Table of Contents
by James W. Cornman, Keith Lehrer, George Sotiros Pappas
https://www.amazon.com/exec/obidos/tg/detail/-/0872201244/
Chapter 2 - The Problem of Knowledge and Skepticism
A) About Skepticism
1) The Motives of Skepticism
2) Skepticism and Dogmatism
3) An Analysis of Knowledge
4) Dogmatism and Epistemism
B) Skepticism with Regard to the Senses
1) The Skeptical Argument from the Relativity of Observation ⇩
- Summary of the Argument
- An Epistemist Objection to the Second Premise: Some Actual Errors Don't Prove that We Could Always Be Mistaken
- A Skeptical Rejoinder: Erroneous and Veridical Experience Are Subjectively Indistinguishable
2) The Modified Skeptical Argument
- An Epistemist Objection to the First Premise of the Modified Argument: A Single Example Cannot Support a General Conclusion
- The First Skeptical Rejoinder: The Argument from Hallucinations
- An Epistemist Objection to the Skeptical Rejoinder: Some Actual Hallucinations Don't Prove that We Might Always Be Hallucinating
- A Skeptical Rejoinder: There Are No Grounds for Holding that Some Experiences Are Immune to Hallucination
- An Epistemist Objection: Coherence and the Testimony of Others As Criteria for Non-Hallucinatory Experience
- A Skeptical Rejoinder: Appeals to Coherence and Testimony Merely Beg the Question
- The Second Skeptical Rejoinder: The Braino Argument
- An Epistemist Objection: The Skeptic's Fallacious Argument from Possibility to Actuality
- A Skeptical Rejoinder: How to Pass from Possibility to Actuality
- A Skeptical Argument: The Evil Operator Argument-Another Route from Possibility to Actuality
3) Summary of the Modified Skeptical Argument
- An Epistemist Objection: The Meaninglessness of the Perfect Hallucination Hypothesis
- A Skeptical Rejoinder: The Charge of Meaninglessness Rests on Confusion
- An Epistemist Objection to the Third Premise: The Skeptic's Failure to Distinguish Probable Judgments from Lucky Guesses
- A Skeptical Rejoinder: The Employment of the Notion of Probability Begs the Question Against the Skeptic
- An Epistemist Objection: An Externalist Account of Justification Undermines the Skeptic's Argument Against Probability
- A Skeptical Rejoinder: The Externalist Begs the Question Against the Skeptic
- An Epistemist Objection: The Principle of Charity Supports Externalism
- A Skeptical Rejoinder: Charity without Knowledge
- An Epistemist Objection: Internalism and Complete Justification
- A Skeptical Rejoinder: Satisfaction of Internal Standards Is Not a Guide to Truth
- An Epistemist Objection: Innocent Justification
- A Skeptical Rejoinder: Appeal to Innocent Justification Does Not Rid the Epistemist of the Burden of Proof
- An Epistemist Objection: An Argument to the Best Explanation Provides the Required Proof
- A Skeptical Reply: Explanation without Truth Is Worthless
4) Another Argument for the Skeptical Conclusion:
- Complete Justification As Excluding the Chance of Error: The Lottery Paradox
- Summary of the Argument
- An Epistemist Objection: The Skeptic's Braino Hypothesis Is a Mere Possibility
- A Skeptical Rejoinder: The Survival Hypothesis - A Realistic Skeptical Hypothesis
5) Summary of the Final Skeptical Argument
C) New Knowledge as Undefeated Justification:
A Revisionist Alternative to the Skeptic and the Epistemist
D) Exercises
1) Skepticism, Knowledge, and Truth
2) Skeptical Arguments:
The Relativity of Observation and the Braino Hypothesis
Adapted from Keith Lehrer - Philosophical Problems and Arguments: An Introduction - by James W. Cornman, Keith Lehrer, George Sotiros Pappas
https://www.amazon.com/Philosophical-Problems-Arguments-Introduction-Cornman/dp/0872201244
We shall now examine the skeptical argument with respect to perceptual belief. By so doing we shall consider seriously and at length a challenge to one of the most fundamental assumptions of the current intellectual milieu, namely, that we obtain knowledge of the world by means of sense experience, by observation and perception. The initial stages of the argument may strike you as weird and disorienting. This is to be expected. When our fundamental assumptions and presuppositions are brought before the court of evidence and pronounced unfit, we feel abandoned in our uncertainty with nothing to sustain us. We may be inclined to repudiate the court as unjust. Such a response is natural but unwarranted. A skeptic claiming that our perceptual beliefs fail short of knowledge need not suggest that they be abandoned. As long as those beliefs remain more probable than those with which they compete, it is reasonable enough to maintain them. But if skepticism wins the day, then we must regard even those quite reasonable perceptual beliefs with an open mind and generously expose them to criticism and debate. Doing so, though initially discomforting, like exposing one's flesh to the elements, soon becomes routine and, moreover, provides an invigorating sense of well-being. With these words of reassurance, we embark on our quest for skepticism with regard to the senses.
Our perceptual beliefs about what we hear, touch, and see are based on evidence. This might not seem obvious at first, because such evidence is rarely formulated in words. We do not ordinarily justify our perceptual beliefs-for example, my belief that I see a red apple in my hand-by appealing to any other belief or statement. However, these beliefs are not without evidence. It is the unformulated evidence of our senses, the direct and immediate evidence of sensory experience, of the way things appear, that seems to justify our perceptual beliefs. For example, imagine a person who believes that she sees a red apple. We generally assume that what justifies this belief is the visual experience she is then having, the visual appearance of a red apple. We do not ordinarily formulate and state such "evidence," because we have no reason to do so. We do assume that this "evidence" justifies our perceptual beliefs, however. The question we must now consider is whether this "evidence" provides complete justification for these beliefs.
1) The Skeptical Argument from the Relativity
of Observation
One argument for skepticism is based on the relativity of the observer. Suppose some object stimulates my sense organs, and I see something red. It might happen that the object also stimulates the sense organs of someone else, who sees an object of a different color, for example, green. Imagine that the object is in fact white, but that there is a transparent red plastic shield between the object and me, and a green one between the object and the other observer. If neither of us knows of the presence of these shields, then each of us might be entirely convinced that he sees the object's true color.
