Truth is obviously a central notion for philosophers. It has seemed natural, nay obligatory, to try to provide a theory or at least a definition of truth. That's what philosophers do, isn't it? Unless you think that truth doesn't exist – and how could anyone think that? – then the word refers to something, and the task of metaphysics (ontology) is to say what there is, and that includes truth. That's the thought, anyway.
So we have the "correspondence theory of truth," the "coherence theory," etc. The problem with these accounts, generally speaking, is that they take us one step forward and two steps back. So truth is "correspondence" between true statement and fact. What does that mean? What sort of mysterious metaphysical relation could this "correspondence" be? And how is it supposed to explain the concept of truth?
On Davidson's view, the concept of truth is "admirably clear" compared with anything we might use to explain it (or, as I like to say, perhaps to show off my Latin, theories or definitions of truth are necessarily per oscurius). Consequently, he is sometimes described as having "taken truth as primitive." I don't think this is quite right. It's true that one of the key moves in his early work on semantics is the switch from a) providing a Tarski-style theory of truth for a language in terms of T-sentences, to b) "holding truth constant and solving for meaning," giving the meaning of each of the quoted sentences on the left-hand side of those T-sentences in terms of their truth-conditions ("grass is green" is true iff grass is green). The result, then, is a theory of meaning, one which has the virtue (for Quineans such as the early Davidson) of being extensional (and thus in principle empirically verifiable).
But "holding truth constant and solving for meaning" isn't the same as "taking truth as primitive," even given Davidson's reluctance to provide a theory of truth. It's not that we can't say anything about this ineffable concept; it's that whatever we do say would not be intelligible to anyone who did not already understand the concept of truth (enough, that is, to know what a "definition" or "theory" or even "statement" is). It is in this sense that truth is "central."
Supposedly following Davidson (but not exactly claiming agreement – see "Pragmatism, Davidson, and Truth"), Rorty sometimes talks as if he endorses a "deflationary theory" of truth, which consists simply in what we like to call the "disquotational platitude": "p" is true iff p. (I really should use Quinean square brackets here, but you know what I mean.) This is supposed to put all that tendentious metaphysics behind us; but while it's certainly true that (say) the English sentence "grass is green" is true just because grass is indeed green, that doesn't mean the "deflationary theory of truth" is of any use to us. Rather than spurning the very idea of providing a "theory" for a concept that cannot support one, the "deflationary theory" makes it look as if we have indeed provided one, without in fact telling us anything. (See also, but without taking what he says there to be that important, Crispin Wright's argument, in Truth and Objectivity, that "deflationism reinflates.")
So what about the "identity theory" of truth? James points us to Stewart Candish's article at SEP, which he quotes as saying at one point that “One such pressure [i.e. in favor of this theory] is the wish that there should be no gap between mind and world: that when we think truly, we think what is the case.” And of course McDowell says things very much like this all the time (James has a typical quote from M & W), and for that very reason: that we not be misled into thinking that there is a metaphysical gap between mind and world, such that (say) we should attempt to construct over it a philosophical bridge. In McDowell's thinking (though he does not put it this way), that would be to try to build a bridge between the two peaks of Kilimanjaro, a (philosophically) suicidal undertaking.
As commenter Tom rightly points out, though, while McDowell is indeed willing to accept the "identity theory" in these terms (Tom points us to McDowell's replies in Willaschek, ed., Reason and Nature, even giving a link to a pdf), his position (as Tom puts it) "seems to be another 'quietistic' one, that the identity theory is a truism and not really a theory in the philosophical sense at all." Why bother affirming it then? McDowell now (p. 94): "Just that keeping it in view helps to prevent unprofitable philosophical anxieties from arising."
So far, so good – the "quietist" refusal even to take on the theoretical requirements seemingly foisted on us simply by employing the Cartesian picture is an important part of McDowell's conception of his Wittgensteinian heritage. But still, when such a "truism" is something of which we must be reminded, it seems that we might indeed take a close look (closer, paradoxically, than one at that of "truisms" usually rewards) at its content (or, better, its (overlooked or misunderstood) consequences) in our context. That is, we might look at it, for the moment, as if it were a "doctrine" or "theory" after all. (It is McDowell himself who reminds Rorty that, and I forget the exact quote, traditional philosophy has resources Rorty overlooks – that is, for combating the confusions of traditional philosophy itself.)
