![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
Commentators regularly attribute to Frege realist, idealist, and quietist responses to metaphysical questions concerning the abstract objects he calls ‘thoughts’. But despite decades of effort, the evidence offered on behalf of these attributions remains unconvincing. I argue that Frege deliberately avoids commitment to any of these positions, as part of a metaphysical separatist policy motivated by the fact that logic is epistemologically autonomous from metaphysics. Frege’s views and arguments prove relevant to current attempts to argue for epistemological autonomy, particularly that of ethics.
Acknowledgements
Oddly enough, this paper is mainly indebted to an anonymous reviewer for a different journal, commenting on a different paper of mine. That reviewer—while generally disparaging the paper—singled out for praise its brief discussion of Frege's metaphysical separatism, and claimed (in effect) that he or she would have much preferred to have read a paper entirely about that topic, rather than the snoozefest to which he or she had in fact been subjected. On reflection, this struck me as pretty reasonable: the bit about separatism was the most interesting thing in that other paper. When it came to actually writing this one, I owe the most to some uncommonly helpful comments from two anonymous reviewers and an editor at this journal, and to discussions with Richard Lawrence, Kirsten Pickering, Gurpreet Rattan, and Yuan Wu. Working through Pickering, Why Answer the Epistemic Challenge?, a doctoral dissertation about metaethics, also had a great influence.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 For example: “The number 4 … is not a spatial object … Not every object is somewhere”. “Numbers … are outside of time”. (Frege, Die Grundlagen Der Arithmetik, §61: Frege, “Le Nombre Entier”, 74).
2 As Burge (“Frege on Knowing the Third Realm”) observes of Frege’s existence claims: “Carnap might have said at least some of these things … Kant might have said them”. (637). See §2 of Balaguer, “Platonism in Metaphysics” on ideal abstract objects.
3 Miller (“Realism”) identifies “Generic Realism” about a range of objects as the claim that they “exist, and the fact that they exist … is … independent of anyone’s beliefs, linguistic practices, conceptual schemes, and so on”. This standard account has well-known limitations. For example, when a carpenter makes a bed, the bed’s existence is partly explained by the carpenter’s beliefs; but this does not imply that idealism is true about it. We might refine the standard account along the lines of Shafer-Landau (Moral Realism) by restricting the sort of cognition that is relevant: the bed is real iff its existence is not explained by cognition affirming its existence: by “ratification from within any given actual or hypothetical perspective”. (15) Alternatively, following Devitt (Realism and Truth), we could refine it by holding that only certain sorts of explanatory/dependence relations are relevant: the bed is real iff its existence is not constitutively explained by cognition. Morton (“Normative Principles” 1154) plausibly argues that “no one has developed a view of the distinction that avoids the problems”, but these subtleties do not affect the issues in this paper. Without much affecting the discussion, anyone who favours a particular refinement of the standard account should be able to, for example, replace every use of ‘explains’ with ‘constitutively explains.’
4 E.g. McDowell: “Quietism does indeed urge us not to engage in certain supposed tasks, but precisely because it requires us to work at showing that they are not necessary … There is no question of quickly dismissing a range of philosophical activity from the outside”. (“Wittgensteinian ‘Quiestism’,” 371)
5 Carnap, Scheinprobleme in der Philosophie, §9.
6 As Joyce (“Moral Anti-Realism”) describes, for certain classical non-cognitivists in ethics, “one of the major attractions of noncognitivism [about first-order moral claims] is that it is a means of sidestepping a number of thorny puzzles”, including metaphysical questions: for if the ontological claim that a thing has a certain property is not literally true, then any claims purporting to explain the thing’s having the property are plausibly also neither true nor false.
7 Frege, “Der Gedanke”, 75–6; see Frege, Die Grundlagen Der Arithmetik (§61) for the reading of “outer” in a “spatial sense”.
8 Frege, “Der Gedanke” (69) claims that “disputing about the truth” and there being “a science … on which many [can] work together” involves the grasping of thoughts.