This rather contrived example has quite general implications. For if we pay close attention to what we see, it becomes plausable to claim that no two people see the same object in exactly the same way. For example, consider a square envelope. Pay very close attention to what you see when you look at this object. If the envelope remains stationary while you move about, or is moved while you remain stationary, what you see will constantly change. When you see the envelope from one angle you will see something perfectly square, but as you move away to one side and see the envelope from a more oblique angle you will see something that is not quite square. Thus two people seeing the envelope from different angles will not see the same thing. These are familiar facts of perceptual experience. How can they be used to serve the purposes of the skeptic?
The example of the envolope may be taken to show that at least one of the two persons involved does not see the envolope as it really is. The envelope cannot be both perfectly square and not square at one and the same time. Thus, if each person comes to have a perceptual belief, one to the effect that she sees some square object and the other to the effect that she sees an object that is not square, then at least one of them will be mistaken. Hence, at least one of them will not know, on the basis of his or her perception, that his or her perceptual belief is true. One person may have a true belief. Thus, the person who views the envelope from directly above and sees a square object, will form the true perceptual belief that he sees a square object. The fact that perception is relative, however, that what one sees depends on a variety of constantly changing factors (position of the observer, lighting, condition of the observer, and the like), is enough to show that any perceptual belief may be a mistake, for our perceptual beliefs are often in error.
To obtain the skeptical conclusion, we need another premise: that if a person knows something and is thus completely justified in believing it. he or she cannot possibly be mistaken in his orher belief. If a person says that he or she believes something but admits that he or she could be mistaken, then he or she has thereby admitted that he or she is not completely justified in his or her belief and does not know that what he or she believes is true. Similarly, if a person could be mistaken, then the person lacks knowledge about the matter. Even when a person is not mistaken, of the person could be mistaken, the person does not know what they believe is true. Thus a person knows something only if the person could not be mistaken.
The relevance of the preceding comments is straightforward. We noted that sometimes a person's perceptual belief will be in error, in part because of the relativity of perception. When in error, surely, a person does not gain perceptual knowledge. This shows that it is possible for a person to be mistaken in his or her perceptual belief. Since people are, in fact, sometimes mistaken in such beliefs, then clearly it is possible for them to be mistaken in those beliefs. Moreover, we are perfectly justified in arguing from the premise that people are sometimes mistaken when they believe what they see (or otherwise perceive) to the general conclusion that this is always possible. For either the premise that a person has a perceptual belief that he perceives a thing with a certain sensible quality, such as being red or round, entails that there is a thing with that quality or it does not. Because we sometimes make mistakes in perceptual beliefs of this sort, obviously the entailment does not hold. If the entailment does not hold, then it is always at least logically possible that a person should have a perceptual belief and that the belief is nonetheless mistaken.
Having established the preceding points, we may now reconstruct the argument for skcepticism. Whenever a person believes what he or she sees, it is possible that he or she is mistaken. If a person knows something then it is not possible that he or she is mistaken. Therefore, when a person believes what he or she sees, he or she does not know it. This conclusion is correctly derived from the premises antecedently defended.
Moreover, the argument employed to show that no one ever knows an object to have some sensible quality, one perceived by the senses, can be cogently generalized to show that no one even knows that the object itself exists. For, just as we sometimes see something as having some sensible qualify, when it does not have that quality, so we sometimes see some sensible object when it does not even exist. Hallucinations are experiences of this sort. A person who has delirium tremens, or one who has taken a heavy dose of LSD, sometimes sees things-for example, pink rats-when there are no such things. Thus a person might believe that what he or she sees exists and be completely mistaken in this belief. Seeing an object does not constitute knowing that that object exists. A person who knows something cannot be mistaken.
- Summary of the argument
We may now formulate an argument for skepticisai as follows:
1. We are sometimes mistaken in our perceptual beliefs.
2. If we are sometimes mistaken in our perceptual beliefs, then it is always logically possible that our perceptual beliefs are false.
3. If it is always logically possible that our perceptual beliefs are false, then we never know that any of our perceptual beliefs are true.
Therefore
4. We never know that any of our perceptual beliefs are true.
This conclusion is validly deduced from the three premises which were just defended.
- An Epistemist Objection to the Second Premise:
Some Actual Errors Don't Prove that We Could
Always Be Mistaken
The skeptic has illicitly adopted the premise
2. If we are sometimes mistaken in our perceptual beliefs, then it is always logically possible that our perceptual beliefs are false.
In those cases in which our perceptual beliefs are false, it is obviously possible that they are false. But how does that prove that it is always possible that our perceptual beliefs are false? It may be that there are some true perceptual beliefs which could not possibly be false. Until we have some reason for thinking otherwise, we are surely justified in asserting there are such true perceptual beliefs and, consequently, in rejecting premise (2) of the skeptical argument. We may also then reject the skeptical conclusion.
- A Skeptical Rejoinder: Erroneous and Veridical
Experience Are Subjectively Indistinguishable
How can we sustain the skeptical claim that it is always logically possible that our perceptual beliefs are false, in the face of the preceding objection? The answer arises from considering the evidence that we have for our perceptual beliefs. The evidence that we have for any perceptual belief, no matter how plausible, can be duplicated by evidence for an exactly similar perceptual belief which is, in fact, mistaken.
To clarify this point, suppose that two people are looking through different windows. The first person reports that there is a sphere on a table outside her window; she sees the sphere to be green. She sees this no matter from what vantage point she views things. Suppose further that the second person, looking through her own window, sees and reports the very same thing. Each person has exactly the same justification for claiming to know that there is a green sphere outside his window. Each is in just as good a position to know this as the other. Surely, the only correct conclusion to reach is that either each person knows there is a green sphere outside her window or that neither of them knows this. It would be entirely arbitrary, and hence unreasonable, to say that one person knows this and the other does not.