But it is not as the "identity theory of truth" that I think we should look at it. Here's what the SEP article says, in an introductory section, about this theory:
The simplest and most general statement of the identity theory of truth is that when a truth-bearer (e.g., a proposition) is true, there is a truth-maker (e.g., a fact) with which it is identical and the truth of the former consists in its identity with the latter. The theory is best understood as a reaction to the correspondence theory, according to which the relation of truth-bearer to truth-maker is correspondence. A correspondence theory is vulnerable to the nagging suspicion that if the best we can do is make statements that merely correspond to the truth, then we inevitably fail to capture the reality they are about and thus fall short of the truth we aim at. An identity theory is designed to overcome this suspicion.This doesn't sound right at all to me. I'm not denying that people do indeed talk this way, as Candish claims they do. But it won't work for our purposes. For one thing, that's not the problem with the "correspondence" theory, or at least that's not the best way to put it. In fact I'm surprised that this slipped past the editors. What's that "nagging suspicion" again? The correspondence theory is an account of what it is in which the truth of true statements consists (assuming, of course, that it is indeed of statements that truth is to be predicated). When we speak truly, then, what we say "corresponds" to how things are; and historically it has been this account which has (e.g. as laid out in Aristotle: "to say of what is not, that it is not, or of what is, that it is, is true"), been regarded as virtually truistic. So when we do "the best we can," what happens is not that we "make statements that merely correspond to the truth"; what happens is that we speak the truth, which "corresponds" not to the truth, which we have, but to how things are.
So what's the problem? I presented it above, as Davidson does, as being that the idea of "correspondence" is impossible to cash out intelligibly (i.e., as an informative theory of truth). Candish presents it here as being an essentially skeptical worry (so that he should say that the worry is that even when we speak the truth, all we do is make statements that "merely correspond" to how things are). That's not exactly wrong, but it deflects attention from the real problem (i.e. with giving a "theory of truth" in the first place) to what is at least a slightly different one. Rorty does this too: he equates the idea of "correspondence" with the dualistic conception it is (usually) meant to illuminate (or at least manifest). That is, he essentially allows the correspondence theorist intelligibly to advocate a theory Rorty believes false or problematic: that when we speak truly, we make statements that describe a reality which is ontologically detached from, or transcends, anything to which it seems that we can be sure of having epistemic access. In other words, he equates it with metaphysical realism.
This makes it look like the skeptical difficulties into which that position "inevitably falls," as Kant puts it, are problems concerning the notion of truth, and that we need therefore to deny or redefine that notion, or downplay its importance. But it is important, and (as Davidson protests) even central – and it certainly should not be redefined as "that which our contemporaries let us get away with"! (How can someone who claims such inspiration from Davidson – even to the point of defining pragmatism in Davidsonian terms, as trying to do away with the scheme-content dualism – say such a thing??) [Actually I have an answer to this one...]
They are instead, these skeptical difficulties, due of course not to the very idea of "correspondence" – which in its truistic form says pretty much what we said before: that when we think truly, we think what is the case – but to its characteristically Cartesian metaphysical perversion. (So to rescue it, all we need do is, as Wittgenstein would say, to retrieve it "from its metaphysical to its everyday use.") At root, then, the problem here is the Cartesian picture itself, with its metaphysical dualism of subject and object. One aspect of our attack will then be, if perhaps not straightforwardly described as a metaphysical theory of our own, at least something which takes place on a characteristically metaphysical battlefield (as in Kant and Hegel; which partly explains McDowell's interest in these thinkers).
