2 Davidson, Donald ‘Truth and Meaning,’ reprinted in Inquiries into Truth and Interpretation (hereinafter IT & I) (Oxford: Clarendon Press 1984), 25Google Scholar

3 Quine's is a purer behaviorism than Davidson's. For Quine, the manifestations of assent are taken purely behavioristically; for Davidson they are direct evidence of intentional states of ‘holding true.’ The distinction is unimportant for our purposes; the content of what is held true is still fixed as a posit in theory building.

4 Davidson, ‘Radical Interpretation,’ reprinted in IT&I, 137Google Scholar

5 Davidson, ‘Belief and the Basis of Meaning,’ reprinted in IT&I, 152Google Scholar

6 I do not mean to imply here that the principle of charity is grounded in practical interpretive need. In discussion, Davidson has recently claimed an independent and direct warrant underlying the principle. That warrant, as I understand it, is that thought is constitutively — conceptually — rational. It is not merely that we could not come to a satisfactory understanding of an irrational cognitive agent, but that to be a cognitive agent is to be in large measure rational.

7 Davidson, ‘Radical Interpretation,’ 137Google Scholar

8 Davidson, ‘Belief and the Basis of Meaning,’ 153Google Scholar

9 Gary Kemp emphasizes this point in a recent paper. He argues, as do I, that much of the critical response to Davidson's arguments fails adequately to keep Davidson's account of belief in mind. Kemp feels that the important matter for dispute is whether a behavioristic account of belief is preferable to introspectivism, and seems to think that, given the behavioristic account, Davidson's conclusions go through unproblematically. I find the behavioristic account congenial for many reasons, but think that Davidson's anti-skeptical conclusions require an independent lemma that begs the question against the external world skeptic. See Kemp, Gary ‘Davidson, Quine, and Our Knowledge of the External World,’ Pacific Philosophical Quarterly 73 (1992) 44–62.CrossRefGoogle Scholar

10 Davidson, ‘Thought and Talk,’ reprinted in IT&I, 168Google Scholar. This passage shows how easily one might take Davidson's argument to be verificationist. His claim about the 'identification’ of beliefs can be taken as an epistemic one about how we can and must individuate beliefs in an interpretive process, or as an assertion that, as a conceptual matter, belief and meaning are holistic. Taking the warrant for holism thesis to be epistemic, Jerry Fodor and Ernest LePore have argued that there is no reason to suppose that interpretive practice is in fact holistic (Holism: A Shopper's Guide [Oxford: Blackwell 1991] ch. 3). This leads them to conclude that Davidson gives us no reason at all to suppose a holism of the mental. If, however, Davidson's argument is a conceptual rather than an epistemic one, Fodor and LePore's claims about interpretation, true or not, do not warrant their conclusion.

11 My mention of ‘conceptual beliefs’ is not to be taken as a retreat to analyticity. Davidson accepts Quine's thesis that the analytic/ synthetic distinction is untenable. This does not mean that there are not some beliefs which are more central than others to the possession of a particular concept. If a speaker does not hold enough of these, there is no basis for supposing that she has the concept in question. This certainly does not imply that these central beliefs are true in virtue of meaning as opposed to the way the world is. And this point is reflected in the fact that, if a given English speaker fails to assent to claims such as ‘that is a dagger’ when an obvious dagger is pointed out to her, it is again appropriate to doubt that the speaker possesses the concept, notwithstanding that such claims are surely not analytic.

12 In this connection it is worth noting a parallel problem, the familiar ‘isolation objection,’ faced by coherence theories of knowledge when they are married to internalist, Cartesian views of mental content. Even supposing, implausibly, that I can grasp and assess the inferential and other relations among my introspectible beliefs, the coherence of my beliefs with one another can give me at best subjective justification for holding any one of them. Unless I have reason to suppose my beliefs caught up with the external world in the right way, causally, for example (but see note 54, infra), I am without a reason to suppose that my subjectively justified beliefs are true of the world, or, indeed, that there is any external, objective world at all. Mental holism, involving as it does the view that our beliefs must cohere, may remove the burden on Cartesians to show that their introspectible beliefs are subjectively justified, but it still does not touch the central problem for Cartesian epistemologists of showing that such justification gives us knowledge of an objective world. Unless connections to the world are built in to the very possibility of belief, holism (and coherentism) will fail to dislodge us from the familiar oscillation between skepticism and idealism that has characterized epistemology since Descartes. Davidson's theory of RI, which invokes both holism and a necessary connection between world and belief, tries to do just this.

