but can occur with at least one argument in nominative case, as in (5). Any additional arguments specified by the verb will be in dative case. Transitivity is altered by derivational affixes which follow the root, for example -elh ‘Mediopassive’ (intransitivizing) or Transitive Verbaliser -elhil (transitivizing). Derivational morphology is further discussed in Chapter 3.

(5) artwa rtn-eyel man:NOM stand-PRES ‘A man is standing’

(6) Arrwekeleny-el renh arlkw-enh ancestors-ERG 3S:ACC eat-IMPP ‘The ancestors would eat it’

Wilkins (1989: 224) reports a further ambi-transitive class of roots in Arrernte, with just two members amp- ‘burn’ and wern- ‘blow', which can be both transitive and intransitive. I have not found ambitransitive verbs in the Alyawarr data. Alyawarr ampeyel ‘burn’ is exclusively intransitive: it does not occur with Ergative or Accusative arguments. It can however occur with a Locative argument marked by -el, as in (7), which is homophonous with the Ergative marker. Since such an argument can co-occur with a Nominative NP but not a distinctively Accusative one, the simplest analysis is Locative not Ergative case. Note that the S pronoun ayeng in (7) is Nominative rather than the Accusative ayenh. The locative analysis is in agreement with Yallop’s glossing of similar examples (1977:51). A further argument that amp- is intransitive, is that transitive amperneyel ‘cook’ is straightforwardly derived from this root with the transitivising -ern.3 See §3.6.3.

(7) Itwern-el ayeng ingkety amp-eyel heat/sun-LOC 1S:NOM foot:NOM burn-PRES ‘My feet are burning from the heat of the sun (on the ground)’ Literally: ‘I feet are burning’.

2.4 Verbal Sub-categories

Intransitive verb roots can be further categorized on morphological grounds. Compounding roots are defined on two grounds: (i) they can occur as the second element in Verbal Compounds §2.5.3.1 (Yallop 1977:62) and (ii) they take Plural allomorphs which are distinct

3 Note that the same derivation occurs in Arrernte where amp- is analysed as ambi-transitive.

41 from the regular -err allomorph. These Compounding roots are subcategorized on two further grounds: (i) the specific Plural allomorph they take: -erl.iw with the two Basic Motion roots and the reduplicative RED.ew with the two Basic Stance roots, and (ii) the ability to occur with Path markers. These classifying properties are represented in Table 5.

Table 5: Verbal Subcategories

Class Compounding roots V2 in VV Verbal Plural Path affixes Compounds Basic Motion alh- ‘go’ Y -erl.iw N alp - ‘move to base’ Basic Stance an- ‘sit, exist’ Y RED.ew Y aynt- ‘lie, exist’ General (Red.plural) Specific roots N RED.ew Y General (Other) Remaining roots N -err Y

2.4.1 Basic Motion verbs

Basic Motion roots alh-‘go’ and alp- ‘move to base’ form verbal compounds that are semantically compositional, that is, the semantics of the verbs contribute to the meaning of the compound in a transparent and straightforward manner. Basic Motion verbs have the same meaning in their compounding role that they have in free verbs, as in (8), contra Yallop (1977:62) who claims that in their compounding role ‘they seem to convey aspectual distinctions rather than their independent senses’.

(8) “Atheperr-el ayeng alp-erl.aynt-eyew.” cool-LOC 1S:NOM back-CONT2-PURP

aylp-ey.alp-enh kwaty-warl enter-BASE&DO-IMPP water-ALL ‘ “I’ll go back to the cool!” It went back into the water’

Understanding the semantics of the Basic Motion verbs is crucial to understanding the semantics of Motion compounds. The discussion in the following sections focuses on the motion parameters that are central in Alyawarr as a preliminary to more detailed discussion of each compound in Chapter 4.

42 2.4.1.1 Base and relationship to the subject

The two Basic Motion roots, alh- and alp- depend on the notion of a base but describe complementary types of motion. Alh- can be explicated as ‘motion away from a base’, and alp- as ‘motion to base to stay’. The distinction between them warrants some discussion since this analysis differs from other sources on these roots in Alyawarr and on the corresponding roots in other Arandic varieties, and because they can be translated in context in ways which do not clearly reflect their meanings.

Like the languages discussed by Elson and Pickett (1988:25), Alyawarr Basic Motion verbs depend on the direction of the subject’s motion relative to a reference point. The frequent use of the Basic Motion verbs reflects the importance of both travel and the ‘base’ notion in Alyawarr culture. Lewis (1976: 262, as cited by Wilkins 1989:283) speaks of a ‘mental map’ used by Aboriginal people in the central desert region which was continually updated ‘so that the hunters remained at all times aware of the precise direction of their base and/or objective’. The two kinds of motion paths expressed by these roots are common in the narrative accounts of Alyawarr speakers. The alh- motion includes hunting day trips where food would be gathered and brought back to the apmer base, which is called ‘camp’ in local English. This movement is referred to as awangk alheyel, a ‘day trip’. The alp- motion path is to a base which the subject has travelled from or to a new base, and it can often be translated as ‘return’ or ‘move to’. Fairly frequently in traditional lifestyles there would be a move to establish a new base, a situation which still exists despite the more sedentary nature of Indigenous lifestyles on the larger communities in recent decades.

In both cases, the critical issue is the motion of the subject in relation to the base. Only in the speaker-oriented Deictic Motion compounds (§2.5.3.2) is the speaker’s position relevant. This contrasts with the corresponding Arrernte basic motion roots as analysed by Wilkins (1989) and Wilkins and Hill (1991) in which Arrernte lh- and alp- both entail motion away from the speaker (1989: 268).

2.4.1.2 Alh- ‘go’

Alh- expresses the subject of the clause moving along, optionally to one or more specific destinations. It does not specify that any destination will be a new base or the base from which the motion started in (9). There is an implied sequel of an alp- motion to a base. The subject of the clause is going away from the base on a trip and reaching a destination with an expectation of returning to its original base in (10).

43 (9) Aker weth-ilkwer rernem arlkw-enh arrwekeleny-el-rnem game that-SEMB 3PL:ERG eat-IMPP ancestor-ERG-NPL

pwety-el alh-erl.iw-enh-el bush-LOC go-PL-IMPP-LOC ‘They would eat game meat like that when they were walking around in the bush’

(10) Kel ra arlewatyerr-ek alh-enh arleng akwerlp-warl then 3S:NOM goanna-DAT go-IMPP distant sandhill-ALL ‘Then the boy would go out for goannas a long way to the sandhill country’

2.4.1.3 Alp- ‘move to base’’

Alp- may be considered to have a single sense ‘move to base’’. Since the base will frequently be the one from which the subject initially departed, alp- is often translated as ‘go back’ or ‘return’, as in (11) where the action follows an alh- movement. AED (Green 1992:27) only lists such ‘return’ senses - ‘go back, go home, return’ - but alp- is not only ‘return’.

(11) Apmer-warl alp-ew home-ALL back-PP ‘went back and stayed at home’

The base that is moved to may also be a new one, as in (12). The verb alp- is also applied when the Subject moves to a place where they die, or in the case of mythological characters, end their time travelling on the earth as in (12). The permanency of such a move can be emphasized, for example as: alpew intem=antey=anem ‘then moved away to stay at another place permanently’.

(12) Ratherr=ap alp-ek =anem apmer ingwer-warl=anem 2D:NOM=CONJ back-PP=THEN place another-ALL=THEN ‘Then those two moved away and stayed at another place’

44 (13) Awey nhaym=an atha il-em altyerr ra kwaty boy This:NOM=FOC 1S:ERG say-POT Dreaming 3S:NOM water

apeynt=ant-warl alp-enh spring=JUST-ALL back-IMPP ‘I shall speak about boy who finished up at a spring’

Alp- can describe certain situations that involve movement away from a reference point such as the speaker.4 In fact, Yallop (1977:143) defines alp- as ‘go, move, walk’ (away from speaker or point of reference, hence often: ‘go away’)’. However this can be related to the basic meaning ‘move to a (new) base’ given that where the participant is at a particular location, moving to a base will necessarily involve movement away from that location and the fact that the destination is a base means that it will be a substantive or even permanent move. In (15), for example, the cattle were grazing in the open and might escape and not return to their owners. Alp- is often used in an imperative form to send someone away. Dogs and children are commonly chased away with the imperative alpay! (Yallop 1977:53,143). This reading is well illustrated in (14), where the traditional owners who have always lived in a region, and belong there, are being chased away from their land.

(14) Alp-enharey wenh! Alp-enharey apmer nhe-they! away-IMP.PL QUOT back-IMP.PL place this-ABL ‘Get away! Stay away from this place!’

(15) Ratherr=ap yarraman-akert=anem alh-enh=anem pwelek 3S=CONJ horse-COM=THEN GO-IMPP=THEN cattle

rernem arleng alp-ekerr 3PL distant back-APP ‘And then they were going out with horses in case the cattle strayed ’

Alp- also has the extended meaning ‘to recede away to a distant point’ which is used of inanimate objects extended horizontally in space such as roads and pipes. Such objects lead away from the reference location and continue receding away to an endpoint which is some

4 Strehlow’s comment, note 36a on page 36 of Diary 17 (1953) seems to support this reading.

45 distance away. This resembles the ‘fictive motion’ described in Talmy (1996) and Levinson and Wilkins (2006:52).5

The critical factor in determining which Basic Motion verb is used is the overall destination of the Subject of the clause. Alyawarr resembles the languages which contrast one-way motion versus round-trip motion (Elson and Pickett 1988:26). In the Zapotec languages there is an affix which occurs with motion verbs that expresses incomplete action, and which resembles the distinction between alh- (a motion which implies a return and a round trip) and alp- (one way motion).

2.4.2 Basic Stance verbs

As we have seen, the Basic Stance roots, an- ‘sit’ and aynt- ‘lie’, are a subclass of Compounding roots defined by their selection of the RED.ew Plural allomorph. They combine two functions: (i) they denote the stative physical configuration (stance, posture, orientation) of the human body and other objects and (ii) they have copular functions in equative and attributive clauses and are used in existential constructions. They share this combination of functions with two other roots, on which basis they together constitute a wider semantic category of Stance-Existence verbs.

Table 6 shows the properties of the Stance-Existence roots. An-, aynt- and rtn- form lexical compounds (§2.5.2.1) with monomorphemic non-verbal elements which have low productivity. Like the Basic Stance roots, rtn- takes an irregular Plural allomorph, in this case -enherr. However unlike them, rtn- doesn’t occur in verbal compounds. The fourth root, altywen- doesn’t occur as a compounding element in either type of compound. In contrast with the Basic Motion roots in compounds, compounds formed with stance-existence verbs are less semantically compositional.

Stance verbs can be applied to all animate and inanimate objects depending upon their physical configuration. Inanimates do not control their configuration though humans can adopt any of the basic postures.

An existential sense of the Stance-Existence verbs has a reading of ‘being’, describing the existence of different Subjects and having a function which is similar to the various forms of

5 The notion of extension along a path may also account for the Arrernte compounding marker for ‘extended action along a path’ which involves a monosyllabic reduplicant and rle-alpe (Henderson 1998:236).

46 English ‘be’. Mountains are said to be altyweneyel ‘crouching’. Various objects which have a vertical orientation are described as rtneyel ‘standing’ such as trees, fenceposts, cars and cattle. Objects which rtneyel ‘stand’ are not necessarily tall. For example, the ground-hugging iylaw plant. Bodies of water such as waterholes are described as rtneyel.

Verb Gloss Applied to Lexical Verbal Root e.g. compounds compounds an- ‘sit’ people Y Y aynt- ‘lie’ snakes, Y Y roads, traditions rtn- ‘stand’ trees, Y N cattle, waterholes altywen- ‘crouch’ hills N N

Table 6: Alyawarr Basic Stance verbs

The aspectual uses of -an and -aynt as ‘continuous’ in compound verbs (Yallop 1977:62) are plausibly related to the ‘being’ existential sense of the independent verb. Similar to the situation reported for Arrernte (Wilkins 1989:221), in clauses like (16) the default interpretation in the absence of an existential verb is present tense. The existence verb is only necessary for a temporal situation which is non-present. An attributive function which predicates a quality can be regarded as copula (Dixon 2002:239-42, Wilkins 1989:221). However, a Stance verb can only be a pure copula when the Subject is a more abstract entity which has no physical orientation.

Stance-existence verbs convey states rather than accomplishments, which have terminal points. This is in contrast to English where, for example, ‘sit’ can have both dynamic and static interpretations (Newman 2002:4). The sentence ‘I sat on the chair’ could mean ‘I sat myself down on the chair’ (dynamic, accomplishment) or ‘I was sitting on the chair’ (and didn’t move) which is stative. Alyawarr uses morphological means to convey a dynamic ‘change of posture or state’ event, as shown in Table 28. This Aktionsart change will be further explained in §4.7.

2.4.2.1 Aynt- lie

This root describes the human ‘lying’ configuration and its equivalent for other species which can adopt a corresponding position, as in (16). More generally it can refer to objects which are physically extended horizontally. Extended senses are ‘sleeping’ or ‘staying over one or more

47 nights’, as in (17). The stative Aktionsart and ‘staying’ sense are related to the use of -aynt to convey an aspectual sense of extended duration in compounds §4.9.

(16) Aherr yanh-ey lyw-el aynt-eyel kangaroo there-EMPH shade-LOC lie-PRES ‘The kangaroo is lying in the shade’

(17) Aynt-ey.alp-ew anwenakerr ingwer-el=anem=arl lie-BASE&DO-PP 1PL.CP.EX other-LOC=THEN =SUB ‘Then we went and stayed at another place’

The verb applies to objects which have a natural shape or posture imagined as long, for example, snakes in (18). Certain inanimate objects are also said to be ‘lying’, for example, roads, tracks, wells and soakages6 as in (19).7

(18) Pwerl-el ar-erl.alp-ek warl ikwer-itwek aynt-enh=antey=arl Pwerl-ERG see-DO&RET-PP house 3S:DAT-LOC lie-IMPP=STILL=SUB ‘Pwerl went and saw it (snake) while it was staying in the house and returned.’

(19) Artarn=arl ikwerenh aynt-enh weth-ilkwer warl-they track=SUB 3S:POSS lie-IMPP that-SEMB house-ABL ‘His tracks were leading away from the house’

Aynt- can also refer to abstract entities which are conceived of as holding over an extended area. Both extension in space and extension in time are possible understandings of aynt-. There appears to be a logical connection between these spatial and temporal senses. Abstract entities like names, languages and laws8 can be seen as extending across the region and apply to those who live in that region, as in (20). Often the temporal extension sense of aynt- is used to refer to ‘tradition’ or ‘the way things always were’. Laws (21) may be thought of as having traditional authority which extends back in time and perhaps their extension is through time and generations rather than just through space.

(20) Pwerlany angka aynt-enh-ey.alp-ew Avon Down-they Pwerlany language lie-POINT-BASE&DO-PP place-ABL ‘The Pwelany language went along from Avon Downs’

6 A soakage is a kind of shallow well or depression in which water collects 7 For roads and tracks at least there is some degree of overlap in aynt- and the ‘recede away into distance’ sense of alp- in that the objects are extended in space. Compounds formed with –alp can also involve spatial extension or ‘fictive motion’ (§2.4.1.3). 8 These entities may be laid down in journeys of mythological beings.

48

(21) Government arrwekeleny-wenh=arl aynt-ew alakenh-anyem=arl government olden days-POSS=SUB lie-PP like-ROUND=SUB ‘In the olden days the government way (law) was like that’

2.4.2.2 An- sit

This root describes the human ‘sitting’ configuration, as in (22) and its equivalent for other species such as dogs which can adopt a corresponding position. More generally it can refer to entities which are not extended either horizontally or vertically, for example frogs. The root also has a more existential sense, describing the normal existence of humans i.e ‘staying’, ‘living’ as in (23). The stative aktionsart of an- contributes to the complex State marker erl.an §4.9.2.

(22) Itwern-wety anantherr lywenty-el an-eyel sunshine-REAS 1PL:NOM shade-LOC sit-PRES ‘ We are sheltering from the sun, sitting in the shade’

(23) Arrwekeleny-rnem alakenh-anyem an-enh alakenh=an ancestor:NOM-NPL like-ROUND sit-IMPP like that=FOC ‘That is how the early day people lived/were, like that’

2.5 Compound Verbs

Compounding involves the combination of lexemes to form a single morphological word. The morphological word consists of one or more phonological words. The major types of compounds are seen in Table 7.

Table 7: Types of Alyawarr compounds

Compound type initial compounding compositional Lexical compound preverb verb No unanalysable Lexical compound preverb verb Yes analysable Verbal compound verb verb Yes

49

The complex verb in Alyawarr consists of:

Initial element------Compound {preverb} {v root} (DERIV) (PATH) (SEQ) (NUM)} (v root){T-A,M, clausal}

The initial element may consist of a preverb, forming a lexical compound. Verbs can be multiply compound and involve more than one type of compound. For example the compound in (24) is formed from both a lexical compound (arntarnt.ar-) and a verbal compound (erl.alp-).

(24) Arntarnt.ar-erl.alp-eyel look after-DO&RET-PRES ‘Look after and then go back’

The transitivity of the initial verb root determines the transitivity of the compound. The number marking of the compound will follow the number marking of the initial root. A number of elements that appear similar to compounding verbs have meanings which are not resolvable into the meanings of their component parts. I have handled them as affixes. While they may have developed from free verbs, this is not the preferred synchronic analysis. They are regarded as single grammatical morphemes without a clear relationship to the free verbs.

2.5.1 Other evidence for compound status

In the following sections I outline the criteria which define compounds in Alyawarr versus both non-compound words and syntactic sequences of distinct words.

2.5.1.1 Grammatical wordhood

A compound is a grammatical word and therefore has the general properties of cohesiveness of its constituent elements, fixed order of those elements and conventionalized coherence and meaning (Dixon and Aikhenvald 2002:19). However compounds often exhibit mixed behaviour that reflects something of their two faces. Compound verbs in Alyawarr satisfy the requirement of a fixed order of the constituent elements relative to each other. This distinguishes them from syntactic sequences because although the preferred constituent order for clauses is for adverbs and object nominals to precede the verb, it is possible for them to follow the verb or for object nominals to be ellipsed.

50 Compound verbs generally meet the requirement of cohesiveness except that certain non-verbal material may intervene between the compounded elements, most frequently in Verbal Compounds (§2.5.2.3) , as in (25) where the clitic =arl intervenes. The possible intervening material is highly restricted, most often a clitic or particle which can alternatively occur at the end of the entire compound. This suggests a word-like boundary within the overall compound and intervening material, for example, clitics, is evidence for two phonological words. There is no constraint on the number of in each of these phonological words. Clitics =arl SUB and =anem THEN occur most commonly. There is likely to be an implicational relationship such as that reported for EC Arrernte (Henderson 1998:275), in that word-final clitics will be permitted to occur if other intervening material can follow the verb stem. More research is required.

(25) Rntern-enh=anem irrtyart-el kwaty-itwew=an antyw-ey=arl-alh-enh spear-IMPP=THEN spear-INS water-LOC=FOC drink-GO&DO1=SUB-GO&DO2-IMPP ‘They would spear (them) when they were going to the water to drink’

2.5.1.2

Stress placement marks the phonological word in Alyawarr and provides evidence for compound status versus a syntactic sequence of words on one hand and a single non-compound word on the other. Yallop (1977:46) notes that stress can differentiate two words when there is no segmental distinction between a single morphologically complex word and a sequence of two words. In non-compound verbs, secondary stress falls on alternating syllables after the primary stress. However, in compounds such as (26) the position of the secondary stress indicates that the second compound element, here alpew, appears to constitute a distinct domain of stress placement as opposed to a continuation of the alternating stress pattern. Brackets indicate the compound boundary in the examples. Where there is intervening material, individual clitics retain their original stress regardless of the number of syllables, as in (27) and the stress placement in the following verb stem of the compound is independent of any intervening material.

(26) Apmer (mpw r-ey)-(alp- w) camp make-BASE&DO-PP ‘Went away and made camp’

51 (27) Artwa ikwer-rnem-wety (aylp ny ant nem arl)-(irr- nh) man 3S:DAT-NPL-AVERS sick of=JUST=THEN=SUB-IV-IMPP

mpwernaynewantherr-wety. their spouses-AVERS ‘Then they were just getting sick of the men, of their husbands’

2.5.1.3 Compositionality

Semantic compositionality constitutes a criterion for recognizing compounds in Alyawarr. Crystal (1997:77) defines compositionality as ‘a hierarchical model of linguistic structure in which larger units are seen as being composed of smaller units’.

Compounding is a gradient phenomenon in both productivity and compositionality. At one end of the cline are lexical compounds which are limited in productivity and non-compositional. In some cases the interpretation will be quite specific and restricted to a fixed combination of verb roots. A number of elements that appear similar to compounding verbs I have handled as affixes. While they may have developed from free verbs, this is not the preferred synchronic analysis. They are regarded as grammatical morphemes without a direct relationship to the free verbs.

In Lexical Compounds, semantic compositionality correlates strongly with productivity, i.e. compositional types are likely to be productive. The largest number of compounds which can be formed with an analysable preverb is the irrtyert ‘steal’ compounds as in (28) of which thirteen have been recorded. Less compositional types are likely to be less productive as is the case for the preverb iytarr in (29).

(28) irrtyert-ineyel hidden-get ‘steal’

52 (29) iytarr-ineyel ?-get ‘drag out’

Verbal Compounds are the most semantically compositional and the most compound-like and are discussed in more detail in §2.5.3.1.

2.5.2 Subclassification of compounding types

2.5.2.1 Lexical compounds

Lexical compounds consist of a non-inflecting and monomorphemic non-verbal preverb which precedes an inflecting verbal element. Lexical Compounds vary widely in their compositionality. According to Yallop (1977: 65), some compounds may be more easily understood as lexical items composed of elements which are unanalysable, having ‘no separate existence as a word’. Certain adverbs and nominals can occur as the preverb element in a similar way to that for Warlpiri (Nash 1980: 42), however there are a large number of preverbs that are not derived from any other category and cannot occur without a verb from a small set selected by the given preverb. The following verb roots occur in lexical compounds:

Table 8: Verbs occurring in lexical compounds alh- go alp- move back, move away akwern- put into aylp- enter rtn- stand aynt- lie an- sit iw- throw, leave behind arrern- put in- get rnkern- stand up ar- see akng- carry athen- lay something down atw- hit angk- say

53 The inflecting elements in lexical compounds occur with the same obligatory final inflection and are analysed as identical with the corresponding free verbs. Evidence that they are related to independent verb roots is that they occur with the same distinctive plural marking allomorphy as a free verb, at least for those corresponding free verbs that take an irregular Plural allomorph. An example is (30). Most of the corresponding free verbs take only the regular Plural allomorph, so that if the inflecting verb in the Lexical Compound also takes the regular Plural allomorph, this may just be due to it being the regular allomorph.

(30) Irrtyert-alh-erl.iw-enh hidden-go-PL-IMPP ‘going along hiding’

Analysable preverbs optionally reduplicate as for example (31) and there are a few examples of unanalysable preverbs which reduplicate. Preverbs are the category most likely to reduplicate in compounds.

(31) rlweth-rlweth-ak-eyel ‘chopped into small pieces’. irrtyert-irrtyert-alh-eyel ‘running away and hiding’

When the preverb also occurs as an independent lexeme such as the examples given in (32), there is an issue of how lexical compounds can be distinguished from an object nominal O+V.

(32) irrtnya ‘skin, hide’ irrtnyeweyel ‘skin something’ ankwa ‘sleep’ ankwatheneyel ‘lay down to sleep’ arlwar ‘circular’ arlwarertneyel ‘swell up’ arlpwart ‘uncovered’ arlpwart-rtneyel ‘uncover, expose’

The preverb always precedes the verb in lexical compounds whereas object+verb or adverb+verb sequences can occur in either order. A lexical compound is shown in (34), with example (33) something close to a paraphrase. In (33), the verbal element precedes the object nominal. In (34) itna appears to be phonologically incorporated into an intransitive mediopassive verb which only requires one argument. It is not possible for the itna part to follow the welhew part. This constitutes evidence that itnewelheyel be regarded as a lexical compound and not as a verb and object nominal. Preverbs can be distinguished from adverbials, again by relative order in relation to the verb.

54 (33) Arrengikw r=an Antywemp-areny arrpemarl renh=arl grandfather 3S=FOC place-ASSOC also 3S:ACC=SUB

rnkern-ew itna. put in-PP name ‘His grandfather was from Antywemp as well and that is why he got the name’

(34) Artwa ra itn-ew Antywemp arrpemarl r=arl man 3S name-DAT name again 3S:NOM=SUB

itn-ew-elh-ew=anem name-leave-MED-PP=THEN ‘He was Antywemp as well and that is how he named himself’

Some preverbs occur as independent lexemes outside of Lexical Compounds, for example those which involve a body part nominal followed by the -el LOC marker with akngeyel ‘carrying’ such as the items in (35).

(35) artep ‘back’ artepel-akngeyel ‘carrying on back’ nthek shoulder nthekel-akngeyel ‘carrying on shoulder’

The compounds act as hyponyms of the free verbs in (36). The preverb doesn’t occur outside Lexical Compounds.

(36) rtwapety-aneyel ‘sitting with legs folded back to one side’ aymparrk-aneyel ‘sitting with legs crossed’

The lexical compounds shown in (37) are less compositional semantically, that is to say the base is not transparently related in meaning to free verb roots.

(37) arntarnt-areyel ‘watch over, care for’ areyel ‘see, watch’ aynterrar-iweyel ‘hang something up’ iweyel ‘throw, leave’

An example of an unanalysable compound is iterl-areyel ‘knowing’. The iterl appears to be a cranberry morph as it does not occur independently. Iterl may be related to iterreyel ‘think’ and areyel ‘seeing’ though there is no regular process that would account for the formation.

55 Speakers don’t seem to identify the two verbs as composed of independent lexemes. An alternative analysis for this is, as in Arrernte, it-el-ar ‘throat-LOC/INST-see’ (Wilkins 1989:267), parallel to another cognition verb it-irr- as ‘throat-IV’. Some clitics are able to intervene between the preverb and the verb, for example =anem ‘THEN’ in (38), indicating its status as a compound.

(38) it-erl=anem-ar-enh

think-DO&SEE1=THEN-DO&SEE2-IMPP ‘Then (they) would know’

The productivity of preverbs varies. Most preverbs are unique to a single compound. The highest number of compounds formed from a preverb found in the data is thirteen which are based upon irrtyert in (28). The compounds which are based on tek ‘dry something out’ are shown in (39).

(39) tek-ampeyel ‘dry out’ tek-arrerneyel ‘hang out to dry’ tek-iweyel ‘put out something to dry’ tek-aneyel ‘sun, sunbake’

2.5.3 Compounds based upon IV and TV

The verbalising morphemes Intransitive Verbaliser (IV) in (40) and Transitive Verbaliser (TV) appear to have the most word-like properties of all the compounds. They are highly productive and compound with a large number of nominals.

(40) Aylpeny-irr-enh=anem=arl ikwer-rnem-wety. get sick of-IV-IMPP=THEN=SUB 3S:DAT-NPL-AVERS ‘Then they would get sick of them’

An indication of high productivity is that they combine freely with loanwords from English:

(41) change-irr ‘change’ gallop-irr ‘gallop’ yardem-il ‘yard’ loademup-il ‘load up’ registerem-il ‘register’

56 These verbalising morphemes often occur with intervening material, similar to the case reported for Arrernte (Henderson 1998:275). Non-verbal material can intervene between a nominal/adverb root and the TV marker, such as the intervention of clitics in example (42) and (43). This suggests that the base and verbaliser parts can constitute separate prosodic words or are prosodically independent in some way. There is more restricted variation in the constituent ordering with lexical compounds than is possible with verb phrases. Verbalizing suffixes occur after certain Obligatory suffixes: ey-angenhil §5.2.5.2 and etyek-irr §5.2.3.

(42) Rernem arlk-elh=ant=anem=arl-il-enh renh-rnem.

3PL call out-TV1=JUST=THEN=SUB-TV2-PP 3ACC-NPL ‘Then they just called out about them’

Further evidence of their compound status is that Verbalisers may carry primary stress and can occur with a reduplicated preverb:

(43) Althart ikwer akarr-akarr=anem-il-ew ceremony 3S:DAT gathered-RED=THEN-TV-PP ‘Then they gathered them up for the ceremony’

2.5.3.1 VV compounds

Verb-verb compounds introduce an additional verb root, nearly always in combination with a specific suffix which precedes the compounding root, and with a compound boundary between the two. The brief account of verbal compounds given here will be expanded upon in Chapter 4.

Table 9: Types of verbal compound

Compound type Initial verb root Compounding verb root Non-deictic various Motion and Stance Deictic ap-ey, anp-erl Motion

Compounding roots carry mostly the same inflections as their free verb counterparts but only the last in a given word bears the obligatory morphology for the entire compound, as in (44).

(44) Arrathen-enh-ey.alh-ew load spear-POINT-GO&DO-PP ‘(We) went and loaded spears into the woomeras along the way’

Although uncommon in the data, the occurrence of intervening material provides evidence of a single compound containing two phonological words. 57 Certain clitics, Plural number -err/-arr and DISC -am can intervene between sequence suffixes and compounding Motion verb roots. See §3.7.1 and §3.8. The insertion is taken to be indicative of a phonological word boundary.

Non-deictic verbal compounds based upon Basic Motion verbs are semantically compositional. I discuss Motion verb compounds further in Chapter 4. Although less compositional, the Basic Stance compounding roots retain some of the meaning that they have as free verbs. Different senses of the free Stance verbs are in focus within different compounded forms. For example, -aynt can denote ‘motion’ or ‘stance’ depending upon the particular configuration of morphemes of the compound.

Table 10: Alyawarr Basic Stance compounds aynt- ‘lie, stay’ -erl-aynt ‘continuous aspect’ -ey-aynt ‘rise and do’ an- ‘sit, live’ -erl-an ‘stative’

2.5.3.2 Deictic Motion compounds

The Deictic Motion Compounds have roots ap- ‘arrive, come’ and anp- ‘depart’ which do not occur independently. They are analysed as verb roots on the basis of their occurrence with a similar range of intervening material as occurs with other Motion compounds. These compounds have a high frequency of occurrence in texts.