It is with this in mind – that our concern is with multiple manifestations of the same fundamental Cartesian confusion – that we should handle the concept of truth. That is, while it is true that one manifestation of that disease is a metaphilosophical compulsion to lay down specific requirements [see PI § 107 and environs, discussed below] or constructive tasks for our philosophical theories to fulfill, another head of that same hydra is the realistic/skeptical metaphysics/epistemology that complements that compulsion at the ground-philosophical level. In other words, I think we can kill two birds with one stone; or, better, show, by killing them with simply the one stone differently construed, that there was just the one Übervogel after all. Going back to our above context, this means seeing what does the work of what might otherwise seem to be (just another) metaphysical theory – a non-dualistic one – as at the same time the straightforward consequences of a mere truism, newly recognized as such. If the "identity theory" of truth isn't even seen as a truistic non-theory, as it seems Candish would have it, then we aren't left with either one of the two things I want us to bring together. It asks us to put just the wrong spin on the dictum "when we think truly, we think what is the case" – however true that dictum may be.
Okay, so let's get back to it. What, then, are the supposedly overlooked, momentous-but-not-necessarily-theoretical consequences of the "truism" that when we think truly, what we think is the case?
While he does resist the call for a "theory of truth," Davidson, unlike McDowell, does not put the point in the Wittgensteinian terms of the salutary effect of reminding ourselves of forgotten truisms. Instead, as I began to discuss above, he trades that purported obligation for his preferred project, a theory of meaning. For Davidson, as we shall see, truth is the transparent conceptual link between what we mean and how things are. We might say, then, that for Davidson truth is a semantic notion, not a metaphysical one. Indeed, a key article of Tarski's to which Davidson refers is called "The Semantic Conception of Truth." But (again) what Davidson says is best considered not w/r/t the form of its treatment of truth, but instead the content of that of meaning. It is when we understand the latter that we will be able to see the former in its properly truistic light (as in McDowell).
Davidson can naturally be read as a semantic externalist. But his externalism is not the metaphysically realistic kind, like that of Kripke et al. As early as "The Very Idea of a Conceptual Scheme," Davidson says (let me quote the stirring conclusion):
In giving up dependence on the concept of an uninterpreted reality, something outside all schemes and science, we do not relinquish the notion of objective truth—quite the contrary. Given the dogma of a dualism of scheme and reality, we get conceptual relativity, and truth relative to a scheme. Without the dogma, this kind of relativity goes by the board. Of course truth of sentences remains relative to language, but that is as objective as can be. In giving up the dualism of scheme and world, we do not give up the world, but re-establish unmediated touch with the familiar objects whose antics make our sentences and opinions true or false. [Inquiries, p. 198]Compare his criticism of Quine in "Meaning, Truth, and Evidence," in which Davidson makes a corresponding move. Here he rejects Quine's explanation of meaning and content in terms of the "firings of sensory nerves" (Truth, Language, and History, p. 47.) Instead of this "proximal" theory, Davidson substitutes a "distal" theory, where the relevant content-bestowing cause of our belief and/or utterance is (say) not the sensations which are (in fact, when they are) caused by a rabbit, but instead the rabbit itself. Just as the content of our thought is not so far removed from us as a potentially, or indeed inevitably, inaccessible world-in-itself, it is also not so near as the excitations of our sensory surfaces. As Davidson points out, "proximal theories, no matter how decked out, are Cartesian in spirit and consequence" (p. 58). The problem with seeing the senses as "epistemic intermediaries," he says, um, somewhere else, is that "we cannot swear them to truthfulness," and skepticism is again the unexpected result.
We must pass over the best part of Davidson's (later) account, which he explains in the often misunderstood terms of a semantic/epistemic "triangulation" among inquirer, interpreter, and world. Let's cut directly to the bottom line in this context. Once the idea of explaining meaning in terms of truth-conditions, and semantic externalism generally, is rescued from the imputation of a perverse Kripkean realism, it becomes clear how the semantic content of what we say can indeed be provided by the "external" world. At the same time, the world to which we answer (semantically and epistemically) is, in McDowell's terms now, nothing more independent of us, in the relevant sense, than "the world as it figures in our world view," a locution meant to fit with its Kantian/Hegelian counterpart in Mind and World, i.e. the "unboundedness of the conceptual realm," which can now (if not before) be seen as having no more idealistic consequences than the former did skeptical ones.