13 Davidson, ‘Radical Interpretation,’ 136, n. 16Google Scholar. The use of ‘reference to’ in this quotation is potentially misleading. It invokes no technical sense of ‘reference.' 'Appeal to’ captures the sense of Davidson's remark without opening up questions as to the reference relation.

14 Davidson, ‘A Coherence Theory of Truth and Knowledge,’ (hereinafter CTTK) in LePore, ed., Truth and Interpretation (Oxford: Blackwell 1986), 317Google Scholar

15 Davidson, CTTK, 314.Google Scholar I italicize ‘depends partly on’ in order to emphasize how much weaker Davidson's claim is than the claim that the content of occasion sentences is ‘determined by’ their causes. The significance of this distinction will become clear in section IV, where I examine the charge that Davidson has violated his holism and coherentism in gearing meaning to causes in methodologically basic cases.

16 Davidson, CTTK, 314Google Scholar

17 Davidson, CTTK, 315Google Scholar

18 Davidson, CTTK, 317Google Scholar. Davidson need not be taken to invoke some strong realist sense of ‘causes’ here; Humean causality will do, for it is a correlation between salient environmental change and assent that the interpreter exploits, not some necessary or even strictly nomic connection.

19 Williams, Michael ‘Skepticism and Charity,’ Ratio 1 (1988), 188CrossRefGoogle Scholar

20 Davidson, CTTK, 317Google Scholar

21 Ibid.

22 ‘The Method of Truth in Metaphysics,’ reprinted in IT&I, 201

23 Evnine, Simon Donald Davidson: An Introduction (Stanford: Stanford University Press 1991), 142Google Scholar

24 Vermazen, Bruce ‘The Intelligibility of Massive Error,’ Philosophical Quarterly 33 (1983) 69–74CrossRefGoogle Scholar

25 Ludwig, Kirk makes this point in ‘Skepticism and Interpretation,’ Philosophy and Phenomenological Research 52 (1992) 317-39.CrossRefGoogle Scholar

26 Vermazen, ‘The Intelligibility of Massive Error,’ 74Google Scholar

27 Davidson, ‘Reality Without Reference,’ reprinted in IT&I, 222Google Scholar

28 Williams, ‘Skepticism and Charity,’ 185Google Scholar

29 See the argument in ‘Truth and Meaning,’ 19.

30 Davidson, ‘Reality Without Reference,’ 225Google Scholar

31 Moreover, even if we could discern the various parts of speech of expressions comprising occasion sentences, we would still be unable to fix reference. Reference is inscrutable. There will always be multiple possible truth theories for a given language, and referential apparatus will vary from theory to theory, though meaning will not. Even within a given theory, reference is fixed only in the sense that the way we talk about reference is fixed ('Reality Without Reference,’ 239).

32 We should note that Williams's reading of Davidson is expressly designed to make Davidson a realist, to at least a minimal extent. Since ‘even the most minimal realist wants more than “truth in a model'” Williams thinks that Davidson must constrain interpretation in a way that ‘connects reference with something “external,” with how things objectively are’ ('Skepticism and Charity,’ 185). If Davidson can call himself a realist, it is because we are constrained to treat occasion sentences as being about objective features of the world; but whatever realism Davidson may claim cannot, in light of his holism and the inscrutability thesis, receive support from a theory of direct reference.

33 Williams notes first that the description of the omniscient interpreter scenario may not even be consistent. A person who is omniscient about the causes of all of another person's sentences, occasion and standing, would not, according to Williams, need to employ the same methodology of interpretation that constrains fallible interpreters to optimize agreement. Williams passes over this point, and we may do so as well, for as I argue in the text, Davidson's omniscient interpreter need not be supposed to have such knowledge.

34 In Williams's words, ‘the internal factor in meaning, inferential relations between beliefs and other beliefs, has dropped out of sight’ (Williams, ‘Skepticism and Charity,’ 190).Google Scholar

35 Ramberg, Donald Davidson's Philosophy of Language: An Introduction (Oxford: Blackwell 1989) 76Google Scholar. The example in the quoted passage is not of an occasion sentence, to be sure, but Ramberg gives no ground for supposing that anything less than knowledge is required for linguistic and interpretive competence in such cases. Hence the challenge to holism and coherentism that stems from causal identification in these cases is implicit in Ramberg's example.

