Descartes' way: Justification requires certainty

To see how we might get to the first way of interpreting the justifi­cation condition—Descartes' way—let's start by examining more precisely what it means to say that the evidence my lawyers present in the hypothetical case we have been considering is consistent with my being guilty.

One way of putting more precisely what I mean by saying the evi­dence is consistent with my being guilty is this:

lang=EN-US style='font-size: 10.0pt;line-height:112%'>a)     there is a true sentence (call it “T”) that reports all the evidence, and

b) T is consistent with a sentence that says I am guilty.

Two sentences are consistent just in case it is possible for them to be true at the same time.

Suppose, then, that we have a sentence, and we're looking at the evidence for it. Let's call the sentence that reports the evidence the “evidence-sentence,” and the sentence for which it is evidence “S.” What we mean, then, by the evidence being consistent with the sentence S being false is that it is possible that the evidence-sen­tence should be true and S should be false at the same time.

Nevertheless, “John is crying and looking downcast” is good evi­dence that John is unhappy. Evidence like this, which is consistent with the falsity of the sentence it supports, is called “defeasible” evidence. (“Defeasible” because it could be defeated by later evi­dence that undermined it.) If, on the other hand, you have evidence for the truth of a sentence, S, that is so good that it is not possible that S should be false when the evidence-sentence is true, then you have what we call “indefeasible” evidence for S. The evidence-sen­tence “It looks red to me,” for example, if true, would be taken by Descartes (and most people) as indefeasible evidence for the sen­tence “I am having a visual experience.”

The jury in my story plainly did not have indefeasible evidence that I was innocent: for, as I said, the evidence was consistent with my being guilty. One possible view, then, would be that what the jury in my story lacked was indefeasible evidence and that, if they had had that, they would have had knowledge. The justification con­dition for knowledge, on this view, means that you must have evi­dence that justifies your belief indefeasibly.

This was, as I say, essentially Descartes' view. Descartes didn't know much about how brains work. But he got to this conclusion by considering problems very much like the one raised by Marie, the unscrupulous neurosurgeon, with which I began. One problem he raised was how we could know that all our experiences were not just a dream. In many ways this is just like asking how we know that we are not Marie's victims.

Here are two passages where Descartes first faces the possibility of the evil demon, and then considers how to respond to it.

I will suppose therefore that there is not a true God, who is the sovereign source of truth, but a certain evil spirit, who is no less devious and deceitful than he is powerful, and that he has set about with all his ingenuity to deceive me. I will imagine that the sky, the earth, and the colors, shapes, sounds, and all external objects that we see, are nothing but illusions and tricks, which he uses to entrap my credulity.

But were I persuaded that there was nothing at all in the world, that there was no sky, no earth, no minds, no bodies; would I also be persuaded that I did not exist? No, surely, I would exist, without doubt, if I was persuaded, or even if I thought anything. “But there is some unknown deceiver, who is very power­ful and very devious, who is using all his ingenuity to deceive me.” Then there is no doubt at all that I exist, if he is deceiving me; and were he to deceive me as much as he wishes, he would never be able to make it true that I am noth­ing, so long as I am thinking that I am something. The result is that after hav­ing thought precisely about it and having carefully examined all things, in the end one must conclude, and hold as sound that this proposition: “I am,” “I exist,” is necessarily true on all occasions that I utter it or that I conceive it in my mind.

This is a very persuasive argument: it is, indeed, one of the most famous arguments in the history of philosophy.

Descartes thought he could escape the demon's tricks if he could find other beliefs that were as certain and indubitable for him as his own existence—the “I am”—and the fact that he had thoughts—the “I think.” So long as he had any such certain and indubitable beliefs at all, he could claim these beliefs as knowledge, however hard the demon tried to confuse him.

Descartes, then, suggested that the right way to explain the justi­fication condition was to insist that the evidence you possessed enti­tled you to be certain of what you believed. And by “certain” he meant that it had to be impossible to doubt it. This, after all, is a nat­ural extension of the idea that we express by asking people who think they know something, “But are you sure?” We want them to consider whether they really have no doubt at all that they are right.

It is only a short step from insisting that a belief that is to count as knowledge must be impossible to doubt to insisting that you must have indefeasible evidence for the belief. For if it is impossible for you to doubt S, then you must have evidence that couldn't be true unless S was true. And I defined “indefeasible evidence” as evidence for the truth of a sentence, S, that is so good that it is not possible that S should be false when the evidence-sentence is true. So Descartes is committed to the view that to know something you must have indefeasible evidence for it—or, equivalently, that your evidence must make the belief indubitable.

a) you must believe it,

b) it must be true, and

c) you must have indefeasible evidence for the belief.

Descartes' view has one surprising immediate consequence. Some sentences—such as “Nothing is both in New York and not in New York at the same time”—couldn't be false, and they are called “necessary truths.” It turns out that, given the way indefeasible evidence is defined, any sentence at all is indefeasible evidence for a necessary truth. Take a sentence, S, which is a necessary truth. By definition, it can't be false. Indefeasible evidence for S is defined as evidence that couldn't be true if S were false. Consider any other sentence at all; say, T. It certainly isn't possible for S to be false if T is true. For it isn't possible for S to be false under any circum­stances. So you have indefeasible evidence for any sentence that is a necessary truth, provided you believe anything at all!