Yallop (1977:65) claims that apey-alhem is not a compound because the apparent root ap- in the compound is not semantically identifiable with the ap- root which occurs in simple verbs and which he claims means ‘rub’ in Alyawarr. However in Arrernte (Koch 2001:80) the root is found in Arrernte petye- ‘come’ (where -tye- contributes the ‘hither’ direction). According to Koch it is also probably found in the frequentive affix -pe- and with -erl-ap (§4.6.1) with motion-related meanings9 so it is clearly inherited from proto-Arandic. Clearly, apey-alp and apey-alh are semantically compositional with the compositionality of alp- and alh- as I will further explain below. Although ap- is analysable, it can be regarded as having a ‘defective paradigm’ consisting of only two forms having a sense ‘arrive at speaker’.

9 The verb root ap- doesn’t occur in the AED (1992) nor in the recent dictionary extensions file (D.Blackman, p.c). The translation which speakers cite for ‘rub’ is aperneyel which appears in Yallop’s wordlist (1977:144) as ‘paint’. Apeme ‘rub, paint’ (Henderson and Dobson 1994: 180) occurs in neighbouring Arrernte dialects (which use apetyeme for ‘come’). I conclude that the Deictic Motion verb apey-alhem and apeme ‘rub’ are unrelated. Further discussion of apeme~aperneme is in Henderson (1998:131). 58 Note that apetyeyel and apey-alheyel appear to be equivalent and are probably lectal variants. Comparing akngety- ‘bring’ and akng- ‘carry, move’, suggests a suffix -ety HITHER. On that basis, apety- ‘arrive’ can be analysed the same as Arrernte ap-ety- ‘go-HITHER’ (Wilkins 1989:269). The -ety suffix attaches only to these two verb roots.

Table 11: Deictic Motion compounds

Motion carrying apety akngety ‘come towards speaker’ ‘bring towards speaker’ alh apey-alh ‘go away from base’ ‘come towards speaker’ anperl-alh akngey-alh ‘depart, move away’ ‘take away from speaker’ alp apey-alp akngey-alp ‘move to base’ ‘arrive back towards speaker’ ‘take back to speaker’ ‘come to speaker at new base’

These compounds are deictic in that their meaning, specifically the direction of the motion, is dependent on the physical context of the utterance, specifically the location of the speaker, as in (45). These compounds and their deictic motion roots in Alyawarr do not correspond exactly to Wilkins’ class of deictic motion verbs in Arrernte because he claims (1989:278 and 1991:224, 241) that Arrernte lhe ‘go’, alpe ‘go back’ and knge- ‘carry’ are ‘inherently deictic’, entailing ‘motion away from speaker’. The corresponding Alyawarr compounds operate in terms of the base as the reference location. See §2.4.1.1.

(45) Akarneng-they ap-ey.alp-enh waylpel ahelengkw-kety Pl.name-ABL arrive-BASE&DO-IMPP white person aggressive-REAS nhenh-el ap-ey.alp-enh ingkety-el-ey Here-LOC arrive-BASE&DO-IMPP foot-INS-EMPH ‘From Akarneng they arrived here because of the aggressive whitefellers’

For the Deictic Motion Compounds, the reference location coincides with the speaker’s location. The location of the old men is regarded as the reference location in (46) where it is the speaker of the reported speech and not the speaker of the overall utterance.

59

(46) Artwa ampw-el-rnem alhengk.aw-enh. “Ah artwa alethang man old-ERG-NPL recognise-IMPP Ah man stranger ap-ey.alh-em-aw!” arrive-GO&DO-POT-EMPH ‘The old men would recognise (its call) “Ah, a stranger is arriving!”’.

The ap- forms express movement along a path and arrival at the speaker’s location. The -alp part of the apey-alp compound apparently has a reading similar to that of its independent form and other compounded forms which mean ‘move to (new) base’. It makes a separate contribution to the meaning of the word. The deictic ap- forms can also be used to describe a narrative deictic centre: movement to a place where a particular set of events happen which is seen as the setting for the narrative.

The root anp- ‘depart’ does not occur outside of this compound in Alyawarr except that the apparent root anpenh- ‘depart’ may be analysable as anp- plus -enh (POINT). There is a corresponding free verb in some other Arandic varieties: E/C Arrernte has arnp- ‘set foot down’ and the compound arnperle-alh- ‘set off on foot’. Apart from the two compounds here, ap- also does not occur elsewhere in Alyawarr although it is relatable historically to -erl.ap ‘do continuously while going along’. However the corresponding root ap- in Kaytetye is the basic ‘go’ verb, equivalent to Alyawarr alh- and corresponding roots also occur non-productively in some derived motion verbs in Southern Arrernte. See discussion in Henderson (1998) and Koch (1984:23). Wilkins (1989:269) argues that (a)pe may have been the original Arandic word for (undirected) ‘go’ and this is evidenced by the need to mark it for direction in Alyawarr. The ap- form may have been applied to general movement whereas alh- and alp- imply directionality and may have largely taken over from the ap- form which became restricted to just these two compounds in Alyawarr. According to Breen (Dixon 1976:624) neighbouring Bularnu has a few verbs of the form V + -baga meaning ‘to V while going’ (baga = go).

2.6 Verbal Reduplication

Yallop (1977: 44) gives four examples of noun reduplication and concludes that “Reduplication is not a productive process in Alyawarra - reduplicated words appear in the vocabulary as separate lexical items”. Contrary to Yallop (1977) and Fabricius (1998:107), Alyawarr has productive nominal and verbal reduplication of several types, all partial reduplication. The purpose of the following sections is to describe the morphological pattern of each verbal reduplication type, as summarized in Table 12. An alternative to affixation is to analyse 60 reduplication as a non-concatenative process. (Haspelmath 2002:22). Verbal reduplication occurs when a template is attached to a base. Sometimes an additional element may occur between the base and the template. The reduplicative marker either precedes (left reduplication) or follows (right reduplication) its base. The relative position of the marker cannot be determined when the base and template are identical, which is the case for the reduplication of preverbs and the whole stem. However, I will not address the analytical issues of an alternative analysis.

Table 12: Verbal reduplication patterns in Alyawarr

Single Base Overall form Marker Reduplicant Reduplicant morpheme label form copies from

Plural Root RED.ew PL monosyllabic right §3.7.4 Attenuative Root RED.elp ATT monosyllabic right §2.6.3 Frequentive Root RED.ep FREQ monosyllabic right §2.6.4 Morphemic complexes Small (-erl).iw REP disyllabic left repeated Do and -ey-RED-alp DO&RET disyllabic left move away Reduplicated -elh-RED-elhil TV disyllabic left TV

2.6.1 Semantics of reduplication and iconicity

Reduplication of all or part of the verb stem corresponds to repetition or continuousness of the verb action and is thus iconic (cf. Moravcsik 1978; Dixon 2002: 210). Some reduplicative markers also have an attenuative meaning, parallel to nominal reduplication in Alyawarr which involves plurality and attenuative or diminutive meaning as can be seen in (47) where the reduplicated form indicates ‘a number of small bank-like things’. Alyawarr thus has some parallels to Warlpiri, where according to Nash (1980), reduplication indicates Plurality of subject or object, distributive, speed and repetition.

61

(47) arnkarr ‘creek bank’ arnkarr-arnkarr ‘small drainage lines, gutters’.

2.6.2 Pre-bases

Three reduplicative morphemes may precede the verb root, Frequentive and Attenuative and Plural -erl.ew (§3.7.3) but discussion in this section is restricted to the two aspectual pre-bases. These pre-bases both involve iconicity where repetition of the form expresses the repetition or extended duration of the activity. Strehlow (1944:173-174) apparently regarded the corresponding RED.elp and RED.ep in Western Arrartna as alternatives in his list of periphrastic verbs. All of his examples are with atw- ‘hit’ but he says that ‘not all of the reduplicated verbs are frequentive in meaning’ and divides reduplicated verbs into regular and irregular verbs. Examples of RED.elp forms in Strehlow’s Aranda grammar (Strehlow 1944) have inceptive meaning as I claim for Alyawarr.

My analysis of these reduplicative types must be taken with caution as only about 20 of each pattern have been found. I have attempted to characterize the kinds of verbs with which each type of reduplication type occurs as shown in Table 13.

In this section I claim that the two aspectual morphemes in Alyawarr are in near-complementary distribution. I have attempted to subcategorise the forms semantically in terms of their Aktionsarten or aspectual types, chiefly according to the property of telicity or boundedness. Frequentive occurs with statives and activity verbs and Attenuative occurs more frequently with liminal verbs. As can be seen in Table 13, FREQ seems to indicate an undifferentiated, repeated and durative activity whereas ATT indicates the beginning of a progression or action involving less intensity. Frequentive tends towards action done to multiple indefinite objects whereas the target of the Attenuative event is a single object. Similar to the situation reported for EC Arrernte, Alyawarr appears to have a number of reduplication templates.Table 13 table shows the distributions of FREQ and ATT with different bases.

62

Table 13: Occurrence of verbs with FREQ and ATT

Root and meaning FREQ ATT an- ‘sit, be’ anepanem ‘live, exist’ anelp-anerl-alhey10 ‘sit a while before going’

rtn- ‘stand, be’ rtnep-rtnem ‘live, rtnelp-rtnem ‘stand for exist’ while’

aynt- ‘lie, be’ ayntep-ayntem ‘live, exist’.

akngan- ‘stay’ akngep-aknganeyel ‘stay’ alh- ‘go’ alhep-alhem ‘going’ alhelp-alhem ‘getting alp- ‘go to base’ alpep-alpem ‘move ready to go’ away’ alpelp-alpem ‘getting ready to move away’. ilkw ‘get shocked’ ilkwelp-ilkweyel ‘twitch’ amparrk ‘shine’ ampep-amparrkeyel ampelp-amparrkeyel ‘glisten’ ‘twinkle’ aytn ‘fall’ aytnelp-aytneyel ‘stagger’ atw ‘hit’ atwelp-atweyel ‘tap, knock, pat’ aw ‘hear’ awelp-aweyel ‘overhear’ ar ‘see’ arelp-areyel ‘catch glimpses’ antelh ‘meet up with’ antelp-antelheyel ‘visiting’ arlw-il- ‘encircle, muster’ arlwelp-arlwileyel ‘round up’

10 This is the best example in the data of an- ‘sit’ with Attenuative. An Attenuative verb that directly contrasts with anepaneme would be better though at least this demonstrates that they can both occur with the same root. 63 ilterr- ‘tear’ iltelp-ilterreyel ‘rip up’ ngan- ‘climb’ ngelp-nganeyel ‘start to climb’ anterr- ‘run’ antelp-anterreyel ‘running around’ angk- ‘say’ angkep-angkem ‘keep saying’ ipar- ‘clean, wipe’ ipep-ipareyel ‘clean fruit’ thak- ‘make hole’ thep-thakeyel ‘make holes’ rak- ‘grab’ rep-rakeyel ‘grab many’ arn- ‘touch’ arnep-arneyel ‘touch continuously’ akalth- ‘break off’ (tr) akep-akaltheyel ‘break off many’

I have tried to elicit other potential forms but have not been able to, e.g. *atwepatweyel ‘knocking’. Although both morphemes occur with Basic Stance forms, it is apparent that different meanings come out with FREQ vs ATT. In the case of the motion verb stems alh- and alp-, these forms only occur with Potential and not the full range of TAM morphology. Each reduplicating subtype is described in greater detail in the following sections.

2.6.3 RED.elp Attenuative

Attenuative has a monosyllabic template involving copying from the right a VC(C) which is followed by -elp. In each case the exact meaning of a reduplicated form will depend upon the base verb. This raises the issue of whether it is more derivational in character inasmuch as it is not regular in meaning, or whether the overall meaning of the verb is predictable given the meaning of the root. I will make a case that the meaning of the verb depends upon the aspectual character of the root. The meaning covers an inceptive sense ‘be at the beginning of the verb event’ and an attenuative sense ‘doing X repeatedly with less intensity’ where a lesser degree of effort or force is involved.11 I will use the label ‘Attenuative’ for the morpheme regardless of which area of meaning might be in focus. Wilkins’ (1989:248) label for the equivalent Arrernte morpheme is ‘continous inception’ which attempts to capture both attenuative and inceptive senses. This is understood as ‘at no time could it be said that the

11 Attenuation is also characteristic of nominal reduplication (see §2.6.1) 64 action has happened’. However, for the Alyawarr data it’s not clear that there is an overall gloss which takes account of both functions of the morpheme.

I have found limited examples of RED.elp with state and activity verbs. With telic verbs the sense is attenuative. With verbs which involve duration the sense is inceptive. Examples of the Attenuative can be seen in (48). The basic meaning of atweyel describes a strong or robust action involving the movement of an arm. Atwelpatweyel describes a repeated action of lesser force or intensity ‘patting, tapping’ as with tapping a damper, or patting someone lightly to wake them up from sleep (Frank Holmes, p.c), an action which involves small movements of hands and fingers.

(48) atweyel ‘hitting’ atwelp-atweyel ‘knocking, patting’ aynteyel ‘falling’ aytnelp-aytneyel ‘staggering’

In the following example text, the Subject begins to stagger and only gets part of the way along the path of travel which is described as close up, angath-antey’ in (49). Later the Subject falls down to the ground which is described by aytn-. With Attenuative, the Subject staggers. Inception is in focus, in that he is continuing the beginning stage of falling but not reaching the terminal point.

(49) Aytn-elp-aytn-eyn-ew=an angath=antey=anem ATT-fall-RET-PP=FOC close=STILL=THEN ‘(He) then staggered repeatedly after only going a short distance’

We see the connection here between attenuation and inception: the beginning stage of an action will often involve a reduced intensity of action. Some instances of Attenuative are strongly associated in translation with the beginning of an activity. For example, speakers report that ngelp-nganeyel means ‘begin to climb’.

In (50) the arrival of the emu is something which is noticed by the miner bird. The ‘seeing’ by the miner bird involves repeated intermittent glimpses or glances. The miner bird glimpses the beginning of the emu’s approach to the water where the hunter waits to spear it.

65 (50) Aympaymp-el ar-elp-ar-eyel-aw “Atnelengkw miner bird-ERG ATT-see-PRES-EMPH emu

ap-ey.alh-eyel kwaty-warl Arrer anem” arrive-GO&DO-PRES water-ALL close=THEN “The miner bird glimpses it! “The emu is coming to the water. Its close now”

In (51), information was heard from indefinite objects, ‘nobody in particular’ in the ‘general area’ of the north, ayerrer-ampeny. Their hearing of the information seems to have occurred over some time and from unstated subjects. There are an undefined and numerous group of people (who are not quantified) who tell the news. The action occurs intermittently. In another sentence example of awelp-aw in the data, people hear a herd of cattle which bellow intermittently.

(51) Ratherr=an aw.elp-aw-ew ayerrer-ampeny-they=anem=arl 3DU=FOC ATT-hear-PP north-direction-ABL=THEN=SUB ‘Then they heard around the north’

The verbs in (52) involve the intermittent repetition of an event. The Attenuative verbs resemble achievements and usually have an attenuative sense ‘doing in a diminished way’.

(52) ilkweyel ‘make a sudden ilkwelp-ilkweyel ‘be twitching, movement’ take a fit’ amparrkeyel ‘shine’ ampelp-amparrkeyel ‘glitter, glisten’

There is a possible parallel with other -elp forms which have inceptive senses involving a change of state or progression. Roots which occur with Attenuative are likely to also have RUN -ey.elp and -erl.elp forms, such as aytn- ‘fall’ and ngan- ‘climb’. (See §4.4.4) A comparison can be made between aytnelp-aytneyel ‘staggering’ and aytneyelpew ‘run and fall’. In the case of aytneyelpew, the subject has been running or moving fast and the verb describes the inception of the falling event. An Inceptive meaning also occurs with RUN&DO ey.elp, for example areyelpeyel (Green 1992:73), indicating the onset of an event. More research is necessary to determine if there is a correlation between verbs occurring with reduplicated -elp and the other - elp forms.

66

2.6.4 RED.ep Frequentive

Frequentive has a monosyllabic (VCC) template which is copied from the right and followed by -ep. Frequentive is of low frequency with approximately only 20 tokens in my data. Stative and Basic Motion verbs with Frequentive have a sense of extended duration and continuous aspect as with example (53). This analysis refines that of Green (1992:134): ‘shows that the action is happening repeatedly or keeps happening’.

(53) Anyent-el=antey ayeng akng.ep-akngan-eyel one-LOC=STILL 1S:NOM FREQ-stay in one place-PRES ‘I am staying in one place’

With intransitive stative verbs FREQ marks continuous aspect as in (54) and (55). One speaker says that the dog in the second example would be ‘sleeping’ (Casey Holmes, p.c.) whereas the dog in the first sentence is lying awake, suggesting that the activity of the second sentence clearly involves an activity of longer duration. The two sentences also involve different verb inflections as FREQ doesn’t appear to occur with the Present tense marker.

(54) arengk nhenh atyek aynt-eyel dog this awake lie-PRES ‘this dog is lying awake’

(55) arengk yanh-el aynt.ep-aynt-em dog that-LOC FREQ.lie-POT ‘the dog lies over there’

FREQ has a ‘backgrounding’ sense which indicates that the Subject usually lives at the location but may not in fact be at the location at the reference time. Subjects are third person and less definite. The distal nhak spatial deixis marker occurs, as with STAT (see §4.9.2). FREQ does not occur with PRES but only modal POT. The source of the backgrounding sense could be the POT morpheme. A person who is located at the speaker in (56) is contrasted with a person who is located at another place in (57).

67 (56) Ayeng nhel an-eyel 1S:NOM here sit-PRES ‘I am here’

(57) Ingwer=an an.ep-an-em other=FOC FREQ-sit/be-POT ‘Another lives (over there)’

In combination with the punctual verbs in Table 13, Frequentive signals iteration or repetition. The following examples are telic verbs and are all transitive. In example (58), the subject repeatedly holed a door with a revolver. The short duration of the events is further illustrated in example (59) involving a repeated process of ‘cleaning’ desert raisins of the toxic chemicals which their skins contain. Processing involves rubbing the desert raisins in earth and the process must be repeated for all of the individual fruits which have been collected.

(58) Revolver-el-tangkwel=anem th.ep-thak-ew revolver-INS-FIRST=THEN FREQ-make hole-PP ‘He holed it with his revolver’

(59) awenp ip.ep-ipar-eyew fruit (stage) FREQ-clean-PURP ‘To process the ripe ones’

The FREQ event referred to could be a more long-term activity such as the removal of children from their mothers in (60), a government policy which lasted many decades. The suggestion is that there are many discrete ‘grabbing’ events involved in the systematic government policy of removal of children which are an indefinite or unbounded group.

(60) R.ep-rak-enh arlengarr-ew=antey ingwerenty=an FREQ-grab-IMPP -LOC=STILL other=FOC

ingwerenty=an rak-enh other=FOC grab-IMPP “They were grabbing (or: they would grab them) them while they were still in their coolamons’

68 In Alyawarr STAT (§4.9.2) and Frequentive reduplication with stative verbs have similar meanings and a ‘distant subjects’ interpretation. These two forms may be interchangeable and reflect dialect variation. Further research is needed. Adjacent Arandic dialects have a corresponding element of the same form, RED.ep, possibly with slightly different functions. In Eastern Arrernte and Eastern Anmatyerr RED.ep appears to occur more frequently with stative verbs. The label ‘frequentative’ to describe RED.ep originates with Strehlow (1944:179). Wilkins (1989:244) labels it Frequentive ‘happens frequently’ and relates it to -rle.pe ‘do continuously while in motion’ and the proto-Arandic form for ‘go’ ape-.

69 3 Derivation and Number

3.1 Introduction

Suffixes of the Verb Derivation category and some suffixes of the Number category occur immediately after the verb root.

3.2 Derivation

There are several types of verb derivation which can be classified according to the category of the base and whether a transitive or an intransitive verb results. Transitive and intransitive verbs can be derived from each other. As I claimed in §2.3, Alyawarr verbs are unambiguously transitive or intransitive. Verb-deriving formatives follow the root. By contrast, nominal- deriving formatives occur in the obligatory position. In the following sections I will describe the verb-deriving morphemes.

Table 14: Derivational Suffixes

Label marker base Intransitivising Intransitive Verbaliser -irr nominals Attribute Intransitive verbaliser -elh nominals Mediopassive -elh transitive roots Reciprocal -err transitive roots Transitivising Transitive verbaliser -il nominals, adverbs -elhil intransitive roots -ern restricted set of roots

3.3 Verbalising suffixes

The compound properties of the verbalising suffixes, IV -irr and TV -il have been introduced in §2.5.3. Verbs are formed by the addition of verbalising morphemes to nominals and adverbs. In all cases the IV verb will have a TV counterpart. The verbs formed by the addition of Transitive Verbaliser -il and Intransitive Verbalisers -irr are somewhat ambiguous in morphological status but are analysed as compounds in §2.5.2.

70 3.3.1 Intransitive Verbaliser -irr

Yallop (1977:66) identified the Intransitive verbaliser –irr. IV productively derives intransitive verb stems from nominals and adverbs, with the meaning ‘to become X’:

(61) ater ‘fear’ aterirreyel ‘become frightened’ mwerr ‘good, well’ mwerrirreyel ‘become well, better’ irrpwerl ‘dark, black’ irrpwerlirreyel ‘darken, become black’

(62) Ra ater-irr-ey.alh-enh arleng-they 3S fear-IV-GO&DO-IMPP distant-ABL ‘He was getting frightened, going a long way’

3.3.2 Attribute Intransitive Verbaliser -elh

The Attribute morpheme attaches to a nominal base. There are eight in my data. Mostly this is a nominal which can be reduplicated and which describes the attributes of the Subject. These are similar to the Physical Attribute Intransitive Verbaliser of Arrernte (Henderson 1998:362). The Attribute Verbaliser seems to co-exist with the IV -irr intransitiviser and is applied to many of the same verbs.

(63) alyelk ‘slippery’ alyelkelheyel ‘slip over’ atherrk-atherrk ‘green’ atherrkelheyel ‘become green’ arlwerr-arlwerr ‘curly’ arlwerrelheyel ‘become curly’ altyek-altyek ‘wizened, dry’ altyekelheyel ‘become wizened, dry’ amper-amper ‘around’ amperelheyel ‘go around’ aylemp ‘crooked, bent’ aylempelheyel ‘become crooked, bent’ inkwer ‘happy’ inkwerelheyel ‘get happy, rejoice’ iylpwer ‘empty’ iylperelheyel ‘become empty’

3.4 Mediopassive -elh

Mediopassive (MED) derives an intransitive stem from a transitive root and hence lowers valency. It has a range of functions dependent upon the individual verb root as shown in Table 16 and discussed in the subsections below. Some verb roots occur with just one of these functions in the Mediopassive, others have more than one. For example, akelheyel (Green

71 1992:9) has the senses ‘break’ (middle) and ‘cut self’ (reflexive). Mediopassive is limited in the number of verb roots with which it occurs. These are listed in Table 15.

Table 15: Verb roots which combine with Mediopassive (including compounds)

atw- ‘hit’ ak- ‘cut’ arrern- ‘place, put down’ tywen- ‘put up’ arertn- ‘tie’ aw- ‘hear, listen’ arntarnt-ar- ‘look after, watch over’ atyerr- ‘shoot’ il- ‘tell, do’ iw- ‘throw, leave’ art- ‘cover, bury, build’ iltew- ‘split, smash, break’ kwarn- ‘hurt something’ arrtyern- ‘shine’ tyerr- ‘extract, remove’ apern- ‘rub on’ ingkwern- ‘paint, write’ il- TV thel- ‘pour’ akngartew- ‘turn over’ akng- ‘carry, take’ alth- ‘pluck’ rlkwen- ‘wash’ lhew- ‘wash’ rnkern- ‘stand something up’ rtern- ‘straighten something’ nak- ‘praise someone’ intwererremew- ‘comb, brush’ ipm- ‘leave’ irrnyew- ‘peel’ irrtyertew- ‘hide something’ iylkw- ‘swallow something’ aylertn- ‘twist’ arreyn- ‘scratch something’

72 arrkern- ‘try’ atak- ‘dismantle’ arlengkew- ‘hide something’ antw- ‘call something’ in- ‘get’

The traditional handling of MED in Alyawarr is to label it as ‘reflexive’ with the central interpretation of ‘do to self’, a characterization which has sometimes gone along with a less detailed analysis of its functions. Strehlow (1944: 198) labelled the -elh verbs in Arrernte as ‘reflexive’. He maintained that there is no clearly passive marking in Arandic languages as there is no cultural or pragmatic need for the passive. Both Yallop (1977:57) and the AED have followed this definition: ‘goes on verbs to show that someone or something is doing the action to themself’ (Green 1992: 133).

Table 16: Mediopassive functions

Label Function Reflexive S does V to self Middle V happens to S spontaneously Passive V is done to S Antipassive S does V of perception or attempt

3.4.1 Reflexive

In a majority of cases Mediopassive has a reflexive interpretation ‘do V to self’, as in (64). The Subject has both Actor and Undergoer roles.

(64) Irrtyertew-elh-eyel ayeng hide-MED-PRES 1S:NOM ‘I am hiding myself’

With a transfer verb root, MED derives a type of motion verb as in (65).

(65) Kel thip-rnem ingkerr arrern-elh-ey.alh-enh then bird-NPL:NOM all put down-MED-GO&DO-IMPP ‘Then all the birds were landing ’

In instances like (66), there is an issue of what case labels should be assigned to the arguments of the clause. The issue arises because of case syncretism of the S and O arguments for full

73 nominals. There are other whole-part constructions for which both NPs appear in the same grammatical function within the clause, so the simplest analysis in (66) is to assume that there are two NPs in the same grammatical function. The same part-whole analysis would also be required for clauses with non-derived intransitive verbs as in (67). This is the simplest analysis and is preferred to one which sees one of the arguments as being in Object function within the clause. This analysis could be seen as one in which both NPs are in the same grammatical function but have different semantic roles.

(66) Artwa akapwert art-elh-eyel itwern-el amp-ekerr man:NOM head:NOM cover-MED-PRES heat-LOC burn-APP ‘the man is covering his head so that he doesn’t get burnt in the heat of the sun’

(67) Akapwert ayeng arnt-eyel itwern-penh head:NOM 1S:NOM hurt-PRES heat-SRC ‘my head is hurting because of the heat of the sun’

With reflexive transfer verbs, the whole may be in the agent role and the part in the source or destination role as shown in (68). An interesting aspect of reflexive transfer verbs in Alyawarr is that a source or destination argument can be in Ablative or Allative case as in (69). The reflexive transfer verb co-exists with the more standard transitive verb. The Mediopassive is not obligatory in reflexive transfer situations as the transitive verb arrern in (70) is also used to describe the process of placing an object on one’s head.

(68) Iyterlarr arrern-elh-eyel akapwert headband:ACC put down-MED-PRES head:NOM ‘(you) put the headband on your head’ (AED:178)

(69) Kwey amerterr arrern-elh-ew akapwert-warl girl:NOM headdress put down-MED-PP head-ALL ‘the girl put the headdress on her head’ (AED: 36).

(70) Iylarnt akapwert-warl arrern-etyek apmer-warl akng-eyn-etyek headring:ACC head-ALL put-PURP camp-ALL carry-RET-PURP ‘You put the head-ring on your head to take the things back to camp’

The distinction between these constructions is lost in some contexts due to argument ellipsis

(and the syncretism of cases for full nominals). For example in (71) two analyses are possible: (i) the spear is both the (purported) agent and the patient and the construction is therefore of the first type, or (ii) the spear is the patient and an ellipsed argument is the agent, in which case the

74 construction is of the second type. This highlights the very important question as to the distinction between these functions. It is most likely that the spear is the S of the clause in (71).

(71) Ingwer irrtyart=an tyerr-elh-enh-ew other spear:NOM=FOC extract -MED-POINT-PP (i) ‘Another spear:NOM pulled itself out’ (ii) ‘(Someone) pulled another spear:ACC out of themself’

Reflexive verbs may be formed by the addition of -ilelh to a nominal. They are not common with around ten examples in the data. Usually S has some control over the action of the verb and the reading can be autocausative ‘make self like the attribute of the nominal’ with an auto- causative as in (72). The two anaynt marked forms in (73) constitute evidence that -ilelh could be composed of Transitive Verbaliser -il and Mediopassive -elh. Further research is required to investigate the identity between TV-MED and the Mediopassive verb ilelh ‘make self into something, change self’ (Green 1992:143).

(72) Ra=arl apmw-il-elh-ew=arl irrtyart=arl athen-ew-el 3S=SUB wrong-TV-MED-PP=SUB spear=SUB lay down-PP-LOC

aketh-warl open-ALL ‘He made a mistake by laying his spear down in the open’

(73) Artwa ra=an athen-ey.elp-ew irrtyart=arl=ap man 3S=FOC lay down-RUN&DO-PP spear=SUB=CONJ

anaynt-il-ew=an. Anaynt-il-elh-ew=an=ap ra=antey. clear-TV-PP=FOC clear-TV-MED-PP=FOC=CONJ 3S=still ‘He moved quickly and laid the spear down and exposed it. He left himself exposed’

75 3.4.2 Passive

In many Australian languages, there is a single verbal morpheme which has both passive and reflexive functions (Blake 1987:61). The focus is on the state of the Undergoer in an agentless passive. Dixon and Aikhenvald (2000:25) claim that ‘the main effects of the passive are drawing attention to the O, downgrading the original A and focussing on the state the original O is in, as a result of the activity reported by the verb’. For example the intransitive verb artelh- in (74) is derived from the transitive verb art- which has a meaning ‘to cover or to build’ and thus has a meaning ‘to be covered’ or ‘to be built’. Mediopassive verbs occur disproportionately with the subordination marker marking a dependent clause. There is a clear correlation between passive and stativity. There is also an association between the passive construction and the - ek=arl PP=SUB morphemes which are similar to past perfective forms of the verb.

(74) Warl nhaym=an art-elh-ek=arl house that build-MED-PP=SUB ‘that house that was built’

3.4.3 Middle

The Middle12 has no suppressed agent, as in (75). With the ‘spontaneous middle’, the action of the verb happens without an agent. The Subject typically refers to an inanimate entity.

(75) Arwerl=an ingkerr-antey-anem ak-elh-ew. tree=FOC all-STILL-THEN break-MED-PP ‘Then all of the trees broke’

3.4.4 Antipassive

The antipassive function involves deletion of the patient or downgrading it to a dative-marked object. Compare the accusative-marked and dative-marked objects in (76) and (77).

12 The term ‘Middle’ has had a wide range of applications (Kemmer 1993:1). I am using the term as a semantic categorization for a limited number of Mediopassive verbs. 76 (76) Atha ngenh aw-eyel 1S:ERG 2S:ACC hear-PRES ‘I am hearing you’

(77) Ayeng ngkweng aw-elh-eyel 1S:NOM 2S:DAT hear-MED-PRES ‘I am taking notice of you’

It can be compared with the dative of attempt (Wilkins 1989: 180) and (78) although there is no change of transitivity in that case.