36 This argument appears in Vennazen, ‘The Intelligibility of Massive Error.’

37 Ludwig, ‘Skepticism and Interpretation,’ and Dalmiya, Vrinda ‘Coherence, Truth, and the “Omniscient Interpreter,“’ Philosophical Quarterly 40 (1990) contain arguments of this sort.CrossRefGoogle Scholar

38 This possibility is flatly inconsistent with the content considerations discussed above, according to which the possession of false beliefs is intelligible only against the background of the possession of many true beliefs. So omni-ignorance, conceived as possession of mostly false beliefs, rather than none at all, is incoherent.

39 Klein, Peter D. offers versions of both of the simple ‘non-sequitur’ objections discussed in the text in his ‘Radical Interpretation and Global Skepticism,’ in LePore, Truth and Interpretation (Oxford: Blackwell 1986) 369-86.Google Scholar

40 If this is right, then it is not enough to ensure that we are in contact with the contents of our beliefs to say, as Davidson has, that ‘whatever is responsible for the content of our thoughts, whether known or not, is also responsible for the content of the thought that we have the thought’ ('Reply to Burge,’ typescript, 1). This ensures that our second order beliefs will accurately capture the contents of the first order beliefs they are about, but not that we have the needed kind of access to the contents of either.

41 Boghossian, Paul A. ‘Externalism and Inference,’ in Villanueva, ed., Philosophical Issues 2: Rationality in Epistemology (Atascadero, CA: Ridgeview 1992), 11–28Google Scholar

42 Dummett, Michael ‘Frege's Distinction Between Sense and Reference,’ reprinted in Truth and Other Enigmas (Cambridge, MA: Harvard University Press 1978), 116-44Google Scholar

43 We should avoid this locution. It encourages the view that thought is some medium through which a believer comes to terms with what the thoughts are about. The demand for ‘transparency’ just sounds like the demand that the medium be unproblematic. Moreover, ‘transparency’ signals a specifically visual kind of unproblematic access, and its use reinforces the imagistic view of the contents of mind rightly decried by Richard Rorty in his Philosophy and the Mirror of Nature. That thought is a sort of representational medium or intermediary, particularly one analogous to a visual medium, is as good a synopsis as any I know of the central idea which has driven what John McDowell calls ‘the dreary history of epistemology,’ and we should avoid terminology that works against our overcoming it.

44 While this may express a necessary condition for grasping meaning, it may well not be sufficient. On some conceptions (though not, I think, on Davidson's), I may know that two tokens of the same syntactic type share meaning, without knowing what that meaning is. I owe this point to Rob Cummins.

45 See Stalnaker's, Robert ‘On What's in the Head,’ in Tomberline, ed., Philosophical Perspectives, 3: Philosophy of Mind and Action (Atascadero, CA: Ridgeview 1989) 283–316Google Scholar, for good discussions of the varieties and consequences of semantic externalism.

46 Among the most significant expressions of this worry are found in Breukner, Anthony ‘Brains in Vats,’ journal of Philosophy 83 (1986) 148-67CrossRefGoogle Scholar; Breukner, Anthony ‘Charity and Skepticism,’ Pacific Philosophical Quarterly 67 (1986), 264-8CrossRefGoogle Scholar; Craig, Edward ‘Davidson and the Skeptic: The Thumbnail Version,’ Analysis 50 (1990) 213-14CrossRefGoogle Scholar; Nagel, Thomas The View from Nowhere (New York: Oxford University Press 1986), 71-3Google Scholar; Williams, Michael ‘Skepticism and Charity,’ 190Google Scholar; Boghossian's ‘Externalism and Inference.’

47 For Davidson's, most complete discussions of the issue, see ‘First Person Authority,’ Dialectica 38 (1984) 101-11CrossRefGoogle Scholar; ‘Knowing One's Own Mind,’ APA Proceedings (1986) 441-58; ‘What is present to the Mind?’ in Villanueva, ed., Philosophical Issues 1: Consciousness (Atascadero, CA: Ridgeview 1991) 197–213.Google Scholar

48 Any manner of expressing length will do, so long as it is commensurable with the others, i.e., so long as linear transformation among the scales is possible. Measurement does require numerical structures, and this might compel us to objectify number; but it does not require us to objectify numbers of inches, meters and cubits, though these are the entities with which we keep track of length.

49 Davidson, ‘What is Present to the Mind,’ 205-6Google Scholar

50 Davidson, ‘What is Present to the Mind,’ 207Google Scholar

51 Davidson, ‘Knowing One's Own Mind,’ 456Google Scholar

52 Davidson, ‘First Person Authority,’ 111Google Scholar

53 Davidson, ‘First Person Authority,’ 110Google Scholar

54 This admittedly glosses some difficulties presented by sentences containing indexicals, as well as tensed verbs, if these latter are analyzed in terms of temporal demonstratives. But this does not compromise the essential point that marks the asymmetry of first and third person authority— that both the left and right sides of T-sentences ascribed by the speaker to her own utterances are expressed in her own language.

