One of the most firmly entrenched beliefs of contemporary philosophy is that the only way to analyze a concept is to state its truth conditions. In epistemology this has led to the search for reductive analyses, to phenomenalism, behaviorism, and their analogues in other areas of knowledge. Arguing that these attempts at reductive analysis have invariably failed, John L. Pollock defends an alternative theory of conceptual analysis in this book.
The author suggests that concepts should be analyzed in terms of their justification conditions rather than their truth conditions. After laying a theoretical foundation for this alternative scheme of analysis, Professor Pollock applies his theory in proposing solutions to a number of traditional epistemological problems. Among the areas of knowledge discussed are perception, knowledge of the past, induction, knowledge of other minds, and a priori knowledge.
Originally published in 1975.
The Princeton Legacy Library uses the latest print-on-demand technology to again make available previously out-of-print books from the distinguished backlist of Princeton University Press. These editions preserve the original texts of these important books while presenting them in durable paperback and hardcover editions. The goal of the Princeton Legacy Library is to vastly increase access to the rich scholarly heritage found in the thousands of books published by Princeton University Press since its founding in 1905.
Read an Excerpt
Knowledge and Justification
By John L. Pollock
PRINCETON UNIVERSITY PRESSCopyright © 1974 Princeton University Press
All rights reserved.
What Is an Epistemological Problem?
1. The Fundamental Problem of Epistemology
Let us begin by looking at an example of a classical epistemological problem:
There is a book sitting on my desk in front of me. But, now, suppose I ask myself how I know that there is, or more generally, how I know that there is anything there at all (regardless of whether it is a book). A sensible answer to this question would be, "Because I see it." We know that there are material objects around us because we see them, feel them, hear them, etc. And the statement that we see something, or feel it, or hear it, logically entails that it is there to be seen, or felt, or heard. But now, I say that I see something (a book) there before me on my desk, but how do I know that I do? Mightn't I be hallucinating, or seeing an after-image, or witnessing some sort of cleverly constructed optical illusion? My experience might be exactly the same as when I really am seeing a book, and yet there might not be any book there, because I am hallucinating. Generalizing this, how do we know we ever perceive the things we think we do? Mightn't we always be hallucinating?
As it has sometimes been denied that it is even meaningful to suppose we might always be hallucinating, let us make this possibility more concrete. Suppose that a group of psychologists, biophysicists, and neurologists have constructed an adequate explanation of the neurophysiology of perception, and to test their explanation they take a subject from birth and wire him into a computer which directly stimulates his brain in such a way as to give a coherent, but completely false, sequence of sensations. In the subject's own mind he would seem to live out a completely normal life, growing up, making friends, going to school, getting a job, marrying and raising a family, etc. And yet all those years he was really sealed into an experimental apparatus in which he was fed intravenously and never had any contact with the outside world. It is true that in the present state of neurophysiology this could not be done, but it is certainly a meaningful hypothesis and a logical possibility.
Now, how do I know that I am not in the position of the subject of the above experiment? Perhaps a group of scientists have me hooked into such a computer, and all of the experiences that I think I have had since birth are really figments of the computer. How can I possibly know that this is not the case? It seems that any reason I can have for thinking that I am not hooked into such a computer must be either a reason for thinking that such a hypothesis is logically impossible, or else an empirical reason, arrived at inductively, for thinking that it is false as a matter of fact. But it is hard to see how this skeptical hypothesis can be logically impossible (it seems to make perfectly good sense — we know what it would be like for someone to be wired into a computer), and it seems that in order to have inductive evidence for an empirical reason I would already have to be able to rely upon some of my perceptions — which I cannot do without simply begging the question against the skeptical hypothesis. How then can I know that the skeptical hypothesis is false?
Faced with this sort of argument, one might be tempted to conclude that the skeptic is right — we really don't know the things we think we know. But such a conclusion flics in the face of common sense. There are many things that I know: I know that there is a desk before me, that I am holding a pen and writing on a piece of paper. I know that the walls of my study are lined with books, and that it is raining outside. All of these things I am certain about. It would be ridiculous to conclude that it is in principle impossible for me to know them.
