Formal semantics of natural language: papers from a colloquium sponsored by the King's College Research Centre, Cambridge Ed. by Edward L. Keenan (review)

REVIEWS663 dashes of the Morse Code, for instance, have for their signata the letters of the alphabet. In drum and whistle systems, signs stand for signs; second-order surrogate signs stand for signs of signs.' Some of S's statements have a perhaps unintended piquancy, as when his glowing praise of U. Eco's Theory of semiotics ends with the sentence: 'In the jargon of the Italian left, the author has been identified as an "idealist"; if so, I admire the company he keeps' (167). This suggests a lack of awareness that Eco is a far from unimportant contributor to the rhetoric of the Italian left, to which S pays a presumably unintentional compliment by praising ' the company he keeps '. In other cases, given the high standard of accuracy of these essays, the formulation is not as precise as one might desire; e.g., it might be useful to point out that the article by M. Corti referred to on p. 150 is not so much devoted to 'fortune-telling by tarot' as to a discussion of I. Calvino's // castello dei destini incrociati. Again, in what sense should one take the word 'appears' in the following passage? 'Wallis, Locke's friend and former mathematics professor in Oxford, appears, in turn, to attribute the term semeiotikè, as the art of musical notation, to Marcus Meibomius, with two references to the latter's Antiquae musicae auctores septem (1652)' (48). In fact, Wallis, in his edition of Ptolemy's Harmonics (1682), p. 286, does refer to Meibom, notes ad Alypium, p. 66, where the term is used. We do not know, of course, whether Meibom invented the term, and Wallis does not state this; it may be interesting to observe, in this connection, that the Latin term used in these texts for musical notes is notae, but the Greek is semeta. Sebeok's writing is punctilious, though the effect sometimes verges on the recherché. The scholarship of his book is impeccable, the information ranges through an extraordinarily varied array of references, and even the purely external accuracy is most remarkable (I noted only a handful of misprints). REFERENCES Sebeok, Thomas A. 1972. Perspectives in zoösemiotics. (Janua linguarum, series minor, 122.) The Hague: Mouton. ------. 1974. Structure and texture: selected essays in Cheremis verbal art. (De proprietatibus litterarum, series practica, 44.) The Hague: Mouton. [Received 10 August 1977.] Formal semantics of natural language: papers from a colloquium sponsored by the King's College Research Centre, Cambridge. Ed. by Edward L. Keenan. Cambridge: University Press, 1975. Pp. xiii, 475. £12.50. Reviewed by Gerald Gazdar, University ofSussex, and Ewan Klein, University ofAmsterdam* Within the last decade, there have been an increasing number of attempts to apply to natural languages some of the techniques originally developed for the semantic analysis of formal languages in mathematics. This book, henceforth FSNL, gives a fairly representative sampling of such work. Thomason 1976 has recently pointed out that there is a lack of 'a healthy give-and-take relation between theory and data' in the field of semantics—a tendency to dwell on either facts or formalism at the expense of the other. The present volume is no exception. The fact-oriented trend is exemplified particularly by the papers of Emonds, Gross, Ross, Seuren—and, to a lesser extent, Biggs and Dahl; but a roughly equal number * The second author wishes to thank the Netherlands Organization for the Advancement of Pure Research (Z.W.O.) for support during preparation of this review. 664LANGUAGE, VOLUME 54, NUMBER 3 (1978) of papers are highly formal, and probably opaque to the non-technically-minded linguist, e.g. those of Altham & Tennant, Heidrich, Jardine, Kutschera, and Stechow. In very different ways, the lucid, well-argued papers of David Lewis and of E. L. Keenan both manage to strike a nice balance between the two extremes. Lewis ('Adverbs of quantification ', 3-15) proposes in effect that adverbs like sometimes, often, and always, in sentences like la-c, are to be treated as unselective quantifiers over the free variables in 2: (1)a. Sometimes, a man who owns a donkey beats it now and then. b.Often, a man who owns a donkey beats it...