Philosophy Blog
Life at the thought-face.
Friday, August 15, 2003
Memory and Knowledge
Sitting in the car during my daughter’s gym class last night I made the decision to completely abandon a paper that has been kicking around and haunting me for a long time. The paper tried to defend the following claim:
(KPK) If someone is to have memory knowledge, then he must now be justified in thinking that he previously knew what he now remembers.
The idea is that this places a slightly stronger condition on memory knowledge than the currently fashionable:
(TP') If someone knew that p at t1, and has an accurate memory that p at t2 (and has no reason to doubt either the accuracy of her memory or the justifiedness of the prior belief), then she knows that p at t2.
After several attempts, I came to the conclusion that there was no example which required (KPK) to explain our intuitions: all of them could be adequately explained by (TP’). But I was undaunted because it should be clear that the method of example and counter-example is not the only way to argue in epistemology. In fact, the apparent conclusiveness of such well-known arguments by example as Gettier’s famous paper might even be an illusion. (Brain Weatherson has recently given an argument to the effect that it is an illusion, but whatever the success of that particular argument, which rests on an objectionable theory of meaning, the moral remains that we should be more cautious about examples in epistemology.) And in fact if you look at arguments for claims like (TP’), they are based not on examples but upon very general considerations about the possibility of rational belief. The other thing you notice about those arguments is that they tend to lump memory and testimony together, as if they are in epistemically the same position (e.g. Dummett, Burge, Owens). Now I agree that the general considerations make the testimony version of (TP’) correct, but I decided to argue for (KPK) by showing that there were significant disanalogies between memory and testimony.
But the arguments I came up with were lousy: not even rhetorically effective. If you have less sense than time, you can read them here (the lousy stuff is in Section III, the rest I still think is quite good).
Anyway, I am pleasantly surprised at what a relief it is to completely abandon a paper.
.: posted by Tom Stoneham 10:43 AM
Monday, August 04, 2003
More on testimony
Over the weekend I was thinking about how to respond to a potential objection to the argument I offered on Foley’s behalf, and realized that Foley cannot accept my argumentative gift.
The objection is that I have cut methods too coarsely. If I look hard enough, there will be some ‘method’ of forming opinions which it is a priori guaranteed any two people will share. Surely I trust myself not merely because I form my opinions by reflective self-criticism, but because when I am critical of my own opinions, I appeal to a particular set of epistemic principles (e.g. Bayes Theorem, or the injunction not to believe a logical falsehood). However, when I know as little about someone as I know about Anonymous, it cannot be an epistemic default position that she has formed her opinions by the same methods which I trust in my own case.
To see why this criticism fails we need first to make clear a terminological move of Foley’s (p.27). His primary thesis is that one should trust one’s opinions in proportion to their invulnerability to self-criticism, and he uses a sense of ‘rational belief’ which allows the paraphrase: one’s opinions are epistemically rational in inverse proportion to their vulnerability to self-criticism. This allows him to say (p.41) that someone who, unlike us, thought that dreams were a better guide to the external world than sense perception, could turn out to be epistemically rational so long as this conviction would survive reflective self-criticism. Such a person should trust her dream judgements in the same way that we trust our perceptual judgements. Similarly, Foley allows the possibility that someone suffering from Cotard’s delusion is ‘not necessarily being irrational’ (p.41). It seems to follow that the only thing which matters when considering whether to trust an opinion of one’s own is that it survives reflective self-criticism. In particular, it does not matter which epistemic principles one endorses, for none are intrinsically better than any others, just so long as those principles also stand up to reflective self-criticism. In other words, rationality and trustworthiness goes with standing up to self-criticism, whatever the epistemic principles endorsed by the subject.