9 See e.g., Skiba “Frege’s Unthinkable Thoughts” on Frege’s “principle … [that] every thought can be grasped”, (2) and Heck and May (“Frege’s Contribution,” 23) on why Frege thinks all thoughts are “possible contents of propositional attitudes”, and that “we must recognize in [cognitive] episodes something … that may be … affirmed by one person and denied by another”. Frege 1918–1919a introduces a notorious complication: thoughts that can only be grasped by one person. Perhaps someone who grasps such a thought still engages in the sort of thinking which admits of significant agreement and disagreement, even if for special reasons nobody else can agree or disagree; or perhaps Frege simply falls into inconsistency here along lines described by Skiba.
10 Burge, “Frege on Knowing the Third Realm”, 638.
11 A commitment to an ontology-to-cognition direction of explanation is such a common route to realism that it is sometimes built into (non-standard) definitions of realism itself, like that of Reck (153). On such non-standard definitions, views on which a third factor explains both ontology and cognition are non-realist, while by our standard definition, they are realist.
12 Frege, “Der Gedanke”, 69. On Frege’s reasons for this claim, see §1 of Ricketts, “Objectivity and Objecthood”.
13 Frege, “Der Gedanke”, 74 and Frege, “Logik”, 145; these passages are cited by Burge, “Frege on Knowing”, 639–64, and by Dummett to rule out certain idealist views (The Interpretation, 514).
14 Weiner, “Realism Bei Frege”, 369.
15 Dummett also does not think these remarks fully commit Frege to realism about thoughts, though for a very different reason. Dummett employs a non-standard account of what realism is, according to which “the antithesis between realism and idealism … [is] an opposition between two accounts of what … an understanding of our language consists in”, where “realism is … the acceptance of a truth-conditional theory of meaning resting on an unmodified two-valued semantic theory”. (Dummett, Frege Philosophy of Language, 684; Dummett, The Interpretation, 442–3.) Dummett sees that this apparently has “little to do with traditional disputes concerning realism”: for on this account, it is “highly dubious whether a classical phenomenalist can properly be described as having held an anti-realist view”; and “a Dedekindian who maintained that mathematical objects are free creations of the human mind … [could] be a species of realist” about them; and even “Berkeley … ends up in a position describable as a sophisticated realism about the material world”. (Dummett, “Realism and Anti-Realism”, 465–7; Dummett, “Realism,” 84; Dummett, The Interpretation, 462.) Since these views are idealist in the standard sense, realism in Dummett’s sense does not imply realism in the standard sense. Since our concern is whether Frege is a realist in the standard sense, then, we need not pursue Dummett’s case for Frege’s realism in his sense. (Dummett himself holds that conceiving of ‘realism’ in terms of dependence or explanation is confused, and that his account captures what is most interesting about traditional discussions using the term—though as Miller (“Realism”) reports, “few have been convinced” by either claim. See Miller, “The Significance” for discussion.)
16 Frege, “Die Verneinung”, 146, (“Man kann unter dem Sein eines Gedankens auch verstehen, dass der Gedanke als derselbe von verschiedenen Denkenden gefasst werden können”); Weiner, “Realism Bei Frege”, 369.
17 Something similar applies when, after denying that logic concerns “minds and contents of consciousness whose bearer is an individual person”, Frege says that “one might perhaps see [logic’s] task as the investigation of the mind”. (In this case though, even endorsing that characterization of logic’s task would not imply idealism about thoughts, since an ‘investigation of the mind’ might require describing things which are independent of the mind but to which it stands in important relations.)
18 One might wonder: how can Frege take this modal idealist view as an open possibility, having written §4 of Begriffschrift? For according to that passage, one who says that P is possible “either … abstains from judgement [of P] … or he says that the generalization of the negation [of P] is false … [making] a particular affirmative judgement”; and if so, to claim that a thought’s existence depends on the fact that it can be grasped is presumably to say that it depends on someone actually grasping it—which view Frege of course rejects. The answer, I think, is that Frege no longer holds that dismissive view of modal talk. For he never mentions it outside of that single passage in Begriffsschrift, and he himself makes significant philosophical use of modal talk that cannot be understood along those lines. (E.g.: two sentences express different senses when someone “could hold the one … to be true, the other false”; and the same sense when “anyone who recognizes the content of [one] as true must also immediately recognize the content of [the other] as true”. (Frege, “Über Sinn Und Bedeutung”, 32; Frege, “Kurze Übersicht”, 213.)
