Unger on Knowledge

Citation preview

Unger on Knowledge Gilbert Harman The Journal of Philosophy, Vol. 64, No. 12. (Jun. 22, 1967), pp. 390-395. Stable URL: http://links.jstor.org/sici?sici=0022-362X%2819670622%2964%3A12%3C390%3AUOK%3E2.0.CO%3B2-V The Journal of Philosophy is currently published by Journal of Philosophy, Inc..

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/jphil.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Sat May 12 00:34:35 2007

THE JOURNAL OF PHILOSOPHY

COMMENTS AND CRITICISM UNGER ON KNOWLEDGE

P

ETER UNGER has recently attempted to refute the empiricist view that knowledge of matters of fact must be based on experience one has or has had.* I accept such an empiricist view and would like to explain why I am not convinced by Unger's attempted refutation of it. On my view whenever one comes to know something one's knowledge is based on inference (in a technical sense, as when one speaks of inferences made by a computer). T h e premises of such an inference either express one's prior knowledge or describe one's present experience. Since I also think one knows only if one has at some time come to know, I hold that knowledge must be based on experience one has or has had. I have described and defended this view elsewhere.=Here I shall restrict myself to Unger's arguments against this and similar views. His main argument consists in the construction of a purported counterexample to the empiricist thesis. He imagines that scientists have deliberately constructed an exact duplicate of some ordinary person and claims that such a duplicate will have much of the knowledge that the original person has; e.g., the duplicate may know that there are more people in China than there are in India, although the duplicate's knowledge is not based on experience he has or has had. Faced with this example, one is inclined simply to deny that such a duplicate has the knowledge Unger supposes him to have, but one is also tempted to wonder whether perhaps the duplicate can be said to know those things. I shall argue that the example does bring out something interesting about our use of the words 'know' and 'knowledge' but that it does not refute the empiricist thesis. T h e empiricist can agree that we might say that the duplicate has the knowledge Unger ascribes to it; but he will deny that we can say this, strictly speaking. Unger claims that his duplicate illustrates what he takes to be a necessary condition of a man's belief being an instance of factual knowledge:

* Peter Unger, "Experience and Factual Knowledge," this J O U R N ~ Z L , LXIV, 5 (March 16,1967): 152-173. 1 "How Belief Is Based on Inference," this JOURNAL, LXI, 12 (June 11, 1964): 353359; "The Inference to the Best Explanation," Philosophical Review, LXXIV, 1 (January 1965): 88-95; "Lehrer on Knowledge," this JOURNAL, LXIII,9 (April 28, 1966): 241-247; "Knowledge, Inference, and Explanation,'' forthcoming in A~neiican Philosophical Quarterly.

COMMENTS AND CRITICISRI

39I

(12.1) For any sentential value of p, a man's belief that p is an instance of knowledge only if it is not an accident that the man's belief is true (1 72).

Because of the care and attention of the scientists who made the duplicate, it is not an accident that certain of the duplicate's beliefs are true; and this may encourage us in the thought that the duplicate has knowledge that is not based on his own present or past immediate experience. However (12.1) does not appear to be correct as it stands, if it is taken to provide a necessary condition on knowledge. If it is an accident that a man knows that p and otherwise he would have believed the opposite, then it is an accident that the man's belief is true. This general sort of counterexample is not met by Unger's remark, "It may be an accident that p, and a man may still know that p," which he takes to be consistent with (12.1). For it may be an accident that p and not an accident that a man knows that p. Part of the trouble with (12.1) is that what counts as an accident depends upon what considerations one has in mind as well as upon how one describes what is to count as the accident. "It is not an accident t h a t . . ." is context-dependent in a way that one would expect given the close relationship (not of course exact identity) between "not an accident" and "necessary." T h e following things all might be said about the same belief: "Given that it's knowledge, it's no accident that the man's belief is true." "Given that he was about to believe the opposite, it is an accident that his belief is true." "Given that he reasoned badly, it is an accident that his belief is true; but given that we carefully selected as our example someone whose belief about this subject is true, it is not an accident that his belief is true." If (12.1) were right we ought to be able to say of a particular person such things as, "Given that he reasoned badly, he does not know; but, given that we carefully selected as our example someone whose belief about this subject is true, he does kno~v."But it cannot be true of a particular person that relative to certain considerations he knows but also that relative to others he does not know, at least in the way that it can be true of him that relative to certain considerations it is no accident that his belief is true whereas relative to other considerations it is an accident that his belief is true. Therefore (12.1) does not offer any support ior Unger's claim that his duplicate knows things. T h e plausibility of (12.1) lies entirely in the fact that one can know only what is true. But because of its context-dependent character, nonaccidentality of the truth of a belief cannot be taken to provide a condition of knowledge.