However, it is perfectly possible that one of these people is mistaken and the other is not. Suppose the first person sees what she does because there is a green sphere outside her window. On the other hand, suppose the second sees what she does because she is being tricked with mirrors and drawings-there is no green sphere outside her window at all. Moreover, the deception is so excellent that from behind the windows no one could detect any difference in what is seen through each. This shows that the first person, who is in fact not mistaken, could have been mistaken. The second person was mistaken, and the first person had no better evidence for what she believed than the second person did. Since having this evidence did not keep the second person from being mistaken, the first person, too, could have been mistaken. What was so in the one case could have been so in the other. The only reasonable conclusion is that neither person has knowledge.
What we have just imagined has perfectly general implications. The experiences a person has, when he or she sees something that really exists, can always be duplicated by the experiences of another person who is being deceived. Because the experiences in question provide the only evidence a person has for believing what she does, if one person fails to know what she believes, so must the other. If one is mistaken in believing something, then another person who has a similar belief based on similar experiences surely could have been mistaken-even if in fact she is not. Since this duplication of experiences is always possible, it is always possible that a perceptual belief based on sensory experience is mistaken. The argument for skepticism requires no other assumption.
2) The Modified Skeptical Argument
We rnay conclude, then, with a slightly modified formulation of the argument for skepticism. The first two premises of the argument, which differ from the initial premises of the preceding skeptical argument, are as follows:
1. The experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false.
2. If the experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false, then it is always logically possible that our perceptual beliefs are false.
The next premise is the same as in the earlier argument:
3. If it is always logically possible that our perceptual beliefs are false, then no one ever knows that any of our perceptual beliefs are true.
From these three premises we can deduce the skeptical conclusion.
4. No one ever knows that any of our perceptual beliefs are true.
- An Epistemist Objection to the First Premise of
the Modified Argument: A Single Example
Cannot Support a General Conclusion
The preceding argument might be challenged on the grounds that a general conclusion is drawn from a particular example. It is true that the two people looking through their respective windows might have almost exaclly the same experiences, even though one of them is mistaken and the other correct. Perhaps this shows that both of these people could have been mistaken and that neither of them knows what she believes. However to concede this point is not to concede the more general conclusion that whenever anyone sees something which is in fact the case, his experiences may be exactly duplicated by the experiences of another person who sees the same thing, though in fact he is mistaken. For example, consider a person holding the object he sees directly in front of him. How could his experiences be duplicated by the experiences of someone who is not confronted with such an object? If they could not, then premise (1) of the preceding skeptical argument may be rejected.
- The First Skeptical Rejoinder:
The Argument from Hallucinations
We noted in our experiment that both people view what they believe to be a green sphere from behind a window. But taking away the window, though it might help those two people to discover the trick that has been played on one of them, will not alter the primary force of the argument. For all of us view the world through the 'window' of our senses, and as a result, a person who sees something that does exist may in general have the same experience as one who sees something that does not exist. The clearest example of what the skeptic is trying to prove is supplied by Lady Macbeth. At one time, after the murder of Duncan, she sees and feels blood on her hands. Her hands seem to be covered with Duncan's blood. She goes mad. Part of her madness consists of seeing and feeling blood on her hands. Of course, this is a hallucination. But the experiences she has when she is hallucinating might be exactly similar to the experiences she had when there really was blood on her hands. Because she could be mistaken, at the later time, in believing there is blood on her hands-indeed, she is mistaken-she could also be mistaken in believing the same thing at the earlier time. She had the same experiences to rely on in both cases. Consequently, it would be arbitrary, unreasonable, and epistemologically undemocratic to suppose that Lady Macbeth knows that she has blood on her hands at the earlier time but not at the later.
In addition to being misled by visual experience, the subject of hallucinations may also be misled by tactual experience and the experience of the other senses as well. We may imagine that, when mad, Lady Macbeth not only sees but also feels and smells blood on her hands, though there is none there. Such a hallucination is complete and systematic with respect to all the senses. Thus, our defense of the premise
1. The experiences of a person who has a true perceptual belief may be duplicated by the experiences of a person whose perceptual belief is exactly similar but false.
is that such duplication of experiences may always result from hallucinations. Erroneous perceptual beliefs based on hallucinatory experience obviously do not constitute knowledge, and correct perceptual beliefs are no better corroborated by experience. Therefore, perceptual beliefs in general, whether true or in error, never constitute knowledge.
Again, we must guard against misunderstanding. By distinguishing between those cases that ivolve hallucinations and those that do not, the skeptic is not contradicting herself. She is not supposing that we know which cases are which. We may, with perfect consistency, both agree that there is a distinction between hallucinatory experience, which evokes false perceptual belief, and ordinary experience, which evokes true perceptual belief, and yet deny that we know which kind of experience we are having. That is the position adopted here by the skeptic.
- An Epistemist Objection to the Skeptical Rejoinder:
Some Actual Hallucinations Don't Prove that We
Might Always Be Hallucinating
The preceding argument suffers from the same defect as the skeptic's argument about the two green spheres: a general conclusion is drawn from a particular example. The skeptic produces an example: Lady Macbeth with and without bloody hands. The experiences she has when she sees something that exists are duplicated in hallucinations when she sees something that does not exist. The skeptic concludes that, when a person sees something that exists, her experiences may well be exactly duplicated in hallucinations when she sees something that does not exist. But how does this general conclusion follow from that one example? By what means does the the skeptic prove that it is always possible to duplicate in hallucinations the sense expenences we have when our perceptual beliefs are true? How does she demonstrate that hallucinations are always possible no matter what our experiences are like? So far no such demonstration has been given; therefore we have no reason to accept the skeptic's conclusion.
