[This document can be acquired from a sub-directory coombspapers via anonymous FTP and/or COOMBSQUEST gopher on the node COOMBS.ANU.EDU.AU] The document's ftp filename and the full directory path are given in the coombspapers top level INDEX file] [This version: 4 August 1993] -------------------------------------------------------------------------- File 1. There are 9 parts of this thesis: from 0 to 8 -------------------------------------------------------------------------- A Grammar of Garadjari, Western Australia by Anna Kristina Sands 1989 Bachelor of Arts Thesis Department of Linguistics, Faculty of Arts Australian National University, Canberra ACT 0200 1 INTRODUCTION The Garadjari language is spoken in an area of Western Australia bounded by the 80 mile beach on the coastal side and extending inland into the desert. Along the coast it extends from Cape Jaubert to Thangoo with the main centre for speakers now being the station of La Grange, South of Broome, where there are some twenty to thirty speakers. The Garadjari language is divided into two dialects, the coastal dialect refered to as nadja 'here' and the inland dialect, nanu 'there' [ Capell, 1962 ]. Though, according to McGregor, [ 1988 ] there are three dialects; Najanaja, the coastal dialect, Nawurtu, the inland dialect, and Nangu, the dialect in between. 1.1 Relation to Other Languages. Garadjari, as stated in O'Grady et. al. [ 1966 ] along with Njangumarda ( its' Southern neighbour ) and Mangala ( its' Eastern neighbour ) form the Marrngu sub-group of languages within the Pama-Nyungan group. This sub-group is most closely related to the Ngumbin sub-group ( including Djaru and Walmatjari ) and the Ngarga sub-group ( including Warlpiri ). To the North Garadjari is bounded by Yawuru a member of the Nyulnyulan family and one of the prefixing languages. 1.2 Previous Work For a comprehensive list of material available on Garadjari see McGregor [ 1988 ] . This list includes both linguistic and anthropological work done on the people. There have been two major linguistic studies done on Garadjari. Capell [ 1962] includes a sketch grammar of Garadjari in his handbook of Australian languages. His work was done on the coastal dialect. He bases his grammar on bases ( roots ) and supplements ( suffixes ) with the supplements divided into numbered series, ie. he uses a very old fashioned method of presenting his data. He did not recognise that there are verb conjugations in Garadjari. His grammar remains fairly incomplete with more concentration on the morphology and very little on syntax, though in the morphology his description of the supplements and their usage and semantics is very brief. McKelson [ ? ] has also done some work on the Garadjari language though his work is based on the inland dialect, ie. it tends more towards Mangala. The major component of McKelson's work is a word list with a very comprehensive range of vocabulary. McKelson has also done a very brief sketch grammar of Garadjari concentrating almost entirely on nominal and verbal inflection for which he gives paradigm charts. As in Capell's work, his description of the usage of each suffix is extremely brief. Thus, previous work on Garadjari is sketchy and based mainly on inflectional morphology and more particulary on the form of the suffixes rather than the usage. Studies in the syntax are minimal and limited to word order. Thus the way remains open for a much more detailed study of the language. 1.3 Laves' Texts For the basis of my work on Garadjari I have used a corpus of texts transcribed by Gerhardt Laves during the nineteen thirties. This corpus contains some forty-five texts - some rather hefty - all relating to myths and legends of the tribe, ie. the texts are all stories, rather than descriptive, and are based on the third person, not the first person. Laves worked with informant(s) from the coastal dialect ( ie. the same dialect that Capell based his description on ) though I do not know how many informants he had. His texts consist of a handwritten phonetic transcription of the Garadjari with English glosses for most of the words underneath. He ignores most suffixes in his glosses with the most notable exception of the ergative suffix. He includes most sentence or clause breaks by indicating a slash. A brief example of a portion of text is as follows : Mirin kadibunu karlikuru / dulkanu mabudjarnu biljgabardinu hero cutting bullroarer split ( chisel ) crack him kalayagarnu / warainj djingarnu / kalaya mabu bunula leave him one make good enough cutting another The second portion of the texts is a typewritten translation of the texts into English. The translations are full of holes and question marks where Laves is not sure of the translation and in some sections whole portions of the text have been left out. Laves' translation of the above portion of text follows : Mirin cut ( gross cutting ) a bull-roarer, he ? split it, [ and ] made it good. ? It split, so he abandoned it. Another one he broke off. Good! He cut ( ie. chiselled ) it . 1.4 Problems Working with photocopies of handwritten texts presents problems in the actual interpretation of the data. It is often very difficult to determine the difference between letters with an upslash or a down slash, ie. g's, y's, ng's, p's etc. become virtually indistinguishable in his more rapid transcriptions and it is often only through knowing the actual word from previous transcriptions that I can tell what is actually written down. Words, therefore, that appear only once or twice in the texts may be interpreted incorrectly, but in a grammatical description, rather than a vocabulary, this is not that important. The greatest problem arising from Laves' style of writing is with the rhotics where there appears to be a great deal of confusion. This problem is more fully described in chapter 2. Apart from the problems associated with the writing are the difficulties involved in the actual content of Laves' transcriptions and translations. Because the translations are full of holes there is much that remains uncertain and I can only guess as to what a translation should entail. In many cases where the vocabulary is totally unknown to Laves ( and therefore to me ) or where the syntax is particularly obscure I cannot make a guess and a hole must remain. Thus, many points of the more complex syntactical constructions may be missed. Laves' translations are very simple and seem to represent the least complicated translation from the given data without representing any of the semantic nuances that one might expect that a slightly different grammatical construction would imply. For example, with verbs of telling, there are various cases that may be used for the agent and recipient ( see section 4.3 ). the various combinations, however, are translated exactly the same in Laves' translations and thus if there is any semantic difference associated with the usage of a different case here it is left unstated. This is a problem, of course, that could arise from having informants with limited English. As I do not know how Laves obtained his translations, whether through informant elicitation, or through his own guess-work, I do not know how accurate his translations actually are. 1.5 Representation of the Texts Throughout the following paper I quote examples of sentences from the Laves' texts. These sentences I have altered to a phonemic, rather than phonetic representaion of the data and they thus differ in this way from the material in the texts themselves. The translations are generally Laves' except where Laves has stated something that in my analysis seems to be wrong and I have altered the translation to fit what I believe to be a more accurate translation. We also note that Laves' Translations are full of question marks. I have omitted these question marks from my interpretation and generally left out from my examples sentences which I do not fully understand. Phonologically, I have basically stuck to Laves' representation of the phonemes. The letter j folowing a d, n or l ( eg. dj ) represents a lamino-palatal sound and an r preceding a d, n, or l ( eg. rl ) represents an apical post-alveolar. The velar nasal I represent by the sequence ng and where we have a velar stop following an alveolar or velar nasal the velar stop is represented by the letter k. This is to eliminate ambiguity that may arise from the sequence ng. -------------------------------------------------------------------------- end of file