Anderson’s general introduction (pp. 17-54) provides detailed justifications for the methodologies to be applied. Then follow four chapters on color terminology, a topic much debated by twentieth-century researchers that provides a thorough workout for Anderson’s approach (55-218). The next six chapters deal with central concerns of human experience: seasons of the year (219-66), geometric shapes (267-308), the five senses (309-26), the folk psychology of mind and soul (327-52), plant life-forms (353-404), and animal life- forms (405-52). The book concludes with a chapter on the problem of universals (453- 84). Back matter includes endnotes (485-506), extensive bibliography (507-75), an index of early English words discussed (577-79), and a general index (581 -87). Specialized word indexes are provided for individual chapters.

Historical semantics is an irreducibly complex field. Anderson responds to its challenges in the only possible way, by diligent collection of diverse evidence. Indo-European precursors of key Old English terms are extensively reanalyzed from the perspective of recent anthropological research. Semantic developments of these words are followed through Middle English to the present day, with a wealth of historical and literary detail that is not only pertinent but interesting in its own right. Theories of Old English semantic development are tested against evidence from the history of non-Indo-European languages. Relations of textual genres to social status are duly considered.

Anderson is particularly adept at untangling popular semantic systems of Old English from systems of specialized crafts and from systems disseminated by the monastic academy. Cognate words from other Indo-European languages indicate plausible starting points for Old English innovations. For the Old English stage, details of morphology provide independent criteria for distinguishing basic terms of the folk taxonomy from secondary or tertiary terms. In general, we would expect a basic term to be simplex rather than com pounded or derived. Discourse constraints also provide valuable diagnostics for Anderson’s analysis. We would expect a basic term to be employed in quite general contexts, like the color term red, as distinct from bay, which is largely restricted to descriptions of horses, or the common Old English psychological term mod ‘mind’, as distinct from hyge, a word of similar meaning largely restricted to poetry in the Old English corpus, though it seems to have been the basic term in Germanic (p. Contextual restriction signals the presence of a cultural specialization (horse breeding, versecraft) with terms of art not in general use.

Because Anderson treats words as elements of systems, his comments on individual Old English texts have special weight. Distinguishing folk taxonomies from religious-academic classification systems reveals a sharp cultural divide between Beowulf and the prose of Christian intellectuals. This result should prove sobering to anyone who imagines the Beowulf poet as an antiquarian academic, “one of us,” so to speak. The section on employment of an archaic two-season system in Old English verse (pp. 255-66) shows how very inappropriate it has been to speak of the Beowulf poet as inspired by Latin academic models. Writers clearly inspired by such models have quite different semantic systems. Awareness of this difference is betrayed by, Ælfric when he defines the four seasons for his readers. If the four-season system had been established in Anglo-Saxon England, Anderson observes, Ælfric would hardly need “to explain that there are four seasons, and to enumerate them for anyone who might have forgotten what they are” (p. 219). A principled approach that keeps overall usage patterns in view seems likely to bear more fruit than the sort of ad hoc word study that too often obscures much of what was clear before.

Anderson’s analysis of psychological terms calls into question a hypothesis put forward by Michel Foucault and Harold Bloom, among others, that the concept of self is a modern inventions or at any rate post—Old English (chapter 9). If we set aside the specialized terms of clinical psychiatry our modern semantic system seems no more elaborate than the Old English one (p. 351). Old English folk taxonomy seems rather scientific in associating the self with the upper half of the body, where all organs required for consciousness can be found, including, of course, the heart and lungs. Our modern popular distinction between mind and heart is a rather simplistic binary that does not delimit the locus of “self” more precisely. If academic discourses are to be compared, a passage by Ælfric giving Latin and Old English terms for subjective manifestations of the soul (p. 339) seems entirely adequate. The simple but common mistake is to compare modern academic discourses with popular Old English ones.

As Anderson correctly points out, renewal of historical linguistics would be greatly facilitated by two-way constructive criticism between anthropological and historical linguists (p. 21). Unfortunately, Anderson adopts no such constructive attitude toward mainstream theoretical linguists, who must surely have something to offer as well. Noam Chomsky’s universalist program, for example, is dismissed as ethnocentricity unmitigated by good intentions: “More insidious is the Chomsky an deduction according to which, because the universal grammar is inherent in all languages, it can be recovered from an examination of any one language” (p. 30). There is no reference for this supposed deduction, and no work by Chomsky appears in the bibliography. I don’t remember encountering such a programmatic statement even in the 1960s, when the pioneers of generative grammar based their arguments on languages they happened to control—necessarily so, since reference grammars available then did not provide the kind of detail they needed. Today, at any rate, Chomsky’s evolving models are checked against evidence from all major language groups according to standard protocols, and the level at which universals reside is approached by working through relevant linguistic differences. In contemporary theory, a construction from a particular language is never represented as normative but rather as the outcome of choices from a universal set of possibilities. Contemporary phrase-structure grammar analyzes all syntactic constituents into an obligatory head, an optional complement, and an optional specifier; but the ordering of those universal constituents varies from language to language. Many predictions of current models are implicational. If a language routinely places the verbal head after the object complement, for example, it is predicted to use postpositions. Exceptions that arise during transition from one basic order to another are acknowledged and actively studied. Research of this kind is obviously not ethnocentric.

I think Anderson is simply mistaken in supposing that he disagrees with Chomsky about the nature of universals. Many generalizations endorsed by Anderson are implicational universals of a type familiar to mainstream syntacticians, for example, the generalization that if a language has only one chromatic color term, the paradigmatic referents of that term will be red (p. 130). Chomsky might fairly regard the work under review as justification from the mouth of a hostile witness with interests quite different from his own.

Anderson’s gratuitous opposition to Chomsky does not invalidate the many significant achievements of his book. The author has found adequate conceptual tools for implementation of his research interests. Mainstream syntax would not be directly helpful. I hope, however that future work of this kind will be situated within the prevailing intellectual framework for general linguistics. To judge from Folk-Taxonomies in Early English, that would not be difficult.

Geoffrey Russom, Brown University