It follows, of course, that, on Descartes' view, we know any nec­essary truths we believe. For necessary truths are, by definition, true under any circumstances, and, as we have seen, we automatically have indefeasible evidence for them.

As far as necessary truths are concerned, then, Descartes' theory is very permissive. The difficulty with the theory is that it is, by con­trast, very demanding when it comes to beliefs about the physical world. Indeed, it is so demanding that it is hard to think of any beliefs about physical objects that Descartes could claim to know. For, after all, as the story of Marie and Albert showed—as Descartes' own story of the demon shows—the evidence we actually have is consistent with our being wrong about almost everything we believe, except (as Descartes saw) what we believe about our own existence and our own thoughts.

Descartes thought he could escape the skeptical consequences of his definition of knowledge. His way of avoiding these consequences depends on the belief that there is an omnipotent, benevolent God who does not want us to be deceived. It is important to state as clearly as possible why this helps, because it allows us to make explicit one of Descartes' assumptions about the way we ought to seek justification for our beliefs. That assumption, as we shall see, is crucial to many philosophical views about justification.

But before we go any further, we must notice another of Descartes' assumptions. Descartes thought that we could not be wrong about the contents of our own minds. He thought, for exam­ple, that if I think I am now thinking about oranges, then I must, in fact, be thinking about oranges. It is worth asking whether Descartes is right about this. For it might seem that sometimes, in fact, we make mistakes about what we are thinking. Certainly, it does not follow from the cogito argument alone. From the fact that, if I am thinking, I must exist it doesn't follow that I can't be wrong about what I'm thinking; it follows only that I can't be wrong in thinking that I exist. Nevertheless, there is at least some plausibility to the thought that I can't be wrong about the contents of my own mind, and many philosophers of his day thought that this was so.

Now, suppose I have a sensory experience that I can describe by saying:

E: It looks to me as though there is a book in front of me.

I call this sentence “E”—for “evidence.” Since E is about my own mind, Descartes will allow that I can know it to be true: according to him, as I have just pointed out, I can have indefeasible evidence of my own state of mind.

But how can I come to know, on the basis of this state of mind, that there is, in fact, a book in front of me? Descartes says that if God is both benevolent and all-powerful, then He can make sure that the experiences we have correspond with the way the world really is. But even if my experience in fact corresponds with reality, because God has guaranteed it, I cannot know that it does unless I have indefeasible evidence. Suppose, however, that I have indefea­sible evidence that God guarantees that sensory experience corre­sponds to how the world is. Then I know that if it looks, sounds, or, in general, seems to me that something is so, it is so. And so I know, in particular, that

R: If it looks to me as though there is a book in front of me, then there is a book in front of me.

Now from the two sentences, R and E, it follows logically that there is a book in front of me. (We shall discuss what it means for some­thing to follow logically in the next chapter; see 3.10.) Furthermore, I know, according to Descartes, that both R and E are true. Suppose that if something you believe follows logically from two things you know, then you know it, too. If that were true, Descartes could say that I knew that there was a book in front of me.

Descartes' claim that God's guarantee of our senses can form the basis of knowledge will be correct, therefore, if both

a) we know about God's guarantee, and

b)     the following principle is correct: for any two sentences, A and B, if you know A and know B, and if from A and B, together, C follows logically, then if you believe C, you know C.

This principle is usually called the “deductive closure principle.” For it says that the class of things you know includes all your beliefs that are logical (or “deductive”) consequences of everything you know already.

Notice that the deductive closure principle is really a consequence of Descartes' definition of knowledge. For, on Descartes' theory, if you know both A and B, then it is true of each sentence that

a) you believe it,

b) it is true, and

c)  you have indefeasible evidence for it.

Suppose you believe C, which follows logically from A and B. Since you do know A and B, it follows that your belief in C is true. (Here's the argument: If a conclusion follows logically from some assump­tions, then the conclusion will be true if the assumptions are. From (b), it follows that if you know A and B, then A and B are both true. As I just said, if C follows logically from A and B, then C is true if they both are. So if you know A and B and if C follows from them, then C is true.) That gives us conditions (a) and (b) for your belief C. So you know C, provided the justification condition (c) is satisfied as well. Does your knowing A and B mean you have indefeasible evi­dence for C, which follows from them? Obviously. For if C follows from A and B, then the evidence-sentence that makes A and B true makes C true as well. (Here's the argument: Suppose E is the inde­feasible evidence for A and E' is the indefeasible evidence for B. Then (E & E') is indefeasible evidence for (A & B). That just means that if (E & E') is true, then (A & B) must be. But if (A & B) must be true then C must be true, too, because it follows from (A & B). So, if (E & E') is true, C must be true. Which means that (E & E') is indefeasible evidence for C.) So the deductive closure principle is correct.

