¹ South Australian Museum Records, VII, 1942, 7.Google Scholar

² Wurm, S. A., ‘The present state of New Guinea (non-Melanesian or Papuan) and Australian historical and comparative linguistics’, Proceedings of the ninth International Congress of Linguists, The Hague, 1964, 579.Google Scholar

³ K. Hale, 'Vocabularies and cognation judgments for 30 Cape York Peninsula languages '(unpublished).

⁴ The writer's major debt is to the Mbabaram informant, Albert Bennett. Mention must also be made of Jimmy Taylor and Mick Burns (Mbabaram), Mitchell Dodd and Jack Brumby (Wagatnan), Willie Richards (Dyangun), Mrs. D. M. McGrath (of Petford), Jack Doolan (for assistance on Palm Island Aboriginal Settlement), Douglas Seaton (of Cairns), and others who assisted in various ways. Major acknowledgement must be made to F. D. McCarthy and the Australian Institute of Aboriginal Studies, for their great co-operation and encouragement throughout and beyond the writer's stay in Australia. And to A. Capell, S. A. Wurm, K. Hale, and La Mont West, Jr. Also, M. A. K. Halliday and R. D. Huddleston read a draft of this article and made some useful suggestions for improvement. Thanks are also due to M. Young of the Department of Geography, University College, London, for his expert drawing of the map.

⁵ cf. the writer's What IS language? A new approach to linguistic description, London, 1965.Google Scholar

⁶ cf. Halliday, M. A. K., ‘Categories of the theory of grammar’, Word, xvii, 3, 1961, 241–92CrossRefGoogle Scholar; Dixon, R. M. W., ‘A logical statement of grammatical theory’, Language, xxxix, 4, 1963, 654–68.CrossRefGoogle Scholar

⁷ In traditional terms, pronoun constructions are of the nominative type, and noun constructions are of the ‘ergative’ type (see, for example, Matthews, W. K., ‘The ergative construction in modern Indo-Aryan’, Lingua, 4, 1953, 391–406).Google Scholar

⁸ cf. Longacre, R. E., Grammar discovery procedures, The Hague, 1964.Google Scholar

⁹ See Dixon, ‘Mbabarain phonology’ (to appear in Trans. Phil. Soc., 1965), section 5.5; and compare Halliday, M. A. K., ‘Categories of the theory of grammar’, Word, xvii, 3, 1961, 254–5Google Scholar; Firth, J. R., ‘A synopsis of linguistic theory, 1930-1955’, Studies in linguistic analysis, Oxford, 1957Google Scholar, 5, 17; Palmer, F. R., ‘“Sequence” and “order”’, Report of the 15th Annual Round Table Meeting on Linguistic and Language Studies, Georgetown, 1964.Google Scholar

¹⁰ The exact status of is doubtful. This ‘word’ occurred three times in the corpus; in (112) and in (113) with gloss ‘take it over there to the east’. It may differ from in implying a different order of distance (‘ a long way over there’, rather than ‘just over there’, say).

¹¹ Although, whereas languages from Mamu down to Dyirbal and Giramay have four noun classes, Mbabaram makes no grammatical distinctions of this sort. A grammatical comparison of Mbabaram with Dyirbal will be included in a projected full description of the Giramay, Dyirbal, and Mamu languages.

¹² This is a quite different unit from grammatical word, lexical word, and graphological or graphetic word, and is set up with regard paid mainly to purely phonological distributional considerations.

¹³ We can usefully think of the phonetic exponent of a term in these phonological vowel systems as a ‘volume’ in three-dimensional phonetic space, where the dimensions are defined by ‘close/open’, ‘front/back’, ‘rounded/unrounded’. Then we are saying that the i-volume for Mbabaram does not fall completely within the u-volume for Dyirbal. In fact, most of both the u-volume and the γ-volume for Mbabaram fall within the u-volume for Dyirbal; but a small part of the γvolume falls within Dyirbal's i-volume. In each instance, the ‘point’ within a particular volume that is ‘chosen’ (i.e. that describes the articulation in this instance) is determined partly by environmental and partly by structural considerations. (Cf. Jones, Daniel, The phoneme, Cambridge, 1962, especially pp. 92–8.)Google Scholar

¹⁴ One Agwamin informant said , the other (now living several hundred miles away) ; the sole Dyangun informant said .

¹⁵ At the level of phonology only systems can be compared, and those only through their phonetic exponents. In order to set up system-system correspondences examples should be given involving all or almost all the terms in each system, showing their correspondents in the other system (and dealing with all kinds of possible structural environment). It can be seen that here the examples given are quite inadequate for the comparisons to have full scientific validity. Similarly, grammatical comparison can only be of systems, and only through the systems'contextual and situational exponents.

¹⁶ Roughly, γ is further forward than . For more detailed phonetic remarks reference should be made to the writer's forthcoming account of Dyirbal.

¹⁷ One informant gave two words for water: gúga and bána; the other only gave bána. bána is very common, as the word for water, in languages of this area.

¹⁸ Tindale, N. B. and Birdsell, J. B., ‘Tasmanoid tribes in North Queensland’, South Australian Museum Records, vii, 1942, 1.Google Scholar

¹⁹ ibid., 2.

²⁰ ibid., 3.

²¹ ibid., 7; no further study has appeared in the intervening 24 years.

²² A.Cape11,A new approach to Australian linguistics (Oceania Linguistic Monographs, No. 1), Sydney, 1962, 111.Google Scholar