55 Akeel Bilgrami has objected to Davidson's account of first person authority on the grounds that the fact that interpreters must treat interpretees as knowing their own minds does not explain how they have this knowledge, and that this needs explaining, especially in the light of externalism's claim about what determines mental content. Bilgrami, ‘Thought and its Objects,’ in Villanueva, ed., Philosophical Issues, 1: Consciousness (Atascadero, CA: Ridgeview 1991), 215-32Google Scholar. Bilgrami's criticism is one example of a broader kind of objection aimed at Davidson's radical division between the internal epistemic realm and the external causal realm which is responsible for contents of the internal states. Faced with arguments that make externalism about content plausible, but unable to see what is external as epistemically significant, Davidson has simply denied that any epistemic link to contents is required. Richard Rorty has been most eager to applaud and encourage this move on Davidson's part, arguing that the burden of coherentism itself is that there can be no epistemic significance to the causal hookup between world and agent. In his Mind and World (Cambridge, MA: Harvard University Press 1994), John McDowell argues that this divorce, far from clearing things up, makes an utter mystery out of the phenomenon of empirical thought. McDowell focuses on the arational nature of the causal interactions which are supposed to fix content, urging that thought which is not rationally constrained by the world cannot properly be thought of as being about the world at all. Both Bilgrami's and McDowell's complaints are, in my view, substantial. But it is not my purpose to defend Davidson here against all comers, but only against challenges to his anti-skeptical views which misconstrue them.

56 Boghossian, ‘Externalism and Inference,’ 18Google Scholar

57 Only on the view that natural kinds are particularly important as external determinants of content will the example of mistaking water for twater seem more troubling than garden variety examples of mistaken identity. But, their historical significance in this context notwithstanding, such views are not compulsory to externalist semantics.

58 Malpas, J.E. Donald Davidson and the Mirror of Meaning (Cambridge, UK: Cambridge University Press 1992), 222 and n. 83Google Scholar. Interpretation of such envatted brains would be, given Davidson's holism, ‘highly implausible on a priori grounds alone.’ Malpas goes on to argue that Davidson's semantic externalism can seem problematic only if we fail to keep his holism in mind. ‘Davidsonian externalism must be read … in conjunction with Davidsonian holism ...’ (223). This is, in fact, just another way of putting the point I have been urging, that all of the constraints on RI, including charity and the directive to identify the meaning of occasion sentences with the causes of assent to them, are absolutely central to Davidson's semantic and epistemological views, and that taking any one of them apart from the others can only result in confused and irrelevant objections.

59 Foley, Richard and Fumerton, Richard ‘Davidson's Theism?’ Philosophical Studies 48 (1985) 83-9CrossRefGoogle Scholar

60 If the environings that correlate with assent to occasion sentences in the languages of that world were very different from their counterparts in ours, then the natives of that world would believe very differently than we do. Davidson's arguments in 'On the Very Idea of a Conceptual Scheme’ (reprinted in IT&I, 183-98) could be read to imply that the possible extent of this sort of difference is in principle very limited.

61 Victoria McGeer makes essentially this point in a paper entitled ‘Davidson's Omniscient Interpreter: An Argument that Could Not Be,’ read at the 1993 Pacific Division Meetings of the APA.

62 Williams, Michael ‘Skepticism and Charity,’ 190Google Scholar

63 Klein, ‘Radical Interpretation and Global Skepticism,’ 386Google Scholar

64 We have seen here several reasons to prefer the Davidsonian view, which integrates content externalism with charity-based interpretivism, to more purely external views of content. Charity ensures that one's beliefs will be largely coherent, which pure externalisms do not. It thus helps to avoid skeptical challenges to our knowledge based on questions of justification. It similarly does justice to the reasonable view that mental content is anatomic, in the sense that it is impossible to be in one mental state without being in (indefinitely) many others. Moreover, while pure externalisms may or may not ensure the truth of most of our beliefs, they provide no account of our knowledge of our own minds, and therefore are open to skeptical challenges based not on the falsity of our beliefs but on our grasp of their contents. And they cannot easily account for the asymmetry implicit in first person authority. As we saw in section VI, however, both first person authority and a general grasp of content fall out of Davidson's interpretivist framework.