Transfer in language production Edited by Hans W. Dechert and Manfred Raupach (review)

BOOK NOTICES 395 it is high back unrounded [i]); thus, following a current preference for using auto-designations of indigenous peoples in scholarly literature. D introduces the term Tümpisa Shoshone' (hereafter TS). His work provides the first substantial body of published data on the language, and perhaps the last: in 1988, 'there were less than half a dozen people who could speak the language fluently, and they were in their 80s, 90s, and 100s' (Grammar, p. 5). It is not surprising to find that TS phonology resembles the system made famous by Edward Sapir in his article on 'The psychological reality of phonemes'; Sapir's Southern Paiute (SP) is a Southern Numic language, and a close cousin of TS. Thus /p t ts k kw/ become voiced [b d dz g gw] after nasals, are weakened to [ß r ? 7 7w] between vowels, and become voiceless [f r ? ? xw] when the following vowel undergoes phrase-final devoicing. TS goes beyond SP in one respect: after front vowels, [rr] are replaced by [d ?]. TP morphology is. again, similar to that described by Sapir in The Southern Paiute language (Boston, 1930-31). But since, in the style of his day, Sapir said little about syntax, D's account of TP sentence structure is especially welcome. TP turns out to display 'most of the typological characteristics of a verb-final language ' (13): however, 'it is by no means a rigid verb-final language', but has 'flexible word order ... different word orders perform different pragmatic functions in discourse' (16-17). One may suspect that the same is true of other languages which linguists have described simply as falling into the SOV category. In his Introduction to the Grammar, D notes that, 'Because this publication is not aimed only at specialists in linguistics or the Uto-Aztecan languages, technical terms are occasionally defined ' (p. 9, n. 1). The effort to make works on American Indian languages accessible to a wide audience is laudable, though I wonder what most nonlinguists will make of a description that begins: 'TP is a synthetic language primarily using agglutination to form words' (p. 10). Nevertheless, in an age when writing descriptive grammars sometimes seems an obsolescent enterprise, as compared to the elaboration of linguistic theory, D is to be congratulated on producing an admirably intelligible, relatively theory-neutral account. Perhaps not all potential readers, but at least linguists of all persuasions , will appreciate its lucidity. The Grammar concludes with six texts—a historical narrative, four myths, and a conversation . D presents each text in an unusual triple format—first in the original language alone, 'to preserve the aesthetic integrity of the TS discourse ' (447), second in idiomatic English, and finally in TS with an interlinear translation. An appendix contains a short 'Basic Vocabulary List', a tape recording of which is available in the Berkeley archives. The Dictionary is basically TS-English; affixes are included as well as words. Most entries , commendably, contain at least one illustrative sentence. Spanish loanwords are identified, e.g., pihyooni 'bean'