(78) Anamerl ar-ey.elp-erl.an-a anwenger-ampeny aker-ew. quickly see-RUN&DO-STAT-IMP.S other side-around game-DAT ‘You quickly go off and look for meat on the other side’

A small number of verbs of perception and experience occur with Mediopassive, examples of which are listed in Table 17.

Table 17: Some verbs with antipassive interpretation

Transitive verb interpretation Mediopassive interpretation aw hear, understand awelh listen out for, obey arntarnt-ar look after, watch arntarnt-arelh look out for artepang-ar look back and see artepang-arelh watch out behind arrkern tempt, try out arrkernelh attempt ayl sing someone aylelh sing

As noted above, Antipassive involves the demotion of the object/patient which is then marked with the dative case. The properties of Antipassives are outlined in Table 18. The emphasis is on the action of the verb not the agent or the object/patient which in some cases will become nondefinite or non-referential. Dixon and Aikhenvald (2000:9) note that the ‘Antipassive focuses on the activity itself (that is, on the agent’s performing the activity).’ Emphasis may be on the ‘activity indulged in’ as opposed to the effect on the patient (Blake 1987:58). In Alyawarr the O in an antipassive clause is similar to that reported for -thadi- (Austin 1981:155) which has a similar range of functions.

77 Table 18: Properties of Antipassives

Avoid mentioning the underlying O Non-definite, non-referential O Lack of success, failed attempt Focuses on activity

The difference between transitive root aw ‘hear’ and derived Mediopassive stem awelh is seen in (79). The men are listening out for something which is unmentioned. After they hear the noise, the action is described in terms of the active transitive verb aw- and the Past Completed tense marker.

(79) Aw-elh-enh=anem rernem artwe-rnem-ey. Awer-ø=anem=ap hear-MED-IMPP=THEN 3PL man-NPL-EMPH wind-ACC=THEN=CONJ rernem aw-ek=an. 3PL hear-PP=FOC ‘The men were listening out. Then they heard the wind.’

In (80), the verb arntarnt-arelh is used in conjunction with a previous intransitive verb. There could be differences in mood indicated by (80) where Mediopassive indicates an unrealized event. Mediopassive interacts with aspect, indicating continuous aspect, that this is an ongoing activity cf. Dixon (1994:148). Lack of completion of an activity can be conveyed by imperfective aspect. Imperfective Past occurs with the Mediopassive awelh in (79) to indicate imperfectivity rather than completed actions.13 The gloss ‘getting frightened by something’ in (81) is close to the passive interpretation although a clearly distinct function can’t be distinguished. There appears to be an aspectual meaning, possibly inceptive or inchoative ‘become’ which indicates that the activity is ongoing and uncompleted. Further research would be needed to establish a correlation between Mediopassive and lack of result.

(80) Ratherr=ap ywarn=ant ey=anem mpwelh-enh=an. Arntarnt.ar-elh-enh. 3DU=CONJ unsuccessful=STILL=THEN wait-IMPP=FOC watch-MED-IMPP ‘They waited without success (for him) They were watching.’

13 Awelheyel is also translated (Green 1992:119) as ‘feel’ but this sense may only occur in combination with a complement such as arntety ‘sick’ meaning ‘to feel sick’. This question needs further investigation. 78

(81) Ra=ap awey ater.anth-elh-eyn-enh 3S=CONJ boy frighten-MED-RET-IMPP ‘The boy was going along becoming scared’

3.5 Reciprocal -err

Reciprocal derives an intransitive stem from a transitive root. An exception is the intransitive root angk- ‘speak’. Like Mediopassive, it lowers sentence valency as in (82). The interpretation is ‘do V to each other’.

(82) Arengk tnhw-err-em dog:NOM bite-REC-POT ‘Dogs attack each other’

Reciprocal is homophonous with the distinct -err Plural allomorph. Strehlow (1944:122) doesn’t mention the reciprocal for Arrernte and some of his clearly reciprocal examples are analysed as number marking. After reading Strehlow’s analysis, Capell (1962:75) subsumed ‘reciprocal’ under number marking. Yallop (1977:60) sees -err as a single morpheme with two uses, reciprocal and plural. However the two morphemes can be distinguished on three grounds: meaning, position in the verb structure and effect on transitivity.

Firstly, the two morphemes are distinguishable on the basis of the number of participants involved in the activities which they describe. Reciprocal involves two or more participants doing an action towards each other. In contrast Plural involves three or more Subjects who are doing an activity which is not directed towards each other. Secondly, Plural can follow the root but can also follow other post-obligatory morphemes including the –enhey element with compounding forms (§3.7.1.) as in (84). Reciprocal immediately follows the verb root only as in (83).

Thirdly, Reciprocal and Plural are distinct because the former derives an intransitive verb while the latter does not alter the transitivity of the stem. Thus, with a transitive root, the Reciprocal verb will take a Nominative-marked subject while the homophonous Plural verb will take an Ergative-marked subject. Compare the following examples from the AED (Green 1992:136). In (83), ergative-marking indicates that the verb is Plural. This sentence would have a reciprocal interpretation if there was Nominative marking on the subject as in (84).

79

(83) Ampe-rnem-el atw-enh-ey-err-alp-eyel arelh-Ø

child-NPL-ERG hit-POINT-BASE&DO1-PL-BASE&DO2-PRES woman-ACC

‘All the kids are hitting the woman as they go along’

(84) Ampa atherr atw-err-enh-ey.alp-eyel child two:NOM hit-REC-POINT-BASE&DO-PRES ‘The two children are hitting each other as they go along’

The Subject of (85) from Yallop (1977:60) has been ellipsed. Without knowing that, aherr could be either the subject or object of the sentence, which is therefore ambiguous. Yallop claims that, in that particular example, the meaning can be distinguished pragmatically because the reciprocal sense could not apply with (herbivorous) kangaroos ‘in many if not all contexts’. In the supposed real-world knowledge of the speaker kangaroos don’t eat each other so the interpretation of the sentence is likely to be plural. A Plural reading of the verb is probable because aherr can signify a generic, mass or non-count noun ‘kangaroo meat’ as in ‘we used to eat kangaroo (meat)’. There is more specification for Number in Alyawarr than Yallop claims. For further discussion see §3.7. There are examples of arlkw-err with a reciprocal meaning, such as (86) which reports the purported cannibalism of the Kalkatungu people of Mt Isa. It is also conceivable that mythological kangaroos might eat each other since mythological characters frequently behave in ways that differ from their real-world counterparts. There is also the possibility that the reciprocal reading is only possible with ament ‘by themselves’ and that ‘they eat’ is the usual meaning of the verb.

(85) Aherr-Ø arlkw-err-enh kangaroo-ACC eat-PL-IMPP ‘(We) used to eat kangaroo’

(86) Ament=antey arlkw-err-enh separate:NOM=STILL eat-REC-IMPP ‘(They) on their own used to eat each other’

Reciprocal occurs with only one intransitive root, angk- ‘speak’. A common context for the use of angk-err is a gathering or a meeting and it is usually understood to mean ‘discussing, talking with each other’. Yallop (1977) analyses angk-err as Plural rather than Reciprocal on the grounds that angk- is an intransitive root. However as demonstrated in (87) from Yallop (1977: 60), angk-err can occur with a dual subject, in which case it is not Plural. Like other Reciprocal verbs, angk-err is compatible with both dual and plural subject arguments.

80 (87) Angk-err-eyel alanth speak-REC-PRES 3DU:NOM ‘The two of them are talking to each other’

3.6 Transitive Verbaliser

The Transitive Verbaliser derives transitive verbs from nominals and verb stems, predominantly intransitive roots. I analyse the Transitive Verbaliser as having two central allomorphs, -il and - elhil and a third more marginal one, -ern. For the sake of consistency, I will write the second of these as -elhil, although it could also be written -elhel.

3.6.1 Nominal bases

Transitive verb stems are created with the addition of TV to a nominal with the resulting meaning ‘make O have the attribute of the nominal’, as for example in (88) which is similar to Factitive. In the majority of cases, the -il allomorph occurs, as in (89), but there are three nominals (recorded so far) which require the -elhil allomorph (90). Further research is necessary to determine which roots occur with -il and which occur with -elhil.

(88) Arrarntenh akngerr-il-enh thip ra arrakwerrakwerr-el-ey bush plum many-TV-IMPP bird 3S honeyeater-ERG-EMPH ‘The bird was increasing the number of bush plums, the honeyeater’.

(89) arlpenty ‘long’ arlpentyileyel ‘lengthen’ . anaynt ‘clear’ anayntileyel ‘to make clear, explain’. iteth ‘alive’ itethileyel ‘make alive, revive’ arternp ‘slow’ arternpileyel ‘slow down’

(90) anyent ‘one’ anyentelhileyel ‘join, make into one’ rlterrp ‘rattle’ rlterrpelhileyel ‘rattle something’ arlwar ‘round’ arlwarelhileyel ‘blow up, inflate’

In two cases where -elhil occurs, shown in (91) , the base does not function as an independent nominal but in both cases the same element also occurs with the Intransitive Verbaliser.

(91) akemelhileyel ‘raise, get up’ ahartelhileyel ‘shift’

81 3.6.2 Verb bases

The Transitive Verbaliser derives a transitive verb from a verb stem which is usually an intransitive root, but is a transitive stem in one case (92). In the majority of cases the function is causative, but is applicative with some intransitive roots. In this context, only the -elhil allomorph occurs. As discussed in Chapter 2, this results in a type of compound verb. The -elh and -il elements can be separated by intervening non-verb material, as in (42) in Chapter 2 and the stem which ends in -elh is also subject to reduplication.

(92) Rwarr-el akngenh-elhil-eyel lywenty=an wind-ERG move-TV-PRES bough shade=FOC ‘The wind makes the bough shade move’ (AED:14)

3.6.2.1 Causative function

In its causative function the Transitive Verbaliser derives a verb in which the A argument causes the O argument to do the action of the verb stem, that is, the O argument of the causative corresponds to the subject argument of the underived stem (Yallop 1977:61). This is an example of causative marking within the verb (Elson and Pickett 1988: 32).

(93) Ntwa aytn-elhil-ek ayenh 2S:ERG fall-TV-PP 1S:ACC ‘You made me fall over’

3.6.2.2 Reduplicated TV

This involves a form of reduplication where the -elh stem is reduplicated to express a repeated action as in (94). See discussion on reduplication in §2.6.1 A similar construction is reported for other Arandic varieties, including Mparntwe Arrernte (Wilkins 1989:246).

(94) Pwelek nh-el-rnem=an arrangk-elh-angk-elhil-eyel

Cattle here-ERG-NPL=FOC cry-TV1-RED-TV2-PRES ‘These ones are causing these cattle to bellow!’

The reduplicated TV is uncommon and I have only recorded it with eight verb stems, as listed in (95). The reduplication patterns involved here are mixed. The first case is ambiguous as to whether the reduplication is of the whole-stem that ends with -elh or reduplication to a

82 disyllabic template. In the second case it is a compound, so it is also structurally ambiguous. The simplest account of its structure is that the reduplicant is disyllabic.

(95) angkeyel speak angkelh-angkelhileyel ‘cause something to make noise’ wey-angkeyel breathe wey-angkelh-angkelhileyel irrper-angkeyel stamp irrper-angkelh-angkelhileyel ‘make thudding sound’ arrangkeyel cry arrangkelh-angkelhileyel ‘make someone cry out’ akngelheyel move akngelhew-akngelhileyel ‘shake, buffet’ pweyel blow pwelh-pwelhileyel ‘cause wind to blow’ atneneyel gut atnelh-atnelhileyel ‘pull guts out’ altheyel pluck althelh-althelhileyel ‘pull hair out in a fight’

It may have a sense ‘distributed all about’, as might be the case in (96) where gusts of wind buffet a sailing vessel. The fifth form listed above is unique in that it appears to have a disyllabic reduplicant plus an unidentfied component -ew. As there is only one example of this - ew, it is unclear whether it relates to the -ew element in the Plural allomorph in §3.7.4.

(96) Awer-el anwenhantherr akng-elh-ew-akng-elhil-ek

wind-ERG 3PL:ACC carry-TV1-PL-RED-TV2-PP ‘The wind buffeted us’ (Acts 27:4 translation)

3.6.2.3 Applicative function

A less common use of transitivising -elhil is the Applicative in which original S becomes A (not O as with Causative) and an original peripheral argument becomes O. For example the applicative arlk-elh.il ‘call out-TV’ in (97) has an Accusative-marked topic of communication yanh. The underived verb would express the topic of communication in a peripheral NP. There are only four such verbs in my data, listed in (98) and only a few tokens of these. Applicatives are also found in Mparntwe Arrernte (Wilkins 1989:259, Austin 2005:12). The first two Applicative verbs in the list are also common in other Australian languages (Dixon 2002:204).

(97) Yanh arlk-elhil-ew waylpel-ew-rnem=anem=arl that:ACC call-TV-PP Whitefeller-DAT-NPL=THEN=SUB ‘(She) then told the whitefellers about that one’

(98) atherreyel ‘laugh’ atherrelhileyel ‘mock, laugh at’ artneyel ‘cry’ artnelhileyel ‘mourn with, cry with’ altyweneyel ‘crouch’ altywenelhileyel ‘sit on eggs’ arlkeyel ‘shout’ arlkelhileyel ‘shout out about’

83

3.6.3 Transitiviser -ern

The -ern suffix is in apparent complementary distribution with the -il and -elhil forms and so is analysed here as an allomorph of the Transitive Verbaliser as in (99). Yallop (1977:124) does not analyse -ern as a suffix but as an element of the root. He avers that most roots ending in -ern are transitive but he doesn’t regard -ern as a morpheme and claims that it has ‘no productive regularity’. Wilkins (1989) analyses the corresponding -ern in Arrernte as a non-productive transitiviser. Compared with the other transitivizing forms -il and -elhil, the -ern morpheme in Alyawarr is of limited productivity. As Wilkins (1989:259) proposes for Arrernte, Alyawarr may have had a more widespread productive transitivising morpheme -ern at some earlier time.

(99) intransitive transitive ampeyel ‘burn’ amperneyel ‘cook something’ ilweyel ‘die’ ilwerneyel ‘extinguish something’ inteyel ‘smell’ interneyel ‘smell something’

transitive transitive arrtyeyel ‘light fire’ arrtyerneyel ‘shine’

The transitive verb arrtyern- ‘shine’ is an apparent exception as it is derived from transitive arrty- ‘burn something’ but the addition of -ern doesn’t change the transitivity of the root. The Arrernte tn- ‘stand’ is plausibly related to tnern- ‘stand holding/with something’ derived from tn- ‘stand’ (Henderson 1998:364). In Alyawarr, rtn- ‘stand’ and tnern- appear to be separate roots which differ in their initial consonant and tnern- is not apparently derived with -ern.

3.7 The Category of number

Subject number is marked in the verb by a diverse set of forms which occur in four structural positions in the verb. See Table 19. The allomorphy is both grammatically and lexically conditioned. Verb number has some parallels with number in the pronominal system but some important differences. It is obligatory in the imperative mood (see Chapter 5) where singular, dual and plural are distinguished. In non-imperative moods, number-marking is not obligatory (Yallop 1977:61). The distinction there is unspecified vs more than two. Often a collective sense is indicated- ‘doing something together’.

84

Yallop (1977:61) claims that plural marking is not obligatory and for example, angkeyel artwa ‘talk-PRES man’ can mean ‘the men are talking’. However, I have found that the pragmatics of a situation describing typical ‘talking’ events require non-singular verbs, for example angkerreyel ‘talking together’. I have identified angkerreyel as reciprocal rather than plural where the context is that the men are having a conversation. See §3.5.

Table 19: Number marking on Alyawarr roots and derived stems

Simple IMPERATIVE SG -Ø DU -enh(err)atherr PL -enh(err)arey NON- most roots -err IMPERATIVE rtn, ilw, w rtn ‘stand’ -enherr ilw ‘die’ Basic Motion alh- ‘go’ -erl.ew alp ‘go away’ action rntw ‘dance’ RED.ew- arlkw ‘eat’ stance/state aynt ‘lie’ an ‘sit’ Certain intransitives rntw ‘dance’ -elhelerr anterr ‘run’ arlk ‘call’ -arlelherr artn ‘cry’ COMPOUND Verbal Compounds -err (after roots and Deictic Motion derived stems) -arr, -am (following V without post- obligatory morphology.

No examples have been recorded of marking in more than one of these positions in a single verb. Since number is obligatory in Imperative verbs, that takes precedence over other positions. In compound verbs the final position of the pre-compound stem appears to be the preferred site but other sites also occur:

85 (100) Ampa akngerr atw.ew-atw-err-erl.alp-eyel child many PL-hit-REC-DO&RET-PRES ‘Lots of kids are fighting each other as they are going back’. (AED:136)

(101) Ampe-rnem arwerl-they aytn-em ngan-err-erl.an-enty child-NPL tree-ABL fall-POT climb-PL-STAT-IRR ‘The children fall from trees when they climb up’

3.7.1 Plural -err and -arr

These two similar allomorphs are in complementary distribution, contra Yallop (1977:66) who claims that they are in free variation. The -err allomorph occurs in (i) the post-root/derivation position and (ii) following -enhey at the end of a pre-compound stem. The -arr allomorph occurs only at the end of a pre-compound stem, except where the preceding stem ends with - enhey on the verb root. The similarity of these two allomorphs suggests that they are historically related, as presumably are the other allomorphs which have an -err element. This may also be related to the -err in plural pronouns (Yallop 1977: 99).

3.7.2 -enherr

There are a few monosyllabic verb roots which select the -enherr Plural allomorph:

(102) ilweyel ‘dying’ ilwenherreyel ‘many dying’ rtneyel ‘standing’ rtnenherreyel ‘many standing around’ weyel ‘shooting at’ wenherreyel ‘many shooting at something’

This form may be related to the non-singular component of the Dual and Plural Imperative in §5.2.1. Plural wenherrenh in (103) ‘were hitting something with a projectile’ is distinct from Reciprocal werrenh ‘hitting each other with projectiles’. Similar forms are also reported for Antekerrepenh (Breen 1982:30), -enirr and Mparntwe Arrernte -ernirr.

(103) Maket-el=anem w-enherr-enh r=arl rtn-ey.aynt-ek-warl. gun-INS=THEN shoot-PL-IMPP 3S=SUB stand-UP-PP-ALL ‘They were shooting with rifles at the point where it had come up’

86 3.7.3 -erl.ew plural with Basic Motion verbs

There are two distinct but homophonous -erl.ew morphemes. SIDE erl.iw occurs with non- Motion verbs and is discussed in §4.5.1. Plural -erl.ew (104) occurs with the Basic Motion verbs alh- ‘go’ and alp- ‘go back’ (Strickland 1998:115).

(104) Thip-rnem alp-erl.ew-ek bird-NPL back-PL-PP ‘The birds went away’

Two of Yallop’s (1977:62) -erl.ew examples involve plural subjects and yet he appears to regard -erl.ew as an aspectual ‘continuous’ marker. While I regard this reading as unlikely, more research may investigate whether there is an additional component of meaning such as ‘dispersed subjects moving’.

3.7.4 RED.ew-

This allomorph occurs only with mono-syllabic roots, It consists of a specified part -ew preceded by a reduplicant which is mono-syllabic and therefore is identical with the verb root. These include state verbs (an- ‘sit, be’ aynt- ‘lie, be’) and activity verbs (arlk- ‘eat’, rntw- ‘dance’):

(105) aneyel ‘sit, be’ anewaneyel ‘ a group sitting around, living’ aynteyel ‘lie, be’ ayntewaynteyel ‘a group lying around, staying’ arlkweyel ‘eat’ arlkwewarlkweyel ‘eating in groups’ rntweyel ‘dance’ rntwew-rntweyel ‘dancing in groups’

Interpretation of individual reduplicated forms depends upon the aspectual character of the base verb. With stative verbs the interpretation is that there are several referents with a distributive reading:

(106) Ngenty=an arrangkw artwerrety=arl. Kwaty-weny arwa soakage=FOC nothing dry=SUB water-PRIV just

aynt.ew-aynt-ew arrangkw PL-lie-PP nothing ‘The soakages were dry. They had no water in them, nothing’

87 Transitive activity verb roots occurring with RED.ew usually involve plural subjects and multiple events:

(107) ntang=arl arlkw.ew-arlkw-enh kwaty artnwep-itwew seed=SUB PL-eat-IMPP water soakage-LOC ‘They were eating seed tucker at the soakage in groups’

In this example, there is a pluractional sense involving episodes of eating which involve separate groups of people which are distributed over an area. In the same text, arlkwerrenh ‘they used to eat’ with plural reference also occurs, referring to the habitual activity of people as a whole. Arlkwew-arlkwenh is not only plural, but distributive, referring to dispersed groups who are involved in discrete eating events in different places on the landscape. More than merely plural, there is also a ‘spatial distribution’ and possibly a ‘separate groups’ component to the interpretation although speakers claim that -err Plural and RED.ew have the same interpretation of multiple subjects.

Plural -ew may be related to an unproductive -ew morpheme which appears with certain reduplicated nominals:

(108) ament.ew-ament ‘in separate groups’

The -ew part may be related to -erl.ew which marks plural in other contexts such as with the Basic Motion verbs alh- and alp- (§3.7.3).

3.7.5 -elhelerr

This form is taken as a distinct Plural number allomorph which has a reciprocal sense and which occurs with less than ten verbs in the data, as in (109).

(109) rntweyel ‘dance’ rntwelhelerreyel ‘all dancing’ anterreyel ‘run’ anterrelhelerreyel ‘run off with each other’ rlkeyel ‘show eyes’ rlkelhelerreyel ‘make eyes at each other’

3.7.6 -arlelherr

The final allomorph, -arlelherr, is the most restricted. The two verb roots involved are similar in number marking, and form Applicative verbs (§3.6.2.3) which allow intervening material to occur between -elh and -il:

88 (110) arlkeyel ‘shout’ arlkarlelherrenh ‘all shouted out’ artneyel ‘cry’ artnarlelherrenh ‘all grieved’

3.8 Discrete -am

The Discrete morpheme appears on nominals and verbs. Discrete and -arr Plural number marking occupy the same position in the morphological structure of compound verbs. Evidence suggests that -arr and -am are in opposition and I have categorised Discrete with number marking in a single category. My analysis is consistent with that of Strickland (1998:104) who claims that it is as though the process is made up of small increments distributed over a period of time.

When -am attaches to a verb stem, the apparent meaning is ‘discrete events’, separate but related events ‘one by one’. Separate subjects each doing a discrete action is illustrated in (111). Although the spearing events seem to have occurred on one occasion, (because the victim of the spearing died), nevertheless the individual events of throwing the spears were discrete. A possibility is that Discrete signals that each of the spear throwers was acting on his own initiative and there isn’t the meaning of ‘doing in a group’ signalled by plural -arr.

(111) Nhanyem rntern-erl.elp-ew menty ahern this spear-DO&RUN-PP leave it region

nhareny-el-rnem=an ap-ey-am-alp-em-el

this one-ERG-NPL=FOC arrive-BASE&DO1-DISC-BASE&DO2-POT-SS ‘They speared this one, then left it for the ones from this region who were coming’

The notion of discreteness may be applied to events or actors: discrete subjects acting independently or discrete events. Applied to a singular S, in (112), it suggests that events occur at intervals.

(112) Aynelikw=an altywen-elhil-eyel rtn.erl-am-ap-eyel

father.ERG=FOC crouch-CAUS-PRES stand-CONV1-DISC-CONV2-PRES

rtern-elh-em-el=an straighten-MED-POT-SS=FOC ‘The father (emu) is crouching and standing up regularly, straightening itself out’

89 There are about 10 tokens of DISC occurring in my data. Most commonly it follows the sequence morphemes -ey and –erl in verbal compounds and precedes the inflecting compound element. In nominals and derived nominals it occurs with a reduplicated base:

(113) anyent-am-anyent ‘one by one’ akngerr-am-akngerr irreyel ‘building up numbers, becoming more’

In the following example DISC is attached to the preverb of a lexical compound:

(114) Aley=arl ra aynt-eyel anwenantherr amperl-am-ar-enh now=SUB 3S exist-PRES 3PL.EX track-DISC-see-IMPP

yard-akerr ratherr art-ey.elp-ew yard-CONJ 3DU build-RUN&DO-PP ‘It’s there now, we have seen the tracks/imprints (on a number of separate occasions) and also the yards which they built’

In this case the Discrete marker –am indicates that there were many occasions on which the tracks or imprints were seen. Another morpheme used to mark discrete events is Distributed - erlenty.akngenh §4.6.2 although the latter seems to occur with different roots from those which occur with -am.

90 4 Motion and Aspect

4.1 Introduction

This chapter includes the category of Motion and the category of Aspect as marked in the non- obligatory morphemes which follow verb roots and derived stems. Firstly I describe the morphemes which occur in simple verbs. Then I describe complex elements which range in semantic compositionality and productivity, from compounds to monomorphemic affixes.

In Alyawarr, complex verbs express a variety of grammatical categories. In this chapter I argue that verbal compounds expressing Sequenced Motion are highly compositional. Aspect, vertical motion and motion path are expressed by suffixes which are less analysable into combinations of smaller units. There is likely to be a historical relationship between the less semantically compositional complexes and free verbs, although I will not make a definite claim here.

There are six sections within this chapter:

Motion 1. Action along a path §4.3 2. Sequenced Motion §4.4 3. Rapid Motion §4.4.4 4. Habitual Action along Path §4.6 5. Vertical motion §4.7 6. Aspect §4.9

4.2 The category of Motion

As I mentioned in §1.2.1, the grammatical marking of motion events is important in Arandic languages. Payne (1997: 248) claims that “many other (than European) languages grammaticalise spatial grounding” and that spatial deixis is often ‘more central to the language than temporal deixis’. Associated Motion was first proposed as a distinct category for Kaytetye. Koch (1984:23) described the category thus: “the motional forms are largely distinguished according to the direction of the motion and the time of the motion relative to that of the main action”.

The category of Associated Motion has been described in other Arandic languages, Central and South Australian languages, for example Breen (1976), Austin (1981), Tunbridge (1988) and 91 Wilkins (1989). Wilkins (1991:212) follows Koch (1984) in regarding Associated Motion as a separate grammatical category, distinct from Aspect.

The grammatical marking of motion is related to the issue of compounding which was introduced in §2.5. Sequenced Motion compounds are semantically compositional forms (§2.5.1.3) consisting of a sequence morpheme and a compounding verbal element. The latter bears the obligatory morphology. I have opted for a compounding analysis in agreement with Yallop (1977:61-66) who described compound verbs in Alyawarr, following Strehlow’s ‘periphrastic verbs’ (Strehlow 1944:172-179).

4.3 PATH morphemes

The two Path morphemes contrast in the same position within the verb and describe motion on a path that is simultaneous with the primary action of the verb stem. Return and Point do not co- occur and have the same ordering relative to other markers in the verb and are therefore taken to occupy the same structural position. They describe complementary motions similar to the Basic Motion roots and thus both involve a motion path which is anchored at one end. Neither of the Path elements occurs with Basic Motion roots.

4.3.1 -eyn Return

Return has a primary sense of ‘to move to a base’, involving returning to a base or travelling to a new base. A range of senses depending on the verb stem to which it is attached can be summarized as follows:

 move (away) to base  extended sense: move to implied endpoint  induced motion / transfer verbs: motion of Object (and Subject)  manual action along an object (without motion of whole Subject)  extended sense (‘base’ and ‘motion’)  physically extend on a path to an implied endpoint  a single object with elongated orientation or  a line of separate objects  change towards recognized end state

Return is similar to the cognate Arrernte Reversive -irtne ‘do while going back’ (Wilkins 1989:277). Return indicates that the action of the verb happens while returning to a base or moving to a new base, as in (115). A close parallel exists between RET and alp- ‘move to and stay at (new) base’ (§2.4.1.3) in that they both convey ‘move to base’. This motion takes its

92 reference location from the base but is not inconsistent with motion towards the speaker. Things which come to a final state of rest are categorised with things that are making a move back to base.

(115) Alkwarrer aylenanth ak-eyn-enh bush banana 1DU.EX cut-RET-IMPP ‘We two cut (were cutting) bananas as we went’ (Yallop 1977:61).

In (116) the particular transitive verb involves that both the Subject and the Object return to a previous base of the Object (who is dragged back angkep to the place where he committed a crime). There is a degree of ambiguity with Return in that it is sometimes not clear whether S or O is moving to a final point. This sense is similar to Kaytetye ‘induced motion’ -eyne (Koch 1984), indicating that something is being carried or caused to move away from its original location.

(116) Ratherr=ap angkep renh iytarr.akng-eyn-enh 3DU=CONJ back 3S:ACC drag-RET-IMPP ‘They (2) dragged him back’

The event of (117) is not necessarily happening all along the return motion path, but happens at a point along the path. The opposite action - exiting the burrow - is described in the same text by tyerrelh-enh- ‘exit-POINT’ which utilizes POINT (associated with an alh- motion away from base).

(117) Ayl-ek=anem renh Kwerrenarr renh aylp-eyn-etyek=anem sing-PP=THEN 3S:ACC rainbow snake 3S:ACC enter-RET-PURP=THEN ‘They sang the Rainbow snake so that it would re-enter (its burrow)’

Manual action occurs along an object with limited motion or without the motion of the whole S:

(118) Alerl atha angern-eyn-eyel, wait 1S:ERG dig-RET-PRES ‘Wait while I dig along’

With bounded (telic) verbs the action happens along the path while the Subject is returning or going away. Example (119) shows the distinction between what are activities (cutting, falling) done at one location and an activity which necessarily involves the manual action along an object: ‘stripping off leaves along the branch’.

93 (119) Artern-em aytn-em alperr-rnem alperr-rnem alth-eyn-em chop-POT fall-POT leaf-NPL leaf-NPL strip-RET-POT ‘They would chop and the branches would fall. They would move away stripping off the leaves’

With State verbs, one or more objects can be configured in a line, as in (120), which recedes away from the reference position. Objects recede away figuratively with a fictive motion which is also mentioned in connection with alpeyel in §2.4.1.3.

(120) arla=nt nhanyem aynt-eyn-eyel. root=JUST this lie-RET-PRES ‘There's only the (horizontal) root here’

RET can also refer to a line of separate objects, for example houses which are in a line, although I don’t have any examples of this use in the text data.

