Proust’s Swann is obsessed by what he doesn’t know about Odette. His anguish has no remedy; finding out more only adds to what he does know about her. Since Kant, lots of philosophers have suffered from a generalised and aggravated form of the same complaint. They want to know what the world is like when they aren’t thinking about it; what things are like, not from one or other point of view, but ‘in themselves’. Or they think that maybe that’s what science aims to know, and wonder whether it’s a project that makes any sense. They are thus worried about ‘the possibility of objectivity’.
A.W. Moore’s book is a sort of defence of the possibility of objectivity (what he often calls the possibility of ‘absolute representations’). He doesn’t argue that objectivity is ever attained or ever likely to be; perhaps not even in our most scrupulous investigations. All he wants is that the goal should be coherent Given the modesty of the enterprise, it’s really shocking the conclusions that he’s driven to: ‘We are shown that absolute representations are impossible ... [But] we do well to remind ourselves ... that what we are shown is nonsense. Properly to replace “x” in the schema “A is shown that x” is a quasi-artistic exercise in which one creates something out of the resonances of (mere) verbiage. There is no reason whatever why this should not sometimes involve making play with inconsistency.’
It’s a fascinating story how Moore gets to this; in fact, a cautionary tale. He takes on board, from the beginning and essentially without argument, what is currently the received view of meaning in philosophy. The rest follows as the night the day. What’s splendid about the book is that, unlike most philosophers who share his premises, Moore is prepared to face the consequents. What’s appalling is the tenacity with which he clings to an account of representation from which it follows, inter alia, that his own philosophical views are nonsense. (‘Inter alia’ because, of course, your views and mine are nonsense, too.) Not just like nonsense, mind you, but the real thing; on all fours with ‘ “Phlump jing ax” ’ since ‘there can be no other reason for an utterance’s failing to be a representation than that certain [of the] words lack meaning.’ I propose to trace, briefly, the course of these events. At the end, I’ll moralise a little about how things stand if Moore’s book is read, as I think it should be, as a reductio of his theory of meaning.
The book starts oddly: ‘a representation need not be objective ... a representation can be from a point of involvement ... I shall call any [such] representation ... “subjective”.’ This makes representations that contain indexicals (like ‘it is humid today’) perspectival and subjective since, ‘if I wish to endorse an assertion I made yesterday of [this] sentence, I have no alternative but to produce a representation of some other type (‘it was humid yesterday’).’ This, to repeat, is an odd way to start. For, while indexical sentences are uttered from a point of view (in the sense that ‘it’s raining’ can be true here and false there at one and the same time; or true then and false now in one and the same place), there doesn’t seem to be anything particularly subjective about them. If what you want is subjectivity, try ‘red is warmer than green’ or ‘Callas was better than Tebaldi.’
Indexicality is about content whereas objectivity is about truth. If you take indexicality as a paradigm of the perspectival, then it’s a mistake to run ‘perspectival’ together with ‘subjective’. What’s interesting about a sentence like ‘it’s raining’ is that the content of the assertion that I make when I say it depends on where and when I am. What’s interesting about ‘red is warmer than green’ is something quite different; viz that (may be) it can be true for me but false for you; or (maybe) there’s no matter of fact at all about whether it’s true. There’s certainly a matter of fact about whether it’s raining; so what on earth does whether a representation is indexical have to do with whether it’s objective? So one wonders as one reads along.
I guess what Moore has in mind is something like this: the content of an indexical representation depends on its context. But so likewise do the contents of representations of every other kind, since having a representation means having a language, and ‘having a language involves having an outlook; or, more specifically ... having a language involves having its own distinctive outlook.’ Moore swallows whole not just Wittgenstein, but also Whorf and Sapir; which is, perhaps, a little quaint of him. So the content of a representation, indexical or otherwise, is fixed partly by its conceptual role; by the inferences it licenses. That one can only represent things from the point of view of one’s inferential commitments is therefore part and parcel of what content is. But representations from a point of view are ipso facto perspectival, hence ipso facto not objective. It seems, in short, that semantics puts pressure on epistemology; there is a prima facie conflict between the possibility of objectivity and the idea that content is holistic. How to resolve this conflict is what Moore’s book is really about, though I doubt he’d want to put the case this way.
