Realism bei Frege: Reply to Burge

Frege is celebrated as an arch-Platonist and an arch-realist. He is renowned for claiming that truths of arithmetic are eternally true and independent of us, our judgments and our thought; that there is a ‘third realm’ containing nonphysical objects that are not ideas. Until recently, there were few attempts to explicate these renowned claims, for most philosophers thought the clarity of Frege’s prose rendered explication unnecessary. But the last ten years have seen the publication of several revisionist interpretations of Frege’s writings—interpretations on which these claims receive a very different reading. In ‘Frege on Knowing the Third Realm’, Tyler Burge attempts to undermine this trend. Burge argues that Frege is the very Platonist most have thought him—that revisionist interpretations of Frege’s Platonism, mine among them, run afoul of the words on Frege’s pages. This paper is a response to Burge’s criticisms. I argue that my interpretation is more faithful than Burge’s to Frege’s texts. Gottlob Frege has long been regarded as one of the least compromising realists in the recent history of philosophy. This assessment of Frege is supported by his renowned claims that the truths of arithmetic are eternally true; that the truths of arithmetic are independent of us and our judgments and thoughts; that not every object has a place; that there is a 'third realm' containing objects which are neither physical objects nor ideas. There have been relatively few attempts to explicate these renowned claims, largely because most philosophers have agreed that the clarity of Frege's prose renders explication unnecessary. But the last ten years have seen the publication of several revisionist interpretations, including interpretations on which even Frege's renowned claims express unfamiliar views. In a recent article, 'Frege on Knowing the Third Realm', Tyler Burge has responded to this revisionist trend. Burge argues that Frege is indeed the Platonist and realist that most have thought him and he argues that several revisionist interpretations, mine among them, go afoul of the words on Frege's pages. On Burge's view, the meaning of Frege's renowned claims is obvious and does not require interpretation. Consequently, Burge devotes a good portion of his paper to what he regards as simple reports of Frege's views along with observations that revisionist interpreters get them wrong. Thus most of Burge's descriptions of Frege's Platonism are supported neither with explanations of Frege's views nor with textual arguments for the accuracy of Burge's interpretation but, rather, with long lists of citations. Burge's point, which he thinks should be obvious to any reader of Frege's writings, is that Frege believes in a 'third realm' consisting of objective, but non-spatio-temporal entities, which are not in any way dependent on human conceptualization or human language. This realm includes, according to Burge, "some objects, and all functions" (1992, p. 637) or, to be more specific, it includes "extensions -including numbers -functions -including concepts --and I am indebted to Cora Diamond, Gary Ebbs, Mark Kaplan and Thomas Ricketts for comments on earlier drafts. The Burge paper to which this is a response was presented at a conference on early analytic philosophy in honor of Leonard Linsky and I have benefited from discussions with participants. I have also benefited from discussions of an ancestor of this paper, titled 'Realism bei Frege', at the University of Chicago, the University of California at Riverside and the University of Wisconsin at Madison. 1 The interpretations to which Burge responds explicitly are those in Hans Sluga (1980) and Thomas Ricketts (1986) as well as my own (1990). thought contents" (1992, p. 636). While Burge thinks that this much is evident, however, he does not think that these claims alone constitute a Platonist view. He writes, Many of these things might be maintained by someone who was not a Platonist. One might make the remarks about imperceptibility, non-spatiality, atemporality, and causal inertness, if one glossed them as part of a practical recommendation or stipulation for a theoretical framework, having no cognitive import -or as otherwise not being theoretical claims or claims of reason. ... Or one might have some other basis for qualifying these remarks, reading them as "non-metaphysical" or as lacking their apparent ontological import. ...Platonism has no monopoly on claims to lawlike or inter-subjective objectivity about nonspatial, atemporal entities. (1992, p. 637) Frege's renowned statements about non-spatio-temporal objects constitute an expression of Platonism only if they are meant to play some theoretical role in the explanation of some phenomenon. Burge's burden -and here he admits he is engaging in a substantial interpretive effort -is to argue that Frege's renowned claims are meant to play such a role and that the phenomenon they are meant to explain is the objectivity of science. One of the interpretations to which Burge objects is the one I set out in my (1990). And, given the above, it is not difficult to see why. I have argued that, while what is objective is, on Frege's view, independent of what actually has been (or will be) expressed in language, it is not independent of what can be expressed in language. Although concepts, for Frege, are not created, but discovered, the discovery of a concept is not anything like the discovery of a law of nature -since the Fregean notion of concept bears no interesting relation to anything like that of a natural kind. Concepts are discovered by the formulation of precise distinctions. Moreover, these formulations of precise distinctions also suffice to give us (at least some) thoughts. Finally, and perhaps most important, none of Frege's views about concepts and functions can be part of a theory. Frege himself states that such predicates as 'concept' and 'function' are defective and he says, of his sentences in which such terms appear, that they are not literally true. The purpose of these sentences in Frege's writings is meant to be elucidatory. For Frege, on my interpretation, no theory can include a statement that functions or concepts are non-spatio-temporal objective entities. And, consequently, Frege would have denied that it was possible even to formulate the sort of metaphysical theory that, on Burge's view, constitutes Frege's Platonism. Further, such terms as 'objective' and 'subjective' are also precluded from playing a role in a theory. Thus Frege would not have been attempting, as Burge claims, to give a theoretical explanation for the objectivity of scientific practice. Now, given what I have said so far, the choice between Burge's view and mine may seem easy. Burge takes Frege at his word. I claim that Frege cannot always be read literally. It looks as if it is Burge who takes fidelity to the words on the page more to heart. But it is not that simple. Frege himself says, "By a kind of necessity of language, my expressions, taken literally, sometimes miss my thought" (1892 p. 204; see also 1979, p. 250 (1983 p. 269)). Thus Burge cannot be taking all Frege's remarks literally. He must, at least, be disregarding the above statement. Burge might respond that, if he is disregarding some of Frege's statements, he is surely disregarding fewer than I am. But significance cannot be measured by counting statements. I will argue that my 2 See, e.g., 1979, pp. 177-178, 239 (1983 pp. 192-193, 257-258) 1980a, p. 141 (1976 p. 224). For a more complete discussion of Frege's view about sentences in which such predicates as "concept" or "function" appear, see my 1990 pp. 246-260. 3 For a discussion of elucidatory statements and the significance of the difference between elucidation and theory, see my 1990, Chapter 6 and 199?b. interpretation is more faithful to the words on Frege's pages than Burge's apparently flat, unvarnished reports are. I will also argue that Burge's substantive interpretive work is no more convincing than his reportage.