- A Skeptical Rejoinder: There Are No Grounds for
Holding that Some Experiences Are
Immune to Hallucination
Where does the burden of proof lie? If one concedes that a hallucination like Lady Macbeths is possible, what reason is there for denying that hallucinations are always possible? When we believe we see something on our own hands, that is the kind of belief that we ordinarily accept with the greatest confidence and equanimity. When we believe we see something at a distance, or when our vision is obscured in some other way, we may have some doubt. But when we see something on our very own hands and feel it and smell it as well, then we have no doubt. Instead we feel certain that the thing exists. If experiences of this kind may be produced by hallucinations, and the resultant perceptual beliefs be in error, then how can we reasonably deny that all experiences can be produced by hallucinations, and consequently be accompanied by mistaken perceptual beliefs? Surely we cannot deny this. The range range of experiences that hallucinations might possibly produce is witnout limit. And the argument for skepticism is perfectly sound.
- An Epistemist Objection: Coherence and the
Testimony of Others As Criteria for
Non-Hallucinatory Experience
The argument for skepticism just stated assumes that complete and systematic hallucinations are always possible. This is the premise we shall now critically examine. Sometimes hallucinations mislead us concerning what is in our very hands. But we are not entirely at the epistemological mercy of such hallucinations, for we do have ways of discovering when our experiences are hallucinatory. But if we have some way of telling whether or not we are suffering a hallucination, then there must be some perceptual beliefs that are accompanied by experiences that rule out the possibility of hallucinations. In this case the experiences on which our belief is based could not be hallucinatory.
Moreover, it is not difficult to explain what kind of experiences rule out the possibility of hallucinations. We need only ask how we in fact discover that we are suffering a hallucination. One way is through the testimony of other people who know that our experiences are hallucinatory. Many people were in a position to tell Lady Macbeth that her experiences were hallucinatory, and although in her madness she would not accept such information, it was entirely available to her. Because we are not primarily concerned with madness but rather with normality, it is quite relevant that a normal person may discover that some of his experiences are hallucinatory through the help of others. Moreover, when one is in the company of others, and all agree concerning what they see, it is altogether reasonable to assume that no hallucination is taking place. Nonetheless one qualification is necessary. Sometimes, in unusual circumstances, we know that a whole group of people is susceptible to hallucinations. For example, suppose they have taken a drug known to produce hallucinations, or suppose they have suffered some acute physical hardship, like being deprived of drink or sleep. Such conditions might produce group hallucinations. However, if people and circumstances are altogether ordinary, as they usually are, then an agreement in perceptual belief rules out the possibility of hallucinations.
Hallucinations are detected in another way, which does not require the testimony of others. It concerns the coherence amoung our experiences. A person who suffers from the hallucination of drinking water, when none is available to him, may see, feel, and even taste water which, because of his hallucination, he mistakenly believes exists. However, if he has been long deprived of water, he will soon notice that his thirst has not at all abated. People have, in fact, concluded from such experiences that they were suffering a hallucination. They were subsequently no longer misled. Thus, in this case there is a failure of experiential coherence. By drinking water one expects to quench one's thirst; and when one has the experience of seeing, feeling, and drinking water, but one's thirst is not quenched, then the experiences seem incoherent, and incomprehensible. Incoherence is therefore a sign of hallucinations, and coherence is, on the other hand, a sign of reality. Indeed, experiences that are sufficiently coherent exclude the possibility of hallucinations.
Finally, these criteria of nonhallucinations, or of veridical experiences, may be satisfied together and may thereby mutually reinforce each other. Often our experiences are entirely coherent, and our perceptual beliefs completely agree with the beliefs of others in our company.
- A Skeptical Rejoinder: Appeals to Coherence
and Testimony Merely Beg the Question
The appeal of the foregoing argument to the testimony of others as a safeguard against being misled by hallucinations begs the question. For to know what the testimony of another is, we must first know that we are confronted by another and know what he or she is saying. But to know these things is to know something by seeing or otherwise perceiving that they are so. The argument rests squarely on the assumption that perceiving such things constitutes knowledge. Of course, this assumption is precisely what the arguments for skepticism are intended to refute.
Moreover, the second argument, based on experiential coherence, has the same weakness. Some hallucinations do make themselves manifest to their victim by some kind of incoherence, but suct a hallucination is partial or incomplete. Some expected experiential feature does not turn up, and the expenence shows itself to be hallucinatory by being a bit too surprising. But why must we assume that hallucinations always expose themselves in this way? What proof is there that coherent and systematic experiences cannot possibly result from hallucinations? There is no reason to suppose that such hallucinations are impossible. Therefore, we are again justified in concluding that the experiences that a person has, when his perceptual beliefs are true, may be exactly repeated in hallucinations when his perceptual beliefs are false. Consequently, such beliefs never constitute knowledge.
- The Second Skeptical Rejoinder: The Braino Argument
However, we must pause long enough over the epistemist's argument to ensure that the skeptic is not again accused of passing too quickly from the particular example to a general conclusion. To rid ourselves of this recurring objection, let us indulge in a bit of science fiction and by so doing prove once and for all that skepticism is the tenable and correct position. Imagine that a superscientis invents a machin-we shall call it a "braino"-that enables him to produce hallucinations in certain subjects. The machine operates by influencing the brain of a subject who wears a special cap, called a "braino cap." When the braino cap is placed on a subject's head, the operator of the braino can affect his brain so as to produce any hallucination in the subject that the operator wishes. The braino is a hallucination-producing machine. The hallucinations produced by it may be as complete, systematic, and coherent as the operator of the braino desires to make them.
What does the possibility of such a machine prove? It proves that the experiences a person has when his perceptual beliefs are true could be duplicated in hallucinations when the same perceptual beliefs are false. This shows that no mark or sign in experience distinguishes true perceptual beliefs from false ones resulting from hallucinations.
An analogy helps to illustrate the importance of the preceding considerations. Suppose that you are confronted with a barrel full of apples some of which are rotten and others of which are not. Usually there will be some sign or mark by which you can tell the rotten apples from the good ones. The rotten apples will be brown of soft, or they will have some other visible defect by which you can detect their condition. On the other hand, the good apples will be firm,red, and otherwise appear desirable. We suppose we can tell the difference between a good apple and a rotten one because we have signs to guide us.