You don’t have to resolve it, of course; you could just try to live with it. Quite a lot of philosophers, including some of our most eminent contemporaries, seem to think that’s indeed the best that you can do. So Thomas Kuhn famously maintained that, because concepts are implicitly defined by the theories that endorse them, radically different theories are ipso facto ‘incommensurable’. There is, to that extent, no such thing as a rational choice between conflicting theoretical paradigms and no fact of the matter which, if either, of conflicting paradigms is objectively correct. Changing your paradigm isn’t changing your mind, it’s changing your life-style.
Variants of this disagreeable idea are on offer from philosophers with as little else in common as, say, Wittgenstein, Quine, Goodman, Putnam, Derrida and Davidson, among many others. Part of what’s wrong with it is that it’s so unstable, Just how major does the disagreement in background commitments have to be before incommensurability sets in? And if, as seems likely, there is no principled answer to this question, isn’t it the moral that all judgment turns out to be more or less subjective? We’re not far from the view that argument and persuasion are mostly power politics in disguise; class politics on some accounts, sexual politics on others. Whatever became of the disinterested pursuit of objective truth?
There is, however, a traditional, more or less Kantian, way to split the difference between what’s perspective and what isn’t; it’s the doctrine called ‘transcendental idealism’, and much of what’s most interesting in Moore’s book is about it. It’s one thing, Moore says, to suppose that representation is ipso facto ‘from an outlook’; it’s another to conclude that representation is ipso facto not objective. The second follows from the first only on the assumption that the perspectival representations aren’t capable of being integrated from a point of view that embraces both.
Suppose one takes for granted that representation is, as Moore puts it, always ‘of what’s there anyway’. That is, all of our points of view are perspectives on the same world, the world to which our beliefs conform insofar as they are true. True beliefs have to be mutually consistent so it must be possible, in principle at least, to integrate all the true perspectives, even if only by conjoining them. At the end of the day, representation is still perspectival, but perhaps the only perspective to which it is intrinsically relative is our point of view as rational inquirers, or cognitive systems, or (as Moore sometimes puts it) as processors of knowledge. Maybe you can’t know what things are like in themselves, but you can, at least in principle, know what they are like insofar as they are possible objects of knowledge. If this sounds a bit truistic, so much the better.
That’s encouraging as far as it goes, but it doesn’t go far. For one thing (a point Moore doesn’t emphasise) the perspectival relativists he is trying to split the difference with are unlikely to grant him his ‘basic assumption’ of a ready-made world. To the contrary, it’s part and parcel of the Kuhnian sort of view that thinkers with different paradigms ipso facto live in different worlds. That’s the metaphysical obverse of the semantical claim that you can’t integrate divergent paradigms. (Since Kuhn takes content to be paradigm relative, what you get if you conjoin representations from different paradigms is just equivocations.) Indeed, the many-worlds story is, arguably, just the incommensurable-paradigms story translated from semantics to metaphysics; so, arguably, you can’t have one without the other. If it’s not clear how literally we’re to take either, or even what taking them literally would amount to, that just shows what a mess we’re in.
And second, a point to which Moore is entirely alert, transcendental idealism looks to be self-refuting. If you really can’t say anything about the world except as it is represented, then one of the things that you can’t say is that you can’t say anything about the world except as it is represented. For, the intended contrast is between how the world is as represented and how it is sans phrase. But how the world is sans phrase is how it is not from a perspective, and it’s part of the story we’re trying to tell that representation is perspectival intrinsically. It’s worth remarking how much worse off Moore is than Kant in this respect. Kant thought that there are substantive constraints on the ‘form’ of the possible objects of knowledge; hence that we can’t know about what fails to conform to these constraints. Moore, following the early Wittgenstein, thinks that there are substantive constraints on the form of the possible objects of representation; hence that we can’t even think about what fails to conform to these constraints. For Kant, transcendental idealism was strictly mere metaphysics (viz it’s unknowable); for Moore it is strictly mere babble (viz it’s unspeakable). It is, I think, a general rule that whenever philosophy takes the ‘linguistic turn’ the mess it’s in gets worse.