That an argument "P1 ..., Pn; therefore ~ Q" is valid does not establish that its conclusion is true. It merely establishes that if the premises are true then the conclusion is true. Or better, it establishes that at least one of the propositions P1 ..., Pn, Q is false. The argument does not by itself determine which is false. A skeptical argument proceeds from prima facie reasonable premises to the conclusion that we do not know things that we are quite certain we do know. But all that such an argument establishes is that either one of the premises P1 ..., Pn is false or that proposition Q (that we do have the sort of knowledge we think we do) is false. In deciding which of these propositions to reject, all we can do is seize upon the one we are least certain about. But we will never be as certain about the premises of a skeptical argument as we are that we do have knowledge. So it will always be more reasonable to reject one of the premises than to accept the skeptical conclusion. A skeptical argument can only be construed as a reductio ad absurdum of its premises. There must be something wrong with any skeptical argument. Presented with such an argument, what we must decide is which premise to reject.
Skeptical arguments generate epistemological problems. Apparently reasonable assumptions lead to the conclusion that knowledge of a certain sort (e.g., knowledge of the physical world, or knowledge of other minds) is impossible. Faced with such an argument, our task is to explain how knowledge is possible. The problem is not to show that knowledge is possible; that much we can take for granted. What we must do is find the hole in the skeptical argument that makes it possible for us to have the knowledge we do. The problems of epistemology are problems of how we can possibly know certain kinds of things that we claim to know or customarily think we know. In general, given a statement P, we can ask, "How do you know that P?" This is the general form of an epistemological problem. The question "How do you know that P?" is a challenge — a demand for justification. The task of the epistemologist is to explain how it is possible for us to know that P, i.e., to explain what justifies us in believing the things we do.
Returning for a moment to the skeptical argument with which this chapter began, we can be confident that it proceeds from a false premise. Its conclusion, which is that knowledge of the physical world is impossible, is certainly mistaken. The task of the epistemologist is to find the false premise. This is not a task that can be undertaken at this point. Preliminary groundwork is necessary. We will return to this skeptical argument in Chapter Two, at which time it will be possible to pinpoint the error.
2. Reductive Analyses
Now let us turn to a second skeptical argument, around which the contents of this book will be organized. The development of this argument is rather involved, occupying this entire section. At the end of the section, the argument will be summarized.
Our knowledge can be separated into areas according to subject matter. These areas will include knowledge of the physical world, knowledge of the past, knowledge of contingent general truths, knowledge of other minds, a priori knowledge, and possibly knowledge of moral truths. The significance of these areas is that each has associated with it a characteristic source of knowledge. For example, the source of our knowledge of the physical world is perception. This is not to say that the only way to know that a physical object has a certain property is by perceiving the object. There are other ways, e.g., we may remember that it does, or we may be told that it does. But these other ways are all parasitic on perception. If we could not acquire knowledge of the physical world through perception, we could not acquire it in any of these other ways either. Analogously, the source of our knowledge of the past is memory; the source of our knowledge of contingent general truths is induction; the source of our knowledge of other minds is the behavior of other bodies. It is not clear just what are sources of knowledge either of a priori truths or of moral truths.
Given a statement P, let us call the conditions under which one would be justified in believing-that-P the justification conditions of the statment P. We can distinguish between two problems. The first is to state the justification conditions for the propositions in each of our areas of knowledge, and the second is to prove that those are the justification conditions. These two problems are not unrelated, but they are distinct problems. The second has generally interested epistemologists more than the first. Epistemologists have usually been content to give only a very rough description of the justification conditions of statements, and then have gone ahead to try to prove that those are the justification conditions. For example, the classical Problem of Induction is one of justifying induction as a way of learning the truth of universal generalizations. Although they've never been very clear about just what those grounds are, few philosophers doubt that we do base knowledge claims on inductive grounds. But what is demanded is a proof that we are justified in doing so. And similarly, the Problem of Perception is the problem of explaining how we can justify basing knowledge claims about physical objects on sense perception. It is not doubted that we do in fact base them somehow on sense perception, but what is wanted is a proof that we are justified in doing so.
The fundamental problem of epistemology is to explain what it is that justifies us in making the kinds of knowledge claims that we do customarily make. This problem has traditionally been construed as requiring a justification for our basing knowledge claims on the grounds on which we do in fact base them (a proof that what we suppose to be the justification conditions really are the justification conditions). In other words, it has been identified with the second of the above two problems. On the face of it, there seems to be a very good reason why we should, in principle, be able to give a proof of the sort desired. If we cannot establish any connection between one state of affairs and another, then we cannot be justified in making claims about the one state of affairs on the basis of the other. Thus if we cannot justify our customary grounds for knowledge claims, then we cannot take them as justifying our claims to knowledge. But if we cannot take them as justifying those claims, then they do not justify those claims, and so they are not really good grounds at all. Therefore, unless we can, in principle, give a proof of the sort desired, we are led to skepticism.
