Explorations in Korean Syntax and Semantics (review)

Book Reviews Explorations in Korean Syntax and Semantics, by Seok Choong Song. Korea Research Monograph 14. Berkeley: Center for Korean Studies, Institute of Asian Studies, University of California, 1988. xiv, 378 pp. Table of romanization, notes, appendix, references, index. $20. The book under review brings together the works of one of the most original and insightful linguists working in the field of Korean syntax over the past twentyfive years. The articles in this volume vary in theme and in quality, but are unified by an adherence to sound linguistic methodology rather than to one particular theory. Song is at his best in his no-nonsense synchronic analyses of controversial problems in Korean syntax and in his devastating criticisms of Korean linguists working within the generative framework in Korea. On the other hand, his ventures into Korean historical grammar are less than successful. This review will characterize each of the chapters, then evaluate the book's overall contribution. Song organizes his book into four parts: Part I presents five chapters under the heading "Markers, Particles, and Grammatical Morphemes," Part II contains six chapters on "Negation: Conjectures and Refutation," Part III is an eclectic blend of five papers on "Causative, Transferentive, Locative and AlIative ," and Part IV finishes up with three papers on "New Approaches to Old Problems." Chapter One, "The Disappearing Nominative Marker," shows how the assumptions that the Nominative Marker (NM) starts out as —i/ everywhere and that both Noun Phrases (NPs) in an affirmative equational sentence are marked with the NM underlyingly, lead to a simpler description of the grammar of the copula in Korean. Thus, the underlying sentence 1) (p. 14) *ku kyoswu-ihakca-ii-ta that professor-NMschoIar-NM copula "That professor is a scholar." 168BOOK REVIEWS will produce the grammatical sentence /ku kyoswu-ka hakca ta/ via rules necessarily independent elsewhere in the grammar. Note also that his analysis in this chapter depends crucially on the existence of a copula in Korean, a problem Song returns to in Part IV. This first chapter sets the tone for the rest of the book: it is well-written (with occasional digs at the English prose of some of his colleagues), cogently argued, and rich in illuminating examples. Chapter Two, "A Ubiquitous Plural Marker," is an attempt to account for cases where /-tul/ appears after elements other than nouns. Song posits a rule of "Plural Marker Copying (PMC)" for those cases where /-tul/ "seems to pop up freely in the least expected positions," such as after adverbs, complementizers, case markers, and so on. For example, 2)ese-tultuleo-key quickly-Plural come in-Imperative "come right in." Song formulates his PMC rule as follows (p. 22): # NP, PL, (X), Y, Z# > 1,2, (3), 4+2, 5 12 3 4 5 Song continues, "After the plural marker is copied, the original marker is usually deleted, as is the subject NP itself." This is an an old-fashioned type of analysis by now in linguistics, and the main contribution of this chapter is that it calls attention to the phenomenon of the roving /-tul/. For example, the position of /-tul/ in 3)cha-na-tulmasi-p-sita tea-deemphasis-Plural drink-humble-propositive "Let's have tea or something." "is a puzzling violation of rules of morphology." To my mind, the behavior of /-tul/ in all these cases is reminiscent of second-position clitics in diverse languages , and I wonder if this would not be a more profitable angle from which to approach this problem in future. Chapter Three, "A Suspicious Analysis of the Suspective Morpheme," is a clever demonstration that what Martin and Lee (1969) considered two different /-ci/ morphemes are in fact four. The following example contains an instance of each different /-«'/morpheme: 4)manwula-lulmusewehanun d-ka wife-accusativefearing1 elmanatoy nunrimalha c/anh keyss ci howlong become 2 say3 will 4 "[He] wouldn't say how long he has been afraid of his wife." Song rejects Martin's analysis of ci-3 and ci-4 above as identical instances of a "suspective morpheme" /ci/. Instead, Song claims (p. 26) that "ci-3 is a variant of the complementizer ki in negative sentences in a specifiable type." Since...