The core of the argument here is expressed in the following principle:

PDJ: If you take any two sentences, A and B, then, if you are justified in believing both A and B, and if from A and B together, C follows logically, then, if you believe C, you are justified in believing C.

The American philosopher Irving Thalberg has called this the “prin­ciple of deduction for justification” (PDJ, for short.) The PDJ is certainly correct if justification means “indefeasible justification.” And, as we just saw, given the PDJ and Descartes' definition of knowledge, the deductive closure principle follows.

Descartes requires the deductive closure principle because, with­out it, even the existence of a benevolent God, attempting to do the opposite of the evil demon, would not allow us knowledge of the world. With both the principle and the knowledge that God guaran­tees that our senses will not deceive us, however, Descartes is able to allow that we have some knowledge of the physical world.

But there is a serious problem with the Cartesian position. It is that Descartes offers no convincing reason for thinking that we know that God guarantees the evidence of our senses. After all, it seems that our senses can sometimes deceive us: sometimes we seem to have hallucinations. And if we sometimes have hallucina­tions, then God doesn't always guarantee that the world is as it appears to be.

It won't help here to say that God sometimes makes sure our senses don't deceive us, because to know anything, on Descartes' view, we would have to know when. Descartes was aware of this problem, and he proposed a solution to it. His idea was that God had given us a way of telling which of our ideas were in fact reliable. For he argued that we would never go wrong if we believed only those ideas that were “clear and distinct.” But it is far from clear that we do in fact have a way of telling, from the character of our expe­riences, whether or not they are reliable, and Descartes' notion of “clear and distinct” ideas is not, in the opinion of many philosophers, a satisfactory solution to this problem. If they are right, then we do not, in fact, have a God-given guarantee that some of the evidence of our senses is correct.

Unless we know that God guarantees at least some of what our senses lead us to believe, then we don't have any indefeasible true beliefs about the physical world. So we know nothing about it. Still, as we saw earlier, we do have some knowledge, since we know any necessary truths we believe. The real reason that Descartes thought we knew necessary truths is that we do not need evidence from our senses to justify belief in them at all. His theory leads to skepticism about the physical world because all the evidence of our senses is defeasible. But we can work out necessary truths without relying on our unreliable senses.

Because Cartesianism lays such stress on certainty, it leads to the conclusion that we know only those things that we can work out by reasoning, without appeal to sensory evidence, even though Descartes tried to avoid this consequence. The position that the most significant elements of what we know are derived by reasoning rather than experience is called “rationalism.” We shall discuss the nature of necessary truths in the next chapter, where we shall see that the rationalist belief that all our knowledge of necessary truths comes solely from reasoning alone is mistaken.

The main objection to Cartesian rationalism, however, is that it leads to skepticism about the physical world. Isn't it just absurd— the worst sort of philosopher's nonsense—to claim that we don't know of the existence of any physical objects at all? The British philosopher G. E. Moore once held up his hands in an expression of exasperation with those who deny the existence of the “external world,” the world “outside” our minds, and said that he certainly knew that his hands existed. He was, in effect, assuming that we should reject a theory that had so absurd a consequence as that he didn't know he had two hands. Very often in philosophy, we argue against a position by showing that it has absurd consequences: a pro­cedure called reductio ad absurdum (or reductio, for short), which is just the Latin for “reducing to absurdity.” Moore's point was that we should reject a philosophical theory of knowledge that leads us to conclude that we do not know that our own hands exist. We should reject such a theory because this consequence reduces it to absurdity.

It is important in a reductio proof that the consequence we draw should not merely strike us as absurd but actually be false. We shall discuss in the next chapter the fact that if you can draw a false con­clusion from a position, the position must be false itself. Because it is the falsity of the conclusion that means that the position must be false, we sometimes refer to an argument as a reductio simply because it shows that a position leads to (what we believe is) a false conclusion.

There is no doubt that we have to be very careful with reductio ad absurdum as a form of argument. This is because it is not always clear that what we take to be absurd really is false. For a long time, for example, it might have been thought absurd to draw the conclu-

sion that God doesn't exist. Nowadays, even many believers agree that it is not absurd to suppose that there is no God (though, of course, they think that it is an error to believe this). So before we reject Descartes' position in Moore's way, we should consider seri­ously the possibility that it is not false that we know nothing of the external world.

But we have at least one strong motive for rejecting Descartes' extremely strict interpretation of the justification condition, if it does have the consequence that we know only of the existence of our own thoughts; namely, that a theory of knowledge that says that we can know nothing about the world in which we live makes the concept of knowledge rather uninteresting. We certainly have beliefs about the world, and some of them seem better justified than others. Even if knowledge is unavailable, we should still need the idea of justified beliefs. And whatever “justified” means, it cannot mean “indefeasibly justified” in this context, because, as we have seen, no beliefs about the physical world are indefeasibly justified.

We have, then, good reason for hoping that Descartes is wrong to insist on indefeasible justification, because this theory of knowledge leads to skepticism. But we may be able to develop a theory of knowledge that does not lead to skepticism if we find another way of interpreting the justification condition. Is there any way of inter­preting the condition that is less demanding?

2.4