However, suppose we are confronted with a barrel of apples that are quite indistinguishable in appearance, though some of the apples are rotten at the core. We are now presented with an apple from this barrel and prohibited from cutting it open. In this predicament, if someone should ask us whether the apple is rotten or good, the only thing to reply is, "I don't know" We might add, "There is no way to tell."
Similarly, because there is no mark or sign in experience by which we can distinguish true perceptual beliefs of ordinary experience from false ones arising in hallucinations, if someone should ask us whether a perceptual belief of ours is true or false, the only thing to reply is, "I don't know." We might add, "There is no way to tell." Just as there is no way to tell whether the apples in the second barrel are rotten, because we have no experiential signs to guide us, so there is no way to tell whether our perceptual beliefs are true, because we have no experiential signs to guide us. With only experience to guide us, we have no way to rule out the possibility of error. Even when our perceptual beliefs happen to be true, we could just as well have been mistaken. Indeed, when a perceptual belief is true, this is more a matter of good luck than good sense. Of course, no belief that turns out to be true as a matter of luck can reasonably be counted as knowledge.
- An Epistemist Objection: The Skeptic's Fallacious
Argument from Possibility to Actuality
In the first arguments we examined, the skeptic argued from a single example of hallucinations to a general conclusion, which is surely a fallacy, but, having now avoided this fallacy, she argues from possibility to actuality, which is no less of a fallacy. The present argument starts from the premise that the braino is a logical possibility, and consequently that it is a logical possibility that there should be hallucinations that are coherent, complete, and systematic in every way. From this premise of logical possibility, she arrives at the conclusion that we in fact have no way of telling whether or not we are hallucinating. By what line of reasoning may such a factual conclusion be derived from a premise concerning mere possibility? Even if it is logically possible that hallucinations should be coherent, complete, and systematic, hallucinations are not in fact so hard to detect. Therefore, experience does in fact enable us to tell whether our perceptual beliefs are true or false.
- A Skeptical Rejoinder: How to Pass from
Possibility to Actuality
The contention that "hallucinations are not in fact so hard to detect" is the very heart of the issue. If the braino is a logical possibility, then how can we tell that hallucinations are not in fact so hard to detect? On the contrary, we may suffer hallucinations that we cannot detect. If it is logically possible that hallucinations should be coherent, complete, and systematic in every way, then there is no way of detecting at any moment that we are not suffering from a hallucination. Our critic supposes we can detect many hallucinations, but this is beside the basic point. The problem is to explain how we can tell that we are not hallucinating. Our braino argument was intended to establish that we can never tall this, even if we can sometimes tell that we are hallucinating.
Sometimes we can tell that we are hallucinating, but we have no way of telling that we are not. Our argument to support this claim is best put in the form of a challenge. Consider some perceptual belief that you would maintain does not result from hallucinations. What expreiences guarantee this? Indeed, what experiences provide you with any evidence of it? Notice that whatever experience you indicate, the braino argument will be quite sufficient to prove that such an experience is no guarantee against hallucination. All we need do is imagine that you have, unknown to yourself, the braino cap on your head. The operator of the braino is producing the very experiences you claim guarantee that you are not hallucinating!
The passage, from the possibility of hallucinations to the conclusion that there is in fact no way of telling that one is not hallucinating, is legitimate, because the former possibility may be used to reject any experience that is so impertinent as to present itself as a sure sign of reality. Any experience can be shown to be unequal to this task, on the grounds that it can be produced in hallucinations.
- A Skeptical Argument: The Evil Operator
Argument-Another Route from Possibility to Actuality
The preceding argument may be reinforced if we let our examples become even more fanciful than we have heretofore. Imagine that all people are controlled by the braino and that the machine is run by some evil being, Dr. O, who plots to keep us completely in error through hallucinations. Dr. O does not wish to be detected, so he supplies hallucinations that are coherent, complete, and systematic. Indeed, the hallucinations he produces in us are a perfect counterfiet reality. Our experiences fulfill our expectations and contain no more surprises than we would expect from reality. But it is not reality we experience; our perceptual beliefs about the world are quite mistaken, for the source of our experiences is a mere machine, the braino, which creates hallucinations. In such a predicament we might have just the sort of perceptual beliefs we now have, based on experiences exactly similar to those we now have. But our perceptual beliefs would be altogether false.
The imagined situation is exactly similar to ours with respect to the reasons or evidence we would have for our perceptual beliefs. Experience is virtually the same in both cases. Consequently, if we lack knowledge in one situation, we must surely lack it in the other. Obviously, we lack knowledge when we are controlled by the braino, for then our perceptual beliefs are false. Hence, we also lack knowledge in our present situation. More precisely, our perceptual beliefs fail to constitute knowledge in either case.
We believe that we are not controlled by such a machine, and if we are fortunate in this belief, then no doubt many of our perceptual beliefs are true. It is however, good fortune and not good evidence that we should thank for the correctness of these beliefs. We are just lucky if there is no Dr. O controlling us with a braino; and from that good fortune may result the further good fortune that most of our perceptual beliefs are true. It is just a matter of luck, however, and nothing epistemologically more glorious than that. If a belief is true as a result of luck, then it is a lucky guess-not knowledge.
3) Summary of the Modified Skeptical Argument
Earlier we defended skepticism by employing an argument the first premise of which was
1. The experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false.
From this premise and the following two premises
2. If the experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false, then it is always logically possible that any of our perceptual beliefs are false.
and
3. If it is always logically possible that any of our perceptual beliefs are false, then no one ever knows that any of our perceptual beliefs are true.
we deduced the skeptical conclusion
4. No one ever knows that any perceptual belief is true.
Premise (1) of the argument was called into question by the opponents of skepticism, and we have now derived this premise from the following premises:
5. The braino hypothesis is logically possible.
6. If the braino hypothesis is logically possible, then any experience may be duplicated in hallucinations.
7. If any experience may be duplicated in hallucinations, then the experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false.
From these premises, premise (1) may be validly deduced.
