Dummett on Frege: Semantic Realism

Review Article DUMMETT ON FREGE: SEMANTIC REALISM Michael Dummett's Frege: The Philosophy of Language is a rich and frequently exciting philosophical work, a searching exploration of important questions. ' It is a generous book-generous in the range of topics that it covers and generous in the level of insight it provides. It is to be recommended not only to those who are interested in Frege's seminal thought but to anyone interested in problems in the philosophy of language. Dummett is continuously concerned with the living history of Frege's contributions. However, it may be that the book is generous to a fault. The reader is sometimes presented with serious difficulties in sorting out the bounty that is offered. These difficulties arise not only in connection with matters of interpretative detail hut also in connection with the general aims and structure of the book. On the one hand, Dummett explicates Frege's positions on a range of central issues. As we would expect, there are excellent discussions of Frege on quantification, on sense and reference, on concepts and predication, on truth and the bearers of truth, on assertion, and on the nature of abstract objects. This list of major topics is partial, and under each heading there fall many subsidiary but important issues in metaphysics and the philosophy of language. Through all these discussions there runs a continuing investigationof Frege's conception of the proper aims and the guiding principles of any acceptable theory of meaning. On the other hand, for each of these topics there is a wide-ranging critical assessment of Frege's views both in light of considerations that Dummett raises himself and in light of later developments in the philosophy of language from Russell to Quine. Indeed, Dummett defends the thesis that Frege, directly through his writing and indirectly through his influence, succeeded in making the theory of meaning, properly conceived, the foundational discipline of philosophy as a whole (see especially Chap. 19). Whatever judgment one makes of Dummett's stimulating defense of this claim, it illuminates an aspect of the structure of this work. It is intended to be read as much as a study of the Fregean heritage in philosophy of language as it is to be read as an exegesis of the special features of Frege's thought. Because this heritage is very much alive in philosophy and linguistics, such a study is highly valuable. What is more, Dummett is at pains to argue that much of this heritage is rightly preserved. Unfortunately, these two broad objectives are often intertwined in a confusing way, and the confusion is exacerbated by the author's mode of presentation. Dummett very rarely appeals directly to Frege's writings, and his book contains almost no quotation from or citation of these texts. As a result, it is often unclear where exegesis of Frege leaves off and critical investigation takes up. In addition, it can be very hard to tell what evidence is intended for various of the interpretative theses. For example, Dummett repeats the familiar assertion that nowhere in his published writings does Frege attempt to argue for the view that predicates have referents (p. 206). And yet in the most obvious source, the essay "On Concept and Object," Frege does argue that in quantified statements such as "All mammals are land-dwellers," the imbedded predicate terms must be treated as the 'logical subjects' of the sentence: Admittedly, the argument is an odd one. The notion of a 'logical subject' is introduced out of the blue and is nowhere subsequently New York: Harper and Row, 1973.Page references to this book appear parenthetically hereafter. 2Gottlob Frege, Philosophical Writings, trans. Max Black and P. T. Geach (London: Basil Blackwell, 1952), p. 48. Similar considerations appear in the essay "On the Foundations of Geometry," in On the Foundations of Geometry and Formal Theories of Arithmetic, trans. Eike-Henner W. Kluge (New Haven: Yale University Press, 1971),pp. 34-35. [457] 458 HISTORY OF PHILOSOPHY explained. Also, the argument seems to assume that every sentence must contain at least one term that counts as its logical subject. Nevertheless, the argument seems worth exploring since it offers Frege a relatively...