As one of my friends recently pointed out, this blog has lain dormant for rather a while (almost a year now!), so I thought it was time to rejuvenate it with some ruminations on a topic I’ve been reading up on lately — namely, disjunctivism. Put simply, this is the thesis that what genuine perceptual episodes (e.g. seeing the lake) and their imaginary or non-veridical counterparts (e.g. hallucinating the lake) have in common is not that they consist of the same internal, phenomenally identical experience, but rather just the fact that both episodes satisfy the description of looking as if there is a lake. On this account, what makes them experiences that they are is different in the two cases: the first is a genuine ‘seeing’ of a lake, and is therefore relational, since it is essentially related to an object in the external world (the lake), whilst the second (the hallucination) is non-relational in any straightforward sense. What’s distinctive about disjunctivism is that it rejects the need for any ‘highest common factor’ that is identical between the two cases. This is supposed to help explain why it is that we can have knowledge of external world objects, and thus form the basis of a response to scepticism, although many (e.g. Crispin Wright) disagree that it has the resources to do so.

Another way to understanding the above conception of disjunctivism is that perceptual experiences are externally individuated. In other words, it’s not just a matter of what’s going on inside one’s head or central nervous system, but that experience literally ‘takes in’ aspects of one’s immediate environment. On this view, the (external) differences between the perceptual and hallucinatory cases are sufficient to make them different experiences, as opposed to the same experience achieved in two different ways, as in the traditional (i.e. internalist) conception of experience. To see how this might begin to address sceptical doubts, we need only think of the standard argument from illusion in which the sceptic argues that since any experience is phenomenally indistinguishable from its hallucinatory counterpart (e.g. being massively deceived by an evil demon or mad scientist — take your pick), merely having that experience is insufficient evidence for the existence of an external world, there being a chair in front of you, or whatever, since you would have had exactly the same experience had you merely hallucinated that the world was that way (thanks to the intervention of the demon/scientist — sorry, it seems that even my examples have become disjunctive!). In effect, the disjunctivist argues that the assumption that the person would have had exactly the same experience is incorrect since the genuinely perceptual experience (the good case) is an entirely different kind of experience than that of the hallucination (the bad case), and so the sceptic’s argument is misguided.

The trouble with this kind of response is that it doesn’t seem to address the main worry. The sceptic’s agument wasn’t that we don’t have any first-order knowledge, but that we cannot know whether we have knowledge or not; i.e. it is second-order. All this talk of taking in the facts and external individuation is all well and good, and may even be right, but it doesn’t do anything to secure our second-order knowledge — i.e. knowledge that we have knowledge (or our warrant to claim knowledge, if you prefer Wright’s terminology), and so, as a response to scepticism, the argument fails. Even worse, there would seem to be a lurking circularity in the disjunctivist’s reasoning since it is only by assuming that there is an external world containing knowable facts that we can grant the possibility of genuine perceptual experience in the first place. It could be that there are nothing but hallucinations, in which case the disjunctive conception of experience fails to get off the ground (except perhaps showing that genuine perceptual knowledge it possible, but that’s hardly a victory for anti-scepticism).

In his 1982 lecture, ‘Criteria, Defeasibility and Knowledge’, John McDowell connects the disjunctivist conception of experience with Wittgenstein’s notion of a criterion. He argues — quite convincingly, I think — against Wright’s interpretation of this term as a necessarily reliable but defeasible indicator of knowledge, and in favour of conceiving criteria as ‘ways of telling’ that something is the case (or that they are ‘thus and so’, as he is fond of saying). This, I think, is a good way to understand the notion of criteria, and certainly fits better than Wright’s defeasible warrants, at least on my reading of the Philosophical Investigations, but it seems characteristic of the disjunctivist response to scepticism that the notion of factivity is portioned off into some other (externally individuated) factor. A criterion, we are told, is a way of telling that something is true such that whenever the criterion is present then we have knowledge, and where there is no knowledge — e.g. we are hallucinating — then there is no criterion either. Since the criteria for a given type of knowledge may vary between cases and what constitutes a criteria is (it seems to me) externally, rather than internally, individuated, there is simply no way of telling whether our experience involves genuine criteria for knowledge or not since second-order knowledge of what constitutes the appropriate criteria is lacking.

In a forthcoming paper, Alan Millar develops this notion of a criterion into what he calls a recognitional ability, which is an externally individuated knowledge-conferring capacity. Such capacity are only said to have been exercised when knowledge is delivered in the appropriate way; i.e. by virtue of their ‘taking in’ the relevant facts. Such abilities are indexed to particular kinds of environment such that one’s ability to discriminate (to take a well-worn example) zebras from cleverly disguised mules may depend upon whether there are normally any such cleverly disguised mules in the vicinity, or (to take an even more hackneyed example) one’s recognitional ability to tell barns from non-barns depends upon whether one is operating in fake-barn country or not. Again, the worry is that this sort of approach, although it locates the factive element in the subject’s ability to tell, rather than in the signs of knowledge themselves, simply portions off the factivity into another externally individuated element — in this case, a recognitional ability. We are left in the position of having (relatively) reliable recognitional abilities, but without necessarily being able to tell when and if we are exercising those abilities; i.e. whether we have knowledge or not. Of course, it is open to Millar to claim (and he does) that we also have reliable ways of telling when our first-order recognitional abilities have been employed — indeed, that’s part of what having the abilities themselves involves — but it’s equally open to the sceptic to cast doubt upon our whether such second-order knowledge is in fact available without getting into some kind of higher-order regress, and so again this seems an unpromising strategy for refuting the sceptic.

Despite the above comments, I actually have a lot of sympathy with the disjunctivist position and the idea that our epistemic capacities make direct content with the objects and/or facts that they supposedly apprehend. It seems much more likely — even from an empirical perspective — that experience should be individuated such that hallucination is a different type of experience to veridical perception, despite the fact that the two may (although by no means necessarily) be phenomenologically indistinguishable. The trouble is, this just seems to shuffle the pieces around such that we can have knowledge of the facts but without us being aware of whether we have knowledge or not, and so scepticism once again gains a foothold. Of course, if you’re inclined (as I am) to think that scepticism is just a kind of philosophical pseudo-problem based upon a misunderstanding of how natural language works, as per Wittgenstein, then this is unlikely to be of great concern, but still, it would be nice to have something more substantive to say about it — especially as I’m hoping to write an essay on the topic in the very near future! In the final analysis, perhaps the enterprise of constructing some conception of knowledge that is fully verifiable by the subject is itself misguided, or appeal to common sense as placing the burden of proof upon the sceptic to give a reason for why we should doubt the existence of such knowledge in the first place. In any case, I doubt that disjunctivism presents a strong argument either way — although this is, of course, an entirely defeasible conclusion.

Postscript: perhaps what disjunctivism has to offer is that it provides a plausible story as to how knowledge is possible, as well as an analysis of what knowledge amounts to; i.e. the ‘taking in’ of worldly facts. Whilst this falls short of providing a satisfactory response to the sceptic, it does at least secure the possibility of knowledge — something that the sceptic not only throws into doubt, but undermines its very basis of by providing an ‘error theory’ as to why we can never know anything due to the subjective indistinguishability of knowing from not knowing. The debate then turns to which model of knowledge is the more plausible, or best fits our everyday linguistic practices, whereby the attractiveness of the sceptical position is undermined, although never actually refuted… Food for thought!