De la distinción entre "estar cierto" y "saber"

Abstract

What kind of relationship is there between “knowing” and “believing”? Does “knowing” imply “believing”? Some of the difficulties arising from these questions could be dispelled if we first ask in which cases it is relevant to introduce the distinction between “knowing” and “believing”. This article is an attempt to answer this question; in doing so, it advances a definition of “knowing” independent of its relation to “believing”.
1. Let’s start by considering an imaginary situation in which it is impossible to refer to someone else’s knowing or believing, or to my own knowing or believing in the future or past. Let us suppose that those verbs are meaningful only when used in the present tense and in the first person. Then the statement “I believe that p” would exclude the possibility of asserting at the same time “I am wrong (now) about p” but it would not exclude the possibility of asserting “I might (now) be wrong”. One can distinguish two different senses of “I believe”: strong when it is used in such a way that it would be contradictory to say “I believe that p but I may be (now) wrong about p”; and weak when the same expression would not be contradictory. In the first sense it could be replaced by “I am sure”, “I am certain”; in the latter by expressions such as “I suppose”, “I would say”, “I think that”.
When “believing” is referred to in the first person singular, two kinds of beliefs can be distinguished: on the one hand, beliefs which are not founded upon good reasons, that is to say, those founded upon insufficient reasons to exclude the possibility of being wrong (now); on the other hand, there are beliefs based on reasons which are sufficient not only to assert something but also to exclude its negation. “I believe that p”, in the strong sense of “believing”, means “I’ve got sufficient reasons to assert that p and to deny that not-p”.
When “believing” is used in the first person singular and in the present tense it is not only possible but useful to distinguish between its strong and its weak senses. Is it also possible to distinguish between those two senses of “I believe” and the meaning of “I know”?
2. When “I believe” is used in its strong sense, it would be contradictory to say “I believe that p but I do not know whether p or not-p”, I cannot be sure that p and at the same time accept the possibility of p’s being false. On the other hand, if I use “I believe” in its weak sense it would not be contradictory to say “I believe that p, but I don’t know that p”. In this latter case “I don’t know” could be replaced by “I am not sure”. “I don’t know” adds nothing to “I am not sure”. Thus, when “knowing” and “believing” are used in the first person singular and in the present tense, it is possible to distinguish between “I believe” and “I know” in the case where “I know” has the same meaning as “I am sure” and “I believe” is used in its weak sense. We have not yet reached the point where it is necessary to distinguish between “believing” in either of its two senses, and “knowing” in a sense that could not be reduced to a kind of belief.
3. Consider now the use of “believing” and “knowing” in the second or third person. Let us imagine a person A who has certain beliefs and another person B who examines A’s beliefs. It would not make any difference, for the present case, if A and B were the same person at two different times.
First suppose that both A and B are sure that p. It would be contradictory for both of them to say “I am sure that p but I don’t know whether p or not-p”. However, B realizes that it is not contradictory to say “A is sure that p but he doesn’t know that p”. Therefore, “knowing” has a different meaning when it is used by B to refer to p asserted by himself than when it is used to refer top asserted by A. It is only then when the need to distinguish a certain sense of “knowing” different from the sense of “being sure” appears. Why is this so? Because A’s reasons to assert p could be considered by B insufficient to exclude not-p. B is able to distinguish between “being sure” and “knowing” only in the case where he can assert at the same time: a) A considers his reasons sufficient to exclude not-p, and b) However, A’s reasons are not sufficient.
Consider now a second case: B does not share A’s belief in p. Then we have:

(1) For A: (a) A believes that p,
(b) A has reasons to assert that p,
(c) Those reasons are sufficient to assert that p,
(2) For B: (a) A believes that p,
(b) A considers himself to have reasons to assert that p,
(c) A thinks that those are sufficient to assert that p,
(d) Those reasons are not, in fact, sufficient.

