Ways around skepticism II: Causal theories of knowledge

We saw that Descartes' definition of knowledge committed him to the deductive closure principle because he had to accept the princi­ple of deduction for justification.

Gettier prepared the ground for his examples by making explicit another important assumption that all empiricists had made. It was that one could be justified in believing what was, in fact, false. This is a simple corollary of Locke's empiricist view that your beliefs can be justified by defeasible evidence. For, remember, to say that defeasible evidence can justify a belief is to say that a belief can be supported by evidence that is consistent with its being false. If—as Locke supposed—what justifies your belief is the evidence, then you could have the same justification in the cases where the belief was false as you have in the cases where it is true.

D: Jones is the man who will get the job, and Jones has ten coins in his pocket.

From (D) it follows that:

E: The man who will get the job has ten coins in his pocket.

Now, Smith knows perfectly well that E follows from D, and accepts E precisely because he believes D.

But now suppose also that, despite what the president said, Smith, not Jones, is going to get the job. Perhaps they decide he is just too impressive to turn down. And suppose, too, that Smith him­self has ten coins in his pocket, even though he does not know it. Then E is true, though D, which was his sole reason for believing it, is false.

In Gettier's example, then, all of the following three conditions clearly hold:

a) E is true,

b) Smith believes that E is true, and

c)  Smith is justified in believing that E is true.

Gettier concludes:

But it is equally clear that Smith does not know that (e) is true. For (e) is true because of the coins in Smith's pocket, while Smith does not know how many coins are in his own pocket, and bases his belief in (e) on a count of the coins in Jones's pocket, while falsely believing that Jones is the man who will get the job.

Because it requires the assumption that a false belief can be justi­fied, this example only works against a theory that allows that justi­fication is sometimes defeasible. It therefore poses no threat to the rationalist who believes that all evidence must be indefeasible. But it is not too hard to show that the PDJ is inconsistent with rational­ist assumptions as well.

Suppose, for example, I believe that some very complicated mathematical theorem is true, just because you told me and I had mistaken you for a very gifted mathematician. Let's suppose that, in fact, you are a very poor mathematician and just made the theorem up on the spur of the moment, but you happened, by pure chance, to come up with a truth.

How are we to react to the discovery that the PDJ is not right? We can begin by noticing that, in each of these cases, it is mere chance that the belief that the person has acquired is true. Though in each case the belief is true and justified, the fact that it is true plays no part in explaining why it is justified. It is the merest chance that Smith is correct in believing E or that I am correct in believing the mathematical theorem you told me. Perhaps, then, we should interpret the justification condition as requiring—as the American philosopher Peter Unger has suggested—that the fact that the belief is true should not be a mere accident.

There are some recent theories, prompted in part by Gettier's problems, that try to say what knowledge is in a way that follows up this idea. And, as it happens, they also allow us to find a sort of solu­tion to the skeptical problem with which we began. These theories are known collectively as causal theories of knowledge.

The basic idea of causal theories of knowledge is that in order to know S,

a) you must believe S,

b) S must be true, and

c)size=1 face="Times New Roman"> your belief in S must be caused in an appropriate way.

The causal theory's interpretation of the justification condition amounts to this: your belief is justified if it is caused in the right sort of way.

Originally it was suggested that your belief must be caused—in an appropriate way—by the fact that S is true. Theories of this sort deal with the example of Gettier's I cited just now.

Once we give up the idea that the fact that makes the belief true should actually cause the belief, the main problem for causal theo­ries is that talk of a belief's being caused in an appropriate way is left rather vague. So we need to answer this question: How, exactly, do we decide which ways are appropriate?

We can provide an example at once that shows that not just any way will do. This example is one from the work of the American philosopher Alvin Goldman, who has played a leading part in devel­oping causal theories. Someone called Henry is out driving and sees a barn. On this basis, he comes to believe correctly that there is a barn. Since there is a barn there and his seeing it is part of the expla­nation for why he truly believes it is there, this might seem to be a clear case of knowledge on the causal theory. Since there is little doubt that in this case, as described, we would say that Henry knew that there was a barn there, the theory does all right so far. But now Goldman expands the story with some extra details.

Suppose we are told that, unknown to Henry, the district he has just entered is full of papier-mache facsimiles of barns.

Goldman suggests that the reason we shouldn't say that Henry knows there is a barn there, is that in this district just looking at a barn from a car is not a way of finding out whether there is a barn there. For, in these special circumstances, just looking out of your car window will lead you to believe that there is a barn on many occasions when there isn't one. Just looking out of your car window is, in these circumstances, an unreliable way of acquiring the belief that there is a barn.

What this story suggests is that the appropriate way of getting a true belief, if you want to have knowledge, is to get it by a method that is reliable in the circumstances. One form of causal theory, then, says that knowledge is true belief produced by a means that is reliable in the circumstances. A view that replaces the phenomeno­logical justification condition with an objective reliability condition, such as this one, is a form of reliabilism. Different forms of relia- bilism spell out different ways in which the belief-forming process must be reliable for the resultant belief to count as knowledge if it is true.

Notice that this theory explains why Smith didn't know that the man who would get the job had ten coins in his pocket and, more generally, why the PDJ is wrong. For Smith came to believe

E: The man who will get the job has ten coins in his pocket

by deducing it from

D: Jones is the man who will get the job, and Jones has ten coins in his pocket.

But, in these circumstances, this was not a reliable way of coming to believe E. For if Smith himself had not happened, quite by chance, to have ten coins in his pocket, E would have been false. We cannot accept the PDJ, because in many circumstances, like this one, deducing a consequence will cause you to have a true belief only by the merest chance. That is possible because you can draw a true consequence from a false assumption, a fact we shall discuss in the next chapter.

2.8