19 Frege, Die Grundlagen, §26
20 Resnik, “Frege as Idealist”, 351–2: Resnik also cites the similar §60 of Frege, Die Grundlagen, in which Frege says what he does not mean by “self-subsistence.”
21 Frege, “Über Begriff”, 200. Frege says ‘concept’ here, but his point applies to functions more broadly.
22 Sluga (“Frege on Meaning,” §5) points out that it is hard to see how someone who regards such expressions as senseless could allow for the possibility of a substantive theory of meaning.
23 See Weiner “Theory and Elucidation” for discussion of what he might be doing.
24 Weiner, Frege in Perspective, 214; Weiner, “Realism Bei Frege”, 364, 376. As we saw, Weiner also makes observations that can support an idealist reading. Her position could be put this way: if Frege holds any traditional metaphysical theory, his idealist-sounding passages indicate that it would be idealism; but since he ultimately holds such claims to be neither true nor false, he is a quietist instead. (This non-cognitivist route to quietism fits with the second example given above.)
25 Weiner, Frege in Perspective, 214. Though Ricketts finds his thinking “very much indebted to Joan Weiner”, (“Logic and Truth”, 128) he sees that she is not to be followed here: “The judgment of the existential generalization that there are Fregean thoughts has the objective status of all our judgments”. (140) Attributing three additional views to Frege may explain what Weiner has in mind: that if the claim that thoughts exist is literally true, then it must be part of some scientific system; that if it is part of any system, it must be part of the axiomatic logical system; and that if it is part of the axiomatic logical system, it must be expressible in Begriffsschrift as it stands. We will later see that the second assumption, at least, is false.
26 Weiner “Realism Bei Frege”, 375–6.
27 On these terms, see Weiner, “Theory and Elucidation” and Ricketts, “Objectivity and Objecthood”.
28 Some classic sources are Chapter 2 of Hardin, Color for Philosophers, Harman, The Nature of Morality, and Street, “A Darwinian Dilemma”. There is a very general point about correlations and explanatory claims here. Nearly every time we find smoke, we find fire: this regularity leads us to expect some kind of explanatory relation between the two. Affirming that fire explains smoke is interesting because it provides the explanatory relation we are looking for; denying it is interesting because it means that we have to keep looking.
29 Shafer-Landau, “Evolutionary Debunking”, 1.
30 Quotes in this paragraph from Frege, Die Grundlagen, v-vi and ix-x and Frege, Grundgesetze Der Arithmetik, xiv and xxiii.
31 Frege, Grundgesetze Der Arithmetik, xvii. This passage clarifies §14 of Frege, Die Grundlagen, whose rhetorical questions can give the impression that he commits to an answer: “Would not all fall into confusion, if one tried to deny one of these? Would thinking then still be possible?” (This impression is strengthened by the standard English translation (Frege, The Foundations of Arithmetic) which puts Frege’s questions—“Stürzt nicht alles in Verwirrung, wenn man einen von diesen leugnen wollte? Wäre dann noch Denken möglich?”—into declarative form: “We have only to try denying any one of them, and complete confusion ensues. Even to think at all seems no longer possible”.)
32 Frege, “Logik”, 157; Frege, “Aufzeichnungen”, 273.
33 Also, before the passage on thoughts, Frege defends his “attaching an unusual sense to the word ‘thought’,” on the grounds that “it is allowed in logic to coin technical terms”. (147–8) He evidently thinks that his technical term ‘thought,’ is used in logic, even though no corresponding term figures in the Begriffsschrift expression of the axiomatic science.
34 Dummett (“Thought and Perception”, 286) accordingly criticizes Frege for “relegat[ing] to psychology” this question. While Conant (“The Search”) and others have sought to uncover Fregean commitments to answers to these psychological questions, what matters here is that finding such commitments would show only that Frege failed to adhere to his expressed intention not to opine.
35 Frege, Grundgesetze Der Arithmetik, xix-xx.
36 Quotes in the next paragraph are from Frege, “Logik”, 4; Frege, “Logik”, 139; and Frege, Grundgesetze Der Arithmetik, xv. While the essentials of the interpretation in this paragraph are increasingly widely accepted, they leave open several controversial issues—including what exactly these prescriptions are, and what it means for them to hold “for all regions of thinking”. For some variations, details and defence, see especially MacFarlane, “Frege, Kant, and the Logic in Logicism”, Taschek, “Truth, Assertion, and the Horizontal”, Steinberger, “Frege and Carnap”, and Hutchinson, “Frege on the Generality of Logical Laws”.