Another function of RET involves a metaphorical return to a previous state of affairs, similar to the re- prefix in some English verbs, for example ‘repay’ and ‘renew’. This sense is illustrated in example where the meaning ‘repay’ arises with the root anth- ‘give’. This sense is similar to that reported for Arrernte -irtne by Wilkins (1989:276, 2006:60). Another example is mwerrirreyneyel ‘getting better again’, a metaphorical return to the state that someone was in before they became sick. See Wilkins (1989: 276). I would not make the claim as Wilkins’ (1989:279) does for Arrernte that the figurative interpretation of RET is derivational.

An extended sense of RET is ‘move away to an (implied) endpoint’ which involves movement away from the reference position which can be that of the speaker or Subject. Related to ‘move away’ is ‘be dispersed or diminished’. This can be applied to objects which move to a final state as with (121). There is ongoing activity, reinforced by the repetition of akely=anem ‘then a lesser amount’ in example (122), which suggests that the flour diminished at a steady rate.

(121) Arrnga=an atyenh thel-elh-eyn-enh aleth=antey knife-penh blood=FOC 1S:POSS pour-MED-RET-IMPP always=STILL knife-SRC ‘My blood was flowing out after (being cut with) the knife’

(122) Flour=anem akely=anem akely=anem irr-eyn-enh flour=THEN less=THEN less=THEN IV-RET-IMPP ‘The flour was becoming less and less’

94 The use of Return with areyirr ‘to build up’ in (123) can be related to the apparently irreversible buildup of the funds, reaching a point of no return, similar to small objects and body parts which are involved in Return motions. Return is used to describe a recognized endstate in (124).

(123) Areyirr-eyn-enh=arl=ap maney angkep accumulate-RET-IMPP=SUB=CONJ money back

anth-eyn-ey.angenh=antey akngerr=anem irr-ew ikwerenh give-RET-NEG=STILL many=THEN IV-PP 3S:POSS ‘It kept building up! He couldn’t repay the money, his bookup (account) kept increasing’

(124) Aylpaty rlengk=ant ra ilkw-irr-eyn-eyel breast now=JUST 3SNOM big-IV-RET-PRES

alakenh-anyem il-eyel aylpaty akarnterrng like-around Call-PRES breast akarnterrng ‘Breasts that have just started to get bigger, they are called akarnterrng’(AED: 7)

Previous descriptions of Return -eyn have followed Strehlow (1944:172), for example Yallop’s (1977:60) description of an ‘iterative’ stem-formative which has a sense of ‘an action performed while moving or an action maintained or repeated’, perhaps over-generalising from interpretations such as (115). However, I conclude that repetitive action is not basic to its interpretation. There are other morphemes which have iterative readings such as Frequentive and Distributed.

4.3.2 -enh Point

POINT describes a distinct event which occurs at a point along a motion path. Typically the motion is resumed after the event. Without POINT the verb is unspecified for location and motion. The difference between the verb with and without POINT for the verb areyel ‘seeing’ is shown in (125) and (126).

(125) ar-ek see-PP ‘saw’

95

(126) ar-enh-ek see-POINT-PP ‘saw while going along a motion path’

The speaker’s choice of whether a POINT form is used may be subtle, depending upon whether the action is judged by the speaker to have occurred along a motion path. When I first recorded the text Atyetyakwerleny (1996), my informant used altyewey-alhemel as in (127). Viewed as a whole motion event, the action can be seen as part of a journey. In 2005 upon reflection the informant changed the verb to include POINT and the sentence was changed to (128).

(127) Parrik=anem altyew-ey.alh-em-el fence=THEN rollover-GO&DO-POT-SS ‘While going and pushing the fence down’

(128) Parrik=anem altyew-enh-ey.alh-em-el fence=THEN rollover-POINT-GO&DO-POT-SS ‘While going and pushing the fence down on the way’

The analysis of the verb in (128) as POINT followed by DO&GO provides evidence that POINT and DO&GO are not in the same structural position in the verb.

In contexts like the above where the motion is separately expressed, the semantic contribution made by POINT may go unnoticed and probably for this reason Yallop (1977:62) follows Strehlow (1944) in claiming that it is ‘impossible to associate any specific function’ with it. Strehlow (1944:177) stated that the putative anama suffix (which I take to be equivalent to - enh-em POINT-POT in Alyawarr) ‘is a common termination in lieu of the -ma suffix of the present indicative tense in the non-W.A. (Western Arrarnta) verbs, although it is not unknown in W.A. It has no special force’. There are similarities with Austin’s (1981:79) ‘prolative’ in Diyari. See also Breen (Dixon 1976:752) and examples from Wunambal (Vaszolyi in Dixon 1976:629-46). POINT is homophonous with IMPP, but is distinguished from it because IMPP occurs in the obligatory suffix slot in the verb and can co-occur with POINT. The analysis taken here is different from the entry for POINT in the AED (Green 1992:134) which reads ‘goes on verbs to show that the action is happening while the actor is going past somewhere or moving along.’ The action and the motion are concurrent (but not necessarily co-terminous)’. A single event occurs at a point within the motion path. Often the site of the POINT activity will be a natural boundary as in (129) where a car passes through water. The boundary is a gateway in (130). A

96 discrete activity occurs which involves a temporary suspension of the motion in (131). In the case of riverbeds and gates, a reference position is obvious but in many narratives the exact location at which POINT occurs is not evident.14

(129) Kwaty-angkwarr alh-ew apeylp.atw-enh-eyel mwetek-el water-PERL go-PP splash-POINT-PRES car-ERG ‘The car went through the water, making it splash’ (AED:66)

(130) Keyt atherr weth=an iterl.ar-eyel. yanh-angkwarr=anem gate two that=FOC know-PRES there-PERL=THEN

rwaylp-enh-ek pass-POINT-PP ‘(We) know those two gates. Then we passed through them’

(131) Mwetek-weny alh-erl.ew-enh nantew-el-ant mail-warl. vehicle-PRIV go-PL-IMPP horse-INS-JUST mail-ALL

Arleng-ey. Anyent.inger apek aynt-enh-em-el. distant-EMPH once maybe camp-POINT-POT-SS

Atherr-inger apek aynt-enh-em-el. two-times maybe camp-POINT-POT-SS ‘Without cars they would go by horse to (get) the mail. A long way! And camping maybe one or maybe two nights’

Some POINT verbs are almost lexicalized as seen in Table 20. An example is anpenh- apparently involving a root anp-, for which there are no other forms except for anperl-alh, DO&GO. See §2.5.3.2.

14 For this reason I have not used the gloss ‘PAST’. There will be no difference between the points of view of the narrator and the protagonists in the story and therefore no need to distinguish between ‘text internal and text external reference points’. 97

Table 20: POINT-marked verbs anpenh ‘depart, start out’ rwaylp pass rwaylpenh ‘pass something on the road’ ywerr finish ywerrenh ‘pass out of sight’ artarnp cross artarnpenh ‘make a creek crossing’. aylp enter aylpenh ‘enter, dive into, (sun) set’

4.4 Sequenced Motion compounds

In this section I describe the compounds which are formed with the Basic Motion verbs which are discussed in §2.4.1 and then the -elp forms which are discussed in §4.4.4. The sequence morphemes convey the relative temporal sequence of the motion and the action of the initial verb. They indicate that the activity occurs before the motion DO&MOTION (-erl) or that the activity occurs after the motion MOTION&DO (-ey). The Sequenced Motion compounds shown in Table 21 are semantically compositional, that is, predictable from the sum of their component parts. Although their meanings are also determined or constrained by their accompanying verb and other morphemes, these do not change the meaning significantly. Non-verbal material may intervene between the stem and the compounded Basic Motion verb. The Motion compounds are highly productive and have a high frequency of occurrence. Yallop (1977: 62) reports that they are common in narrative style.

Sequenced Motion compound gloss reference -ey.alh go and do §4.4.2.1 -ey.alp go to base and do §4.4.2.2 -erl.alh do and go §4.4.1.1 -erl.alp do and go to base §4.4.1.2 -ey.elp run and do §4.4.4.1 -erl.elp do and run §4.4.4.2

Table 21: Sequenced motion compounds

Motion before action Motion after action -ey.alh -erl.alh ‘travel away from base and do V’ ‘do V at base and travel away’ -ey.alp -erl.alp ‘move (back) to base and do’ ‘(move from base to endpoint) to do V and move (back) to base’

98

4.4.1 DO&MOTION -erl

DO&MOTION indicates that the activity of the verb occurs either before a motion or a change in stance. There is no evidence to relate -erl to participial or continuous aspect functions.Yallop (1977:63) stated that ‘Neither is it entirely clear what distinctions, if any, are conveyed by the different ligatives’, where his ‘ligatives’ include what I analyse as the sequence morphemes. One possibility, in his view, is –erl having a participial function e.g. atwerl-alpek would be atwel alpek ‘hitting, went away’. However, the motion and the verbal event are not happening concurrently which is the case for the concurrent activity marker Same Subject -el15. See for example (265) in §5.3.5. Yallop has apparently selected the wrong participial form in the translation of the example. The similarity would be with English past participles in a complex sentence with an interpretation ‘having hit, went away’.

In contrast with activities marked by MOTION&DO, the DO&MOTION activity often appears uncompleted and seems to refer to unrealised states or events. Yallop (1977:62) has taken a ‘present continuous’ analysis16 over from Strehlow (1944:178) who understands this morpheme as ‘to be doing’.17

I disagree with the analysis of continuous aspect for -erl with Motion verbs. At the reference time, the action of the -erl stem is concluded whereas the motion is not concluded (because the verb action precedes the motion), giving the impression that the activity of the verb is ongoing and continuous. The activity described may appear more continuous where it occurs with stance-existence verbs which have an inherently continuous aspect. Some complex markers such as -erl.iw, -erl.ayn and -erl.arrern involve activities of short duration and always occur with PP in past reference. Other -erl complexes such as -erl.ap and -erlenty.akng only occur with IMPP and refer to events of longer duration.

15 Yallop (1977:130) regards -erl as a participle and gives examples which are inconsistent. 16 Wilkins (1989: 253) regards Arrernte +rle as meaning ‘do continuously’ but that distinct +rle occurs in continuous verbs only and not the +rle that occurs in Motion verbs. He regards them as distinct but homophonous elements. 17 Wilkins (1989:287-288, notes 23-25) comments upon Strehlow’s analysis of these morphemes, finding that they are insufficiently differentiated. For example, three forms attached to the verb tweme ‘hit’ which have distinct interpretations in present day Western and Mparntwe Arrernte are glossed ‘hit while going along’. 99 A possible interpretation of -erl is ‘begin Ving at the reference time’. This differentiates - erl.aynt ‘start doing V continuously at the reference time’ from -aynt ‘do V continuously at an indefinite time’.

4.4.1.1 -erl.alh

‘do V and then move away from base’

The motion is subsequent to the verb action as in (132). The form is comparable to the account of +rle non-concurrent verb action in Wilkins (1989: 287). A change in stance can also be marked by -erl.alh, specifically a change from the sitting position to the standing position as given in (133) where the crippled man is told to ‘get up!’ from a sitting position.

(132) Il-erl.alh-ek renh Iylpakepey an-erl.aynt-enh.atherr tell-DO&GO-PP 3S:ACC Iylpakepey stay-CONT2-IMP.DU

mpwelanth=an aherr-ek=arl alh-em-kety 2D=FOC kangaroo-DAT=SUB go-POT-REAS ‘They told Iylpakepey before going, “You two stay here because we are going hunting for kangaroo!” ’

(133) rtn-erl.alh-a! stand-DO&GO-IMP.S ‘Stand up!’ (Alyawarr translation of Bible, book of Acts 3:6)

4.4.1.2 -erl.alp

‘do V and then move to base’

The Subject does an activity and then moves to a base:

(134) Kwaty=an kel=arl anwenantherr ar-erl.alp-ew apmer water=FOC OK=SUB 3PL.EX see-DO&BASE-PP place

Arlkenty-el=an place name-LOC=FOC ‘Then we went and saw the waterhole at Arlkenty (and came back)’

100 4.4.2 MOTION&DO -ey

MOTION&DO indicates that the action of the verb occurs after the motion. The motion is completed at the reference time. Strickland (1998:107) explains it as ‘the actor goes somewhere before doing an action’ or the action should be done simultaneously. Wilkins (1989:287) analyses the corresponding Arrernte tye- as ‘prior motion’.

4.4.2.1 -ey.alh

‘move from base and do V’

The sense of -alh here is similar to that of the free verb. The relation to the base is evidenced by for example, ayntey-alhek ‘go and stay’ (Simon Ross, pers. comm.) which means ‘camped for only one night’ in (135) and the location where the Subject stays will not be the final destination. If the Subject stays in the new location without returning to the original base as part of an extended journey, the motion is referred to with -alp. The -alh verb only refers to the initial movement away from the base or deictic centre. The -alp verb indicates that the Subject is either returning to the original base or taking up a new base, thereby shifting the deictic centre to the new location, cf §2.4.1.3.

(135) Irrwelty-itwek apmer atyenh-itwek aynt-ey.alh-ek ra place-LOC camp 1S:POSS-LOC lie-GO&DO-PP 3S:NOM ‘It went and stayed at my camp at Irrultja’

In (136) the Subject (Rainbow Snake) goes away from its base to a point where it then dries itself in the heat of the sun.

(136) Tek.an-ey.alh-enh ra atwerrp=an. dry out-GO&DO-IMPP 3S evening=FOC ‘He went and dried out in the evening’.

A number of the -ey.alh compounds have a ‘suddenly do upon arrival’ reading which probably led to Yallop’s comment that they were ‘punctiliar’ with arey-alh which means ‘spotted, discovered’, evident in (137). The lack of a fixed destination for compounding -alh can be understood in the difference between arey-alpew ‘go away to or back to a final point and see’ and arey-alhew ‘go and see’ but also ‘discover’ where the ‘finding’ part of the activity can occur anywhere along the journey and is not restricted to any fixed spatial reference point.

101 (137) Atwaty-atwaty=an rernem ar-ey.alh-enh gorge-RED=FOC 3PL see-GO&DO-IMPP ‘Then they would spot gaps’.

I concur with Wilkins (1989:287 footnote 23) that Strehlow (1944:172) doesn’t seem to indicate ‘prior motion’ for -tye which means that ‘three distinct forms seem to have the same meaning’. Yallop followed Strehlow’s analysis and not surprisingly claimed that -ey had no independent meaning.

4.4.2.2 -ey.alp

‘move to base and do V’

The -ey.alp compounds indicate ‘return to base or move to a new base and do an activity’. The compounding verb -alp appears to have the senses of the independent Basic Motion verb. The use of -ey.alp to indicate ‘move to a new place’ is evident in (138) where the Subjects referents are making a new camp and have never lived in the area before.

For Arrernte, Wilkins (1989:286) correctly maintains that this means ‘arrival back at a place’ not ‘upon arrival’ as Strehlow (1944:172) claims, but for Alyawarr this must be qualified by saying that the Subject’s final point of rest or destination is the critical factor here as alp has a reading of ‘go (away) to a new base’.

(138) Waylpel ahelengkw-el-rnem apmer mpwar-ey.alp-ew whitefeller aggressive-ERG-NPL camp make-BASE&DO-PP

yanh-ew-ey alalew-ey.alp-ew there-LOC-EMPH shift-BASE&DO-PP ‘The aggressive whitefellers went and made their camp there. They went and shifted to there’

In §2.4.1.3 I indicated that the free verb alp is used for situations where the Subject goes away. It occurs in situations where there is an ‘exiting, moving away’ as some form of final ‘movement to a point of rest’ as in (139). A Rainbow Snake comes out of its burrow and then continues its motion. The -ey.alp complex also occurs where objects are said ‘to be carried away’ in a line of movement away from their point of origin or along the line of movement of the Subject, in (140), the Rainbow Snake.

102 (139) Arratyerr-ey.alp-ek=anem akapwert=anem ra akapwert-tangkwel=anem straighten-BASE&DO=THEN head=THEN 3S head=FIRST=THEN ‘Then it straightened up, its head first’

(140) Awer-el awerrk-enh-ey.alp-ek=anem arwerl ingkerr wind-ERG twist-POINT-BASE&DO-PP=THEN tree all ‘The wind went away, twisting up all the trees’

Yallop (1977:63) claims that the -erl.alh and -erl.alp compounds seem to indicate events that last some time. Most of his examples consist of -erl.alp forms. Where the Subject of example (141) was getting other horses along a motion path, the activity would continue. This example seems to be different from my translation and analysis of DO&BASE in §4.4.1.2.

(141) Yarraman ingwer-rnem anterrkw-erl.alp-ek horse other-NPL get-DO&BASE-PP ‘Then (we) went off getting other horses’

4.4.2.3 -enhey compounds

The MOTION&DO sequencing morpheme can be affixed to POINT, which was examined in §4.3.2. Yallop (1977:63) notes the use of -enhey- but doesn’t assign any meaning to the -enh POINT element. The Subject performs an action while moving along a motion path from a base to a destination -enhey.alh (142) or to a base -enhey.alp

(142) Ampen-enh-ey.alh-enh nantew-akert atherr=an track-POINT-GO&DO-IMPP horse-COM two=FOC ‘The two horsemen were tracking as they travelled along’

In (143) the boy presumably sees the birds before he arrives at his camp, but they are seen on the return journey and ‘seeing’ is regarded as a discrete and distinct activity, before arriving home. Ar ‘seeing’ is probably a special case in that it denotes an action which can be done without interrupting the motion.

103 (143) Ra awey ar-enh-ey.alp-enh Aley thip-rnem akarrerr-eyel 3S boy see-POINT-BASE&DO-IMPP now bird-NPL gather-PRES

arelh ampwa ikwer-itwek old woman 3S:DAT-LOC ‘The boy saw (them) as he was returning, “Now the birds are gathering around the old woman” ’

4.4.3 Reduplication and Motion compounding

The sequence morpheme reduplicates with Basic Motion and Deictic Motion compounds:

(144) Arrwekeleny-el arwa nantew-akert=an lat anth-erl-anth-erl.alp- enh ancestor-ERG just horse-COM=FOC letter give-RED-DO&BASE-IMPP ‘The ancestors would go around with horses handing out letters’

The context of the reduplicated verb here is that the letters were given to a number of different recipients in different places during a trip by the same Subjects, and the temporal ordering of - erl ‘do and move’ applies. This meaning seems to be enhanced in combination with Imperfective -enh with a habitual reading ‘to do V habitually’. In my data, reduplication only occurs with Sequenced Motion compounds, SIDE -erl.iw and DO&RUN -erl.elp. In (146) it can be seen that while monosyllabic roots involve the addition of –erl, disyllabic roots have no joining morpheme. This could be a more general pattern. Evidence for this comes from examples such as atnarnp-atnarnp-alp, ‘hopping’ a disyllabic root reduplication. At this stage there are few examples and more evidence would need to be found to confirm that this pattern is general.

Reduplication occurs with the Sequenced Motion forms in (145). A monosyllabic root and a sequence morpheme are copied to the base. With telic roots the interpretation is ‘Subject does the verb action in a number of places while going along a motion path’. The killers were going from place to place killing people. The temporal sequencing of -ey ‘Motion and Do’ applies and involves multiple instances of killing by the same Subjects in different places. Either -alh-or - alp follows the base.

(145) Atw-ey-atw-ey.alp-ew

kill-BASE&DO1-RED-BASE&DO2-PP ‘(They) went and killed, went and killed (at Arrtyeler)’. (Nugget Smith, Lake Nash history)

104

The sense with atelic verbs is ‘to keep doing something continuously while going (away)’. The reduplicated forms in (146) are from ampen ‘follow’ and intern ‘smell’ (trs).

(146) aleyart ra atwarr-enh-ew ampen-ampen-alp-em=anem there 3S descend-POINT-PP follow-RED-back-POT=THEN

intern-intern-alp-em arrpemarl renh smell-RED-back-POT again 3S:ACC ‘There it descended. They would follow it along, sniffing as they as they went along’

The reduplication of apey- in (147) indicates that many Subjects were involved in arriving.

(147) ap-ey-ap-ey.alh-ew=anem rernem. “Kel wenh!

arrive-GO&DO1-RED-GO&DO2-PP=THEN 3PL “OK QUOT

aley=an rernem ayl-elh-eyel=anem” wenh! now=FOC 3PL sing-MED-PRES=THEN QUOT ‘They arrived and (others) arrived. Hey! now they are singing!’

4.4.4 RUN

-erl.elp ‘Do V and then move quickly away along a path’ -ey.elp ‘Moving quickly along a path and then do V, culmination’

These two forms involve the -elp marker which has a variety of interpretations including ‘attenuative’ and ‘inceptive’ with verbal reduplication in §2.6.3. It marks the beginning of an event and has the bounded quality of an Achievement. There is no evidence that a single -elp morpheme occurs in Attenuative and RUN constructions.

There is no indication of the overall direction of the movement in terms of a deictic centre, source or endpoint of the motion. Motion between base and endpoint is not relevant or not related to the action. The point which the Subject moves to is neither a base nor the endpoint of travel. There is no major ‘turn’ or diversion from the motion path (as with SIDE -erl.iw) where the activity occurs. The verbal event is preceded by rapid motion, in contrast with the vertical motion complexes in which the Subject moves from one stationery stance to another stance (Frank Holmes, p.c). Unlike the -enhey forms the motion is either prior to or subsequent to the verbal action. There is a low number of tokens for these two forms (12). Both complexes collocate only with the Past Perfective marker and describe a completed activity. So far I have 105 not found stance/existence verbs with -erl.elp. Most of the -erl.elp forms involve transitive verbs but there are also Basic Motion verbs. There is very limited overlap of the two forms occurring with the same verb roots.

4.4.4.1 -ey.elp RUN&DO

‘Run along motion path and begin to do V’

Table 22: RUN&DO verbs and their interpretations

-ey.elp an-ey.elp-eyew ‘rapidly move and sit’ rtn-ey.elp-eyew ‘rapidly stand, stop suddenly’ ak-ey.elp-eyew ‘run and pick’ aytn-ey.elp-eyew ‘move and fall’ ipmelh-ey.elp-eyew ‘cease doing activity’ an-ey.elp-eyew ‘rapidly move and sit’ apat-ey.elp-eyew ‘become amazed’ mwerrirr-ey.elp-eyew ‘become better’ akat-ey.elp-eyew ‘become quiet’

The -ey.elp complex marker appears to indicate ‘to move and do’ and that an activity is just beginning which are shown in Table 22. The last three examples don’t involve actual motion but a change of state. The focus is upon suddenness. The marker indicates inception, the initial stage of an event.

The Subject has run or driven to the point where the action of the verb suddenly occurs. Strickland (1998:109) reports that it has ‘the aspect of something occurring suddenly or instantaneously’. With stance verbs there is a rapid movement to the point where a stance is adopted. The vehicle travelled along to the position and then stood in (148) indicating ‘quickly move and stand’. A motion leads up to a state. The motion activity may lead to an event described by a transitive verb, indicating ‘quickly move and do’ in (149).

(148) Yanh Toyota rtn-ey.elp-ek that Toyota stand-RUN&DO-PP ‘There the Toyota came to a halt’

106 (149) Apwert atherr=anem ratherr in-ey.elp-ek stone two=THEN 2DU get-RUN&DO-PP ‘Then they went quickly and got two stones’

The sense of ‘quick movement’ is perhaps reinforced by the fact that RUN&DO occurs with the PP tense marker -ew or -ek which suggests fast, completed actions.

In (150) the Subject falls down before completing the journey. He falls off a galloping horse angath-antey ‘close, not far’ without completing the motion. A hunter began to sit in the bird hide waiting for emus to arrive in (151). The Subject in (152) runs to get the bush fruit before the other children can get them.

(150) Aytn-ey.elp-ew angath=antey fall-RUN&DO-PP close=STILL ‘He fell down after not running very far’

(151) Ingwepenh-antey an-ey.elp-ew intart artartey-ey. morning-STILL sit-RUN&DO-PP bird.hide Granny-EMPH ‘In the morning he went quickly to sit in his bird hide. My granny’

(152) Reggie Camphoo-ey anterr-ew re-tangkwel ak-ey.elp-eyew Reggie C.-EMPH run-PP 3S-FIRST pick-RUN&DO-PURP

altwerr-ey bush.orange-EMPH ‘Reggie Camphoo ran so that he could start picking the bush oranges first’

4.4.4.2 -erl.elp DO&RUN

‘Do V and quickly move away’

Table 23: DO&RUN verbs and their interpretations

-erl.elp arrty-erl.elp-ew ‘burn and move away’ iw-erl.elp-ew ‘leave and move away’ anth-erl.elp-ew ‘give and move away’ atw-erl.elp-ew ‘hit and move away’

107 There are a limited number of tokens of this complex marker, which are shown in Table 23. All examples suggest an action followed by a rapid move away along a motion path. In (153) the action continues to have consequences which are relevant at the time of reporting. Although the killing is a completed event, the ongoing drama and consequences are still unfolding from this event at the time of the speaker’s reporting. The Subject has made a journey from the point where he killed the victim, to the point where he has been apprehended. The event began with the activity of the -erl.elp verb. The contrast with SIDE -erl.iw is that SIDE indicates a ‘turn’ in the motion path which is not conveyed by -erl.elp. The brief motion is directly to the point where the action occurs. As it doesn’t refer to a completed motion to another location, Basic Motion verbs are usually needed to complete the picture of where the Subject has moved to.

DO&RUN occurs frequently with Achievement verbs, including anth- ‘give’. In (154) Frank Dean is a policeman who is relieving Jack Kennett, a police colleague who is going away on leave. Dean is at the beginning of his tenure after being appointed by his predecessor.

(153) Nh=an ngenh irrkaty-warl akng-eyn-eyel this=FOC 2S.ACC same-ALL take-RET-PRES ntwa=rl atw-erl.elp-ek-warl. 2S:ERG=SUB kill-DO&RUN-PP-ALL ‘Take you back to where you killed (the other)’

(154) Angwenh-el? Jack Kennett-el anth-erl.elp-ew renh Frank Dean=anem who-ERG Jack K.-ERG give-DO&RUN-PP 3S:ACC Frank Dean=THEN ‘Who? Jack Kennett then appointed (literally: gave) Frank Dean’.

The DO&RUN marker is not mentioned in other accounts of the Alyawarr language. There is a phonological similarity to -erl.alp ‘do V and then move to base’ which can lead to these forms not being heard and recorded as a distinct form by researchers. The reason for accepting this form as a distinct form is that it is a counterpart to the well-attested -ey.elp with the temporal sequencing reversed, where the -erl component indicates ‘to move away after doing the verb action’.

4.4.4.3 -erl.elp DO&RUN Reduplication

Basic Motion roots alh- and alp- form compounds with reduplicated initial elements and -erl.elp (§4.4.4.2). The disyllabic reduplicant copies both the monosyllabic root and the sequence 108 morpheme as in (155). There are no examples in the data of unreduplicated -erl.elp with Basic Motion roots. The translation is ‘multiple Subjects move away from X rapidly’, and is typically applied to groups of people and things like sheets of roofing iron and letters. I have not found the corresponding RUN&DO with reduplication.

(155) Ater-penh ingwerenty alh-erl.alh-erl.elp-ew Wave Hill-warl fear-SRC other go-DO&RUN1-RED-DO&RUN2-PP Wave Hill-ALL ‘There were others who fled in fear to Wave Hill’

4.4.5 Compositionality and motion compounds

I have argued that Alyawarr Motion compounds are compositional. I now consider the handful of specific verb forms which Yallop (1977:65) claims are more opaque in meaning and therefore better analysed not as compounds but as single ‘opaque’ lexical items. The perceived lack of compositionality could just result from translation effects though, because the meanings of these compound verbs are predictable from the senses of their components if all the senses of the individual lexemes are considered. I discuss arey-alhem, as in (156) ‘go and see, discover’ as a semantically predictable compound in §4.4.2.1. Iwerl.alp in example (157) has senses of ‘leave’ in common with other iwe- forms which Yallop recorded e.g for iw-enh-ek ‘left (a place)’. Yallop (1977:164) has ‘drop, forget’. The ‘leave and go away’ interpretation is semantically compositional. I discuss the compositionality of Deictic Motion verbs in §2.5.3.2

(156) ar-ey.alh-em see-GO&DO-POT ‘come upon, find’

(157) iw-erl.alp-em leave-DO&BASE-POT ‘leave behind on a trip’

109 4.4.6 Compounding vs. morphemic complexes

In §2.5.1.3 I introduced semantic compositionality and argued that Alyawarr compounds are compositional. In §4.4.5 I dealt with what appear to be non-compositional compounds in Yallop’s analysis (1977:65). Now the arguments for compounding will be considered.

In relation to ostensible compounding elements within his category of Associated Motion in Arrernte, Wilkins (1989:277) claims that ‘there is nothing to distinguish the functioning of the complexes from the functioning of the monomorphemic suffixes’ - in terms of position within the verb structure. He prefers the term ‘morphemic complex’, avoiding a verb compounding analysis (Wilkins 1989, 1991). In his view ‘the meaning of these morphemic complexes is ‘not the mere addition of the meaning of the morphemes but is more idiomatic’. These forms are ‘unified forms deserving their own definitions and their own entries in the lexicon’. Morphemic complexes ‘convey a singular (idiomatic) meaning’ and therefore are not semantically compositional’. His chief evidence is the interpretation of intye.lhe ‘DO COMING THRU’ and intye.alpe ‘DO COMING BACK’ He compares these with the suffix -intye ‘DO COMING’. According to Wilkins this suffix means ‘motion towards speaker’ but in the complex intye.alpe, alpe ‘return’ would add a contrary ‘motion away from speaker’ component to the complex, and therefore the complex cannot merely be the sum of intye and alpe. Wilkins’ further argument is that ‘all the suffixal complexes in the category of Associated Motion’ are morphologically equivalent to single suffixes. The morphemic complexes ‘take up the same position in the verb as certain suffixes do, with the same inflection types following and the simplest analysis, therefore, is to say they are both of the same category regardless of morphemic complexity’.

Alyawarr lacks morphemes which are equivalent in meaning to intye ‘come to speaker’. In Alyawarr the ‘return to base’ translation of alp- is compatible with ‘come to speaker’ (§2.4.1.3) in the situation where the location of the base and the speaker are the same. Thus there is no negation of alp- ‘return to base’ as there is in Arrernte intye.alpe. Therefore alp- and alh- retain their uncompounded meanings in compounds and are semantically compositional.

The degree to which the motion forms in Arrernte assume an ‘idiomatic’ meaning unrelated to that of the uncompounded forms has been overstated. Even if the intye.lhe and intye.alpe forms were non-compositional, there would be no reason why other combinations of morphemes couldn’t be regarded as compositional and regarded as compounds, given that compositionality is a cline phenomenon.