Given this, we need to distinguish the direct argument I offered for the prima facie trustworthiness of testimony from Foley’s own argument. The direct argument goes: if surviving self-criticism is sufficient to make an opinion of mine trustworthy, whatever epistemic principles I endorse, then surviving criticism by someone else, whatever epistemic principles she endorses, is sufficient to make an opinion of hers prima facie trustworthy for me (with lots of qualifications). I think there is something quite plausible about this argument, but it falls foul of Foley’s requirement that the trustworthiness of testimony be contingent. So Foley’s argument must use as a premise the allegedly contingent truth that in general other people’s epistemic principles would stand up to reflective criticism by me. And it must be reasonable to believe this without any evidence. So what Foley needs to appeal to is the thought that it is an epistemic default position to take it that if someone else’s epistemic principles have survived scrutiny (by them) then they would survive scrutiny by me. And now I am inclined to agree with Owens: it is hard to see how this could be contingent and yet reasonable to believe without recourse to evidence.
.: posted by Tom Stoneham 11:39 AM
Friday, August 01, 2003
Setting the issue of whether Foley is committed to some form of transparency thesis to one side (I think he is, but I also think he would initially deny that he is), I want to look at his account of testimony. The reason I started reading Foley’s book was a review by David Owens in Mind which roundly rejected the account in less than two pages. Now, Owens’ reading of Foley is a perfectly acceptable interpretation of what he says, and on that interpretation, the account is as flawed as Owens says. But I want to argue that there is another, more subtle, reading of Foley which avoids Owens’ objection.
Foley’s starting point is the claim that:
1. It is perfectly reasonable to trust one’s own intellectual abilities.
By this he means that, though we cannot defeat the sceptic and prove (1), we are being rational when we take those of our opinions which survive reflective self-criticism to be true. If we are brains in a vat, this trust is misplaced, but it is not irrational. The argument for this involves some large claims about the nature of epistemology, the relation between rationality and knowledge, and the impossibility of defeating the sceptic, so I am just going to grant it for present purposes.
Foley’s next move (that is ‘next in my rational reconstruction’) is to argue for the contingent empirical claim:
2. My intellectual faculties and the contexts in which I have exercised them are not significantly different from those of others.
He concludes:
3. It is inconsistent not to trust the intellectual abilities of others.
That is, in the absence of any evidence about either the proposition in question or the reliability of the other with respect to such issues, we are reasonable in taking someone’s opinion that p as being correct. Foley graphically illustrates this with a character called Anonymous. All we know about Anonymous is a list of some of the propositions he or she believes. We cannot infer from these anything about when or where Anonymous lived, and what topics he may have been expert or unreliable on. Foley thinks that if all we know is that Anonymous believed p, and we do not have any opinion of our own about p, it is prima facie reasonable to accept p on that basis alone. Foley thinks that, given (1), the only way to deny the conclusion is to deny (2), but while possible, that is very hard to do, for it involves treating myself as special in a way which is not warranted by the evidence.
Owens’ criticism is simple. The notion of inconsistency being used in (3) is epistemic, i.e. one would be in a position which would not survive reflective self-criticism. But that means that (2) must not only be true but also known to be true. And Owens rightly points out that if (2) is an empirical claim, then it is hard to see how we could know it to be true without recourse to the social and biological sciences, and all such scientific knowledge inevitably involves testimony. The solution, Owens suggests, is to give an a priori defence of (2).
What Foley writes is:
Given that it is reasonable for me to think that … my intellectual faculties and my intellectual environment have broad commonalities with theirs, I risk inconsistency if I have intellectual trust in myself and do not have intellectual trust in others. (p.106)
So Owens is assuming that if (2) is a contingent, empirical claim, it can only be reasonable for me to think it (Foley is careful not to talk of knowledge here, given his views about the structure of epistemology) if I have a posteriori evidence for it. And when one asks what such evidence would be, we immediately see how it relies on testimony.
There are two things we can say in reply: First, we might challenge the assumption and argue that there are empirical claims which it can be prima facie reasonable to believe in the absence of evidence to the contrary. Secondly, we can say that the alleged circularity is not vicious, for Foley is not trying to show that we can find reason to trust testimony which does not itself rely on testimony, rather, he is trying to show that if we are to be epistemically rational, we cannot avoid relying on testimony.