Then B can say that A is sure but that he does not know.
Let’s refer to as first order statements those statements which refer to the reasons to assert p (lb, lc, 2d in our preceding scheme), and second order statements those referring to first order statements (2b, 2c). The distinction between “being sure” and “knowing” is relevant only when the truth value of a first order statement and that of its corresponding second order statement may be different. In the case we have just considered, the second order statement “A thinks he has sufficient reasons” is true for B, and the first order statement “A has sufficient reasons” is false. That is why B may say that A is sure but that (in fact) A does not know. On the contrary it is not possible for a first order statement and its corresponding second order statement to have a different truth value when used in the present tense and first person. That is why it is not possible to distinguish between “I am sure” and “I know”.
If A is sure that p, A considers his reasons sufficient to assert p, but B may well ask: “Are these reasons in fact sufficient?” What B means by this question is whether any person in the position to evaluate those reasons, capable of doing so and willing to examine them carefully, would also feel compelled to accept them and then to assert p. For a reason to be sufficient it must be possible for anybody to accept it. Only in the case where B verifies that A’s reasons are sufficient not only from A’s point of view, but independently of A, can B then say that A knows. The distinction between “knowing” and “believing” is appropriate where it is possible to evaluate the sufficiency of someone’s reasons for asserting a proposition, independently of the judgement of the person who asserts it at the moment of the assertion.
Then, “A is sure that p” (when asserted by B) means:

(a) A believes that p,
(b) A considers himself to have reasons to assert that p,
(c) A considers that those reasons are sufficient.

“A knows that p” (when asserted by B) means:

(a) A (in fact) has reasons to assert that p,
(b) Those reasons are (in fact) sufficient,

In our definition of “knowing” there is no reason to take into account any statement about A’s dispositions and statements, In fact, if B states that A knows, this means that A has reasons, which can be tested by anyone, to assert p, and this leaves aside the questions of whether A considers his reasons to be sufficient and of his being in a disposition to assert p (his believing that p).
4. Let’s now go back to the case of the first person. If B thinks that A knows that p, wouldn’t he be in a position to think “I know that A knows that p”, or even “I know that p”? The answer is: only in the case where “I know” had the same meaning of “I am sure”. If A knows that p, then A’s reasons to hold p must be sufficient, but from this it does not follow that B’s reasons to assert “A knows that p” must be equally sufficient; what really follows is that B considers his own reasons to be sufficient. Therefore, when B verifies A’s knowledge, he cannot say “I also know that A knows”, but only “I am sure that A knows”.
However, there is a case where one can say “I am not only sure, I know it also”. This would be the case when one says to someone else: “Here is the evidence of what I assert, you can verify it yourself, thus you must accept that I really know”. The sign that I know is that my reasons can be verified by anyone, not only by myself. In this case I can distinguish between “I am sure” and “I know” because I am considering my reasons as another person would consider them. I am using “I know” in the same way I usually use “he knows” or “I knew”. I apply to myself that sense of the verb which it usually acquires when applied to others.
5. Descartes’ reasoning can be adduced as an illustration of the central point of this paper. In a solipsistic situation, Descartes’ “knowing” cannot be distinguished from what he thinks (or believes) he knows. It is not until he offers a demonstration of another’s (God’s) existence, that he can make the distinction.
6. From what has been said, the insufficiency of the standard analysis of “knowing” can be inferred. The standard procedure holds that knowing should be analysed into three components. “A knows that p” would mean: (a) A believes that p, (b) p is true, and (c) A has sufficient reasons (evidence) to assert that p.
First of all, note that (b) could be omitted. For, if (b) is not analitically implied by (c), then (c) must mean “A considers himself to have sufficient reasons...” But this is not enough to assert that A knows. B could demonstrate that the reasons which A considers sufficient to assert p are not actually so (even though p may be true). In this case, B could only say that A believes he knows that p, but B could not properly say that A knows that p. In order for it to be proper to say this, B must admit that A’s reasons are in fact sufficient, and in asserting this he is implicitly asserting that p is true.
On the other hand: Who is the one who asserts that p is true? If it is A himself, then the statement “p is true” means the same as statement (c). If it is stated by someone else, the statement “p is true” only means that the statement (c) must be understood as “A has in fact sufficient reasons...” and it would not mean anything else.
Secondly, in the analysis of “A knows that p”, one must not include “A believes that p”, since we have already shown that the distinction between “knowing” and “believing” is only relevant when someone wants to point out that the reasons for asserting a proposition are sufficient, independently of the dispositions or considerations of the person asserting it.
Of course, there remains the problem of whether, in fact, somebody can know that p without believing that p. But this is a psychological problem, not a philosophical one. For it is different to ask whether the concept of “knowing” logically implies the concept of “believing” than to ask whether the activity of knowing without a mental state of believing is possible. The former is a philosophical question which can be answered after a conceptual analysis, the latter is a psychological question which should be resolved empirically.
When the definition of “knowing that p” is reduced to “having in fact sufficient reasons to assert that p and to deny that not-p”, the dependence of knowing upon any mental state of the subject is definitively ruled out. The problem is reduced to that of stating the criteria necessary to determine when it can be said that a certain reason (or piece of evidence) is in fact sufficient to assert a proposition.