110 4.5 Rapid actions

I have grouped the Rapid complex markers together because they include an -erl element and describe an overall activity that is rapid in some sense. The Subject moves away rapidly from the site of the verb action. The base and destination are unspecified. Rapid action complexes are less semantically compositional and I regard them as complex markers rather than compounds. Rapid morphemes haven’t been described in detail in previous accounts of Alyawarr. The Rapid complex markers are shown in Table 24:

Table 24: Rapid complex markers complex label meaning -erl.iw SIDE Move rapidly to side and do V -erl.ayn AWAY Do V rapidly and move away

Rapid complexes tend to combine with PP rather than IMPP in past reference (where -enh occurs it will indicate past habitual). I will describe SIDE and AWAY in this section.

4.5.1 -erl.iw SIDE

‘move from the path rapidly to do V and then move back to the path’

Applied to non-Motion (usually transitive) stems, SIDE indicates a diversion from a point located along a pathway between the base and destination. According to speaker Simon Ross (pers. comm.) the motion involves ‘turn’ which I interpret as ‘deviate from the motion path’. My analysis contrasts with the ‘continuous’ reading of Yallop (1977:62) which follows Strehlow (1944:178) who analysed it as an eastern form equivalent to Western and Northern Arrernte -la nama, usually described as Continuous18. The sense of ‘continuous’ is not applicable to SIDE. The semantics of the related root iw- ‘throw’ or ‘leave’ appears to be related to the rapidity of the event. The same form iw- occurs in a number of lexical compounds. Alyawarr speakers describe the event as ‘quick one’ and there is a typically short distance between the path and the point where the verb action occurs. The related Arrernte -rliwe is often translated as ‘like lightning, quick as lightning’ (Wilkins 1989:254). No specific prior or subsequent motion is involved. The lack of duration of the activity is not a sufficient defining property of SIDE as there are other complexes which involve lack of duration -‘quick’-

18 The -erl.iw marker in Yallop’s examples is analysed here as a homophonous allomorph of the Plural (§3.7.3). 111 movements which contrast with it. Wilkins (1989:269) notes the high frequency of the corresponding morphemic complex in Arrernte narrative.

In (158), there is a kind of zig-zag motion from one waterhole to another to inspect them and see whether they contain water. After quickly assessing their state, the Subject diverts back to the motion path.

(158) Arrangkw=arl. aynterrk arrpemarl aylenak ar-erl.iw-ew nothing=SUB dry too 1DU.EX.cp see-SIDE-PP ‘Nothing, we saw that it (waterhole) was dried up as well’

It is even possible to return to a base ‘camp’ as in (159), where the participant gets the rifle then returns to the motion path, but unlike -erl.alp ‘do V and then move to base’, the location of the base is not of concern to the speaker. SIDE relates the Subject to the motion path rather than home or destination points as with the Sequenced Motion suffixes. The contrasting inerl.alp would suggest that he was returning to a base and remaining there. In (159) for example, iw- erl.alp ‘leave-DO&GO’ and iw-erl.iw ‘leave-SIDE’ may be described by speakers as ‘same’ but events described by verbs inflecting with -erl.iw occur at an unspecified point along a motion path. While iwerl.alp means ‘to go and drop something off and then return’, iwerl.iw can be translated ‘go back to the motion path after making a brief diversion in another direction’.

(159) Waylpel anyent yarraman-enp alp-erl.ayn-ek apmer-warl=anem. whitefeller one horse-WITH back-AWAY-PP camp-ALL=THEN

Maket=anem in-erl.iw-ek. rifle=THEN get-SIDE-PP ‘One whitefeller went back to camp on a horse. He got a rifle’

There is also a figurative extension of SIDE from a spatial notion to a temporal sense to mean a brief or temporary state:

(160) Itna=n atha iylpel-erl.iw-ew name=FOC 1S:ERG forget-SIDE-PP ‘I forgot (its) name’

4.5.2 (-erl).iw Small Repetitive Reduplication

‘X does V repeatedly with small movements’

Reduplicated Repetitive occurs with verbs which are listed in the table below and are different from those which occur with unreduplicated SIDE. I only have eleven tokens of the 112 reduplication in the data and conclusions are drawn cautiously. The verbs describe bodily movements which are repeated and in quick succession. Typically these movements are made by a limb, involving the repeated extension or movement of the limb which is reflected in the iconicity of reduplication. In the reduplication, the -erl.iw form occurs with monosyllabic roots in which case only the single syllable of the root is reduplicated. See Table 25 and compare with Reduplicated TV in §3.6.2.2. There is no evidence that the reduplicant is either prefixed or suffixed to the base. Where the stem is disyllabic, the entire root is reduplicated and the -erl element does not occur, as shown in (161) where the apparent meaning of the verb is ‘tie something repeatedly, a number of times in quick succession’.

(161) Belt atyenh arertn-arertn-iw-ew atha renh belt 1S:POSS tie-RED-REP-PP 1S:ERG 3S:ACC ‘I tied my belt around it (leg)’

Another use describing processes such as the processing of seeds is illustrated in (162). The reduplication is formed from a monosyllabic root ath- ‘to grind seeds or make a damper’. Reduplication is used for repetition of an event. The Repetitive reduplication is similar in meaning to FREQ and ATT reduplication in describing small movements and attenuated activities.

(162) Alerl ath-ath-erl.iw-eyel amern-tangkwel wait grind-RED-REP-PRES damper-FIRST ‘Wait, I'll just prepare a damper first’ (AED2)

Table 25: Small repetitive Reduplication

Verb interpretation Reduplication Interpretation Intransitive itnwarnp- heart beat Itnwarnpitnwarnp.iw- heart beat fast, racing atnarnp- jump atnarnp-atnarnp.iw- hop around weyangk- breath wey-angk-angkerl.iw breathe fast arerr- see each other arerr-arerr.iw- keep glancing at each other rlkerl- eyes show rlkerl-rlkerl.iw- flash eyes at someone

Transitive arertn- tie arertn-arertn.iw- tie around quickly irrarlk- clean irrarlk-irrarlk.iw- clean quickly alth- pluck althaltherl.iw- pluck rapidly ath- make damper athatherl.iw- make a damper rapidly

113 arreylp- chew, munch arreylp-arreylp.iw- chew, munch quickly kwern- swallow kwern-kwern.iw- swallow quickly

There is also a similarity with one of the senses of RET -eyn (§4.3.1) which occurs with small movements: the limbs are involved in making rapid repeated movements. This SIDE+reduplication can be compared with atnarnp-atnarnp-alpeyel ‘hop along, skip along’ (Green 1992:110) and with E/C Arrernte (Henderson 1998: 240) ‘hurried repetition’ although the E/C morpheme occurs with different roots.

4.5.3 -erl.ayn AWAY

‘do V and rapidly move away’

AWAY involves rapid motion ‘do something quickly and move away’. According to Strickland (1998:113) this suffix ‘shows that the action happens suddenly, perhaps immediately or spontaneously’. This analysis is tentative as there is a limited number of tokens (<10) within the data. The exact difference between AWAY and DO&RUN is not clear. Like some of the other - erl compounds the -erl here probably indicates that the action of the verb occurs before motion, i.e it is DO&MOTION -erl. In (163) the cattle are described as nhantey ‘here’ and the implication is that the action will happen at the location where the Subjects are. In view of the illicit nature of the activity, there is a rapid move away from the scene.

(163) Rernem-ap angkerr-erl.ayn-enh “Ngay pwelek rntern-erl.ayn-ey 3PL-CONJ discuss-AWAY-IMPP Hey cattle spear-AWAY-HORT arrpemarl-aw! Pwelek-rnem nh-antey rtn-enherr-eyel!” again-EMPH bullock-NPL here-STILL stand-PL-PRES ‘They used to discuss and move away, ‘Hey, lets spear cattle again! All the cattle are here!’

AWAY is plausibly related to Return with a similar sense of ‘move away’ but without the sense of ‘return’. AWAY can combine with at least one Basic Motion verb and at least one Deictic Motion verb:

114 (164) Waylpel anyent yarraman-enp alp-erl.ayn-ek whitefeller one horse-WITH back-AWAY-PP apmer-warl=anem maket=anem in-erl.iw-ek Ap-ey.alp-erl.ayn-ek camp-ALL=THEN rifle=THEN get-SIDE-PP arrive-RET-AWAY-PP waylpel maket-akert-el whitefeller rifle-COM-INS ‘One whitefeller went back quickly to camp on horseback. Then he got a gun. He came back quickly with the gun’

Alyawarr anpenh- ‘depart’ anperl.alh and ‘depart and go somewhere’ usually refer to a rapid departure. There are cognate verbs for ‘run’ or ‘hurry along’ in other languages which Wilkins lists eg. arntpe ‘to run’ in Kaytetye. The -ayn part of AWAY could be cognate with E.C Arrernte -artne- which ‘adds a component of speed’ (Wilkins 1989:291) except that -artne is the first element of a morphemic complex in Arrernte. The Alyawarr form is consistent with this analysis in having prepalatalization, corresponding to the retroflex consonant in Arrernte.

4.6 Action along a path

The markers in this class combine with verbs other than the Basic Motion verbs. They occur exclusively with the IMPP tense-aspect marker and have an Imperfective aspectual reading.

Table 26: Habitual action along a path complex markers -erl.ap CONV X does V, while moving along a path -erl.enty.akng DISTR Distributed motion

4.6.1 -erl.ap Convey something along

‘X does V, while moving along a path’.

Convey has a range of senses involving motion in the data. The analysis here differs from that of the AED (Green 1992: 136) in which -erl.ap has the meaning ‘do continuously while in motion’. The problem with the AED definition is that other markers have the same or similar glosses, e.g. RET in §4.3.1 and the actual differences are not clearly distinguished there. Unlike RET which indicates ‘return’ or ‘away’, CONV appears to have no direction of motion. With intransitive Stance verbs, CONV refers to activities which occur when the Subject is being conveyed in a container of some kind as in a vehicle or a dish as seen in (165), since that allows the essentially static stance to be maintained during motion. 115

(165) Aley-anyem kwey ankwaynt-erl.ap-enh there-ROUND girl sleep-CONV-IMPP ‘From there the girl slept (being conveyed along in a car)’

The contrast with Return §4.3.1 and Stance verbs can be seen with:

(166) aynt-eyn-ek lie-RET-PP ‘Camped on the trip home’

With transitive verbs, there are multiple instances of a verbal event which happen along the motion path according to the aspectual character of the verb root. Activities indicated by unbounded verbs will usually occur all along the motion path and achievements indicated by bounded verbs will happen several times along the path. CONV is inherently imperfective and combines with Imperfective or Present tense-aspect marking.

In (167), Aboriginal children are being picked up by the whitefellers as those whitefellers travelled around. By contrast the direct object ‘food’ is ellipsed in (168). The spears were being used to procure food as they were being carried by their owners and appear to be in Instrumental function.

(167) Apekath=arl in-erl.ap-enh-aw. Waylpel-el atherr-ey halfcaste=SUB get-CONV-IMPP-EMPH whitefeller-ERG two-EMPH ‘It was the halfcastes that the two whitefellers went around getting’

(168) Ikwer-el-ant anyent-el arlkw-erl.ap-enh=an maket-weny-el 3S:DAT-INS-JUST one-INS eat-CONV-IMPP=FOC rifle-PRIV-INS ‘A person would just eat with it (spear), (taking it along) before we had rifles’

Abstract entities such as stories are also conveyed:

(169) Re tangkwel=arl alh-enh arrwekeleny angka=arl il-erl.ap-enh 3S first=SUB go-IMPP first story=SUB tell-CONV-IMPP ‘He went around telling stories- the first one to do that’

116 4.6.2 -erlenty.akng Distributed motion

‘X habitually moves back and forward along a path doing V’.

DISTR collocates only with the IMPP Imperfective past marker. It applies to activities which involve travelling along a motion path doing V on a regular basis. Unlike reduplicated forms which involve stationary iteration such as FREQ (§2.6.4), in this case the S regularly moves backwards and forwards along the motion path. Distributed Among a list of ‘miscellaneous’ (i.e. under-described or little understood) verbal operators Payne (1997:257) has a ‘distributive’ verb operator found in some languages of the world (specific languages are not mentioned) which has a sense of ‘all over the place’ or ‘with a back-and-forth motion’ which is comparable to DISTR. A plausible link exists between Distributed and transitive akng- ‘carry’ as the action of the verb is concerned with a whole motion path from one point to another, although DISTR does not derive a transitive verb. Another possible link is with Irrealis -enty §5.3.3.

There are many situations involving such a distributed motion, for example where animate Subjects travel along a pad19 or regular route to water and back again. The activity in (170) involves the routine daily activity of going and getting horses and leading them to water. Distributed Motion is plausibly related to the number of horses as much as to the number of occasions of movement.

(170) Nantew-rnem in-erlenty.akng-enh kwaty-warl akng-ey.alp-enh horse-NPL get-DISTR-IMPP water-ALL take-BASE&DO-IMPP

atwerrp-atwerrp evening-RED ‘They would get the horses and lead them to water in the evening’

In (171) the action of the verb is being done repetitively and to a degree that is annoying. Given this meaning, DISTR is possibly related to Arrernte ntye-iknge ‘do too much, too often’ (Henderson and Dobson 1994:351).20 Wilkins (1989:357) says of this that the speaker is unhappy or sick of something. However the Arrernte examples of this construction have no sense of ‘going back and forward doing X’ and may be marked by imperative and present tense marking which doesn’t occur in the Alyawarr data.

19 A ‘pad’ is a track that has been worn by animals such as cattle. 20 Wilkins (1989:357) refers to iknge as a clitic to nominals and adverbs and says that the speaker is unhappy or sick of something. 117

(171) Awey-ay! ilek-ek anngernt-irr-erlenty.akng-eyel boy-EMPH what-DAT nuisance-IV-DISTR-PRES

kwaty-ek=an aylekenh-ek water-DAT=FOC 1DU:POSS-DAT ‘Boy! Why are you always going back and forward and being a nuisance with our water?’

4.7 Vertical movement and ‘do while Y approaches’

The pair -ey.aynt UP and -erl.arrern DOWN indicate changes of posture, position in the vertical plane and doing an action while someone else approaches. They appear to be only partially compositional and I treat them as complexes. The ‘V change in posture’ sense of UP occurs with intransitive stems while the ‘do while Y approaches’ sense of -ey.aynt occurs with transitive verb stems as in Table 27.

The complexes account for a gap in the Arandic language data identified by Yallop (1977: 63), in which Alyawarr appeared to lack ways to express verb action occurring while climbing or descending. The number of tokens found in texts in my data is relatively low (less than 20) and therefore caution is used in their interpretation. Vertical motion forms are also apparently rare in Arrernte.21 Wilkins (1989: 286) admits that the two corresponding markers ‘are not attested in my textual corpus at all.’22 These forms are not widely reported for other Arandic languages. Goddard and Harkins (2002:215) mention the -katinyi ‘bring, take, fetch’ forms in Yankuntjatjara as a morphological means of marking ‘active verbs marking a change of position’.23 Table 27 shows the vertical motion markers.

21 Partly this is due to the semantic transparency between –ty.antye and the non-compounding form antye which means ‘climb’. The corresponding Alyawarr verb is cryptic: aynt means ‘lie’ and the connection between the stance verb and the compounding form is not as clear. 22 Wilkins (1989:286, 294) admits that for Arrernte ‘examples of –tye.kerle and –tye.antye have been gleaned from conversational snippets and through elicitation. See my comments about elicitation in §1.7.5. 23 Harkins claims that iwelheme, akemirreme and arrernelheme mean ‘lie down’, ‘stand up’ and ‘sit down’ respectively. It’s not clear that they have this meaning in Alyawarr although the words have very similar meanings. Yankuntjatjara is a language which is not closely related to Alyawarr. 118

Table 27: vertical movement markers complex sense Verb stems transitivity -ey.aynt ‘V happens in upward akemirr ‘rise’ intransitive direction, rtn ‘stand’ ‘X do V while Y ar ‘see’ transitive approaches’ tnhw ‘bite’

-erl.arrern ‘V happens in a downward iwelh, aynt ‘lie’ intransitive direction’ aytn ‘fall’

Alyawarr uses vertical motion complexes to indicate a change of stance or body posture in the vertical plane. The result is to change event classes (Aktionsarten in §4.9). These are bounded as Achievements have an ‘idealised lack of duration’ (Kearns 2000: 203). Both vertical complexes make Accomplishments from States and Activities and collocate with the Past Perfective tense marker describing a rapid, bounded activity.

State verbs usually represent the event class of State as they are unchanging, atelic/unbounded and have duration. Alyawarr takes the posture forms as basic, with the vertical movement suffixes modifying the basic forms to indicate the adoption of a posture. Table 28 shows the corresponding State, Posture and Agentive verbs.

Table 28 State, Adopt posture and Agentive forms

State Adopt posture Agentive an ‘sit’ anerl-arrern ‘sit down’ arrern ‘put down’ aynt ‘lie’ aynterl-arrern ‘lie down’ athen ‘lay down’ rtn ‘stand’ rtnerl-alh ‘stand up’(from sitting) rnkern ‘stand up’ rtney-aynt ‘stand up’ (from lying)

4.7.1 -ey.aynt ‘UP’

‘V happens in an upward direction’.

With intransitive verbs UP involves a change of stance or posture ‘get up and V’. The action involves a rapid upward movement and a change of posture from lying to standing as in (172). I will take a different analysis from the AED (Green 1992:129) and Strickland (1998:106) who 119 label -ey.aynt ‘continuous’, probably relating it to CONT1 -aynt and CONT2 -erl.aynt. Apparently a different sense of aynt is in focus in CONT. The complex is not easy to recognize and further analyse because, unlike the Kaytetye and E/C Arrernte forms, uncompounded aynt- ‘lie’ has no association with ‘rise, move upwards’, leading Yallop (1977: 63) to conclude that there were no vertical motion complexes in Alyawarr. However, it may be a case of sequential ‘lie and then V’. The effect can be that of ‘springing up’, ‘striking’ or ‘bursting forth’ from a more horizontal position to a more vertical position. In all examples recorded so far, UP occurs in situations in which the Subject is ‘lying’ in its usual posture, such as snakes in (173).

(172) Ingwe-penh anantherr akemerr-ey.aynt-ew night-SRC 1PL:NOM rise-UP-PP ‘We rose early in the morning’

(173) Alpentew-ey.aynt-ew nhak=anem antekerr=anem burst-UP-PP there=THEN south=THEN

tyerrelh-ey.aynt-ew exit-UP-PP ‘Then it (snake) burst out on the south side, it came out’

The aynt part of UP is likely to be historically related to the other Arandic language markers. Holmer (1963:45) and Strehlow (1944:76) comment upon the variation between Southern Arrernte irnteme ‘climb’ and other Arrernte dialects ntyeme, which is recognized as the regular correspondence between prepalatalised and palatal consonants across varieties. The prepalatalised Alyawarr aynteyel ‘lying’ closely resembles Southern Arrernte compounding root -irnteme ‘do while ascending an elevation’ (Strehlow 1944:76). This usage is comparable with Arrernte +tye-antye 'DO UPWARDS' (Wilkins 1989:285 and Henderson and Dobson 1994: 563). UP in Alyawarr corresponds to Eastern and Central Arrernte +tye-antye ‘do something while moving upwards’ and Kaytetye -y.ayte- ‘get up’ which ‘have different nuances depending upon the main verb’. According to Koch (1984: 26), ‘with stance verbs like atne- ‘stand’, it contributes to the notion of vertical movement, hence ‘get up’’. The compounding form [ie ayt-] is ‘identical to the verb root that by itself means ‘rise, come up, and grow (of hair and plants)’.

4.7.1.1 ‘Do V when X arrives’.

Another sense of UP involves the subject doing an action to someone who is approaching or arriving24. It occurs only with transitive verb stems and intransitive verbs that select a dative

24 Compare with English ‘come up’. 120 marked core argument. In (174) the two Subject referents of the sentence had been waiting to grab the other man. Their previous stance doesn’t seem as relevant as for the UP sense of ey.aynt although it apparently occurs with Subjects which exist in a horizontal or extended manner as for two snakes waiting to strike in (175).

(174) Ratherr=ap renh anterrkw-ey.aynt-ew=anem 2DU=CONJ 3S:ACC grab-UP-PP=THEN ‘Then they got up and grabbed him when he arrived’

(175) “Alerl!” wenh “Artnarp-ey.aynt-eyew ikwer-warl” “wait!” QUOT jump-UP-PURP 3S:DAT-ALL ‘Wait! (we will) jump up on him when he arrives!’

The most common TAM inflection taken by UP is Past Perfective.

Upwards motion is not a necessary component of the meaning, noticeable in the Alyawarr Bible translation of Acts 3:10 where the Subject is lame and yet would beg for money from the people who are arriving: arlkarew-ey.aynt-enh ‘ask, humbug-UP-IMPP’. Again there is a correspondence between Alyawarr UP with Arrernte +tye-antye ‘do something to someone as they are coming towards you or when they arrive’ (Henderson and Dobson 1994:565). Wilkins (1989:293) says that the equivalent Arrernte form -ty-intye ‘do on Z’s arrival’ indicates that the object or focus moves to the place where the Subject is and then the Subject does the verb action to the object/focus on his/her/its arrival. This construction ‘differs from all other associated motion forms because it attributes motion to the object or focus dative-marked core argument rather than the Subject’.

As Wilkins (1989:294) says ‘in all of these examples the subject of the action is human and s/he has been consciously waiting for the object/focus in order to do the verb action to them’. A similar situation prevails in Alyawarr where all of the examples of UP involve animate Subjects and a ‘conscious waiting’, evident in (176). I asked an informant about the difference between - ey.aynt and the verb without the complex -ey.aynt. The informant indicated that the difference is that when the complex is used, the implication is that the action is intentional; the snake was waiting to strike the man.

121

(176) Rtnwem-el artwa renh tnhw-ey.aynt-ek snake-ERG man 3S:ACC attack-UP-PP

iwerr-el=arl aynt-enh-el track-LOC=SUB lie-IMPP-LOC ‘The snake attacked the approaching man when it was lying on the road’

4.7.2 -erl.arrern ‘DOWN’

‘do V in a downward direction’

DOWN appears to be partially compositional as the transitive free verb arrern has a downward motion component meaning ‘put down, place down’ (and is in contrast to tywen ‘place up’). Transitive arrern occurs in this construction with intransitive stems, where the Subject changes position, but it does not affect the transitivity of the stem. It’s not clear why -arrern occurs in DOWN rather than its intransitive Mediopassive derivation arrern-elh. Given the limited semantic compositionality and the transitivity, I treat -erl.arrern as a monomorphemic complex. DOWN can combine with verb roots which also convey downward motion such as aytn ‘fall’, evident in (177). DOWN with State verbs conveys ‘change posture downwards’. The combination creates Achievements, enabling the speaker to describe an Achievement of moving from a standing posture to a lying posture, as with iwelh- in (178). This partially accords with Strickland (1998:116): ‘the actor puts himself in a set body position’ but my analysis also includes a downward component of motion. DOWN can occur with derived intransitive stems with IV and the Mediopassive.

(177) Pwelek aytn-erl.arrern-enh altywen-erl.arrern-enh cattle fall-DOWN-IMPP roll-DOWN-IMPP

rterrng=anem rernem atw-enh neck=THEN 3PL.NOM hit-PP ‘The cattle used to fall down and roll over. They would spear them in the neck’

(178) Ankw-ew-arrpantey iwelh-erl.arrern-ew atwerrp-atwerrp sleep-DAT-MOCK lie.down-DOWN-PP evening-RED ‘He lay down as if he was going to sleep in the evening’

In (178) DOWN seems to involve an involuntary sudden activity of downward movement, or where the Subject has less agency. This is the case even when the action seems to have been

122 initiated by the Subject. Often DOWN is associated with final states and death. A number of Antekerrepenh examples in Breen (1982) have been glossed with ‘finally, in the end’.

The Alyawarr examples appear to be different in function from Wilkins’ (1989: 284) examples of Arrernte tye.kerle which involve the Subject’s change in stance. An exception is (179) which involves the ingan ‘climb’ and DOWN.

(179) Ingan-erl.arrern-ew apwert ikwer-warl arrpemarl Wakerl-warl climb-DOWN-PP hill 3S:DAT-ALL again place-ALL ‘He climbed down the hill to Wakerl’

Compare with Arrernte tye.akerle (Wilkins 1989: 285) which indicates the downward motion of the Subject in (180).

(180) Artwe-le arrwe re-nhe twe-tye.kerle-ke. man-ERG wallaby 3sg-ACC hit-DO DOWNWARDS-pc ‘The man killed the rock wallaby on his (the man's) way down (the hill)’

Many of the forms in Carl Strehlow’s (1909) Aranda (Arrernte) wordlist (Table 29) have senses including motion in the vertical plane and changes of posture, similar to those recorded for Alyawarr.25

Table 29: Downward motion forms in the C. Strehlow wordlist

C Strehlow wordlist gloss Current Arrernte orthography renalitjikalama ‘sit down’ arrernelhetye-kerleme tnanbutjikalama ‘jump down’ atnarnpetye-kerleme irbutjikalama ‘go into the water, irrpetye-kerleme sun set’ (N. dialect)

25 In a footnote Wilkins (1989:286 footnote 19) says that T Strehlow 'claims that there is a verb form 'kalama' meaning 'to descend from a height'. Wilkins claims "that the verb does not presently exist in any of the Arandic languages" and "I have not found anywhere else in Strehlow's writing an example of this verb's use". Clearly Wilkins hadn't considered the C Strehlow wordlist (C Strehlow 1909). I have found at least ten examples of kalama as a free verb in the wordlist. There is also an entry akerleme in ECAED (Henderson and Dobson 1994:63). T.G.H Strehlow analysed kalama as a bound form or 'periphrastic verb' as he labels them in his 1944 work (Strehlow 1944:172). On the available evidence, it would appear that akerleme is more compositional than Wilkins claims.

123 4.8 Cultural explanation for complex motion verbs

The motion elements can only be understood in terms of the typical lifestyle and movement of Alyawarr speakers, or at least the traditional lifestyles that are presumed to be the setting for the historical development of the language. The contrastive oppositions which exist within the verb motion category characterise typical movements which are made by speakers of the language.

As I discussed in §2.4.1.1, the location is unrelated to the position of the speaker. There is a more important factor for Alyawarr and that is the location of the base which is the hearth or place conceived of as a base. Alyawarr are highly mobile and make temporary base camps, apmer. In the Alyawarr Region where there is a low rainfall and a low resource base, it was necessary to range over a very wide area to obtain food. They would shift camp regularly. Referring to another Central Australian language group, Ngaanyatjarra, Glass (1980:123) observes that ‘travel was an intrinsic part of their daily life’. I say ‘a’ base camp because in traditional hunter-gatherer societies the base camp shifted frequently. Alyawarr views the endpoint of the Subject’s journey as the determiner of what verb is used. The semantics of Basic Motion verbs alh ‘go’ and alp ‘move (away) to a base’ §2.4.1.3 relate to essentially two kinds of motion. One was the day trip awangk where food would be gathered, that is described as alheyel. Another kind of motion alpeyel involves moving to a base, either the original base or a new one. The move to a new base is a frequent occurrence despite the more sedentary nature of modern Indigenous lifestyles on the larger communities. The language accommodates this reality with the semantics of alpeyel.

Wilkins (1991:228) claims that in Arrernte ‘one is only compelled to represent the present spatial perspective on the event and its associated motion path if they choose to use one of these associated motion forms’. In Alyawarr, there is little evidence that the use or non-use of the forms has anything to do with the present spatial perspective. Non-use of the forms is not ‘uncooperative’ or ‘children’s speech’, but the additional information encoded in the compound is often redundant and not required to describe the event. In Alyawarr the choice of which form is to be used to describe a particular motion cannot be said to be ‘dependent on the spatial relation between the place where the report is being made and the motion path of the Subject of the action’ as suggested by Wilkins (1989:268) and diagram 6-1 (Wilkins 1989:281). Consequently, there are not ‘several possible and mutually exclusive reports which make use of the ‘concurrent motion’ forms’. Rather the forms used are dependent upon the base and the motion path of the speaker.

124

4.8.1 Example of Motion

The following is a description of an actual event, showing how the motion forms are used in Alyawarr to describe a variety of movements26. The motion events from this description are seen in Figure 2.

I stayed at the Ampilatwatja clinic and went to see one of the senior men. We talked briefly and then I took him to the council office. While I was talking with the town clerk of the Aherrenge Association at the council office, I indicated to him that we needed to attend a meeting of the health service board and that I should take him there. He went to the arnkenty single men’s quarters, saying to me before he went,

(181) Ayenh ntwa in-enh-etyek 1S:ACC 2S:ERG get-POINT-PURP ‘You pick me up on the way’

He wanted to be picked up from the single men’s quarters’ in order that we attend the health service meeting. POINT -enh represents a discrete and ‘bounded’ activity that occurs along a motion path.

When I arrived at the arnkenty I met a Kemarr man whom I call mwerey (mother in law generation) who asked me if I wanted to buy a pmwar, a wooden dish which was traditionally used as a shovel to excavate soakages to obtain water. My response was:

(182) Store-warl ayeng alh-em man atha in-erl.alp-etyek store-ALL 1S:NOM go-POT money 1S:ERG get-DO&BASE-PURP ‘I might go to the store, get money and bring it back’

I decide to go quickly anamerl to the store mwerey arralty aneyelarl-wety arwerl arterneyew ‘because mwerey is anxious to go and cut more (artefact) wood’.

I go to the clinic and stop. I tell the clinic administrator where I’m going -ilelhenhetyek- which is the POINT form of ilelheyel ‘say what you are doing, what you’re plans are’. Here POINT - enh is used because I am doing the activity of the verb on my way.27

26 I confirmed the appropriateness of all of my utterances with native speakers. 27 This has nothing to do with the speaker’s perspective, contra Wilkins 1989. It has to do with the movement of the Subject in relation to their base.

125

At the clinic another Kemarr asks me for a lift to the store, to go shopping before returning to the clinic meeting. We go to the store. I want to divert (solid line in Figure 2) back to the arnkenty because I have promised to return quickly so I say to Kemarr

(183 ) man atha mwerey renh anth-erl.ew-ey money 1S:ERG relation 3S:ACC give-SIDE-HORT ‘I will go and give money to my relation’

Figure 2: Example of motion suffixes

In this case I use the -erl.iw SIDE form of anth- ‘give’. In this case the trip back to the arnkenty is a deviation from the motion path from the base to the store so I don’t use –alh or -alp compound forms which mark the overall direction of travel. A deviation represents a move to do an activity and return to the motion path. This can even involve a deviation to the original starting point or even the termination point but there is a return to the original motion path. The 126 critical distinction between the transitive -erl.iw forms (non-number marking) and Motion compounding forms is that -erl.iw motions involve a movement from a pathway or non-base and don’t end at the base camp. Rather, there is a verb activity at a point away from the main motion path which involves the Subject’s movement to that point, the doing of the activity and the return to the motion path. If the journey is a brief move away, AWAY or simply the simple verb forms would be used to indicate brief activity at the base.