37 See, e.g. Frege, “Logik”, 144 and surrounding discussion; see also Ricketts “Objectivity and Objecthood” on connections between thoughts and normativity.
38 The present issue is the core of the so-called ‘logical question’ that occupied nineteenth-century German philosophy: how to characterize logic without winding up with the extravagant “metaphysical understanding of logic developed by Hegel”. (See, e.g. Peckhaus, “Language and Logic”, 4.) Frege’s remarks place him in the tradition that characterizes logic in terms of its normativity and ultimately invokes a distinction like this one among kinds of intelligibility. (See, e.g. the distinction between goal-directed necessity and other kinds in the Introduction to Lotze, Logik and the distinction drawn by Kroner “Über Logische Und Ästhetische Allgemeingültigkeit” in response to the argument of Husserl’s discussed in footnote 42 below. See also Hutchinson “Frege’s Critical Arguments” for a discussion of the role given by Neo-Kantians to this sort of goal-directed necessity in justifying logical axioms.) Of course, a full account of this conception of logic must characterize this distinction more exactly than can be done here. I will add, however, that something can plausibly be made intelligible in this sense partly by its consequences: things which it explains or causes. If so, this conception of logic does not rule out ‘third-factor’ metaphysical views that give values or norms—including logical ones—explanatory primacy over cognition and ontology: views which, with Lotze’s Metaphysics, find “in that which should be the ground of that which is” (Volume III, Conclusion).
39 Frege, Grundgesetze Der Arithmetik, xix; Frege, “Der Gedanke”, 71, 69.
40 The claim that “it is impossible for any one to believe the same thing to be and not to be” is key to the elenchtic defence of the principle of non-contradiction in 4 of Aristotle [Meta].
41 See, for example, the defence of modus ponens in Hale “Basic Logical Knowledge”, on the grounds that its “reliability must … be assumed in any demonstration we can give of the … unsoundness … of any … rules”, since it is employed in any such demonstration (297). See also Kripke’s discussion in Berger, “Kripke on the Incoherency of Adopting”.
42 Attention to what Stern (“Transcendental Arguments”, §2) calls the “dialectical situation” between challenger—often a “skeptic”—and defender has characterized discussions of such arguments in the wake of famous criticisms from, e.g. Stroud, “Transcendental Arguments”.
43 Husserl (Logische Untersuchungen, §§19–20) influentially argues that we cannot “base a sharp [epistemological] separation of [logic and psychology] on … the normative character of logic”, because psychological claims must be relevant to whatever normative matters logic is concerned with: internal challenges must be possible. But Husserl’s key claim is that psychological truths can help to establish necessary conditions on, for example, doing what a prescription says to do; and some such necessary conditions are irrelevant to such normative matters as making the prescriptions intelligible. Kroner, “Über Logische Und Ästhetische Allgemeingültigkeit” makes a version of this response to Husserl.
44 Compare the attempt of Dworkin, “Objectivity and Truth” to show that “non-moral discoveries cannot undermine or structurally change morality without morality’s help”. (127–8)
45 McGuire, who notes the similarity with Frege, argues that while various theses have been discussed under the heading of the “autonomy” of ethics, “the real motivation for [these theses] arises from the importance of an epistemological thesis”: that “non-ethical propositions”—especially those of metaphysics and psychology—“are irrelevant to the justification of … ethical propositions”. (“The Autonomy of Ethics”, 431–2, 437.) By contrast, the closest lively logical discussions concern “exceptionalism,” and as Hjortland notes, “the central exceptionalist claim is that the justification of logical theories is a priori” (“Anti-Exceptionalism”). Even if exceptionalism is true, then, presumably an a priori metaphysics or psychology would remain potentially relevant to the justification of logical claims. (On exceptionalism, see, e.g. Field, “The a Prioricity of Logic”, Field, “Apriority as an Evaluative Notion”, Maddy, “A Naturalistic Look”, and Williamson, The Philosophy of Philosophy.)
46 See McPherson (“Metaethics & the Autonomy of Morality,” 1–3) for an introduction to the issues along the lines given here.
47 See, for example, Street, “A Darwinian Dilemma”.