One thing that’s become increasingly important to my thinking is the idea of priority in account. As I think of it, the basic idea is just this. In any theory (including any philosophical theory), some terms will be defined in terms of others, while others will be left undefined. The undefined terms aren’t, of course, left unexplained. In particular, in teaching your theory to others, you’ll need to do something to communicate the meanings of the undefined terms. In philosophy, at least, one of the most important ways of communicating the meanings of these terms is to give examples of the phenomena to which the terms apply. But, in principle, anything that successfully communicates those meanings is allowed.
Importantly, the same points apply, not only to the terms of a theory, but also to the phenomena the theory theorizes. Indeed, it’s in its application to the phenomena that the notion of priority in account becomes truly important. The basic idea, however, can be put similarly: we simply transfer the above ideas from their application to conventional definition (the definition of the terms of a theoretical language) to real definition: the definition of the phenomena themselves. The same points will then apply: some of the phenomena are defined in terms of others, while others are left undefined – though, again, the undefined phenomena aren’t left unexplained. (In fact, on my view, the whole theory itself will constitute the final explanation of the undefined phenomena. But that’s a story for another day.) But I don’t want to go into these issues about the undefined phenomena here. What interests me here is what happens if you don’t pay enough attention to priority relations.
So here’s an application of these ideas in the philosophy of perception. I’m currently reading Bill Brewer’s Perception and Its Objects (OUP, 2011). In chapter 4, Brewer provides an extended argument against what he calls the Content View, according to which “perceptual appearances are characterized as mental representations of specific ways a mind-independent world might be” (48). Importantly – and not at all atypically, I think – many of his crucial objections to the Content View are circularity objections: he argues that the defender of the Content View has no way of explaining the contents of the relevant mental representations except in terms of perceptual relations to worldly objects themselves. This means, though, that the view Brewer is objecting to isn’t the Content View per se, as Brewer characterizes it in the above quotation, but, instead, the view that perceptual appearances are characterized as mental representations of specific ways a mind-independent world might be, where the nature and content of those representations can themselves be explained without appeal to any perceptual relations to object in the mind-independent physical world. In other words, Brewer’s objection is to a project on which we begin with an account of mental representations that makes no reference at all to perceptual relations to objects. Against that project, his objections are, I think, successful. But notice that the failure of that project does nothing to show that perception isn’t representational. It shows only that either perception isn’t representational or perceptual representations are to be explained in terms of perceptual relations to objects.
The latter sort of view – on which perception is representational, but the contents of perceptual representations are to be explained in terms perceptual relations to objects, and, in particular, in terms of perceptual relations to objects in cases of successful perception – is, I think, developed by (for example) both Tyler Burge and John McDowell (though, to be sure, in rather different ways). And yet Brewer counts both Burge and McDowell as targets of his criticisms.
For my own part, I’ve found it frustrating, reading the book, that Brewer describes himself as objecting to a representational view of perception as such, and yet never so much as considers the possibility that the Content View, as he officially describes it, and his own Object View are actually compatible. On the latter view, “perceptual appearances are most fundamentally cases of standing in an essentially experiential relation of acquaintance with specific mind-independent physical objects themselves” (48). But philosophers like Burge and McDowell seem to me to accept this view – as well as the Content View. Their idea – or, if not theirs, then certainly my own – is that, in a theory of perception, the fundamental case is one in which a perceiver is indeed standing in a relation of acquaintance with mind-independent physical objects. On the view I myself prefer (which I think is at least close to McDowell’s), for example, the fundamental thing here (the thing ultimately left undefined in the theory of perception) is an act of acquiring knowledge on the basis of perception. It is in terms of such acts that we then explain the representational contents, first, of (the purely perceptual components of ) such acts themselves, and then, second, of similar acts in which, however, no knowledge is acquired (illusions and hallucinations, in particular). Importantly, such an account will not be subject to any of the circularity objections that Brewer presses against the Content View, because there’s just no pretension whatsoever to provide an account of successful perception in terms of (possibly unsuccessful/non-veridical) representations plus worldly relations between such representations and worldly objects.
The view I just sketched is, I think, quite similar to Brewer’s own Object View. (Just how similar these views are, I don’t yet know – I haven’t read the positive part of the book yet!) The difference is that it doesn’t dispense with representations, but holds on to them for purposes of explaining perceptual error. But it certainly sides with Brewer in treating as fundamental those cases in which the subject stands in a relation of acquaintance to a mind-independent object (most fundamentally, I think, a relation that makes knowledge of that object available to the subject). So it’s unfortunate that he lumps this sort of view together with a version of the Content View on which possibly unsuccessful/non-veridical representations are prior in account to (knowledge-ground) acquaintance-relations to mind-independent objects.
(I’ll also mention, briefly, that many of Brewer’s other – i.e., non-circularity – objections to the Content View seem to me to rest on an unacceptably narrow conception of representational content, one on which representational contents are essentially language-like, at least insofar as they have truth-conditions – as opposed to, say, accuracy-conditions more generally. It seems to me, however, that, if perception is representational, the representations it involves will be basically pictorial. And whether Brewer’s other objections apply to a pictorial-representationalism is, I think, a good question.)
There’s a fairly simple methodological lesson here that’s worth stressing. If you press a circularity objection against an opponent, you’re necessarily assuming that your opponent is making a priority claim. And so you should check to be sure that they’re actually making that claim – and not, for example, its converse.