I returned to the store. Kemarr asked me for a lift to her home which is between the store and the clinic, “Ayenh iweyna!”- ‘drop me off’ which is the Return form of iweyel ‘drop off’. In this case RET is used to indicate that we are going back towards my base on my return journey.

4.9 ASPECT

Aspect is a grammatical category which is marked in a number of places within the Alyawarr verb. The main aspectual distinction in the verb is between completed and imperfective which is not marked more than once in a single verb. Separate aspectual markers exist for past tense: Past Perfective §5.1.3 and Imperfective §5.1.4. The morphological processes of reduplication FREQ §2.6.4 and ATT §2.6.3 convey iteration and inception respectively. A property relating to aspectuality and interacting with it is Aktionsart, a semantic categorisation of verb roots as distinct from grammatically marked aspect. Aspectuality is part of the intrinsic lexical meaning of verb roots (Talmy 2007:108).

The overall interpretation of verbs is determined by the Aktionsart of the verb root together with grammatical aspect. The critical Aktionsart distinction for Alyawarr is between telic and atelic roots. Telic or liminal verbs represent Achievements and Accomplishments. Atelic or non- liminal verbs represent States and Activities. Aktionsart determines how verb roots interact with grammatical morphemes. For example the verb root an- ‘sit, be’ has the Aktionsart of a State and most frequently occurs in past reference with the IMPP imperfective aspect marker. The Aktionsart characterization of Alyawarr verbs is preliminary.

CONT and STAT (Table 30) are imperfective and have ‘continuous or durative aspect’ (Yallop 1977:64). They are partially compositional and I have handled them as complexes. The CONT and STAT complex markers have senses which are extensions of the respective independent copular stance-existence verbs, aynt- ‘lie’ and an- ‘sit, be’ as in §2.4.2. e.g. CONT -erl.aynt represents extension through time as aynteyel ‘lying’ represents horizontal extension through space. The continuousness of ‘lie’ and ‘sit’ is based upon the static behaviour of humans. The relative abstractness of these senses is probably why the aspectual compounds are less

127 semantically compositional than the Sequenced Motion compounds. I will discuss in detail the aspectual complexes in the following sections. Table 30: Imperfective Aspect markers

CONT CONT1 -aynt ‘V continuously’ CONT2 -erl.aynt ‘start V and do continuously’ STAT -erl.an ‘State’

The complex markers shown in Table 31 occur with a restricted subset of TAM elements. CONT combines with Intentional mood markers and the future tense marker. STAT occurs with mood markers and non-realised tense markers. STAT is individual, subjective and used more of third persons whereas CONT is used of first and second persons. Further research and more data are necessary.

4.9.1 Continuous Aspect

All examples of (-erl).aynt in the data combine with Intentional Mood markers and the Future tense marker as outlined in Table 31. My analysis of Continuous Aspect markers is speculative. I have not been able to determine the value of -erl in CONT2 in order to distinguish between CONT1 and CONT2. For this reason I have treated them as monomorphemic complexes and labelled them CONT1 and CONT2 as a tentative analysis. I outline the apparent differences between the two complexes and suggest a possible difference in the following sections.

Table 31 Continuous Aspect with modal markers.

IMP POT PURP HORT SUBS FUT CONT1 -aynta -ayntem -aynteyew -ayntey -ayntey.alkenh -aynteyenh CONT2 -erl.aynta -erl.ayntem -erl.aynteyew -erl.ayntey -erl.ayntey.alkenh -erl.aynteyenh

128 4.9.1.1 CONT1

CONT involves extension through time of an event ‘continue doing V’. A reference temporal indicator appears not to be evident with -aynt in the data, with the exception of temporal expressions like ingwer-inger ‘another time’ as in (184) where -aynt indicates that the action will continue over a period of time. Generic events can occur with CONT1, involving non- referential or less referential arguments and Imperfective aspect.

(184) Rernem ilpatil-enh. Ingwer-inger akatyerr lyap-aynt-eyew 3PL burn-IMPP another time desert raisin grow-CONT1-PURP ‘They would burn so that desert raisins would grow’

4.9.1.2 CONT2

CONT2 appears to have a meaning of ‘start doing X and to continue doing it’ although this analysis is tentative and requires further research. A contrast between CONT1 and CONT2 complexes is shown in example (185). The addressee is already staying at the location. Semantically the critical distinction appears to be that -erl involves the commencement of an activity. Critically, the -aynt suffix enables a stem because a distinction to be drawn between “sit!” and “stay!” The form anaynta! indicates ‘stay, as you are!” The apparent meaning is ‘keep doing V’. The form anerl-aynta would indicate “move to stay here!” The latter has a reading of ‘start doing V and continue doing it’. This contrast could be one of ‘inception’, -erl.aynt indicates that the speaker intends for the action to commence from the reference time and to continue. CONT2 can occur with Basic Motion verbs as in (186). CONT2 also occurs with temporal indicators such as rlengk ‘today’, atwerrp ‘evening’ and =anem THEN. The activity of going in (186) commences at the reference time, atwerrp ‘evening’. The speaker intends for the activity of the verb to start and to continue.

(185) “Nga=an an-aynt-a artwamp! Nhenh akar-erl.aynt-a! warl 2S=FOC sit-CONT1-IMP.S old man this care-CONT2-IMP.S house

atyenh rlwenh-angketyarr. Ingkerr.” An-enh ratherr 1SPOSS tucker-LOTS everything sit-IMPP 3DU:NOM ‘ “You stay here old man! Look after my camp and all of the tucker! Everything!” Those two stayed there’

(186) Artwe-rnem=an angk-err-enh “atwerrp-ant alh-erl.aynt-eyew wenh!” man-NPL=FOC talk-REC-IMPP evening-JUST go-CONT2-PURP QUOT ‘The men were just talking together, “Let’s go in the evening”!’

129

CONT1 has the sense that the speaker intends for the action to happen sometime but there is no definite start to the verbal event. Compare the forms in (187) and (188).

(187) Ilkelh-enh arlkw-aynt-eyew mwerrangker pwelek. want-IMPP eat-CONT1-PURP nice beef ‘(They) wanted to keep on eating nice beef’

(188) Pwelek=anem=arl rernem arrkern-ew mwerrangker arlkw-erl.aynt-eyew beef=then=SUB 3PL try-PP nice eat-CONT2-PURP ‘Then they tried the beef, to find out if they would start to eat it and keep eating it’

4.9.2 -erl.an State

The primary function of STAT is to indicate imperfective aspect. The effect on the verb Aktionsart is to create states. I characterise the inflectional markers which occur with STAT as a class of tense, modal and clause connection markers (see Table 31) given in detail in §5.4.2. These include non-past tense, Habitual2 and all mood and clause connection markers.

The functions of STAT are unlike those reported for the corresponding forms in other Arandic languages, often referring to states of affairs, habitual actions and characteristic behaviours. Yallop (1977:130) analysed Arrernte rle-an (Wilkins 1989) as equivalent to the Alyawarr PRES -eyel (§5.1.1.). Alyawarr -erl.an has different senses from the cognate Arrernte rle-an. Although there are similarities in that both involve imperfectivity, in Alyawarr STAT also encodes modal remoteness (Lyons 1977: 719) and spatial remoteness. The uncertain status of objects is seen with STAT in combination with modal POT (§5.2.4.4). in which the Subject of the clause is remote from the speaker and there is no absolute certainty about the Subject’s location. The Subject is not visible or accessible at the reference time. Subjects are not in the vicinity of the speaker and are nhak ‘over there’ (distant) as shown in (189). STAT appears to be incompatible with both first and second persons in the present, as in (190). STAT can refer to prohibited or hidden objects or activities which are secret as in (191), where the Subject of the sentence will take the meat from the meathouse back to a distant place where he can cook it in safety.

130

(189) Nhak aynt-erl.an-em over there lie-STAT-POT Lying over there’

(190) *Nhenh aynt-erl.an-em here lie-STAT-POT ‘Lying here’

(191) Aylpwer.akng-eyn-enh salt meat ampern-erl.an-eyew ater-el-ey. carry.shoulder-RET-IMPP salt meat cook-STAT-PURP fear-ADV-EMPH

Arlkw-enh ra arleng-ew. eat-IMPP 3S distant-LOC ‘He was carrying the salt meat away on his shoulder so that he could cook it as he was frightened. He was eating it at a distance’

131 5 Obligatory verb marking

In this chapter I discuss the morphemes which occur in the final obligatory position of the verb: tense, aspect, mood (TAM) and clause joining morphemes. The obligatory morphemes have a range of functions, summarized in Table 32. The labels which I have adopted aim to capture the most central function of each morpheme or the most general to describe diverse functions. Glosses for these remain tentative – caveat lector.

Table 32 Alyawarr Obligatory Verb Markers

Label form Section Tense-Aspect Present -eyel §5.1.1 Future -eyenh §5.1.2 Completed -ek, ew §5.1.3 Past Imperfective -enh §5.1.4 Past Habitual -eyartingkerr §5.1.6 Past Habitual2 -ey.inkwern §5.1.7 Mood Imperative -Ø, -enh(err)-atherr, -enh(err)arey §5.2.1 Hortative -ey §5.2.2 Purposive -etyek §5.2.3 Purposive2 -etyek.an §5.2.3.1 Potential -em §5.2.4 Negation -ey.angenh, -eyang §5.2.5 Clause connection Purposive in complex clauses -etyek §5.3.1 Subsequent -ey.alkenh §5.3.2 Irrealis -enty §5.3.3 Subjunctive -emer §5.3.4 Same subject -el §5.3.5 Apprehensive -ekerr, -ewerr §5.3.6 Before -ekerr-awaty §5.3.7 Characteristic -ey.angker §5.3.8

5.1 Tense and Aspect

Tense can be understood as relating an event to the ‘deictic centre’, a reference time which is usually the present or the time of utterance. In the following subsections I will outline the basic system of oppositions within the Alyawarr tense system. This involves getting at the essential meaning of the tenses and distinguishing tense from implicature (Comrie 1985: 28). Tense and aspect are closely associated in the final inflections in Alyawarr, making it necessary for me to discuss both categories together in this section.

132 5.1.1 -eyel Present

The primary function of PRES is to indicate the event is occurring at the time of utterance which is the reference time. The events ‘hold at the present moment’ (Comrie 1985:38). In Yallop’s (1977:52) account Present and Potential (§5.2.4.1) are analysed as present tense. A consideration of the full range of uses of Potential -em in §5.2.4 reveals that it is a modal rather than a temporal marker. Yallop analyses -eyel (PRES) as Present continuous and notes that it is preferred for the description of present happenings and -em (POT) for statements concerning the future and as the default citation form for verbs. The latter was labeled ‘non-continuous present’. His ‘present’ analysis of -em (POT) was probably based upon pseudo-similarity with the Arrernte -me marker, labeled ‘present’ (Yallop 1977:58). Whether the temporal extension of the event relates to PRES or whether it’s a property of the verb stems themselves is critical to determining the temporal range of this marker. Rather than analyzing PRES as continuous aspect (Yallop 1977:49), I support Stanham’s (1972:46) analysis of present tense without an aspectual contrast. Comrie (1976:66) states, ‘as the present tense is used to describe rather than to narrate, it is essentially imperfective’. It is most natural with states and activities but not liminal predicates (Timberlake 2007:315). Present perfective doesn’t occur in languages except in unusual circumstances (Langacker 1991:252). The label ‘continuous’ is therefore unnecessary for -eyel.

PRES can refer to the extension of an event until after the reference time. The reference time is the time of utterance: for example in the sequence of clauses in (192) and (193) it is a Friday. A temporal qualifier or adverbial may specify temporal extension beyond the immediate present as with (193), but is not necessary for this, as demonstrated by (194). The temporal range of the woman’s stay in town is extended to the following Monday by the use of an adverbial.

(192) Ayeng an-eyel 1S:NOM live-PRES ‘I’m sitting/ staying’

(193) Ayeng an-eyel Monday-ek.atwety 1S:NOM live-PRES Monday-UNTIL ‘I’m staying until Monday’

PRES can refer to an expected event in the immediate future. Yallop (1977:51) regards PRES in example (194) as referring to the immediate future, like the English present progressive.

133

(194) Alheyel ayeng Mount Is’-warl go-PRES 1S:NOM Mount Isa-ALL ‘I’m going to Mount Isa’

5.1.1.1 Generic events in the present

Another sense of PRES is to describe events which typically happen. Most often reference is to indefinite Subjects. In (195) PRES contrasts a generic or omnitemporal statement with POT which indicates a possible event. Sometimes the sense is closer to a general or gnomic statement and is very similar to POT generic, as in (196). Most instances of PRES generic refer to customary activities. Often a contrast is being made between events which happened in the past and current ways of doing things where the reference time is ngwangk-ngwangk or rlengk- rlengk ‘nowadays’.

(195) Thip-rnem-el il-eyel. Kerlakwek, itey.itey, alkwarrer.arrpwernenh bird-NPL-ERG say-PRES dove mudlark stormbird

rernem il-eyel “Kwaty apety-em” 3PL say-PRES rain come-POT ‘The birds, the dove, the mudlark and the stormbird say, ‘Rain will come’

(196) Artwa irrpwerl alp-eyel rim-el-ant tay atw-enty-el=an Aboriginal man back-PRES rim-INS-JUST tyre hit-IRR-LOC=FOC ‘Aboriginal people get back on a rim whenever their tyres are punctured’

5.1.1.2 Historical present

PRES can also be a historical present (Timberlake 2007:313) as in (197). When the reference time frame of the event is already established, it’s acceptable to use PRES to describe a past event. The time of the event is not the same as that of the time of utterance. Cross-linguistically there is nothing unusual about the use of present tense markers to describe habitual events and the historical present. The corresponding Non-past Progressive -me in Arrernte has a similar function (Wilkins 1989:230).

134 (197) Artwamp=an apety-enh ilep-akert awey aylelh-eyel apal-antey old man=FOC approach-IMPP axe-COM boy sing-PRES unaware-STILL ‘The old fella was approaching, carrying an axe. The boy is singing and still unaware’

5.1.2 -eyenh Future

FUT indicates that the action will occur at a time after the reference time. Only 10 tokens appear in my data. Neither Stanham (1972) nor Yallop (1977) recorded FUT. Stanham (1972:46) analysed POT, HORT and PURP as future. Yallop (1977:58) labels HORT as future tense in his table of Arandic forms. The similarity between FUT and modal morphemes which have also been labelled ‘future’ by these previous researchers reflects the similarity between FUT and the category of Mood. The state of affairs described by FUT hasn’t occurred at the time of utterance and so the future describes an unreal state of affairs, similar to irrealis mood (Fleischman 1982). The analysis is somewhat tentative, but from the examples below it is clear that the use of FUT occurs when the speaker is referring to an event which occurs in the non-immediate future, but as early as later on the same day as in (198) and (199). Turtle (n.d, sentence 887) labeled FUT as ‘future continuous’ apparently in opposition to PURP which was glossed ‘future’ and also ‘future abbreviated’. Turtle (n.d. sentence 717) also analyses FUT as ‘intentive’. Wilkins (1989:227-230 footnote 8) claims to have found an aspectual contrast between future-marking morphemes in Arrernte: ‘the tense suffix -tyenhe 'non-past-completive (npc)' contrasts with -me 'non-past progressive (npp)’. I have found no contrast between a progressive and completive aspect for the future tense in the Alyawarr data. Stative imperfective is expressed by the use of STAT -erl.an. In Alyawarr the difference between FUT and POT is one of mood vs tense, i.e a tense-mood distinction of likelihood that the event will occur is more apparent than an aspectual distinction.

(198) Atwerrp-atwerrp wenh rlengk=arl rernem akarrerr-eyenh-warl wenh evening QUOT today=SUB 3PL gather-FUT-ALL QUOT ‘Later in the evening today, to where they will be gathering’

135 (199) Irrtywerr-em-el=anem “Irrepern nga alp-eyenh tell lie-POT-SS=THEN sorry 2S:NOM back-FUT arleng-el wenh!” long-LOC QUOT They said dishonestly, ‘Sorry to hear that you will be going away to a distant place!’

Other events such as those of (200) are probably more temporally distant from the reference time, which is the time of the reported utterance. A letterstick has been sent ordering the execution of two murderers. Presumably there is some time duration before the men act on the instructions of the letterstick.

(200) Nhatherr-anyem anwantherr atw-erl.alp-eyenh wenh those two-ROUND 3PL kill-DO&BASE-FUT QUOT ‘We will kill those two!’

An example of ambiguity between future and mood is shown in (201). FUT appears as the apodosis of a conditional sentence (usually only POT occurs). A possible interpretation is that Irrealis is being used for ‘when’ rather than ‘if’ and in that case the FUT tense morpheme occurs with the verb in the second clause. Also -ap CONJ occurs in the second clause which indicates a sequence of activities so that these may be seen as sequential rather than strictly conditional. This example shows the ambiguity of IRR which can indicate ‘when’ or ‘if’ and in this case it is the former.

(201) Ayeng alh-enty Alice Springs-warl atha=ap ngkweng 1S:NOM go-IRR Alice Springs-ALL 1S:ERG=CONJ 2S:DAT

in-ey.alh-eyenh get-GO&DO-FUT ‘When I go to Alice Springs I will get it for you’ (Turtle n.d, sentence 875)

5.1.3 -ek/-ew Past Perfective

The Perfective is in contrast with IMPP, as reported by Yallop (1977:49-50). The choice of which past marker depends upon the whether the events described are perfective or a continuing state of affairs at the past reference time. The events in example (202) may have been of long

136 duration as the narrator indicates that there was ‘a lot of rain’ so presumably rain fell for a long time but had stopped falling at the reference time.

(202) Kwaty=anem rntw-ey.alh-ew apmer Ipmangker-el-ey water=THEN rain-GO&DO-PP place Murray Downs-LOC-EMPH ‘Rain fell at Murray Downs’

The adverbial awangkarl ‘day trip’ in (203) indicates an event with the duration of one day whereas IMPP in the preceding sentence indicates a habitual movement. There is apparently a high degree of correlation between Past Perfective and liminal or telic verbs. Punctual actions which are completed such as ‘spearing, grabbing and placing something’ are only or typically described with Past Perfective. Some verb stems, such as that in (204), only occur with Past Perfective, because they are achievements.

(203) Arrtyeler-angkwarr alh-enh. Arrwekel awangk=arl alh-ew Arrtyeler-PERL move-IMPP first day trip=SUB go-PP

ra apmer Pwerrperreny-warl-tangkwel 3S camp Pwerrperreny-ALL-FIRST ‘He used to go through Arrtyeler. Then he would go on a day trip to Pwerrperreny first’

(204) akngan-ew originate-PP ‘originated, was created’

5.1.4 -enh Past Imperfective IMPP

IMPP indicates past tense with imperfective aspect and has progressive and habitual senses. Yallop (1977:50) claims that it means‘an action continued for some time or has continued up to the present.’ I claim that the action continues to a reference time, which may be the present. Two different suffixes could be used to describe the same event, depending upon how the event was viewed by the speaker, whether Past Perfective or IMPP. Unlike other related languages and dialects, cf Arrernte Past Habitual -tyarte (Wilkins 1989:227), Alyawarr marks no formal distinction between progressive aspect and habitual aspect in currently spoken language. The situation for Alyawarr appears to be that described by Comrie (1976:26) in which a language often has a single category to express ‘imperfectivity as a whole’.

137 With the Progressive sense of IMPP, an event continues and is uncompleted at the reference time. Some atelic verbs characteristically occur with IMPP:

(205) nthwenh ‘were searching for’ mpwelhenh ‘were waiting for’

IMPP is especially common with stative verbs which describe existence. For example, the state described in (206) continued up to the time of the discovery of the body. Within the story, events of lying, bloating, waiting and tracking utilize IMPP. Ongoing and repetitive processes such as cleaning fruit are described with IMPP. When a telic verb such as arrern- ‘place down’ occurs with IMPP, multiple instances of the verb action took place. IMPP and PP are contrasted in (207), an example from Yallop (1977:50). Yallop says that whether the speaker uses -ek or - enh depends upon their ‘focus of attention.’ In (207) the interval of time described by IMPP is the motion event and the action occurs throughout the interval.

(206) Artwa ra aynt-enh. Arlwar.rtn-enh man 3S lie-IMPP bloat-IMPP ‘The (dead) man was lying there and was bloating’

(207) Ap-ey.alp-enh-anem anwenantherr ampep-el anwenantherr iltyerr-ek arrive-GO&RET-IMPP-THEN 1PL.EX mid-LOC 1PL.EX rest-PP ‘Then we were coming back and on the way we had a rest’

The reference time may be the time of the utterance or an antecedent time, already defined in (208) and marked by Past Perfective. The telic root arlkw- in combination with IMPP suggests that there were a number of ‘eating of conkerberries’ events, distributed over a period which occurred up to the reference time.

(208) Nhanyem=anem atha atw-ek awey renh this=THEN 1S:ERG kill-PP boy 3S:ACC arnwekety=arl ra ingkerrenh arlkw-enh conkerberry=SUB 3S all eat-IMPP ‘This then is the boy that I killed, the one who had been eating all of the conkerberries’

138 5.1.4.1 -enh Habitual

Another sense of IMPP is the habitual, including customary and characteristic behaviour. Yallop (1977:58) uses the term ‘usitative’. Turtle (n.d) and Strickland (1998:75) use the term ‘customary’. Although other markers -eyartingkerr and –ey.inkwern also signify habitual aspect, the majority of speakers use IMPP to express habitual aspect with past tense.

IMPP describes characteristic behaviour in (209). The Subject is described as characteristically moody and aggressive. Another use of IMPP is to describe traditional or customary activities which occurred in the remote past as in (210). The adverbial awank ‘long ago’ defines the reference time as the remote past. IMPP may alternate with POT which also may have a habitual reading and is usually describes non-past but may also be used to describe past habitual events- see example (222) and discussion in §5.2.4. Both are concerned with events which are not restricted to one point in time.

(209) Ra=n ahelengkw innga=rl artwerrety irr-enh 3S=FOC vicious really=SUB moody IV-IMPP ‘He was really vicious. He used to get moody’

(210) Wimperr-warl tywen-enh awank=an arrwekeleny-el=an platform-ALL place high-IMPP long ago=FOC ancestor-ERG=FOC ‘Long ago, the ancestors used to put (dead) people on tree platforms’

IMPP is plausibly related to Reduplicated -enh which creates nominals from reduplicated verb roots §5.3.8.

5.1.5 Relative Past Tense

SUB marks the finite verb of a dependent clause in a complex sentence and in this context the tense is relative. Past tense markers occur with SUB to indicate the time at which an event happened relative to the reference time of the matrix clause. For example, the event of the dependent SUB-marked clause is antecedent to the reference time which is established by the main clause in (211). My analysis agrees with that of Yallop (1977:51) who describes -ekarl as perfect and possibly a combination of PP and SUB. For example, when SUB occurs with IMPP -enh as in (212), the event continues to the reference time of the main clause and -penh further indicates that the relationship between the clauses is one of the temporal succession of the main clause.

139

(211) Ahart-el renh ar-enh-ey.alh-ew akemerr-ey.aynt-ew=arl=an Ahart-ERG 3S:ACC see-POINT-GO&DO-PP rise-UP-PP=SUB=FOC ‘Ahart went and saw where (another) had previously arisen’

(212) Irrtyart ra=n amety=anem-irr-ew ra=rl angern-enh=arl=penh. spear 3S=FOC blunt=THEN-IV-PP 3S=SUB dig-IMPP=SUB=SRC ‘His spear became blunt after he had been digging’

5.1.6 -eyartingkerr Past Habitual

HAB occurs in older language and is used by older speakers with a low frequency of occurrence. The number of instances in my data of -eyartingkerr is 10. IMPP has replaced it in many if not all contexts. An informant said of a recently deceased person arlkeyartingkerr ‘used to shout out’ and then rephrased this to arlkenh (Jacky Beasley, p.c.). Hale (cited in Yallop 1977:58) reported -eyart and -etyart for the Macdonald Downs (southern) dialect of Alyawarr. The AED (Green 1992) has one example of -eyart without -ingkerr following. I would expect - eyart to be the equivalent of Arrernte Past Habitual -etyart (Wilkins 1989) and yet neither form occurs in my data. The longer form here may derive from -eyart and -ingkerr ‘all’. I conclude that HAB is a single suffix, especially as ingkerr doesn’t carry the primary stress of an independent word. The single instance of Past Habitual -eyart in the AED may be a transcription error. What is interesting in (213) is the juxtaposition of IMPP with HAB which presumably involves an event with the same regularity of occurrence.

(213) Horsetailer=an ra=arl nantew arlwerr-eyartingkerr=an horsetailer=FOC 3S=SUB horse round up-HAB=FOC

ingwepenh apelem-il-enh=an morning hobble-TV-IMPP=FOC ‘The horsetailer used to round up the horses in the morning and hobble them’

5.1.7 -ey.inkwern Past Habitual2

HAB2 indicates habitual action and appears to be an allomorph of HAB and functionally equivalent to the habitual use of IMPP. There are less than 10 examples in my data. Two verbs in the data are examples of HAB2 occuring with aspectual STAT. In addition, most examples of HAB2 occur with first person Subjects who are talking about activities which they have done. There is a possibility that HAB2 is less about the remote past and describes events which

140 happen until recent times, judging from examples such as (214). My analysis agrees with that of the AED (Green 1992:153) in which HAB2 ‘shows that the action used to happen’. Arrernte +tyunkerne (Henderson and Dobson 1994:578) and Antekerrepenh past habitual tye.nkwern (Breen 1982:48) are probably related.

(214) Inmenth-el=ap nhelkwer arlkw-ey.inkwern ngkwerlp-apeny=antey ashes-INS=CONJ like this eat-HAB2 tobacco-SEMB=STILL ‘(We) used to chew this one with ashes, like wild tobacco’

5.2 Mood

Mood is concerned with the degree of realisation, desirability or knowledge of the state. In writing this chapter, I am aware of the difficulty of defining mood and it is application to Australian languages along with a lack of standard terminology for modal markers of verbs (Sutton and Walsh 1979:29). One of the problems of establishing modality as a cross-linguistic grammatical category is that the semantics is often unclear and ‘there is no single semantic feature with which modality can be correlated in the way that tense can be regarded as the grammatical expression or grammaticalisation of time’ (Asher and Simpson 1994: 2535).

5.2.1 Imperative

Imperative morphemes are affixed to the verb stem. The singular, dual and plural forms of the imperative are arranged in the following table:

Table 33: imperative markers singular dual plural Ø -enh(err)atherr -enh(err)arey

The singular imperative is marked with Ø to the stem of the verb as in (215). Yallop (1977:53) regards the singular as an imperative but apparently didn’t record the dual and plural forms. Number is discussed in §3.7.

(215) Nga=n mpwelh-erl.aynt-a aker ngkweng-akert atw-erl.alp-ey.akenh 2S=FOC wait-CONT2-IMP.S meat 2S:DAT-COM kill-DO&BASE-FIRST ‘You wait! I will kill meat for you first’ (PM, hunting story)

141 5.2.2 -ey Hortative

HORT occurs only in main clauses. Yallop (1977:58) and Hale (n.d) list the morpheme under ‘future tense’. Yallop (1977:54-55) also analyses HORT as ‘permissive’ with a number of uses as first, second and third person imperative and also has uses as a ‘declarative future or as an intentional suffix’. Yallop claims that part of the difficulty with understanding the use of -ey is that there are differences among the Arandic dialects and possibly within Alyawarr. I disagree with the analysis of HORT as tense and agree with that of Strickland (1998:87-89) who analyses it as imperative and permissive mood. The speaker seeks agreement as in (216).

(216) Nh-angkwarr aylanth=an alh-erl.aynt-ey this-PERL 1DU=FOC go-CONT2-HORT ‘Let’s keep going this way…’

In some cases, such as (217), there is an interrogative use in which HORT seems to be interchangeable with POT (§5.2.4.3).

(217) Nthakenh il-ey anantherr nhenh-atherr-ey? what do-HORT 1PL this-two-EMPH ‘What are we going to do to these two?’

5.2.3 -etyek, -eyew, -eyek Purposive

PURP has allomorphs -etyek, -eyew and -eyek, reflecting regular Arandic sound correspondences in suffixes.28 It has different functions in main and dependent clauses (§5.3.1). Earlier researchers such as Stanham (1972: 46) regarded PURP as indicating future tense, along with -em and -ey. Turtle (n.d) glosses the morpheme as FUT ‘future’, INT ‘intentional’ and DES ‘desiderative’ My analysis agrees with that of the later researchers who analysed PURP as modal. Yallop (1977:53): analysed its meaning as ‘want to’ or ‘must’: AED (Green 1992:136) has “‘must’ or ‘ought’ to happen, in order to (do something), so that (something can happen)”. Strickland (1998:83) distinguishes between intentional, obligative and purposive senses of PURP claiming that, ‘it is uncertain whether native speakers distinguish notionally between them.’ Purposive is the third of Wilkins’ broad semantic functions (Wilkins 1989:237) of the cognate Arrernte morpheme. Wilkins says that the construction arises historically from the dative suffix +ke with the nominaliser +tye. Purposive is a label also used by other Australianists e.g. for Yidiny (Dixon 1977:345-6, Palmer 1986:180).

28 A fourth form, -etyew, that might be possible given these sound correspondences has not been recorded. 142 5.2.3.1 -etyek an- Time of intention

An auxilliary verb an- can occur with PURP. In (218) a state of affairs nearly occurred but didn’t eventuate, similar to Wilkins’ ‘time of intention’ (1989:236) construction with the Purposive in EC Arrernte. The an- may be related to the root an- ‘sit, be’ (§2.4.2.2). The latter is seen in an Arrernte example (219) ‘where the -tyeke 'purposive' indicates that the S/A intended to do the verb action but, for some reason, never got around to doing it.’ I have no example in the data in which auxillilary an- can be used with a non-past tense.

(218) Lake Nash-warl alp-etyek an-ek r=antey waylpel Lake Nash-ALL back-PURP be-PP 3S=STILL white person

apmw-il-elh-ek wrong-TV-MED-PP ‘The whitefella himself made a mistake and failed to get back to Lake Nash’

(219) Re alhe-tyeke ne-rne (Arrernte) 3sgS go-PURP be-p.immed ‘She was just about to go (but she got held up)’

5.2.4 -em Potential

POT has a range of senses. The primary sense is that that an event is possible or likely but not inevitable. This often involves the speaker’s intentions. In the following sections I discuss the function of POT in Alyawarr and differentiate its function from that of the temporal marker PRES.

