Scepticism

1 What is knowledge?

So far we have been doing metaphysics. We have been concerned with problems concerning what there is. In this next part of the course we will be concentrating on epistemology: on problems concerning what, if anything, we can know.

Most of us do take ourselves to know something about the world. In fact, most of us take ourselves to know an awful lot about the world. If we have that belief about ourselves then we must also have a grasp of what knowledge amounts to. What, then, is knowledge?

What we are going to be particularly interested in here is propositional knowledge: knowledge that such-and-such is the case. For example, I now take myself to know that I am word-processing a webpage somewhat after my deadline for doing so. You take yourself to know that you are reading webpage. In each case what we take ourselves to know is something that may or may not be the case. What we take ourselves to know are propositions about the world. These are typical cases of knowledge-that.

There are other kinds of knowledge. For example, I know how to operate this word processor. I also know how to ride a bicycle and how to ski. Such knowledge amounts to the possession of various skills and abilities. It is usually called knowledge-how (or know-how for short). The problem of knowledge we are mostly interested in in this segment of the course is the problem of knowledge-that: do we have any genuine knowledge-that and if so what does it amount to? This is not to say that knowledge-how is not interesting—on the contrary it is very interesting. (Indeed, we have already been investigating aspects of knowledge-how in the problem of freedom and determinism.) But let's keep to one problem at a time.

What, then, does it take for someone to know that something is the case? What do we mean when we say for example, that Norman knows that it is raining?

1.1 Knowledge and belief

It would be very strange if someone were to affirm both that Norman knows that he will win lotto this week and that Norman does not believe that he will win lotto this week. Norman cannot know something without believing it. Knowledge implies belief. We can state this in a very general way. Let X be any being that might possess knowledge, and let P be any proposition that such a being might know. Then we can say, quite generally, that for X to know that P it must be true that:

(1) X believes that p

But belief is not the same thing as knowledge. Norman can believe that he will win lotto this week without Norman knowing that he will win lotto this week. People can have beliefs which do not amount to real knowledge. What more must a belief have before it amounts to knowledge?

1.2 Knowledge and truth

Suppose Norman has just bought a lotto ticket and, wishful thinker that he is, believes that this time he will win. But he is wrong. Then clearly Norman does not know that that he will win lotto this week. In order for a belief to count as knowledge the belief must be true. If Norman says "I know that I will win lotto this week" and it turns out that he doesn’t, then Norman is not only wrong about lotto: he is also wrong to claim that he knows that he will win lotto. A claim to knowledge is shown to be quite wrong if the proposition that is claimed to be known turns out not to be true. So we have a second condition on knowledge: for X to know that P it must be that:

(2) P is true.

Still, (1) and (2) together are not sufficient for defining knowledge. Suppose that by a stroke of sheer luck, it turns out that Norman does win lotto. His belief was true. So Norman believes that he will win lotto and it is true that he will win lotto: but we would be loath to say that Norman's true belief about beimg the winner amounts to knowledge. If this is right then there must be more to knowledge than merely having true beliefs.

1.3 Knowledge and justification

Norman's true belief about winning is just a lucky guess. He has no real justification for his belief that he will win, even if, as it happens, he will win. That is why he does not possess real knowledge in this case. It was just a lucky guess. It seems that for X to know that P it must be that:

(3) X 's belief that P is justified.

For X 's belief that P to be justified X must have some good reason for believing P to be true. Without such a reason it would seem that, even if the belief is true, it is just a lucky guess—like Norman's belief concerning lotto.

1.4 Does knowledge = justified true belief?

So far we have three conditions which are necessary for X to know that P. That is, if X knows that P, then (1) X believes P, (2) P is true, and (3) X's belief that P is justified. Most of what we will say in this course depends only on the fact that these are necessary conditions on knowledge. (One cannot be said to know unless all three conditions are satisfied.) In particular, the central problem of knowledge, of how it is that we can know anything at all, arises from these three conditions.

Some philosophers think that these three conditions are also sufficient for knowledge. That is, if X believes P, P is true, and X 's belief that P is justified, then X knows that P. This is called (for obvious reasons) the JTB account of knowledge. (There are some good arguments against the JTB account.)

2 What is the problem?

2.1 Scepticism

The problem of knowledge is basically this. There are various sceptical arguments which seek to show that we do not, or can not, possess knowledge of various sorts, or knowledge in general. Can these sceptical arguments be defeated?

One can be a sceptic with regard to any area of purported knowledge. For example, one can be a sceptic with regard to the senses, or perceptual knowledge. A sceptic about perceptual knowledge maintains that we cannot gain real knowledge from sense experience (sights, sounds, tastes, smells, feels). For a presentation of a variety of such argument see the extract from Descartes. A more radical sceptic might maintain that we cannot gain any knowledge at all. I will present a very powerful regress argument for complete scepticism. Then I will show how different epistemologies (theories of knowledge) can be treated as attempts to deal with this single argument. One of those responses is empiricism: the doctrine that all real knowledge is gained, ultimately, from our sense experience. Finally we will turn to a third kind of sceptical argument, concerning the justification of induction.

2.2 A sceptical regress

Suppose I claim to know that P. Then I must believe P, P must be true, and I must be justified in believing P. To be justified in believing that P I must have a good reason for believing that P. Now, the key to this sceptical argument is the following general principle, which we will call :

THE PRINCIPLE OF NOT BEGGING THE QUESTION:

P is never a good reason for P itself.

For example, suppose we are with Norman when he announces that he will win lotto this week, and we ask him, "Norman, why do you believe that you will win lotto? What reason do you have?" Suppose he just insists "Because I am going to win, damn it!". We would be less than impressed. We would want Norman to produce some evidence for his claim, not just reiterate the claim itself. To reiterate the claim would be to simply beg the question.
So for it to be true that I know P, I must have a good reason for believing P, and that reason cannot just be P itself (that would be begging the question). Suppose that reason is Q. Now Q is not much good to me unless I know that Q. We will call this the principle of

THE ACCESSIBILITY OF GOOD REASONS:

Q is not a good reason for me to believe P unless I know that Q is the case.

Now, for me to know that Q all three JTP conditions must be satisfied. In particular, I must be justified in believing that Q. So:

knowledge that P -> justification for P -> knowledge that Q -> a justification for Q ...

By now you will see the pattern emerging. To be justified that Q I must have some reason for believing Q which supports Q. The reason for believing Q cannot be Q itself (that would be begging the question) and so I must know some further proposition R. To know R I must be justified in believing R .....etc etc.
It seems that in order to know just one thing I have to have an infinite chain of supporting evidence. But since I never have such evidence, it follows that I never really know anything. Simply put, the sceptic just keeps on asking "How do you know that?" whenever an answer is given to that very same sceptical question. The sceptical regress is thus a sophisticated version of the child's reiterated, and very irritating, "why?"