After re-reading Alvin Goldman’s influential 1976 paper, ‘Discrimination and Perceptual Knowledge’ (Journal of Philosophy, 73, pp. 771–91), I found myself questioning the intuition that is supposed to lead to a problem for the traditional causal account of knowledge. As per the original Gettier cases, the scenario that Goldman describes involves an individual (Henry) arriving at a justified true belief (JTB) that we would (allegedly) nevertheless consider to fall short of knowledge. However, in contrast to the Gettier cases, the reason that he has for arriving at the true belief is exactly the same (at least according to traditional accounts of reasons and experience — see my previous post for an alternative view) as the reason that he has in the good case. This makes it even more challenging for post-Gettier accounts of knowledge, such as Goldman’s causal theory, to explain why Henry doesn’t have knowledge since he seems to have done everything right, just as he would have in the good case, giving rise to many different ways of accounting for this discrepancy.

The story goes like this: Henry is driving through the countryside with his young son, pointing out objects of interest, such as trees, fields, cows, tractors etc. Pointing at a barn, Henry says ‘That’s a barn’, which is a true statement, arrived at by recognising the appearance of the building that appears some distance away as a barn. However, unbeknownst to Henry, in this part of the country there are many structures that are indistinguishable from barns when viewed from the road, but which are in fact only wooden frontages, or façades, made to look exactly like barns. (Perhaps it’s a local tourist attraction, or the farmers are engaged in some kind of practical joke, it doesn’t matter.) Now, it just so happens that Henry pointed to the one and only real barn in the area, and, since he is unaware of the presence of the other fake barns, he has true justified belief that what he sees is a barn. But does he know that it is a barn? Goldman thinks not, and many philosophers would seem to agree with him. However, I’m not so sure. My intuition is that he really does know that it is a barn, and that the presence of the barn façades does nothing to undermine this.

Of course, if (ex hypothesi) Henry knew that there were lots of fake barns in the area then he could not be said to know (or perhaps, would not claim that he knew) that this particular object was a barn, since he has no way of determining which of the barn-looking things are real barns and which are fakes. However, according to the case in hand, he does not know this, and so his knowing that the object in front of him is a barn (if indeed it is such) is entirely down to his luck in pointing at the right barn quite by accident. Can we acquire knowledge in such ways? Can luck contribute to our epistemic standing, or is a recognitional ability for telling things from their relevant alternatives (in this case fake barns) a necessary condition for knowledge? I certainly don’t have a clear intuition that it is, and I don’t see why some of our knowledge shouldn’t be lucky in this way.

Note that the above example involves quite a different sort of luck to the one within the original Gettier job interview case. There, Smith thinks that the person who will get the job has $10 (or whatever) in his pocket on the basis of a mistaken belief about who will get the job (i.e. Jones, rather than himself), which undermines his epistemic standing. Henry, on the other hand, has done nothing wrong in terms of the way that he arrives at his belief, and so is epistemically blameless on any account.

My concern is that basing an entire theory of knowledge upon the intuition that Henry doesn’t have knowledge in the above and other similar types of cases may is to start out with a very shaky foundation, since the intuition is at least disputable and could be attributed to an altogether different source. Apart from contamination by philosophical theory — for example by tying the idea of knowledge very closely to the exercise of a recognitional ability, as per Millar (forthcoming) — we may be begging the question against attributing knowledge in this case. There is also the fact that, were Henry to find out about the barn façades — perhaps he reads it in a guidebook — then he may himself be inclined to withdraw his claim to knowledge, since he realises that it is likely that what he saw was in fact a barn façade, and not a real barn, since he has no way of telling the two apart. Consequently, Henry may no longer believe that it was a barn, and so, since knowledge requires belief (or so the story goes), can no longer be said to have knowledge. This seems odd since we do not normally expect additional information to reduce the amount of knowledge that someone has, but rather to add to it, since acquiring information is one of the ways that we gain knowledge. However, this may just be a peculiarity of the case, as it is certainly plausible that I can have knowledge of something — a complex mathematical truth, for example — but then lose it due to someone convincing me otherwise (although in this case we may want to say that I still have knowledge but that I am no longer aware of it, perhaps because I wasn’t very confident about my reasoning). In any case, there certainly seem to be sufficient factors at play to generate what could seem, prima facie, to be an intuition that Henry doesn’t have knowledge, when in fact we are just being distracted by, for example, what we would claim to know in that situation given the way the example has been set up (i.e. the existence of the fake barns).

The above considerations raise some difficult methodological issues for any theory of knowledge, such as what should such a theory answer to, and which of our linguistic intuitions genuinely express our understanding of knowledge, and which are due to mere analogy or convention. I was first led to thinking about this by speculating about what McDowell’s epistemological disjunctivism would have to say about Goldman’s fake barns case, which is something that (to my knowledge at least) McDowell himself has never responded to in print. I suspect that he would be inclined to say that, since we take in the fact that the object is a barn in the relevant way via normal perceptual channels, we do in fact know that the object is a barn. Indeed, we may even know that we know, although that knowledge is based upon our ignorance of a salient fact: namely, that the area is littered with identical looking barn façades. He could say this on the basis of the justification for our higher-order knowledge being just the same as the justification for our first-order belief: namely, that we see that it is a barn, which, since it turns out to be true, warrants us in having knowledge. Of course, if the thing we pointed to wasn’t a real barn, then we would have been wrong to claim that we knew otherwise, but that’s a matter of an erroneous claim to knowledge, rather than there not having being any knowledge in the good case where there is in fact a barn. The one doesn’t necessarily entail the other.

I’d be interested to hear what other people think about the above, and whether your intuitions agree with Goldman’s that there cannot be any knowledge in such cases, and exactly why you think that should be so. My view is that we have no firm ‘pre-theoretical’ intuition to this effect, and perhaps have an intuition to the opposite, that Henry has knowledge in the good case. When prompted to explain why this is so, we might be tempted to say “Well, he just sees the barn, which is a barn, so he knows that there’s a barn”, which would tend to favour the McDowellian hypothesis, although as I’ve argued above, it’s always dangerous to base one’s choice of theory upon a single intuition — especially one which can be modified when put under pressure.