Yallop (1977:52) notes that PRES -eyel is preferred to POT for the description of present happenings, e.g ‘he is sleeping’. In his table of verb suffixes (1977:58) POT corresponds to present tense in the other Arandic languages. There is a similarity in form alone with the Arrernte present tense marker -me (Strehlow 1944) of the ‘false friend’ kind that I mentioned in §1.3.1. Consequently, Yallop analysed -em as ‘present tense’ while labelling -eyel as the ‘present continuous’ tense marker in order to differentiate the two markers. In §5.1.1 I argued that the aspectual ‘continuous’ label is unnecessary.

143 The morpheme is glossed variously by previous researchers: ‘future’ (Stanham 1972: 46), ‘future, ‘desiderative’, ‘present’, ‘customary’, ‘intentive’, (Turtle n.d) 29 and ‘normative’ (Strickland 1998:78). The disparity and lack of a uniform interpretation for this form shows the need to examine its meaning more closely. POT appears to be similar in meaning to the Kaytetye Potential -me/-mere (Turpin 2000:115).

The distinction between POT and PRES can be subtle and in some cases they can appear to be almost interchangeable, as in (220), especially where both markers can indicate generic senses. In some instances this may be dependent upon the speaker’s preference and should be tested further to see whether this is the case.

(220) Aley=arl aynt-em ngenty-ey atha=arl renh now=SUB exist-POT soakage-EMPH 1S:ERG=SUB 3S:ACC

ar-eyel-angker. Aley=arl aynt-eyel ngenty ra=n see-PRES-CHAR1 now=SUB exist-PRES soakage 3S:NOM=FOC ‘The soakage that I see would be there. The soakage is there now’

With questions about the future, POT is used in preference to PRES according to Yallop (1977:52). POT frequently occurs in future temporal environments30, as in (221), owing to the close relationship between future tense and the category of mood. Future is a semantic category where tense and mood merge because of the inherent uncertainty about whether events in the future will happen (Lyons 1977, 2:181ff; Chung and Timberlake 1985: 243). Similarly Comrie (1985: 44) regards the future as more speculative and less definite than the other tenses.

(221) Malewik atherr iw-enh-em pay day two leave-POINT-POT ‘After two pay days from now’ (literally ‘leaving two pay days behind’)

POT is independent of the temporal context, describing potential rather than actual events. It is non-deictic and therefore not a part of the system of tense oppositions of the language (Lyons 1977:688). POT occurs in all temporal environments. For example, POT would seem to be expressing habitual in the past in (222).

29 The translations are not consistent. Turtle’s sentence 800 for example uses ‘present’ for ‘will tell’. 30 Bybee et al (1994) claim that Alyawarr has two ‘present tense grams’. Bybee bases this claim on Yallop (1977) in which -em and -eyel are both analysed as present tense. 144 (222) Artern-em=an=anem ar-em-el=anem renh awey. alperr chop-POT-FOC=THEN see-POT-SS=THEN 3S:ACC boy leaf

ingkerr alth-em “Nhaym-ip ra alperr-el aynt-eyel awey!” all strip-POT this-there 3S leaf-LOC lie-PRES boy ‘Then they would see him as they were chopping. They would strip all of the leaves, ‘There’s the boy lying in the leaves!’

Further evidence that POT is modal is that it can occur with CONT aspectual markers (Table 31) which occur with other markers of Intentional Mood. POT collocates with a number of suffixes and adverbs such as apek ‘maybe’ and nthakenh ‘how’ which reinforce the non- actualness of an event. The juxtaposition of apek ‘maybe, possibly’ enables POT to substitute for other modal suffixes such as IRR.

5.2.4.1 Generic

Generic clauses are less realis and refer to events which are ‘instantiated from time to time by actual events’ (Payne 1997:245). A generic or habitual reading is a part of the system of mood interacting with that of aspect. Comrie (1985:40) states, ‘habitual meaning lies on the boundary of tense, aspect and mood’. Another term for ‘generic’ is ‘gnomic’. Lyons (1977:681) maintains that the gnomic is related to mood and matters of opinion rather than fact. POT tends to involve non-definite Subjects in expository and hortatory texts, similar to Arrernte Present -me and the - rle suffix which is often used to describe a ‘generic event in oratorical-hortative texts’ (Wilkins 1989:231).

Yallop had doubts about his analysis of -em as present tense when he realized that it occurs as the default citation form of verbs, for example when speakers are asked “how do you say V?” This reflects the generic nature of POT which can occur in statements involving non-definite entities or events as in (223) and (224). rather than ‘specific referential objects’ (Payne 1997:245).

(223) Thip irrarnt-el arlkw-em arlkerl. black cockatoos-ERG eat-POT prickles. ‘Black cockatoos eat prickles’

145 (224) Ingwer.inger=an tyerrelh-em=anem ra yanh-el-ey another time=FOC exit-POT=THEN 3S there-LOC-EMPH ‘Sometimes it comes out at that place’ OR ‘another time it might come out at that place’

This meaning of POT refers to possible activities, accompanied by apek ‘maybe’ in (225), similar to the Conditional function in §5.2.4.5.

(225) Ingwerenty apek pwerlperr-eyew alh-em=an artewentyel=antey another maybe swim-PURP go-POT=FOC a lot=STILL

pwerlperr-eyang swim-NEG ‘If someone goes to swim, he shouldn’t stay in for too long!’

Related to the habitual or generic sense is the customary sense found in procedural texts such as in (226) which outlines the typical or normal way to do an activity.

(226) Arwerl kwenkart ak-em Ikwer-they=an kwenkart renh turkey bush pick-POT 3PL:DAT-ABL=FOC bush 3S:ACC

arrern-em pwarlerr-eyew=anem place-POT boil-PURP=THEN ‘(You) pick some turkey bush. After that you put it there for boiling’

5.2.4.2 Authority

POT conveys potential in the form of social conventions relating to the right to tell stories. A frequent occurrence of POT in Alyawarr discourse is the formula at the beginning of a story where the speaker indicates that they are qualified or allowed to tell a story as with (227).

(227) Angka nhaym=an atha il-em story this=FOC 1S:ERG tell-POT ‘I’ll tell this story’

146 5.2.4.3 Interrogative sentences

Alyawarr uses POT rather than PRES in interrogative clauses about the future and possibility, a use of mood which is common in languages of the world (Payne 1997:245). The only interrogative clauses of this type in the data involve wh-interrogatives. POT and HORT (§5.2.2) appear to be interchangeable in interrogative sentences which are about future events. POT occurs where people are undecided or trying to make up their minds. The speaker is enquiring about the time that an event will occur in (228). POT also occurs where people are uncertain about something as in (229).

(228) Ilenger nga alp-em? when 2S:NOM back-POT ‘When are you returning?’

(229) Nthakenh il-em anwantherr? althen-em=an what do-POT 1PL:ERG exterminate-POT=FOC

anwantherr renh-rnem 1PL:ERG 3PL:ACC ‘What are we going to do? We should wipe them out’

5.2.4.4 Spatial proximity and evidentiality

POT-marked verbs refer to less certain and less actualized situations with a low level of certainty on the part of the speaker as to the factivity of a state of affairs (Palmer 1986:140). I discuss POT with STAT in §4.9.2. STAT encodes modal remoteness and spatial remoteness, involving third person and non-referential Subjects. Subjects are commonly described as staying in a place distant from the speaker’s location. The Subject may in fact have gone on a brief trip away from the remote location where they are said to be residing and the exact location of the Subject may not be pinpointed at the reference time. In (230) the men are trying to catch the boy. He has eluded them and is beyond their reach. He is remote from the perspective of his pursuers and the use of STAT and POT reinforces that remoteness. The particles weth ‘that one, that we know about or aforementioned’ and nhak ‘that one, distant’ are used in contexts where a referent is not present, not visible or unreachable. The hill referred to in (231) is in the distance and is only just visible.

147

(230) Nhak=arl=ap ra awey weth-ey rtn-erl.an-em there-SUB=CONJ 3S:NOM boy that-EMPH stand-STAT-POT ‘Over there is that boy’

(231) Arekapengk aley altywen-erl.an-em place now crouch-STAT-POT ‘Over there is Arekapengk’ Literally: ‘Arekapengk (a hill) exists/crouches over there’ (stance-existence verb)

5.2.4.5 Conditional sentences

Further evidence that POT signals mood in Alyawarr is that it can occur in the apodosis of a conditional construction, in which Irrealis -enty occurs in the protasis as in (232). PRES doesn’t occur in this type of conditional clause. A common pattern in languages of the world is for information in the protasis and apodosis to be non-factual according to Palmer (1986:189).

(232) Nga alh-enty ntwa atyeng in-em 2S:NOM go-IRR 2S:ERG 1S:DAT get-POT ‘If/when you go, you might get (something) for me’

5.2.5 -ey.angenh, -eyang Negation

NEG attaches to verbs to form negative clauses (Payne 1997:282). According to Yallop (1977:56) NEG expresses a ‘neutralisation of all other inflectional categories of the verb’, that is, it is in the same structural position as TAM markers and therefore it is in opposition to TAM. I claim that, similar to EC Arrernte (Wilkins 1989:235), NEG indicates general negation including the category of mood. The indicative function of NEG is shown in (233). Yallop gives examples of negated forms which include tense and imperative readings. Strickland (1998:90) agrees with Yallop and gives examples of the negative imperative function which is seen in (234). The AED (Green 1992:45), definition has ‘the action is not happening, did not or will not happen’ which restricts its function to the tenseless indicative. Modal PURP and POT co-occur in (235), and NEG has a modal sense of ‘shouldn’t’ with a temporal adverbial or habitual marker as with STAT in §4.9.2. Two allomorphs -ey.angenh and -eyang are used by the same speakers and are difficult to distinguish on any grounds. The difference may reflect dialect variation. The -eyang variation of NEG appears in all examples from Yallop (1977) which

148 represents the Southern or Lake Nash dialect. David Blackman (p.c) says that -eyang is rare in recordings of Alyawarr from Epenarra (northern dialect).

(233) Atnert=akerr renh w-ew. Arrangkw ilw-ey.angenh gut=CONJ 3S:ACC shoot-PP nothing die-NEG ‘They shot it in the stomach. Nothing happened. It didn’t die’

(234) Arralty.an-ey.angenh kwang arternp-el il-eyn-a rush-NEG obvious slow-ADV do-RET-IMP ‘Don’t rush, go along slowly’

(235) Ingwerenty apek pwerlperr-eyew alh-em=an artewentyel another maybe swim- PURP go-POT=FOC a lot

=antey pwerlperr-eyang STILL swim-NEG ‘Someone going for a swim shouldn’t swim around for too long!’

The use of the post-inflectional clitic -antey STILL seems to give this construction the sense in (236), ‘still hasn’t done V even though they should have’. The suggestion is that the Subject should have known what Iylpakepey was like, the post-inflectional clitic -antey STILL suggesting characteristic behaviour and an enduring state of affairs, not merely an isolated incident of aggression. He was in contact with Iylpakepey over a long period of time but had failed to realise that Iylpakepey was a psychopath.

(236) Iylpakepey renh=an iterl.ar-ey.angenh=antey Iylpakepey name 3S:ACC=FOC know-NEG=STILL name

ra=n=ap alem pweth=arl ahelengkw 3S=FOC=CONJ aggressive=SUB dangerous

‘He still didn’t realize (or: should have been aware) that Iylpakepey was aggressive and dangerous’

5.2.5.1 -erl.aney.angenh Negative STAT

A bounded verb is transformed into a habitual or typical situation with STAT marking. With NEG the result has a sense of ‘shouldn’t keep doing X….’ as evident in (237) and (238) where

149 it indicates the speaker’s disapproval of a state of affairs. The Subject didn’t persist in doing the verbal activity.

(237) Angath-angath-irr-enh-ey.alp-ay arleng an-erl.an-ey.angenh close-RED-IV-POINT-DO&BASE-IMP distant be-STAT-NEG ‘Come back in close, you can't keep sitting a long way away’ (AED:44)

(238) Pwelek=arl rntern-enh irrtyart-el apmwil-elh-ew cattle=SUB spear-IMPP spear-INS do wrong-MED-PP

areyneng=ant arlkw-erl.an-ey.angenh euro=JUST eat-STAT-NEG ‘They made a mistake by spearing the cattle and not keeping on eating euro meat’

5.2.5.2 Negative causative construction

A negative causative construction consists of NEG and TV -il. The following is an example of ‘to make someone not do something’ with a series of negative causative verbs:

(239) Weth-ilkwer alh-ey.angenh-il-enh arrpemarl tyerrelh-ey.angenh-il-enh that-SEMB go-NEG-TV-IMPP again exit-NEG-TV-IMPP

weth-ilkwer outside alh-ey.angenh-il-enh that-SEMB outside go-NEG-TV-IMPP ‘Causing them to not go anywhere, to not go out like that and not go away like that’

5.3 Complex Sentences

Markers used to indicate complex clause types are dealt with in this section. Some subordinating elements also have distinct functions in independent clauses. In Alyawarr as with other languages there are links between complex sentences and the coding of mood e.g the connection between mood and subordination (McGregor 1988:37, Verstraete 2006). Complex sentence markers include Same Subject doing separate activities, where the activity is contingent upon that of the main verb, and Different Subject where the state of the subordinate clause is contingent on the perception of the Subject of the first clause.

150 5.3.1 Purposive in complex sentences

Purposive in independent clauses is described in §5.2.3. A complex sentence may be formed with a complement or adjunct dependent clause where the verb is marked as PURP, indicating the event which is the intended purpose, focus or endpoint result of the main verb action, as in (240). The outcome is what is intended by the Subject of the main clause, commonly translated as ‘to’ or ‘in order to’. The dependent clause does not need to have a co-referential subject, as shown in (241).

(240) Kel rernem=ant angk-ew pwayn anth-eyew=an then 3PL=JUST say-PP poison give-PURP=FOC ‘Then they talked about giving (him) poison’

(241) Aylepelengkw anantherr ilkwelhel-ey arwa-arrpantey young man 1PL initiate-HORT nothing-pretend

apwelh-warl ratherr ap-ey.alh-eyew ceremony-ALL 3DU arrive-GO&DO-PURP ‘We will pretend that we are initiating a young man so that those two (men) will come!’

5.3.2 -ey.alkenh, -ey.akenh, Subsequent

SUBS has allomorphs -ey.alkenh and -ey.akenh and conveys temporal succession of clauses. SUBS marks a dependent clause, as in (242). There does not appear to be a restriction on the marking of the verb of the main clause. HORT and PP occur in the examples below. Previous researchers of Alyawarr have analysed SUBS as mood. Turtle (n.d) labelled -ey.alkenh as ‘obligative’. AED (Green 1992:24) has ‘an action which will happen soon, going to do something soon’. SUBS is regarded as an emphatic form of the suffix -ey in Yallop (1977:54) and used in a future or intentional sense in ‘announcing an intention’. Yallop appears to regard - ey.alkenh as interchangeable with -ey in the ‘permissive’ use. Yallop only has examples of SUBS in simple clauses, whereas it is strictly limited to dependent clauses in my data and it is not possible to tell whether any of his examples had an antecedent clause in the original source.

There are markers in Arrernte which have similar senses to SUBS. In Strehlow (1944:124) phonologically similar eikanha (Strehlow’s orthography) is labeled ‘optative’ a modal suffix, meaning ‘may’. Optative expresses a desire, a hope or a wish (Crystal 1997:271).

While the S/A argument may be shared between the two clauses, often the two activities are unrelated. Alternatively, the same Subject can be involved in two activities which are not

151 naturally or conventionally connected with each other, as in (243). Wilkins (1989:239) argues that the closest Mparntwe Arrernte equivalent, -tyenhenge ‘subsequent’, occurs where the two events are related in a very general sense, there are no necessary causal or conditional relations between them. Other apparently related morphemes such are +yame, +yamenge (Henderson and Dobson 1994:633) which have the sense that the Subject or Agent is hoping to do the verb stem action, do something so that this will happen or to do something so that you can do something else. Further research needs to be done with more data to understand the difference between SUBS and PURP.

(242) Apwert-warl anwantherr alh-erl.iw-ey aker-then amern-then hill-ALL 1PL:NOM go-PL-HORT meat-CONJ damper-CONJ

amp-ey.akenh cook-SUBS ‘Let’s go up the hill, until the meat and damper is cooked’

(243) Waylpel ilenh-el-atherr apmer-they anenhantherr whitefella that-ERG-two place-ABL 1PL:ACC

irrent-ey.alp-ek ratherr an-ey.alkenh apmer nhel-ey evict-DO&BASE-PP 3DU stay-SUBS place this-EMPH ‘The two whitefellers have evicted us from our place so that they could stay here!’

Further evidence of SUBS as a temporal sequence marker is that it collocates with other temporal markers for example -tangkwel ‘first’ and the verb ahayleyel ‘do something first’ in (244).

(244) Athe-tangkwel ahayl-ey ntw-ant=anem antyw-ey.alkenh 1ERG-FIRST do first-HORT 2ERG-JUST=THEN drink-SUBS ‘I will do it first and then you can drink it after’(AED:1)

5.3.3 -enty Irrealis

IRR marks hypothetical statements in independent clauses and functions as a clause connector and marker of irrealis in dependent clauses. There is a connection in Alyawarr between subordination and mood marking which has been made for other Australian languages by McGregor (1988). This type of clause connection involves action nominals, which I will explain in the following sections.

152

5.3.3.1 Hypothetical

IRR can describe a hypothetical possibility in complex clauses. Strickland (1998:93) has analysed this morpheme as an irrealis marker which can function in a conditional clause in addition to participial and nominaliser uses. McGregor’s (1988:64) view is that the occurrence of non-indicative (modal) forms in certain types of subordinate clauses can be explained by saying that the verb activity of the subordinate clause is a matter of the speaker’s perception and not of factual propositional content. In a conditional sentence, the usual pattern of verb-marking is IRR in the protasis and POT in the apodosis as in (245).

(245) String ntwa arertn-enty aleth=arl=ap nga string 2S:ERG tie-IRR forever=SUB=CONJ 2S:NOM alp-em=an apmer=ant-warl back-POT=FOC camp=JUST-ALL ‘If you tie string on (to a tyre tube), you will be able to go all the way home’

5.3.3.2 ‘If’ conditional and ‘when’ temporal clauses

Complex clauses may be conditional or temporal and the ambiguity in their interpretation is quite common in languages of the world (Thompson et al 2007:256). Morphemes other than POT can appear in the verb of the apodosis for example FUT with a reading of ‘when’ rather than hypothetical ‘if’. An adverbial time clause occurs in (246). Sometimes IRR is used to denote a seasonal occurrence with an omnitemporal interpretation as in (247).

(246) Antywer arrty-eyenh arrpem kwaty apety-enty-el grass burn-FUT again rain come-IRR-LOC ‘We will burn grass whenever it rains’

(247) Arwengerrp anter-irr-em kwaty rntw-enty-penh=arl turkey fat-IV-POT rain rain-IRR-SRC=SUB ‘Turkeys fatten up after rain’

153 5.3.3.3 Perception complements

IRR functions as a Different Subject (DS) marker, marking the complement of a perception verb. The Subjects of the two clauses are not co-referential. The O of the main clause is the S or A of the dependent clause. The Object of the main clause controls the S/A of the dependent clause. The events of the main clause and the subordinate clause occur simultaneously, that is, the temporal reference of the verb in the subordinate clause is the same as that of the main clause. One of Turtle’s (n.d) labels for IRR is ‘participle formative four’. Yallop (1977: 131) describes it as a participle which functions as a clause connector. Wilkins (1989:138 footnote 38) says that Alyawarr can ‘mark the complements of perception verbs’ quoting Yallop (1977:131-32) although this is not Yallop’s term as he labels the morpheme a ‘participle’. (Wilkins regards the Arrernte equivalent complement perception to be -rlenge but I can’t find any examples of this usage elsewhere- maybe it is -marlenge).

The verb in the main clause is a perception verb such as in (248) and is marked with a tense marker, most often Past Perfective. The ‘seeing’ and the activity of the dependent clause take place concurrently. (See Wilkins 1989:279 and the comment on angkintyeke). My data and nearly all of Yallop’s -erl.anenty examples (10 out of 11) involve main verbs of perception related to ‘seeing’ e.g areyel ‘seeing’, intwareyel ‘see something moving away’. Often -ek- athen ‘mistaken belief’ and -arrpantey ‘thwarted intention’ occur with -erl.an-enty STAT-IRR to indicate that something was wrongly perceived, as in (249). The addition of these suffixes highlights the irrealis nature of this construction (Strickland 1998:93).

(248) Arengk-rnem iytwar-enh-ek anterr-err-enty dog-NPL see across-POINT-PP run-PL-IRR ‘(We) looked across at the dogs (which were) running’ (Yallop 1977:131)

(249) Rernem=an arrthaw-enh=arl waylpel-el ilt-erl.an-enty-arrpantey 3PL=FOC mishear-IMPP=SUB whitefella-ERG growl-STAT-IRR-thwart ‘They would mishear the whitefellow, as though he was growling at them’

5.3.3.4 Nominalisation

IRR can give a verb stem nominal-like properties, that is, it can occur with nominal case marking, nominal modifiers and other nominal inflections. Previous analysis focuses upon these nominalising properties, though as we will see, there are no sharp distinctions either in the

154 syntactic category involved or in the associated functions. Strehlow (1944: 62) regarded Arrernte as deficient in abstract nouns and lists ‘secondary nouns’ which are abstract nouns derived from verbs by the addition of -intja, for example tjontintja ‘beginning’ derived from tjontama ‘to begin’. This and a number of other nominalisations have apparently come from the Hermannsburg Arrarnta church context and are not consistent with the native Alyawarr functions of the corresponding IRR -enty form. Yallop (1977: 89), following Strehlow (1944:62), claims that -enty forms a nominal from a verb stem. He then translates IRR, as in (250), in a way that suggests he sees them as a type of action nominal. The AED (Green 1992:135) similarly describes this function of -enty as ‘goes on the main part of a verb and turns it into a noun’ but contains only one example of a verb nominalised with IRR, aylelhenty ‘song’ from the derived stem aylelh- ‘sing’. However, as we see in (251), it is not straightforward to distinguish aylelhenty as a nominalised argument of a clause versus marking a subordinate clause in a complex sentence. The latter analysis is proposed here. The Subject of aylelhenty in the subordinate clause in (251) is ellipsed, which makes it indistinguishable from an O of a simple clause. An explicit subject with the appropriate case-marking is however possible.

(250) Arnar-a ap-ey.alp-enty watch-IMP.S arrive-DO&BASE-IRR ‘Watch out for (his) return, look out for (his) coming back’ (Yallop 1977: 89, re-presented in modern orthography but preserving the original translation.)

(251) Aw-etyek akwel aylelh-enty. hear-PURP apparently sing-IRR ‘Apparently they will hear (them) sing’

Warlpiri has constructions which are similar to those in Alyawarr, corresponding to examples like (251). An appropriate label for these deverbalised nominals is ‘action nominals’ (Nordlinger 2002). Koptjevskaja-Tamm (1993:6) reports that the ‘intermediate nature’ of action nominals of the kind found in Warlpiri is reflected in their syntactic behaviour. They occupy a position which is intermediate between prototypical verbs and prototypical nouns. According to Comrie and Thompson (2007:344) the action nominal is a ‘non-finite’ verb form without TAM markers. Action nominals generally take a limited range of case marking, less than for regular non-derived nominals. The nominalised verb in Alyawarr exhibits some prototypical nominal properties, because it can bear a limited range of case marking, for example, the dative in (252) which appears to parallel the dative marking of non-derived nominal ‘goals’.

155 (252) Waylpel-el ar-enh=arl alh-enty-ew=an whitefellow-ERG see-IMPP=SUB go-IRR-DAT=FOC ‘The whitefellows were looking out to see if they were going’

5.3.3.5 -erl.anenty

The combination of STAT and IRR indicates that the event of the subordinate verb was a state or habitual activity at the time that it was perceived by the S of the main clause. STAT thus serves a similar function in the dependent clause as IMPP serves with a free verb. My analysis differs from that of Yallop in assigning an aspectual value to the -erl.an component of the construction. Yallop (1977:132) appears to regard ayntenty ‘lie-IRR’, ayntanenty ‘lie-?-IRR’ and aynterl-anenty ‘lie-STAT-IRR’ alike as ‘lying’ with no distinction in meaning between the three forms. However he also claims that ‘the incorporation of the auxiliary root an- overtly marks the participle as stative or non-completive’ marking ‘ongoing activity’. The most common forms in my data are -enty and -erl.anenty. There are less than five examples of -an affixing directly to a verb root (Yallop 1977:132) in my data.

I also differ from Yallop (1977:136) where he maintains that there is no difference between a SUB marked clause with an IMPP verb (253) and the -erl.anenty STAT-IRR in (254). The SUB subordinate marker is used where the actions of the two clauses do not occur concurrently (§5.1.5). I understand IMPP=SUB in (253) as ‘I found a possum which had (previously) been sitting in a hollow tree until now’. The action of the dependent verb is antecedent to that of the main verb. The IMPP event continues to the reference time.31 Therefore -enharl does not indicate ‘concurrent activities’. By contrast, in (254) the verb actions are occurring concurrently. Further testing and analysis is necessary.

(253) Antenh=anem ar-ey.alh-ek iylpwer-el-arl an-enh possum-THEN see-GO&DO-PP hollow-LOC=SUB sit-IMPP ‘I found a possum which had been sitting in a hollow tree’

31 The free verb an can be interpreted as either ‘sit’ or ‘live’, the latter having a more stative character.

156 (254) Antenh=anem ar-ey.alh-ek iylpwer-el-arl an-erl.an-enty possum=THEN see-GO&DO-PP hollow-LOC=SUB sit-STAT-IRR ‘‘I found a possum which was living in a hollow tree’

STAT refers to a state which holds at the reference time. In (255) the draught horses were tame and would keep approaching people. STAT suggests their tameness and their habit of approaching humans. In (256) it is a state of affairs in which there are a lot of cars compared to earlier times.

(255) Nantew atherr=anem-art rernem arnar-ew-awel nantew horse two=THEN-EMPH 3PL see coming-PP-SUPP horse

atherr ap-ey.alp-erl.an-enty=anem two arrive-DO&BASE-STAT-IRR=THEN ‘Then they saw the horses that would keep on coming back’

(256) Ngwangk-ngwangk atha=n ar-eyel mwetek these days-RED 1SERG=FOC see-PRES car

anterr-err-erl.an-enty akngerr=anem=arl run-PL-STAT-IRR lots=THEN=SUB ‘These days I see lots of cars driving around’

5.3.4 -emer Subjunctive

SBJV appears in main and subordinate clauses. It is uncommon in text, with less than 10 tokens in my data. Strehlow (1944) labels this morpheme ‘conditional’ for Arandic dialects. Agreeing with Strehlow, Yallop (1977:56) sees this suffix as potential or conditional in a general sense with the reading, ‘something might happen’. The AED (Green 1992:134) has ‘if only (the action had happened or could have happened)’. Strickland (1998:89) analyses it as ‘should’ i.e the speaker thinks something ought to happen or is desirable. With the exception of Turtle, previous researchers of Alyawarr don’t analyse this as a contrafactual in complex sentences. Examples 610, 686, 687 and 756 in Turtle’s (n.d) section on complex clauses have examples of this kind. Turpin (2000:115) describes +mere in Kaytetye as having a meaning ‘if something happens then.’ There is apparently no distinction between +me and +mere in Kaytetye (Harold Koch, p.c). Wilkins (1989:233) reports that in Mparntwe Arrernte -mere may occur either as a simple utterance or in a complex conditional construction.

157 5.3.4.1 Main clauses

SBJV occurs on verbs in simple clauses to indicate that something might or possibly should happen. This is a mild form of intention, evident in (257). The informant also said that he might lywa artetyek using PURP which suggests that there is some overlap between Purposive and Subjunctive. SBJV occurs with nthakenh ‘how, what’ in (258) as if to emphasise the tentative nature of the activity. The interrogative with SBJV in (259) highlights the hypothetical and uncertain nature of the action.

(257) Lywa new one art-emer shade shelter new one build-SBJV ‘(I) could build a new bough shelter’

(258) Alakenh.anyem=anem rernem doem-il-enh ikwer like that=THEN 3PL do-TV-IMPP 3S:DAT

Artwamp-ey nthakenh rernem w-emer=an old man-EMPH how 3PL shoot-SBJV=FOC ‘That is what they were doing about the old man, how they could shoot him’

(259) Artwa ra-amathen=anem aylpeny=anem-irr-ew “Ilek-ew man 3S-doubt=THEN sick of=THEN-IV-PP what-DAT

ayeng irrtyert-irrtyert alh-erl.an-emer?” 1S:NOM hidden-RED go-STAT-SBJV ‘Then the man apparently got tired of it, “Why do I have to keep going along hiding?”’

5.3.4.2 Complex Clauses: Contrafactual

Subjunctive appears in conditional sentences which describe a past state of affairs. Subjunctive does not itself mark contrafactuality. The past contrafactual involves an implicature of contrafactual meaning involving potentiality and past tense (Verstraete 2006:72). A course of action could have been taken but it wasn’t. Strehlow (1944:108) discusses contrafactuality in the Aranda (Arrernte) New Testament text of John 11:21 and the difficulty with expressing contrafactual sentences in Arandic languages. As shown in (260), contrafactual sentences in Alyawarr may have both verbs marked with Subjunctive.

158

(260) Ra=rl irrtyert.iw-emer irrtyart-then mwerr=arl ra an-emer 3S=SUB hide-SBJV spear-CONJ good=SUB 3S be-SBJV ‘If he had hidden his spear away, he would have been all right’

The Arrernte equivalent of Subjunctive functions in a contrafactual clause in (261) from Wilkins (1989:233).

(261) Unte apmwerrke peke petye-ke arratye unte re-nhe are-mere 2sgS yesterday maybe come-pc true 2sgA 3sg-ACC see-HYPO ‘If you had come yesterday , then you certainly would have seen her’

5.3.5 -el Same Subject

Same Subject marking on a verb indicates that a clause is dependent and that its S or A is co- referential with the S/A of the clause that it is dependent upon, as in (262). Same Subject follows POT or tense. POT (§5.2.4) appears to have a habitual sense ‘would happen’. STAT+SS -erl.anemel occurs when the activities are habitual or characteristic. IMPP marks the verb of the main clause in (263) in harmony with the stative or habitual aspect of STAT.