How might we set about justifying our basing knowledge claims on some particular source (such as perception)? It seems that there can be only two ways in which this might be done. We could either justify it inductively, showing that it does in. fact tend to lead to true knowledge claims, or else we could justify it logically, showing that there is some sort of logical connection between the source and the knowledge of which it is a source. But an inductive justification is impossible. We could only inductively justify a source of knowledge if we had independent access to both the source and the knowledge of which it is a source, and then could compare them and see that there is a correlation. But we do not have independent access to the knowledge that these sources are supposed to provide. They constitute the sources of this knowledge. For example, we do not have access to the physical world except through perception, and so there is no way to compare the physical world with perception to see that perception is a reliable guide unless we beg the question and assume in the beginning that it is.
It seems then that the only way to justify a source of knowledge is by establishing some sort of logical connection between the source and the knowledge it is supposed to give us. A logical connection must arise from the meanings of the concepts or statements involved in the knowledge claims. And (and here is the step which I shall deny) it has traditionally been supposed that the only way to analyze the meaning of a statement or concept is to give its truth conditions — to say what conditions must be satisfied in order for the statement to be true or for the concept to be correctly ascribable to an object. Furthermore, not just any statement of truth conditions will suffice. Starting from the truth conditions of a statement, we could never establish a logical connection between that statement and the source of knowledge which is supposed to yield the statement unless those truth conditions were stated in terms of the same concepts as are used in describing the source. Thus, for example, we could never establish a logical connection between perception and statements about the material world unless we could state the truth conditions of the latter in terms of the concepts used in describing perception.
An analysis of the truth conditions of a statement in terms of the concepts used in describing the source of our knowledge of that kind of statement is what philosophers have called a reductiveanalysis. Since Descartes, epistemologists have been concerned almost exclusively with giving reductive analyses of statements, and we now see why. Given assumptions that philosophers have traditionally accepted, it follows that the only way to prove that the purported justification conditions of a statement really are the justification conditions is by giving a reductive analysis of the statement in which the truth conditions are stated in terms of the same concepts as the justification conditions.
Philosophers have commonly supposed that they know more or less how we are justified in making the kinds of knowledge claims we do. Although they have not generally been able to state the justification conditions precisely, they felt that they could at least pick out the general sources of our knowledge in different areas. For example, our knowledge of the physical world comes from sense perception. And our knowledge of other minds comes from people's behavior. And our knowledge of right and wrong (if we can properly speak of "knowledge" here) comes from nonmoral states of affairs in the world. Thus if we are to justify these sources of knowledge, it seems we must seek reductive analyses of these statements in terms of these sources. Working within this traditional framework, phenomenalism becomes the only possible theory of our knowledge of the material world, behaviorism becomes the only possible theory of our knowledge of other minds, and some form of naturalistic ethics becomes the only possible theory of our knowledge of moral truth.
It seems that in order to justify sources of knowledge we are driven inexorably to reductive analyses. This appears to be the only way to derive the justification conditions of statements from the meanings of those statements. And it seems that if the justification conditions are not derivable from the meanings of the statements — if there really is no logical connection between them — then they cannot be the justification conditions, because if we cannot justify our sources of knowledge, then they do not justify our claims to knowledge and so are not really sources of knowledge at all. We seem forced to conclude that we must have either reductive analyses or skepticism. They are the only two possibilities.
Excerpted from Knowledge and Justification by John L. Pollock. Copyright © 1974 Princeton University Press. Excerpted by permission of PRINCETON UNIVERSITY PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
- Frontmatter, pg. i
- Preface, pg. vii
- Contents, pg. ix
- Chapter One. What Is an Epistemological Problem?, pg. 1
- Chapter Two. The Structure of Epistemic Justification, pg. 23
- Chapter Three. Theories of Perceptual Knowledge, pg. 50
- Chapter Four. Incorrigibility, pg. 71
- Chapter Five. Perceptual Attributes, pg. 80
- Chapter Six. The Reidentification of Physical Things, pg. 134
- Chapter Seven. Memory and Historical Knowledge, pg. 175
- Chapter Eight. Induction, pg. 204
- Chapter Nine. The Concept of a Person, pg. 249
- Chapter Ten. Truths of Reason, pg. 300
- References, pg. 341
- Index, pg. 347