- An Epistemist Objection: The Meaninglessness of
the Perfect Hallucination Hypothesis
Let us examine the situation in which we are all deceived by the evil genius, Dr. O, who is busy supplying hallucinations to us with his braino. Is this situation really possible? Consider it in very concrete terms. Imagine Dr.O contriving to deceive us. Moreover, let us concentrate on the experience of one man, whose name is Tom, as Dr. O goes about deceiving him. Dr. O will supply our experiences by using the braino. Because it is the braino that produces hallucinations, we need not suppose that there are any of the usual things around in the world at all when Dr. O deceives his victims. Therefore let us suppose that all those things are destroyed. We may thus imagine that Tom stands alone on a barren world, a braino cap firmly attached to his skull. He is, of course, quite oblivious to his predicament, for Dr. O, true to his plan, is supplying Tom with hallucinations that exactly duplicate the experiences of his usual existence.
The success of the evil genius is quite complete. What he has destroyed he deceives Tom into believing exists by causing Tom to hallucinate. Tom's hallucinating experiences exactly duplicate the experiences he would have if the things destroyed still existed. It is an extraordinarily clever deception, but in spite of this apparent success, Dr. O has more nearly succeeded in fooling himself than Tom. Why is this so? Let us return to the supposedly barren scene of the drama.
Imagine that Dr. O, having deceived Tom, grows weary of his unacknowledged success and wishes, while continuing the deception, to have Tom acknowledge this accomplishment. The next day, when Tom is, for example, having the experience of entering a room where flowers stand, Dr. O. tactfully suggests to Tom that there are no flowers. Tom then has the experience of raising the flowers to his nose, smelling them, and touching them. Tom is reassured. He denies there is deception. After all, having looked, sniffed, felt, with satisfying results, what could Dr. O. possibly mean by saying he is deceived? To this Dr. O. replies, "Your flowers are nothing but a hallucination." But Tom is unpersuaded. His flowers are perfect. A hallucination? Never!
At this point in the drama Dr. O is sorely tempted just to throw the switch on the braino and expose the fraud to Tom. But to do so is to give up the deception, which he does not want to do because of all his labors. So the evil genius, not feeling so ingenious now, is faced with a dilemma. Either he can maintain the deception-in which case Tom, refusing to acknowledge the deception, will deny that it exists-or he can alter his plan-in which case Tom, detecting that he has been fooled, will convert a perfect deception into a perfect farce. Either way Dr. O will fail to attain his goal. For if the deception continues, how is Dr. O to make any sense of the idea to Tom?
From Tom's standpoint, the perfect deception is no deception at all. Once all the usual experiences are made to occur, the suggestion that Tom is suffering a hallucination is made senseless. Tom can make no sense of it. It is just as senseless to suppose such a deception is being worked on us. When a person is hallucinating, there must be ways to detect that this is so, even if the victim, like Lady Macbeth, fails to detect the hallucination in these ways. Once all the appropriate tests are made in order to ensure that one is not hallucinating, the suggestion that one might still be hallucinating is meaningless. What we mean by saying that a person is hallucinating is that some such tests will fail. The perfect hallucination is a bubble of semantic incongruity that disappears under the pressure of semantic scrutiny.
The defect in the skeptic's argument is that she assumes that a perfect hallucination is possible. Imagination, when it works overtime, might convince one this is so, but it cannot be. For something to be possible it must be meaningful. However, imagination is not restricted by bonds of meaningfulness. There is little difficulty in imagining something when the very idea of it is completely meaningless. We can imagine a cat in a tree with the parts of the cat disappearing one by one, first the tail, then the paw, then the body, until finally all that is left of the cat in the tree is a feline grin. We can imagine this, but the idea of the grin without a head is perfectly meaningless. The remarkable scope of the human imagination is a joy in life, but it is a trap for philosophical reflection. This is true precisely because it is so easy to suppose that whatever is imaginable is possible. The worldless hallucination of a world is precisely like the catless grin of a cat. We can imagine both but neither is literally meaningful.
Thus our reply to skepticism is a semantic one. The evil genius argument of the skeptic proceeds from the premise that a perfect hallucination is possible. But the idea of a perfect hallucination is meaningless, and hence such a hallucination is not possible. Because the argument of the skeptic proceeds from a false premise, it must be rejected.
- A Skeptical Rejoinder: The Charge of
Meaninglessness Rests on Confusion
Why does it seem plausable to contend that the perfect-hallucination hypothesis is meaningless? The explanation stems from an ambiguity in such terms as 'nonsense' and 'meaningless.' All these terms are used in both a semantic and an epistemic sense. A sentence is nonsense, or meaningless, in the semantic sense of the term only if the sentence asserts nothing, and consequently is neither true nor false. A perfect example of such a sentence is 'Pirots carulize elactically.' That sentence asserts nothing; it is neither true nor false. It is not made up of meaningful words. An example of a sentence that is meaningless in the semantic sense, but that is made up of meaningful words, is 'Verb at do fog Joe.' The defect of this sentence is that it is ungrammatical. We can even have a sentence that is grammatical and composed of meaningful words, but that is nevertheless meaningless in the semantic sense. An example is 'Worms integrate the moon by C# homogeneously when moralizing to rescind apples.' This sentence, like the preceding ones, asserts nothing.
All the sentences we have considered are nonsense and meaningless in that they are either semantically or grammatically defective in such a way that a person who uttered them would, in ordinary circumstances, be asserting nothing. In contrast to this semantic sense of the terms 'nonsense' and 'meaningless,' there is an epistemic sense. Sometimes we say that a sentence is meaningless because, though it asserts something, what is asserted is preposterous. If a man says, "Everyone has died," we might reply, "Nonsense," or alternatively, "What do you mean?" or possibly even, "That is meaningless." It is not that the sentence asserts nothing; on the contrary, it is because the sentence asserts something patently false that we reply as we do. The sentence uttered is perfectly meaningful; what is nonsensical and meaningless is the fact that the person has uttered it. To put the matter another way, we can make sense of the sentence; we know what it asserts. But we cannot make sense of the man uttering it; we do not understand why he would utter it. Thus, when we use terms like 'nonsense' and 'meaningless' in the epistemic sense, the correct use of them requires only that what is uttered seem absurdly false. Of course, to seem preposterously false, the sentence must assert something, and thus be either true or false.