In other words, again, the thought is perfectly symmetrical. Let us not be distracted by the qualifications I have added; they merely reflect the holistic and contextual constraints on interpretation and inquiry revealed by the triangulatory account.
What X means (in my mouth at a certain time) is how things are when X (so construed in that context) is true.
By the same token, again, turning the thought around: just as what I mean cannot be conceptually detached from how things are, how things are cannot "transcend" my meanings. How things are can, of course, "transcend," in a perfectly ordinary sense, my knowledge; but that's different. All that means is that sometimes I speak or believe falsely, and sometimes I don't know how things are; and sometimes I simply don't know what to say. But even when I speak falsely, what I mean is how things would be if I had instead spoken truly. There's no reason to think that what I don't know must be thought of in this metaphysically transcendent way – even if there are things I can never know. That impossible sense of "transcendence" is a gratuitous Cartesian addition.
So while there is a certain sort of unobjectionable realism here, it cannot lead to the sort of skepticism we feared, due to the lack of the necessary metaphysical "gap." (Ignorance, even necessary ignorance, is not a metaphysically loaded state of affairs.) McDowell himself calls his view "naturalized platonism" rather than "domesticated realism," perhaps because of the plethora of similarly modified "realisms" of the past – "empirical" realism, "internal" realism, "external" realism, "scientific" realism, "pragmatic" realism, "common sense" realism – none of which have panned out, and some of which simply repeat the errors of metaphysical realism in a new register. And in fact this lines up with my own usage, as I tend to use the term "realism" only for the bad kind, and for the same reason. (Still, "platonism," ugh; and "naturalized," no less: double ugh.)
And if there is no gap, then there is no need for philosophy to construct theoretical bridges over them. Let me conclude by once again recommending the attempt to bring about in oneself a "Magic Eye" sort of shift in perception here. When Davidson is criticizing Quine, there is no question of (what would in that context seem to be) a limply quietistic retreat to the mere reaffirmation of non-theoretical platitudes. Davidson thinks Quine is wrong and he is right, about, well, meaning, truth, and evidence. He refers, after all, to the combating positions as proximal and distal "theories," and gives arguments (convincing ones, even) for the truth of the latter. But just as in "Very Idea," Davidson's ultimate opponent is the metaphysically dualistic Cartesian picture – the very one which, in the metaphilosophical context, it can be effective to combat by resisting the characteristically Cartesian temptations to "constructive" philosophy against which (McDowell's) Wittgenstein warns.
So seeing (a properly "theoretical" reason) why there is no metaphysical gap can be the same thing, in the end, as seeing a reason to reject the seeming requirement to cross a metaphilosophical gap: i.e., that between naïve pre-philosophical ("truistic") intuition over here and solidly grounded philosophical results over there. Even better: it is the very rejection of scheme and content which Davidson's theory of meaning and interpretation demands which allows us to see it, qua theoretical, as not inconsistent with (as if we had to choose a single "scheme"), but simply another way of grasping, the anti-dualistic confluence of mind and world which licenses rejection of Cartesian philosophical requirements. It is no accident that a favorite quotation of McDowell's from Philosophical Investigations (§95, excerpted: "When we say, and mean, that such-and-such is the case, we—and our meaning—do not stop anywhere short of the fact, but we mean: this—is—so") comes a mere page before Wittgenstein indicts those – including his former self (§114) – who would read the forms of our language back into philosophy as supplying an ideal which must be found in our philosophical accounts of reality as well; see e.g. §101: "We want to say that there can't be any vagueness in logic. The idea now absorbs us, that the ideal 'must' be found in reality. Meanwhile we do not as yet see how it occurs there, nor do we understand the nature of this "must". We think it must be in reality; for we think we see it there." §107 contrasts the actual use of language with the picture implied by the requirement (and vice versa); and soon we are into the heart of the Investigations' call for a rethinking of the entire idea of thinking philosophically.