Batad Ifugao dictionary with ethnographic notes Comp, by Leonard E. Newell and Francis Bon'og Poligon (review)

BOOK NOTICES 427 case with Li's recently published dissertation, which brings into focus the question of vowel harmony (VH) in the Tungusic languages of Siberia, China, and Mongolia by means of dependency phonology (DP) and optimality theory (OT). The first three chapters provide background in the typology (4-24), phonological models (25-34), and previous work in the areas ofTungusic VH (35-49) and VH in general (50-81). Data particular to the analyses are generous and presented well in the following two chapters (82-212), approached systematically (213-313), and used tentatively to suggest further applications (314-40). L was faced with data from varied and often unreliable sources (despite extensive fieldwork); as a result the quality of material differs throughout the book. Formant frequency values for vowels were available for Solon (107), but elsewhere vowel quality had to be extrapolated from detailed descriptions (e.g. Baiyinna Orochen, 86-91) Not surprisingly we find that attested forms can vary widely (e.g. Sibe, 188-89). The analyses themselves were occasionally problematic . DP is based on a generative, nonlinear tree structure with monovalent articulator-based feature nodes (contra Chomsky & Halle's SPE binary features and the Halle-Sagey tree model). Under DP, markedness increases with more features under a node; thus, the primary harmonizing feature of Tungusic VH, 'retracted tongue root' (RTR) according to L, marks a segment. Yet several languages (Literary Evenki, Amur Evenki, Oroch, Udehe, and Classical Manchu) have in their phonemic inventory the marked [o] to the exclusion of unmarked [o], an unnatural situation (114). Further, the monovalent RTR appears to be binary in Xunke Orochen (141-44): L opts to analyze [i] and [i:l as phonologically 'neutral ', that is, neither 'RTR' nor 'nonRTR', presumably assigned 'nonRTR' phonetically by fill-in rules. Given these idiosyncrasies, the choice of RTR over ATR as the harmonizing feature might be called into question (cf. 109-10). The OT analysis suffers in places as well: Vocalic epenthesis in Sibe uses one of two 'disjunctive and independent' sets of constraints based on the presence of a rounding feature in the input (304-6). Such disjunctive subsets are also found in L's analysis of the past tense suffix alternations in Sanjiazi Manchu (277). Under OT, constraints are either ordered or equally ranked but in any case are always available to the phonology. 'Disjunctive and independent' sets assume that different constraints are available for different words, antithetical to the present understanding of OT. On the whole I can recommend this book more as a research reference than a cohesive argument for Tungusic VH as RTR-harmonizing, an otherwise unattested typology among language families (318). The concluding remarks include a brief view into applications within the Altaic hypothesis (337-40), perhaps the most interesting of possible outcomes of this study; yet the discussion yields no resolution. [Steven Berbeco. Harvard University.] Batad Ifugao dictionary with ethnographic notes. Comp, by Leonard E. Newell and Francis Bon'og Poligon. Manila: Linguistic Society of the Philippines , 1993. Pp. xxviii, 744. $42.95. Dictionaries of Philippine languages derive from two primary sources. The University of Hawaii PALI language text series has published dictionaries of Bikol, Hihgaynon, Ilokano, Kapampangan, Maranao , Pangasinan, and Tagalog, and the Linguistic Society of the Philippines has published dictionaries of Cebuano Visayan, Binukid, and now this dictionary of Batad Ifugao. They are some of the most comprehensive dictionaries available, and they have set a high standard for the lexicographer. The Batad Ifugao dictionary is the result of nearly four decades of work by Newell, and the breadth of his knowledge of the language and its speakers is reflected in its detail. The dictionary proper is introduced with a grammatical sketch (92 pages). The Batad Ifugao—English portion occupies 449 pages. The value of N's long contact with the language is most apparent in the comments which frequently accompany the English glosses of the Batad Ifugao entries and which ground them in contexts of cultural usage. This practice extends from entries of cultural significance to items of minor import. For example, accompanying the entry for tantán 'needle', we find 'Traditionally a needle was made of a 5 - 8 cm bamboo piece sharpened at one...