These remarks are directly relevant to the perfect-hallucination hypothesis. The perfect-hallucination hypothesis seems to be 'nonsense' and 'meaningless' in just the epistemic sense of these terms: the hypothesis seems preposterously false. And even the skeptic may concede, it should be remembered, that the hypothesis seems false. However, for the hypothesis to seem false, it must be meaningful; it must assert something in order to seem false.
If the perfect hallucination hypothesis is meaningful in the semantic sense-that is, if it asserts something-then the fact that it is not meaningful in the epistemic sense is irrelevant to the contentions of skepticism. As we pointed out earlier, the skeptic may, with perfect consistency, concede that the perfect-hallucination hypothesis seems false. She may even believe that it is false. All people, whether skeptics or not, presumably believe this. But though we all believe the hypothesis is false, the skeptic argues that no one knows it is false. She concludes that because no one knows it is false, no one knows that any of his perceptual beliefs are true. Therefore, such beliefs do not constitute knowledge.
- An Epistemist Objection to the Third Premise:
The Skeptic's Failure to Distinguish Probable
Judgments from Lucky Guesses
Suppose that epistemist concedes, for the sake of avoiding an impasse, that the perfect-hallucination hypothesis does assert something, that it is either true or false. Further, we shall concede that the evidence we have from sense experience is always such that it is logically possible that our perceptual beliefs based on this evidence are false.
If we make these concessions, must we also concede the day to skepticism? Let us examine once again the logical structure of the argument for skepticism and see if we can find a defect. The argument is as follows:
1. The experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false.
2. If the experiences of a person who has a true perceptual belief may be exactly duplicated by the experiences of a person whose perceptual belief is exactly similar but false, then it is always logically possible that any of our perceptual beliefs are false.
3. If it is always logically possible that any of our perceptual beliefs are false, then no one ever knows that any of our perceptual beliefs are true.
Therefore
4. No one ever knows that any of our perceptual beliefs are true.
By defending the possibility of a perfect hallucination, the skeptic has substantiated premise (1). Moreover, let us concede the truth of premise (2) as well. To extricate ourselves from the consequences of skepticism we shall now direct our attack against premise (3).
The skeptic's defense of premise (3) can be found in her statement of the evil-operator argument (pp. 56-57), where she attributes the truth of perceptual beliefs to luck. This enables her to derive premise (3) from
8. If it is logically possible that any of our perceptual beliefs are false, then a perceptual belief that turns out to be true is nothing more than a lucky guess.
and
9. If any perceptual belief that turns out to be true is nothing more than a lucky guess, then no one ever knows that any of our perceptual beliefs are true.
Although the deduction of premise (3) from (8) and (9) is valid, premise (8) should be rejected. In this way we can avoid accepting premise (3). In premise (8), the skeptic has assumed that if it is logically possible for a belief to be false, then, when the belief turns out to be true, it is nothing more than a lucky guess, and hence not something we know. For convenience, let us label as corrigible any belief of such a kind that it is logically possible for a belief of that kind to be false. Almost all our beliefs about subjects outside logic and mathematics are corrigible.
The skeptical contention that corrigible belief leaves us entirely at the mercy of luck-as premise (8) asserts-is not difficult to refute. Evidence that does not exclude the logical possibility of error may greatly reduce the probability of error. Moreover, when our evidence reduces the probability of error to a point where it is negligible, it is preposterous to say that escaping from error is nothing but luck. To believe what is extremely probable, and to disbelieve what is immensely improbable, is completely reasonable. When a person believes something that is rendered exceedingly probable by the evidence on which he bases his belief, then it is no mere matter of luck if he is right and his belief is true. Beliefs that are sufficiently probable, even if corrigible, should be considered knowledge whenever they are true.
- A Skeptical Rejoinder: The Employment of the Notion
of Probability Begs the Question Against the Skeptic
Every attempt to escape skepticism knocks, sooner or later, on the door of probability. But there is no help behind the door. Indeed once we pass through it, we shall find ourselves securely locked in the very den of skepticism. Let us consider how to bolt the door.
If the appeal to probability is to succeed, the epistemist must go beyond claiming that our corrigible beliefs are based on evidence that renders them highly probable. It is not enough that a belief be highly probable; the one who has the belief must know this to be so, or the belief will again, if true, be so merely as a matter of luck. To see this, let us turn to the example of the gaming table Imagine that a person is invited to play a game of "Millee," which is played as follows: There is a machine with a window, which, when a button is pushed, closes and subsequently opens to display either a red or a green square. Moreover, part of the definition of the game involves the following rule: the machine must be set so that the green square appears only once in a million plays. Thus the odds are one in a million that the red square will fail to appear when the window opens.
Imagine that a person. is invited to play Millee but is not told the odds. He might choose to bet that the red square will appear when the window opens, and of course he is correct. In spite of the odds in his favor, we would be entirely justified in saying that his belief that the red square would appear was, from his point of view, a lucky guess. The reason is that he does not know that the odds are a million to one in his favor. Indeed, for all he knows, the odds might be anything at all. In the absence of,such knowledge, his being right is nothing more than luck.
Now suppose that a person believes that there really is a tomato in front of her, and that this belief is based on the evidence of sense experience, or any other inductive evidence you please. The odds might be a million to one that her hypothesis will turn out to be true when based on such evidence. But if the person does not know that these are the odds, and if, moreover, for all she knows the odds may be anything at all, then were she fight, this would be nothing more than luck. In both this case and the case of the person at the gaming table, being right is a matter of luck, even though the odds are fantastically in favor of both being right. It is a matter of luck because both of them are ignorant of the odds.