Taking the second first, we should note that Foley has another argument, which only establishes the weaker:
3’. It is inconsistent not to trust the intellectual abilities of anyone else at all.
The argument for this is that my opinions have been thoroughly and unavoidably influenced by others already, so if I trust myself I must also be trusting those others who have influenced me. Whatever the merits of this argument, it makes clear what Foley’s objective is, namely to show that it is very hard, though not completely impossible, to trust myself and not trust others, without falling into epistemic inconsistency. He clearly thinks that to trust oneself is nearly always already to trust some others, so he does not accept that there is some testimony-free starting point from which the question of whether to accept the testimony of others can be raised and needs to be answered without accepting testimony. When describing the main argument, he repeatedly says that ‘intellectual self-trust creates a pressure’ (e.g. p.107) to trust others. The thought is not that intellectual self-trust gives us an independent basis from which we can see that trust in others is justified, but that there is a certain sort of instability in the position of one who trusts himself but not others.
What then of the first point? It is clear that Foley thinks we can have empirical reasons to deny (2), so in one sense it is not a priori. But when he argues against the view that the prima facie trustworthiness of testimony is a priori (e.g. Burge), his objection seems levelled at the thought that it is necessary (p.98), rather than the thought that we can come to that opinion without appealing to empirical data. Perhaps the key idea here is that thinking one is, intellectually speaking, no different from the average person is reasonable in the absence of evidence to the contrary. It is, as it were, the epistemic default position. Whether this is Foley’s view or not, I cannot tell from reading his book, but it is a possible position which maintains most, perhaps all, of the attractive features of Foley’s position and deftly avoids Owens’ criticism.
Perhaps it is just too implausible to think that a contingent, empirical proposition could be such that it is reasonable to believe it in the absence of evidence either for or against it? One way of making it plausible is to appeal to the epistemic asymmetry between the claim that one is special and the claim that one is ordinary, no different from the others. It might seem that the claim of specialness always needs more evidence than its denial, and (2) is just the denial that I am intellectually special.
The claim of specialness wears the epistemic burden of proof only against the background of expected homogeneity. But what prior reason is there to expect humans to be homogenous with respect to their intellectual faculties and environments? It is hard to see how there can be an answer to this which does not undermine the status of (2) as a contingent empirical claim.
Perhaps, instead, we should look at (2) given (1). If we accept that our own intellectual faculties and past are such as to give prima facie grounds for taking our own opinions to be true, then to deny (2) is to say that the intellectual faculties of others are not such as to generally get things right. Now, the ‘others’ we are talking about here are reflective and self-critical reasoners, so the denial of (2) given (1) entails that my reflective, self-critical reasoning achieves its goal (truth) whereas other people’s does not. So the situation is this. There is a end (true opinion), a means for achieving it (reflective self-criticism), and an assumption (when I use that means I achieve the end). The question is, then, whether in the absence of evidence to the contrary, it is reasonable to think that other people using the same means will succeed in achieving the same end. And it does seem that reasonableness favours this opinion over agnosticism, because the assumption already commits me to the thought that the means is the right sort of thing to produce the end.
Now we can see a minor difficulty for Foley. What (1) and (2) get him is that the opinions of others which are such as to survive reflective self-criticism, what he calls deeply and confidently held, have prima facie credibility for me. But that does not entail that Anonymous is credible, because we do not know if the list represents deeply and confidently held opinions or the ‘doxastic counterparts of whims’ (p.26). Perhaps in real life cases we can usually tell on a case by case basis whether someone is expressing a deeply held opinion or a mere hunch or inkling, and that is certainly what Foley needs, because the general claim that most of what people sincerely assert expresses opinions that would survive self-criticism seems palpably false.
.: posted by Tom Stoneham 2:28 PM
|
