In lieu of an abstract, here is a brief excerpt of the content:

REVIEWS435 Giacalone-Ramat, Anna. 1979. Lingua, dialetto e comportamento lingüístico: La situazione di Gressoney. Florence: Licosa. Gubert, Renzo. 1978. La città bilingue. Bolzano: Athesia. Hill, Jane, and Kenneth Hill. 1977. Language death and relexification in Tlaxcalan Náhuatl. In Dressler & Wodak-Leodolter, 55-69. MacWhinney, Brian. 1978. The acquisition of morphonology. Chicago: University of Chicago Press. Milroy, Lesley. 1980. Language and social networks. Oxford: Blackwell. Schule, Rose Claire. 1971. Comment meurt un patois. Colloque de Dialectologie Francoprovençale, 195-207. Genève: Droz. [Received 4 August 1981.] Linguistic diffusion in Arnhem Land. By Jeffrey Heath. (Research and regional studies, 13.) Canberra: Australian Institute of Aboriginal Studies; Atlantic Highlands, NJ: Humanities Press, 1978. Pp. viii, 146. US $11.75. Ngandi grammar, texts, and dictionary. By Jeffrey Heath. Canberra: Australian Institute of Aboriginal Studies; Atlantic Highlands, NJ: Humanities Press, 1978. Pp. xii, 297. US $24.75. Reviewed by Ian Smith, National University of Singapore The books under review result from Heath's investigation of a previously unstudied sprachbund in Australia's 'Top End'. ' The four languages involved— Ngandi (Ng.) Nunggubuyu (Nu.), Warndarang (Wa.), and Ritharngu (Ri.)—are spoken in Eastern Arnhem Land (Northern Territory) around the Walker and Rose Rivers. It is customary to classify Australian languages as Pama-Nyungan or non-Pama-Nyungan, according to cognate densities and typological characteristics (for detailed accounts of the differences between the two groups, see Blake 1977, 1979, Dixon 1980). Of the languages in H's study, Ri. belongs to the Yuulngu group of Pama-Nyungan; the other three are non-Pama-Nyungan languages. Among the latter, Ng. and Nu. form a subgroup, separated by substantial time-depth from Wa. H worked on these languages, along with others of the area, while he was a research fellow for the Australian Institute of Aboriginal Studies from 1973 to 1977. Linguistic diffusion in Arnhem Land, originally a 1976 University of Chicago dissertation, is mainly concerned with (comparatively) recent phonological and morphological developments; H hopes to complement this with a study of lexical diffusion among the four languages. Similarly, Ngandi grammar, texts, and dictionary is to be matched by similar descriptive studies of the other three languages. 1. I shall look first at the Ng. monograph, so that the non-Australianist reader may get some feel for the type of language being described. H spent only about six weeks gathering his data—a fact which, even allowing for similarities with other languages of the area, identifies him as a field worker of enviable talent. 1 [See now also Heath's recent article in Lg. 56.335-67, 1981.] 436LANGUAGE, VOLUME 58, NUMBER 2 (1982) 1.1.Part One contains the grammar proper. Ng. has a reasonably typical phonological inventory for Australia: its distinction between lamino-dental, apico-alveolar, lamino-palatal, and retroflex coronals is widespread; two (rather than one) contrastive stop series and a glottal stop are less common, but not unknown elsewhere; a five-vowel system is frequent in non-Pama-Nyungan languages. H wisely chooses to use surface transcriptions throughout the book, except in the discussion of morphophonemic derivations. The morphophonology of Ng. is simple enough to be subsumed in fifteen P-rules. Moreover, at least five of the alternations are so restricted that one wonders why H went to the trouble of formalizing general rules for them. However, the phenomena are dealt with honestly, in that the generality of each alternation is carefully set out. However, morph-by-morph glosses of the examples are generally not provided; and while this is perhaps unimportant in the discussion of phonology, it hampers a careful reading of the morphology and syntax sections. One of the most interesting aspects ofAustralian languages is their case-marking systems, which often involve marking on nomináis as well as cross-referencing on the verb (as in Ng.) or on some other sentential element. Ng. has ergative case-marking for all nomináis, and a cross-referencing system which is neither ergative nor accusative in nature. Ng. and other non-Pama-Nyungan languages employ verbal prefixes to cross-reference the person, number, and class of those nomin áis which have major sentential roles. Ng. contains two sets of such prefixes—one for intransitive...

pdf

Share