(262) Yarraman-warl ngan-erl.alh-em alwern-em-el pwelek horse-ALL mount-DO&GO-POT chase-POT-SS cattle

arternp-il-em-el quieten-TV-POT-SS ‘I would mount a horse, chasing the cattle and settling them down’

(263) Arrwekeleny=arl an-enh yanh-ilkwer arlkw-erl.an-em-el=arl ancestor=SUB sit-IMPP there-SEMB eat-STAT-POT-SS=SUB

‘That is how the ancestors were living and eating’

According to Yallop (1977:131), it may be the case that -emel is equivalent to -el, which he labels a participle marker. He identifies -el with the locative which he says can occur on verb

159 forms. Yallop’s analysis is plausible because -el the Locative marker LOC forms temporal adverbial phrases. So, a possible alternative analysis is that it is not a distinct SS suffix but that the Locative marker follows POT and IMPP.32 SS marks the verb of an adverbial clause in (264). Two events happen simultaneously and are done by the same subjects. SS indicates that the events of the verbs occur concurrently and probably have the same temporal range.

(264) Arengk renh atnwen-enh arrpemarl arwengalker=arl dog 3S:ACC lead-IMPP again honey=SUB

artern-enh-el=an. chop-IMPP-SS=FOC ‘(We) would take the dog as well, whenever (we) were chopping for honey’

I have not found an example like that of Yallop (1977:130) seen in example in which he equates -el with a participle formative, quoted in Goddard (1988:185). His example is anomalous. Usually -el is not affixed directly to the verb stem but occurs after tense suffixes. Although ayntila could be analysed as aynterl, the -arlkw part has not been found as a compounding element - see footnote 15 in §4.4.

(265) antimirna ayntila alkuka. wild.honey-NOM lie-participle eat-PAST 'I ate the wild honey while lying down'

5.3.6 -ekerr/-ewerr Apprehensive

APP is found in independent clauses such as in (266) which function as warnings. In dependent clauses APP indicates a potential or actual negative situation to be avoided, as in (267). The Reason marker -wety occurs with nominals and has a similar function to the Apprehensive. Yallop (1977:55) and Turtle (n.d) labels APP as ‘admonitive’ because it generally suggests a warning or risk although Yallop notes that there ‘are some contexts where it suggests no more than an unpleasant possibility or a warning’.

32 I have not found examples of -el occurring with the Past Completed suffix or other tense suffixes. Although Yallop (1977:131) analyses -ikala as ‘past + locative’, he also appears to label similar constructions as either a single past perfect suffix (page 51) or as a combination of past -ika and the clausal connective -ala (page 137). 160 (266) aytn-ewerr ‘might fall’ fall- APP

akelh-ewerr ‘might break’. break- APP (Stanham 1972: 46)

(267) Artwe-rnem warrkirr-enh menty=arl maket-wety man-NPL work-IMPP leave=SUB rifle-AVERS

ater-el w-ewerr fear-ADV shoot-APP ‘They would keep working, afraid of the rifles, of getting shot’

5.3.7 -ekerr-awaty Before

BEFORE is similar in form to APP plus -awaty ‘while, during’ and is probably diachronically related to APP. The event is usually something that the speaker doesn’t want to happen, similar to Apprehensive. The interpretation is ‘do something before something else happens’ which seems to parallel the -ketye-athathe (Wilkins 1989:374) construction in Arrente. Breen (p.c) indicates that a similar construction occurs in Western Queensland languages:

(268) store mart-elh-ekerr-awaty store close-MED-APP-BEFORE ‘before the store closes’

5.3.8 Characteristic -ey.angker and -enh.RED

Characteristic nominalises or deverbalises a verb and gives it some nominal properties, similar to Agentive –er nominalisation in English and similar to IRR §5.3.3. Although -enh.RED occurs in Alyawarr texts, it is limited in productivity and less common than in other Arandic languages. CHAR is used to describe characteristic behaviour as the examples in (269) show. In (270) CHAR is also used for inanimate objects. (270) CHAR occurs with a limited range of nominal case marking as in (271).

161 (269) anterr-ey.angker ‘escaper, escape artist’ from anterreyel ‘run (away)’ tnakelh-erl.an-ey.angker ‘one who always boasts, skites’ from tnakelheyel ‘boast’ (Alyawarr Bible: Genesis 37:14 )

(270) lher-weny=arl iylpay-weny=arl kwaty ilw-ey.angker=arl nhenh-ey river-NEG=SUB creek-NEG=SUB water finish-CHAR=SUB this-EMPH ‘Not a river and not a creek- the water keeps disappearing’

(271) Anterr-ey.angker-wety ikwer-warl arrern-ew ingkety-warl tyeyn=anem run-CHAR-REAS 3DAT-ALL put-PP foot-ALL chain=THEN ‘They put chains on his feet because he was an escape artist’

-enh.RED is plausibly related to IMPP (§5.1.4), and creates nominal. It is marked by an -enh suffix plus disyllabic reduplication:

(272) Arriper=an ahelengkw im-ek=antey tnhw-enh-tnhw-enh taipan=FOC dangerous corpse-DAT=STILL attack-enh.RED ‘The Taipan (snake) is dangerous, a killer’ (AED:92, my translation)

An Arrernte morpheme (Wilkins 1989:139)33 in (273) has a habitual activity meaning similar to Alyawarr CHAR and probably also DISTR §4.6.2. The formally similar IRR nominalisation in Alyawarr can’t be modified in the same manner.

(273) Re angke-ntye kngerre. 3sgS speak-NMZR big ‘He's always speaking (or he's a big talker)’

5.4 An alternative classification

A tentative alternative classification is proposed for the non-obligatory markers. This forms the basis for future research. Aspect (§4.9) can only be marked once in a verb, with the exception of Past Habitual2. In Table 36 the obligatory markers are divided into three categories on the basis of their occurrence with specific aspectual markers in the pre-obligatory positions in the verb. Only the past tense categories of Category One include aspectual distinctions. Category Two inflections co-occur with both CONT and STAT markers. Category Three is less well-defined.

33 Wilkins appears to analyse -kngerre synchronically as an ‘adjectival nominal’ which modifies the deverbal noun, e.g. ‘big eater’. 162 The markers combine with STAT and all other modal and clause- joining morphemes, but not CONT.

Table 34: Aspectual classes of Alyawarr verbs

CATEGORIES OF OBLIGATORY CO-OCCURING PRE-OBLIGATORY MARKERS ASPECTUAL MARKERS

Category One None

Category Two: Intentional -aynt, (CONT), -erl.aynt (CONT2) future -erl.an (STAT)

Category Three: Alternative -erl.an (STAT)

5.4.1 Aspectual Category Two: Intentional Mood

Intentional mood categories form a distinct class based upon their distribution, combining with both CONT and STAT. The markers in Table 35 express a speaker’s intention. Semantically, they are all similar to the Imperative which is the most typical member of the group.

Table 35: Intentional Mood markers IMPerative -ø, -enherr-atherr, -enherrarey §5.2.1 HORTative -ey §5.2.2 PURPosive -etyek §5.2.3 SUBSequent -ey.alkenh §5.3.2 POTential -em §5.2.4

5.4.2 Aspectual Category Three

The morphemes listed in this section are those which can combine with aspectual STAT (§4.9.1.1) in addition to all non-past tense and Intentional Mood morphemes. STAT codes stative or habitual aspect in clauses. The Alternative State morphemes are given in Table 36.

163

Table 36: Alternative State Markers

Non-actual state (NEG) Possible states (IRRealis, SBJV Subjunctive, POTential, PURPosive, HORTative) States which the speaker judges to be undesirable (APP) Complement clauses involving different Subjects (DS) where the actuality of state of

affairs V2 is contingent upon the perception of S1.

States involving a Subject doing V2 which is contingent upon the occurrence of a

habitual activity V1 (SS) State of affairs in which characteristic behaviour is unusual or different from that of others (CHAR)

On functional grounds, I propose a modal category of Alternative State as a tentative analysis which needs to be confirmed with more research. Alternative State involves polarities which are based upon a variety of criteria, emphasising discrete Subjects, events or realities. Negation and subjunctive are prototypical examples of Alternative State, involving unreal states, uncertain states, separate realities and alternative realities. Alternative State is similar to Contingency which has been described by Timberlake (2007:329) who claims that Contingency mood ‘examines the conditions under which a situation has one polarity or the other as a function of some other situation’. Contingent states are closer to irrealis along the continuum from realis to irrealis. A question which remains to be answered is whether Alternative State marks an explicit relation between clauses, especially between the protasis and apodosis in conditional clauses.

There needs to be explanation of why STAT is used in particular contexts, as illustrated in the following two examples. PURP combines with STAT in (274), describing alternative states of affairs, indicating a situation which is undesirable and against the Subject’s wishes. The use of the interrogative indicates the speaker’s uncertainty about the current state of affairs.

(274) Ilek anwantherr=an arlkw-erl.an-eyew irrkaty=an wenh why 1PL:NOM=FOC eat-STAT-PURP same=FOC QUOT ‘ “Why should we keep eating the same (food)?” ’

HORT combines with erl.an STAT and has a reading of ‘S does something which is away from other people’ in (275). The S in the example is a fugitive and wants to remain in a secretive state, away from his pursuers.

164 (275) Nh-el=antey ayeng aynt-erl.an-ey arlkarl-irr-em-el wenh!” here-LOC=STILL 1S:NOM lie-STAT-HORT cool-IV-POT-SS QUOT ‘ “I will stay here and keep cool”’

165 6 Future Research

This thesis has explored a number of issues in the description of the Alyawarr verb. I have confirmed that the Alyawarr verb differs from that of other Arandic languages. A number of conclusions have been tentatively drawn and further work needs to be done. The precise functions of a number of verbal morphemes still require further investigation. Some morphemes have been described which were not in previous descriptions of Alyawarr grammar. Verbal Reduplication has been described in Chapter 2. The relationship between Attenuative and RUN needs to be investigated.

Additional derivational and number morphemes have been described in chapter 3. The full range of derivational and number suffixes could be further investigated through the use of more data. The functions of the Mediopassive require further attention.

Morphemes which were previously undescribed appear in Chapter 4, for example the Vertical Motion suffixes. Other motion suffixes could be researched further through the use of texts and elicitation.

Additional TAM morphemes have been described in chapter 5. Further research needs to be done of aspect in Alyawarr and its relationship to the categories of tense and mood.

I would hope to refine and extend research as more data becomes available, particularly for those morphemes which have a low frequency in texts. Further elicitation could be conducted to get a more nuanced understanding of the function of verbal morphemes.

166 References

Arandic Languages Dictionary Programme, Institute for Aboriginal Development (1990). Alyawarr Word-list. Alice Springs: Institute for Aboriginal Development.

Asher, R.E and J.M.Y. Simpson (1994). The Encyclopedia of language and linguistics, Oxford; New York; Pergamon Press.

Austin, Peter (1981). A grammar of Diyari, South Australia. Cambridge: Cambridge University Press.

Austin, P.K (1989). Verb Compounding in Central Australian languages. La Trobe Working Papers in Linguistics, 1-16.

Austin, Peter (2004). Language Documentation and Description, volume 2. The Authors.

Austin, P.K (2005). Causative and applicative constructions in Australian Aboriginal languages. London: SOAS.

Austin, Peter (ed.) (1988). Complex Sentence Constructions in Australian Languages. Amsterdam/Philadelphia John Benjamins.

Austin-Broos, Diane (2009). Arrernte Present, Arrernte Past. Chicago: University of Chicago Press.

Bible Society in Australia. Angka Mwerr-angker (2010). Good News: Portions of the Old Testament in Alyawarr together with portions of the New Testament in Alyawarr and English. Sydney.

Bierbaum, Kathy (1987). The function of suffixes -ane and -ape in Alyawarra discourse. Darwin: Summer Institute of linguistics.

Binford, L (1984). An Alyawara Day, Flour Spinifex Gum and Shifting Perspectives. Journal of Anthropological Research 40, 159-171.

Binford, L (1986). An Alyawara Day: Making Men's Knives and Beyond. American Antiquity 51, 547-62.

Blackman, David, Jenny Green and David Moore (2009). Alyawarr to English Dictionary- in progress. Alice Springs.

Blackman, David, David Moore with members of Alyawarr communities (2004). Alyawarr Picture Dictionary. Alice Springs: Institute for Aboriginal Development.

Blake, Barry (1987). Australian Aboriginal Grammar. London ; Wolfeboro, N.H: Croon Helm.

Breen, Gavan (1976). Ergative, locative and instrumental case inflections: Wangkumara. In R.M.W Dixon (ed.), Grammatical categories in Australian languages 336-339. Canberra: Australian Institute of Aboriginal Studies.

Breen, J. G. (1977). Andegerebenha Vowel Phonology. Phonetica 34, 371-391.

Breen, J. G. (1982). Antekerrepenh Grammar, unpublished paper.

167 Breen, Gavan. (1986). Alyawarre Picture Vocabulary. Alice Springs: Institute for Aboriginal Development.

Breen, Gavan. (2001). The wonders of Arandic phonology. In D. Nash J. Simpson, M. Laughren, P. Austin and B. Alpher. (ed.), Forty Years On: Ken Hale and Australian languages. Canberra: Pacific Linguistics.

Breen, G. (2005). A short history of spelling systems in Arrernte. In S. Mitchell. A. Kenny (ed.), Strehlow Research Centre: Occasional paper number 4. Alice Springs: Northern Territory Government.

Breen, Gavan and Jenny Green (1995). hyphens and final vowels in Arandic orthographies. Australian Journal of Linguistics 15, 91-94.

Bybee, Joan, Perkins, Revere and Pagliuca, William (1994). The Evolution of Grammar: tense, aspect and modality in the languages of the world. Chicago: University of Chicago Press.

Capell, Arthur (ed.) (1962). A New approach to Australian Linguistics. Handbook of Australian languages, part 1. Sydney: University of Sydney.

Chung, Sandra and Timberlake., Alan (1985). Tense, aspect and mood. In T.A. Shopen (ed.), Language Typology and Syntactic Description,. Cambridge: Cambridge University Press.

Comrie, Bernard (1976). Aspect. Cambridge: Cambridge University Press.

Comrie, Bernard (1985). Tense. Cambridge: Cambridge University Press.

Comrie, Bernard and Sandra Thompson (2007). Lexical nominalisation. In T Shopen (ed.), Language Typology and Syntactic Description 334-381, Volume III. Cambridge: CUP.

Crystal, David (1997). Dictionary of Linguistics and Phonetics- 4th edition. Oxford: Blackwell.

Denham, W. (1975). Population properties of physical groups among the Alyawara tribe of Central Australia. Archaeology and Physical Anthropology in Oceania 10, 114-51.

Denham, W (1978). Alyawara Ethnographic Data Base Vol. 1 A Guide to Contents, Structure and Operation. New Haven: HRAFlex Books.

Devitt, Jeannie (1988). Contemporary Aboriginal Women and subsistence in remote, arid Australia PhD thesis, University of Queensland

Devitt, Jeannie (1994). Apmer Anwekantherrenh: Our Country. An introduction to the Anmatyerr and Alyawarr people of the Sandover River region, Central Australia. Alice Springs: Urapuntja Health Service.

Dixon, R.M.W (1977). A Grammar of Yidiny. Cambridge: Cambridge University Press.

Dixon, RMW (1980). The . Cambridge Cambridge University Press.

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

Dixon, R.M.W (ed.) (1976). Grammatical categories in Australian languages Canberra: AIAS.

Dixon, R.M.W (2002). Australian Languages: Their Nature and Development. Cambridge: Cambridge University Press.

Dixon, R.M.W, Alexandra Y Aikhenvald (eds.) (2000). Changing Valency: Case Studies in Transitivity. Cambridge Cambridge University Press.

168 Dixon, R.M.W and A. Y. Aikhenvald (eds.) (2002).Word: a cross-linguistic typology. Cambridge: CUP.

Elson, B. F and V. B. Pickett (1988). Beginning Morphology and Syntax Dallas: Summer Institute of Linguistics.

Eylmann, Erhard (1966 (1908)). Die Eingeborenen der Kolonie Suedaustralien. New York Johnson Reprint Corporation.

Fabricius, A.H (1998). A comparative study of reduplication in Australian languages. Munich: Lincom Europa.

Fleischman, Suzanne (1982). The Future in Thought and Language: Diachronic evidence from Romance. Cambridge: Cambridge University Press.

Ford, Margaret (1966). Beyond the furthest fences. London: Hodder and Stoughton.

Glass, Amee (1980). Cohesion in Ngaanyatjarra Discourse. MA thesis Linguistics, Australian National University.

Goddard Cliff, (1988). Verb serialisation and the circumstantial construction in Yankunytjatjara. In Peter Austin, Complex Sentence Constructions in Australian Languages 141-175. John Benjamins.

Goddard, Cliff and Jean Harkins (2002). Posture, location, existence and states of being in two Central Australian languages. In John Newman (ed.), The Linguistics of Sitting, Standing and Lying 213-238. Amsterdam: John Benjamins.

Green, Jennifer (1992). Alyawarr to English Dictionary. Alice Springs: IAD Press.

Green, Jenny (2001). 'Both sides of the bitumen': Ken Hale remembering 1959. In David Nash Simpson J, Mary Laughren, Peter Austin and B. Alpher. (ed.), Forty Years On: Ken Hale and Australian languages. 29-43. Canberra: Pacific Linguistics.

Green, Jenny (2010). Anmatyerr to English Dictionary. Alice Springs: IAD Press.

Hagan, R. and M. Rowell (1978). A claim to areas of traditional land by the Alyawarra and Kaititja. Alice Springs: Central Land Council.

Hagan, R. and M. Rowell (1979). A claim to the Utopia Pastoral lease by the Alyawarra and Anmatjirra. Alice Springs: Central Land Council.

Hale, Kenneth (1962). Internal relationships in Arandic of Central Australia. In A Capell (ed.), Some linguistic types in Australia 171-83 (Oceania Linguistic Monograph 7) Sydney: University of Sydney.

Harkins, Jean (1994). Bridging Two worlds: Aboriginal English and crosscultural understanding. St Lucia: University of Queensland Press.

Hartman, Deborah and Henderson, John (eds.) (1994). Aboriginal languages in education. Alice Springs: IAD Press.

Haspelmath, Martin (2002). Understanding Morphology. London: Arnold.

Heath, Jeffrey (1984). Functional grammar of Nunggubuyu. Canberra: Australian Institute of Aboriginal Studies.

169 Hellwig, Birgit (2006). Field semantics and grammar-writing: Stimuli-based techniques and the study of locative verbs. In A. Dench, N. Evans and F.K. Ameka, Catching Language: The Standing challenge of Grammar Writing. Berlin: Mouton.

Henderson, John (1998). Topics in Eastern and Central Arrernte Grammar. Ph.D thesis, Linguistics, University of Western Australia.

Henderson, John. (2002). The word in Eastern/Central Arrernte. In R. M. W. Dixon and A. Y. Aikhenvald. (ed.), Word: a cross-linguistic typology 100-121. Cambridge: CUP.

Henderson, John and Veronica Dobson (1994). Eastern and Central Arrernte to English Dictionary. Alice Springs: IAD Press.

Henderson, John. and David Nash. (2002). Language in Native Title. Canberra: Aboriginal Studies Press.

Holmer, N (1963). On the history and structure of the Australian Languages. Lund Carl Bloms.

Hoogenraad, Robert (1984). Grass-roots Aboriginal language and culture programs in schools in the Barkly and Sandover regions of the Northern Territory. In Deborah Hartman and John Henderson (eds.), Aboriginal languages in education Alice Springs: IAD Press.

Hoogenraad, Robert (2001). Critical reflections on the history of bilingual education in Central Australia. In Jane Simpson, David Nash, Mary Laughren, Peter Austin and Barry Alpher (eds.), Forty Years on: Ken Hale and Australian languages Canberra: Pacific Linguistics.

Hoogenraad, R. and B. Thornley (2003). Aboriginal languages of Central Australia and the places where they are spoken. Alice Springs: Jukurrpa.

Hopper, Paul J and Thompson, Sandra A (1980). Transitivity in Grammar and Discourse. Language 56, 251-299.

Kearns, Kate (2000 ). Semantics. Hampshire: Palgrave.

Kemmer, Suzanne (1993). Middle Voice, Transitivity and the Elaboration of Events. In Barbara Fox and Paul J Hopper (eds.), Voice: Form and Function Amsterdam: John Benjamins.

Kempe, A. (1891). A Grammar and Vocabulary of the language of the natives of the McDonnell ranges. Transactions of the Royal Society of South Australia XIV, 1-54.

Koch, Harold (1984). The category of 'associated motion' in Kaytej. Language in Central Australia 1, 23-24.

Koch, H. (1993). Review of J. Green Alyawarr to English Dictionary. Aboriginal History 17, 167-169.

Koch, H. (2001), 'Basic Vocabulary of the Arandic languages: from classification to reconstruction', in Simpson J, Nash D, Laughren M, Alpher B, Austin P (ed.), Forty Years On: Ken Hale and Australian Languages, Pacific Linguistics, Canberra, Australia, pp. 71-87.

Koptjevskaja-Tamm, M. (1993). Nominalizations. London: Routledge.

Langacker, Ronald W. (ed.) (1991). Foundations of Cognitive Grammar. Stanford: Stanford University Press.

Lee, Richard B and Irven Devore (1968). Man the Hunter. Chicago: Aldine.

170 Levinson, Stephen C and David Wilkins (2006). Grammars of Space: Explorations in Cognitive Diversity. Cambridge: Cambridge University Press.

Lewis, D (1976). Route Finding by Desert Aborigines in Australia. The Journal of Navigation 29, 21-38.

Liberman, K. (1985). Understanding Interaction in Central Australia: An ethnomethodological study of Australian Aboriginal people. Boston, Mass: Routledge, Keegan and Paul.

Lindsay, D (1889). An expedition across Australia from south to north between the Telegraph line and the Queensland boundary in 1885-6. Royal Geographical Society, London.

Lyons, John (1977). Semantics. Cambridge: CUP.

Lyons, P. and M. Parsons (1989). We are staying: the Alyawarre struggle for land at Lake Nash. Alice Springs: Institute for Aboriginal Development Press.

McGregor, William B (1988). Mood and subordination in Kuniyanti. In Peter K. Austin (ed.), Complex Sentence Constructions in Australian languages 1-67. Amsterdam, Philadelphia: John Benjamins.

McGregor, William B. (2002). Verb Classification in Australian Languages. Berlin: Mouton de Gruyter.

McGregor, William B. (ed.) (2008). Encountering Aboriginal Languages: studies in the history of Australian linguistics. Canberra: Pacific Linguistics.

Meggitt, M.J (1962 ). Desert People. Sydney: Angus and Robertson.

Monaghan, Paul (2008). Norman B Tindale and the Pitjantjatjara language. In William B. McGregor. (ed.), Encountering Aboriginal languages. 251-272. Canberra,: Pacific linguistics.

Moore, David (2008). TGH Strehlow and the linguistic landscape of Australia 1930-1960. In William B. McGregor (ed.), Encountering Aboriginal Languages. 273-300. Canberra: Pacific Linguistics.

Moravcsik, Edith A (1978). Reduplicative Constructions. In J.Greenberg (ed.), Universals of Human Language, Volume 3: Word Structure Stanford: Stanford University Press.

Moyle, R . (1986). Alyawarr Music: songs and society in a central Australian community. Canberra: Institute for Aboriginal Studies.

Myers, Fred R. (1986). country, Pintupi self. Washington: Smithsonian Institution.

Nash, David (1980). Topics in Warlpiri Grammar. PhD thesis, Linguistics, Australian National University.

Newman, John (ed.) (2002). The Linguistics of Sitting, Standing, and Lying. Amsterdam: John Benjamins.

Nordlinger, R. (2002). Non-finite Subordinate Verbs in Australian Aboriginal Languages: Are Nominalised verbs Really nominalised? In Cynthia Allen (ed.) Proceedings of the 2001 Conference of the Australian Linguistic Society. http://www.als.asn.au

O'Connell, James F (1979). Room to move: Contemporary Alyawarra settlement patterns and their implications for Aboriginal housing policy. In M. Heppell (ed.), A Black Reality: Aboriginal camps and housing in remote Australia 97-120. Canberra: Australian Institute of Aboriginal Studies. 171 O'Shannessy, Carmel (2005). Light Warlpiri: A New Language. Australian Journal of Linguistics 25: 31-57.

Palmer, F.R (1986). Mood and Modality. Cambridge: Cambridge University Press.

Payne, Thomas E. (1997). Describing Morphosyntax: A Guide for field linguists Cambridge: Cambridge University Press.

Richardson, N (2001). Public Schooling in the Sandover River Region of Central Australia during the Twentieth Century – A critical historical survey. M.A. Thesis, Department of Cultural Studies, Flinders University, Adelaide.

Roth, W. (1897). Ethnological studies among the north-west Central Queensland Aborigines. Brisbane: Government Printer.

Rumsey, A. (1993). Language and Territoriality in Aboriginal Australia. Language and Culture in Aboriginal Australia. In Michael Walsh and Colin Yallop. (eds.), 191-206. Aboriginal Studies Press. Canberra.

Shopen, Timothy (ed.) (1985). Language typology and syntactic description. Cambridge University Press: Cambridge.

Simpson, Jane, David Nash, Mary Laughren, Peter Austin and Barry Alpher (eds.) (2001). Forty Years On: Ken Hale and Australian languages. Canberra: Pacific Linguistics.

Spencer, Baldwin and Francis J. Gillen (1899). The Native Tribes of Central Australia. London: Macmillan

Stanham, Janet (1972). Notes on the grammar of Alyawara, including a small dictionary (Murray Downs Area). Darwin: Summer Institute of linguistics.

Strehlow, Carl (1907). Die Aranda und Loritja Stamme in Zentral Australien. Frankfurt am Main.

Strehlow, Carl (1909). Vocabulary of the Aranda and Loritja Native Languages of Central Australia with English equivalents.: Unpublished manuscript.

Strehlow, T. G. H. (March 1936). Notes on Native evidence and its value. Oceania VI, 331.

Strehlow, T. G. H (1944). Aranda phonetics and grammar. Sydney: Australian National Research Council.

Strehlow, T. G. H. (1947a). Anthropology and the Study of Languages. Report of the Twenty- sixth Meeting of the ANZAAS. Perth.

Strehlow, T. G. H. (1947b). Aranda Traditions. Melbourne: Melbourne University Press.

Strehlow, T.G.H (1962). Aboriginal Australian languages and literature Hemisphere 6.

Strehlow, T. G. H. (1965). Culture, Social Structure and Environment in Aboriginal Central Australia. In Ronald. M. and. Catherine Berndt. Berndt (ed.), Aboriginal Man in Australia. Sydney Angus and Robertson.

Strickland, David (1998). Alyawarr Grammar: an interim report. Darwin: Summer Institute of Linguistics.

Sutton, Peter and Walsh, Michael (1979). Revised linguistic fieldwork manual for Australia. Canberra: Australian Institute of Aboriginal Studies. 172 Talmy, Leonard (2007). Lexical typologies. In T Shopen (ed.) Language typology and syntactic description 66-168. Cambridge: CUP.

Terrill, Angela (1997). The development of antipassive constructions in Australian languages. Australian Journal of linguistics 17, 71-88.

Thompson, S. A., R. E. Longacre, et al. (2007). Adverbial clauses. In T. Shopen (ed.), Language Typology and Syntactic description Cambridge: CUP.

Timberlake, A. (2007). Aspect, tense, mood. In T. Shopen. (ed.), Language Typology and syntactic description. 280-333. Cambridge: CUP.

Tindale, Norman B. (1974). Aboriginal tribes of Australia: their terrain, environmental controls, distribution, limits and proper names. Berkeley and Los Angeles: University of California Press.

Tunbridge, Dorothy (1988). Affixes of motion and direction in Adnyamathanha. In P.K Austin (ed.), Complex Sentence Constructions 267-283.

Turpin, Myfany. (2000). A Kaytetye Learner's Guide. Alice Springs: IAD Press.

Turpin, Myfany and Alyawarr speakers (2009). Thip-rnemel ileyel-akert: Things that birds let you know about. Alice Springs: Charles Darwin University.

Turtle, N. J (1977). Alyawarra Phonology. Workpapers of SIL-AAB, 1-56.

Turtle, Nancy (n.d). Alyawarr Grammar. Darwin: Summer Institute of Linguistics.

Vaszolyi, E (1976). Wunambal. In R.M.W Dixon (ed.), Grammatical Categories in Australian Languages 629-46. Canberra: Australian Institute of Aboriginal Studies.

Vaux, Bert and Cooper, Justin (2003). Introduction to linguistic field methods Muenchen: Lincom Europa.

Verstraete, Jean-Christophe (2006). The Nature of Irreality in the Past Domain: Evidence from Past Intentional Constructions in Australian Languages. Australian Journal of Linguistics 26, 59-80.

Wallace, Noel (1979). Pitjantjatjara wiltja or white man’s house. In M. Heppell (ed.), A Black Reality Canberra: AIAS.

Walsh, Fiona and Josie Douglas and Alyawarr speakers from Ampilatwatja (2009). Angka Akatyerr-akert, A Desert raisin report. http://www.desertknowledgecrc.com.au/publications/downloads/DKCRC_Angka-Akatyerr- akert_A-Desert-raisin-report.pdf.Alice Springs: Desert Knowledge Cooperative Research Centre, CSIRO and Charles Darwin University.

Walter, Karen R (1988). The proper breadth of interest, Norman B. Tindale: the development of a fieldworker in Aboriginal Australia 1900-1936. M.A Thesis, Anthropology, Australian National University.

Wilkins, David (1988). Switch Reference in Mparntwe Arrernte (Aranda): Form, Function, and Problems of Identity. . In Peter Austin (ed.), Complex Sentence Constructions in Australian Languages. Amsterdam: John Benjamins.

Wilkins, David (1989). Mparntwe Arrernte (Aranda) Studies in the structure and semantics of grammar. Ph.D thesis, Linguistics. Australian National University.

173 Wilkins, David. P (1991). The semantics, pragmatics, and diachronic development of "associated motion" in Mpwarntwe Arrernte". Buffalo Working Papers in Linguistics 91, 207- 57.

Wilkins, David P. and Hill, Deborah (1995). When GO means COME: Questioning the basicness of basic motion verbs. Cognitive Linguistics 6, 209-59.

Yallop, C.L (1969). The Aljawara and their territory. Oceania 39, 187-197.

Yallop, Colin (1977). Alyawarra: An Aboriginal language of Central Australia. Canberra: Australia Institute of Aboriginal Studies.

Yallop, Colin (1982). Australian Aboriginal Languages. London: Andre Deutsch.

174