392

THE JOURNAL OF PHILOSOPHY

A more direct argument against empiricism is provided by Unger's example of the man "who is a total personal amnesiac but whose memory is otherwise quite all right" (163). Unger points out correctly that we would take such a person to know many things of a general sort; and it would seem that this knowledge can be based neither on his present experience nor on his past remembered experience. This does not refute the view put forward above that whenever one comes to know something one's belief must be based on inference from prior knowledge and/or present experience. However, it may seem to refute those philosophers who claim that present knowledge must be warranted or warrantable in terms of present or available experience. Whether it does in fact refute these philosophers is a point I shall return to in a moment. Consideration of the partial amnesiac leads Unger to accept the following general principle:

p is a n instance of knowledge a n d if the man retains his belief that p from t through a later time t' only because h e remembers that p from t through t', then a t t', the m a n knows that p (168).

(10.2) If a t a time t a man's belief that

One might argue that this principle is trivially true, on the grounds that whenever one remembers that p one knows that p. But the way Unger uses (10.2) suggests that, for him, to remember that p is simply to continue to believe that p as the result of the normal operation of memory. So interpreted, (10.2) is only approximately correct; and its failure to hold universally is relevant to Unger's claim about accidentality and knowledge and, therefore, also relevant to his views about the duplicate. Suppose that a man comes to know that p but then is presented with faked evidence that ought to cast doubt on his reasons for believing that p. Suppose, however, that the man ignores this evidence and continues to believe that p. In such a case he could not be said to know that p, although according to (10.2) he should know that p. For example, he comes to know that p by reading it in the newspaper. Then we come by and present him with strong but faked evidence that the newspaper he read was specially printed by us to fool him. He wants what he read to be true, however, and ignores the faked evidence we present. At this point, he does not know, although he should know according to (10.2). This case shows that the experience a person has can be relevant not just to whether he comes to know what he comes to believe but also to whether he continues to know what he continues to b e l i e ~ e . ~ Here perhaps I oversimplify; for it is clear from Unger's discussion (10.2) also fails because of what I have elsewhere called the "social aspect of knowledge"; cf. "Knowledge, Inference, and Explanation."

C O M M E N T S AND CKITICISM

393

of the amnesiac and of another example, the gypsy discussed below, that he will deny that the man in my example fails to know. I am going to suggest that an ambiguity is involved here; but for the moment I restrict myself to pointing out that it is quite natural to say of the man who ignores relevant evidence in my example that he does not know that p. I can hint paradoxically at why I think Unger would describe this case differently, as follows: T h e man who ignores the faked evidence does not lose his knowledge that p, but he no longer knows that p. I have also oversimplified Unger's amnesiac case; and now I want to consider it more carefully. As I have described the case, a man has lost all memory of his personal experience but still has lots of general knowledge. One might take such a case to refute any empiricism that holds that present knowledge must be based on present or available experience, for there does not seem to be any present or available experience on which this man's knowledge could be based. But some empiricists count the experience of being aware that one believes that p to count as a prima facie reason to (continue to) believe that p. As I have described the amnesiac, he does not provide a counterexample to such an empiricist.3 Nor is Unger's description of the amnesiac case directly relevant here. (His gypsy case does offer a purported counterexample to this kind of empiricism.) However, Unger's description is indirectly relevant. H e points out that many empiricists believe or ought to believe that, when one knows by virtue of certain reasons, then those reasons must be true. So he imagines that his amnesiac has come to believe that he is a recently constructed duplicate and, furthermore, that the amnesiac bases his beliefs in the various general propositions he still remembers on his belief that scientists constructed him as an exact duplicate of someone who had considerable general knowledge. According to Unger, the amnesiac continues to know various general truths even though his belief in these truths is based on reasons that are false; so the empiricist is therefore indirectly refuted. But the situation here is considerably more complicated than it may seem at first glance. T h e empiricist will argue that the amnesiac's awareness of his belief that p will, in the absence of countervailing reasons, provide adequate grounds to (continue to) believe that p; 3 Furthermore, what I have said about (10.2) accords with this interpretation of the empiricist view. Notice that one might allow such a prima facie reason to count as sufficient for knowledge in the absence of competing considerations. Alternatively, one might require additional strength from, e.g., the "coherence" of the belief in p with other beliefs. I see no reason not to adopt the first of these alternatives, as long as one remembers that strength of reasons is not the only factor in whether or not one knows. Another possibility, compatible with either of the others, would take one's reasons for p to be specified by an inference to the best explanation of his believing as he does (see the references in note 1 above).