Transcendental idealism is nonsense by its own standards. So be it; ‘this leaves scope for all sorts of distinctions ... We can think of the Nonsense Constraint as offering the following guideline when it comes to making sense of the schema “A is shown that x”: namely to prescind altogether from considerations of content and to think more in aesthetic terms ... To say of some piece of nonsense that it is the result of attempting to express the inexpressible is something like [sic] making an aesthetic evaluation. It concerns what might be called, justly, if a little grandiloquently, the music of words.’ One is powerfully reminded of Frank Ramsey’s salubrious wisecrack about the embarrassing bits of the Tructatus: ‘What can’t be said can’t be said and it can’t be whistled either.’
I’ll spare you the details, which are complicated and, in my view, much less than convincing. Roughly, we have lots of ‘inexpressible’ knowledge. It’s being inexpressible is somehow connected with its being knowing how (e.g. knowing how to process knowledge) rather than knowing that, though I find this connection obscure. Sometimes we’re driven to try to express the inexpressible (that we are has something to do with our aspiring to be infinite); when we do so, we talk nonsense. For example, we say such things as that representations can (or can’t) be absolute. Since this is nonsense, saying it can, of course, communicate nothing. But that we are inclined to say it shows all sorts of things. For example, the nonsense we’re tempted to talk about value shows that ‘things are not of value tout court. Nothing is. However, another thing we are shown is that they are of value tout court. Our aspiration to be infinite, precisely in determining that these things are of value for us, leads to our being shown this.’ It’s tempting to dismiss this sort of thing as merely irritating. But if you propose to do so, you have to figure out how to avoid it, compatible with both relativising meaning to perspective and retaining a respectable notion of objectivity. It is to Moore’s credit that he has faced this squarely.
But where does having faced it get him? According to enemies of objectivity like Kuhn, it’s false to say that the world chooses among our theories. According to friends of objectivity like Moore, it’s nonsense to say that the world chooses among our theories. Objectivity might reasonably complain that with such friends it doesn’t need Existentialists. What’s remarkable is that neither side of the argument considers, even for a moment, abandoning the premise that’s causing the trouble; viz, that content is ipso facto perspectival and holistic.
Perspectival theories of meaning make objectivity look terribly hard. So much the worse for perspectival theories of meaning; objectivity is easy. Here’s some. My cat has long whiskers. He doesn’t (as Moore would be the first to agree) have long whiskers ‘from a perspective’ or ‘relative to a conceptual system’ or ‘imminently’; he has long whiskers tout court. That’s what my cat is like even when I am not there (unless someone should snip his whiskers while I’m not looking; which is not the sort of thing that Moore is worrying about.) Prima facie, the ontological apparatus that’s required for my cat to have long whiskers is extremely exiguous; all that’s needed is my cat, long whiskers, and the state of affairs that consists of the former having the latter.
Likewise, it is possible for me to represent my cat’s having long whiskers. Indeed, I just did. The semantic apparatus required for this is also exiguous, and for much the same reasons. All I need is to be able to represent my cat, long whiskers, and the state of affairs that consists of the former having the latter. First blush at least, neither my cat’s having long whiskers nor my representing it as having them would seem to imply a perspective or an outlook or a desire for infinity, or anything of the sort; the objectivity of the fact is about all that the objectivity of the representation requires. So whence the angst?
That all this is so, and objectively so, is far more obvious than any general principles about the relativity of content to perspectives or conceptual systems. If, therefore, your perspectivist semantics leads you to doubt that it is objectively so, or that it can coherently be said to be objectively so, it’s the semantics and not the cat that you should consider getting rid of. That this methodological truism should have escaped so many very sophisticated philosophers, Moore among them, seems to me among the wonders of the age.