154LANGUAGE, VOLUME 74, NUMBER 1 (1998) understood (including diagrams), gives examples of the verbs with which it is most commonly associated, and identifies metaphorical uses of the terms in situations of actual language use. Thus, for example, the term polang, which occurs frequently in the anaconda story, is included in the category labeled contrasts of sensible experience; it has two senses, only one of which will be exemplified here. Polangi 'describes the moment of emergence from underwater to the surface. Aspectually punctual, completive, and resultative'. The accompanying diagram shows a line with an arrow passing upward through a line representing the rippled surface of water, with a few splash drops to emphasize the movement. The iconic properties ofpolangi are related to its disyllabic structure that shows contrast and movement, the liquid -1- indicating effortless movement as through water, and the 'word-final nasal stop, -ng, creates, by the lowering of the velum, an obstruction in the airstream, which is iconic of the cessation of movement that occurs when something has emerged to the water's surface and has no medium left in which to move' (156). The imaginai and digrammatic aspects of the sound symbolism of polangi are thus displayed so that their application can be easily retrieved from the numerous contextualized examples and understood in relation to the verbs with which they are most commonly associated. This exciting and impressive work has important implications for a number of different directions in linguistic anthropology, including further exploration of sound-symbolism, the relationship between grammar and culture, what happens to languages that were once exclusively oral when they become written, and the general process of synaesthesia. N's analytical implementation ofLakoffs concept of grammatical ecology is highly effective. The iconic or nonarbitrary aspects of language have been much neglected in general. This work will doubtless provide a model for future investigation of this phenomenon, which is found in many languages. REFERENCES Feld, Stephen. 1982. Sound and sentiment. Philadelphia: University of Pennsylvania Press. Mannheim, Bruce. 1991 The language ofthe Inka since the European invasion. Austin: University ofTexas Press Department of Geography and Anthropology and Interdepartmental Program in Linguistics Louisiana State University Baton Rouge, LA 70803 The phonology and morphology of Kimatuumbi. By David Odden. (The phonology of the world's languages.) Oxford: Clarendon Press, 1996. Pp. vi, 315. $80.00. Reviewed by Larry M. ??µa?, University of California, Berkeley Writing a phonology and morphology of a Bantu language is a daunting task. First there is the issue of how to cover everything: Bantu languages are complex at every level—nominal and verbal morphology, lexical and postlexical phonology, prosody, interactions between 'modules', etc. Second, there is the problem ofhow to present one's findings: organize chapters by grammatical category (nouns vs. verbs), phenomenon (vowels, consonants, tone), levels and/or prosodie domains (lexical strata, lexical vs. postlexical phonology), theoretical issues, and so forth. Finally, there is the additional problem of determining what the relationship should be between description and theory. Assuming that one does not ignore theory outright, one could, on the one hand, let the theory be one's guide, whereby focus is on phenomena of theoretical interest. Alternatively, one could let the language guide the work, whereby focus is on the description with possible references to theory. Perhaps these difficulties account for why there have been so few attempts to provide broad or exhaustive coverage of the phonological and morphological phenomena that recur with interesting variation in one Bantu language after another. Part of the interest of the present work, then, is to see how such a task might be undertaken. Odden, a senior phonologist working on Bantu, has an impressive record of theoretical and descriptive accomplishmentthat few can match. His previously published articles on Kimatuumbi, REVIEWS155 an otherwise understudied language, have had great impact on the field, particularly on our conception of the syntax-phonology interface (e.g. Odden 1987, 1990, 1993). In the present study O offers, as one reader of the original manuscript put it to me some years ago, 'everything you ever wanted to know about Kimatuumbi'.He begins in Ch. 1 ('Introduction', 1-19) by identifying this Bantu language spoken in Tanzania, its segmental inventory, and the...