394

THE JOURNAL OF PHILOSOPHY

and here the amnesiac bases his belief on true reasons. The fact that he may also base his belief on false reasons is irrelevant. Unger overlooks this possibility because he forgets that the empiricist can count awareness that one believes that p as prima facie evidence for the truth of p. And, if the amnesiac has the false reasons for believing that p that Unger ascribes to him, then he certainly must have the true reasons that the empiricist would ascribe to him. Furthermore, this raises a problem about Unger's use of a duplicate to establish his anti-empirical thesis if he wants real generality, since the duplicate may have empirically acceptable reason to continue to believe what it already believes, at least on certain conceptions of empiricism. Nevertheless Unger's claims about his duplicate do conflict with certain other forms of empiricism, in particular with the view I stated at the beginning of this paper. Unger's strongest and most relevant argument in defense of his claims about his duplicate is provided by his gypsy example. According to Unger, the gypsy has a crystal ball that can be used only a limited number of times. The ball on occasion gives him information about the lives of other persons. T h e gypsy believes this information to be true because of his upbringing. Unger also supposes that the gyspy has no good reason to believe the ball and knows that he has no good reason; yet he does believe because of his upbringing. This should border on unintelligibility, of course, because of the strange character of the gypsy's belief: "p, although there is no reason whatsoever to believe p." I n any event, I along with most empiricists would deny that the gypsy knows the truth of the propositions he believes as a result of looking into his crystal ball. Unger comments as follows: But is it really true that the gypsy knew nothing at all about the lives of other persons, both near and far away? No, it is not. Despite his science and skepticism and as a result of his early training, the gypsy was able to know more than any man alive about the lives of certain famous men (166).

I believe that it is fair to say that, when one thinks about the gypsy in a certain way, one is inclined to agree with the empiricist that, e.g., the gypsy does not know that a friend lost a fortune on a bad investment (one of the things he believes as a result of looking in the ball), and that, when one thinks about the gypsy in another way, one is inclined to agree with Unger that the gypsy does know this. T h e problem then is how to account for these contrary inclinations. What leads to this difference in view about the gypsy? I am not sure I have the answer to this question; but I think that it may be as follows. I t seems to me that the gypsy knows more than any man alive about the lives of certain famous men only in a special sense of 'know'. I

COMMENTS AND CRITICISM

395

say this because it doesn't seem to matter to Unger's example whether or not the gypsy actually believes the crystal ball. If the gypsy is imagined to suspend belief in what the ball reveals to him, it seems to me that we would not change our description of him as knowing more than any man alive about the lives of certain famous men. That is why we aren't really bothered by the fact that the gypsy believes things he knows it is irrational to believe; we aren't bothered because it is not central to the case that the gypsy believe these things. But, if this kind of knowing does not require belief, it is likely that the word 'know' is being used in a special sense. This suspicion is confirmed by the fact that it is also very natural to say that the gypsy does not know that his friend has lost a fortune. I cannot quite say exactly what sense of 'know' is involved in Unger's example, but it seems to be related to the following uses: We say of someone who hasn't yet got the answer to some problem, "Come on; you know the answer; you can do it." Here 'know' is close to 'ought to know' or 'can know'. A magician does a trick that depends on sleight of hand. At the required moment he reveals that he is dealing a card from his sleeve, although he is so good that his audience does not believe him. Afterward he tells the audience that they already know how he did the trick, and in some sense he is right. These uses of 'know' and the use in which the gypsy "knows" are different from that use which the empiricist is discussing. Strictly speaking, as the empiricist would say, the gypsy does not know. Another complication here is that the gypsy, like Unger's duplicate, does have knowledge, if only in the way that a library has knowledge. The duplicate has knowledge in the sense that, knowing how he has been constructed, we can obtain knowledge from him by discovering what he believes. similarly, if we know that the crystal ball really works, then we can treat the gypsy as a source of knowledge. But this does not mean that the gypsy or the duplicate knows anything; for again it is not important that they believe anything, as long as they can be sources of knowledge for us. Perhaps they have been trained or built to repeat certain things when asked; perhaps these things are in a foreign language, so that they do not even know what they are saying. They have knowledge only in the sense that what they have is knowledge. They do not have it in the sense that they know that it is true. This may not be completely right; but it does suggest how the empiricist can reply to Unger's arguments against him. He can say, "Strictly speaking, Unger's duplicate does not know anything." I have argued that Unger has not refuted this reply. GILBERT HARMAN

Princeton University