Of course, both the gambler and the perceptual believer would be in an entirely different position if they knew the odds. Then, neither would correctly be described as being right merely through luck. The question to be answered by the the skeptic is the following: Need we suppose that the perceptual believer is ignorant of the odds in favor of her belief? Could not a person who based her belief on evidence that rendered her belief highly probable also know how probable her belief is? To establish the case for skepticism, we must prove that the perceptual believer is inescapably ignorant of such probabilities.
As a first step, let us consider briefly how we ever know anything about probabilities. The term probability is interpreted in a number of different ways, but the idea that seems most relevant in this context is concerned with truth frequency. If a person is to convert his perceptual beliefs to epistemic gold by his knowledge of probability, he must know that his belief, based on the evidence he has, is the kind of belief that is more frequently-indeed, much more frequently-true than false, when based on the evidence he has. The perceptual believer must know that perceptual beliefs based on the usual evidence of sense experience are much more frequently true than false.
But we cannot know any such thing; for consider the problem a person faces who wishes to establish that perceptual beliefs based on the evidence of sense experience are much more frequently true than false. To fond this out he wouid have to find a sample of perceptual beliefs and determine how many of them were true. The information about such a sample would be absolutely essential to his finding out that beliefs of this kind are much more frequently true than false. But how is he to acquire this information? To obtain such knowledge, he must be able to determine how many of the beliefs in his sample are true. To do this, he must know which of them constitute knowledge and which do not.
However, this requirement is calamitous, because we have already shown that a perceptual belief can constitute knowledge only if the person knows that beliefs of this kind are much more often true than false. We have now concluded that in order to know that beliefs of this kind are frequently true, we must first know which beliefs of this kind are true and which false. Therefore, before we can know that any perceptual belief is true, we must first know that certain perceptual beliefs are true. This is an altogether pernicious epistemic situation. Moreover, the only alternative is skepticism. In short, either we know that certain perceptual beliefs are true before we know any perceprua beliefs are true, which is absurd, or we do not know that any perceptual beliefs are true. It is obvious the latter alternative must be accepted.
Let us review the argument briefly, In order to escape skepticism concerning corrigible beliefs, it must be shown that such beliefs are based on evidence which renders them highly probable, and also that we know those beliefs are highly probable. To know the latter, we must know that such beliefs, when based on evidence of a specified sort, are much more frequently true than false. However, to find out that such beliefs are much more frequently true than false, we must consider a sample of such beliefs and determine what percentage of the beliefs in the sample is true. To determine what percentage of the beliefs is true, we must know which of a certain sample are true.
Therefore before a person can know that any corrigible belief based on inductive evidence is true, he must know that a certain probability statement is true. But he cannot know such a statement is true unless he already knows that certain corrigible beliefs based on inductive evidence are true. Therefore, no one can know that any corrigible belief based on inductive evidence is true. The escape route via probability is, in fact, an expressway to skepticism.
- An Epistemist Objection: An Externalist Account of
Justification Undermines the Skeptic's
Argument Against Probability
The skeptic assumed the premise
8. If it is always logically possible that any of our perceptual beliefs are false, then a perceptual belief that turns out to be true is nothing more than a lucky guess.
We attacked the premise on the grounds that if a perceptual belief, like my belief that there is a book in front of me, is of a kind that is much more frequently true than false, then that belief is more than a lucky guess. If the odds are extraordinarily in favor of my being right, as they surely are then my belief, assuming it to be true, is no mere lucky guess. That is the epistemist argument.
What has the skeptic replied? She has said that if a person does not know the odds are in her favor, then the person's belief, if true, remains a lucky guess. But is the reply plausible? There is one kind of epistemist doctrine, externalism, that affirms, contrary to the skeptic, that a true belief that is externally related to the truth in the appropriate way is knowledge. It is not merely a lucky guess, even if we are ignorant of the existence of the external relationship. To see the point of externalism, suppose for the moment that our experiences are veridical rather than hallucinatory, and that the external world is what we suppose it to be. Suppose, as I believe, that there is a book in front of me. My belief that there is a book in front of me arises because of this truth. The fact that there is a book before me causes me to have the experience of seeing a book, which in turn causes me to believe that there is a book in front of me. This process is a perfectly reliable one which gives me information about the existence of an object in the external world, the book before me. When I am correct in believing that there is a book before me, that is not merely luck! My belief is the product of a very reliable process. To say that the process is reliable is to say that the process most frequently produces true beliefs and very rarely produces false ones. A belief resulting from such processes, when it is true, is no mere lucky guess. The process connects us with the external world.
The critical point, contrary to the skeptic, is that I need not know that the belief in question arises from such a process to avoid mere guessing. It suffices that the process reliably produces true beliefs. In the case of the game of Millee, luck favors my belief, but the process producing my belief is not reliable. I see a red square on the first play of the game, and believe that I will see a red square on the second play of the game. I am right, but the process of forming a belief that the previous outcome in a game of chance will repeat itself is, in general, not a reliable belief-forming process. Note the contrast with perception. Perception is a reliable belief-forming process, even it it is always logically possible that our perceptual beliefs are mistaken, for, in fact, our perceptual beliefs are reliably produced and not mere guesses. The logical possibility of being mistaken in our perceptual beliefs fails to yield the consequence that such beliefs, if true, are mere lucky guesses. They may be the products of reliable processes which connect us to the external world in a way required to obtain knowledge. Premise (8) is dissolved the application of the solvent of externalism.
- A Skeptical Rejoinder: The Externalist Begs
the Question Against the Skeptic
The epistemist, embracing the doctrine of externalism, has argued that beliefs arising from a reliable belief-forming process are, when true, not lucky guesses but knowledge. The problem with the argument is that the epistemist has assumed that there are reliable belief-forming processes when he is not entitled to that assumption in our dispute. If the story about Dr. O were true, then our belief-forming processes would be unreliable, and we would have no way of knowing that this is so. Since the story about Dr. O represents a logical possibility, our processes may be unreliable and our beliefs, if true, may be mere lucky guesses. We must, however, concede one point to the epistemist. The argument from the logical possibility of perceptual error to the conclusion that true perceptual beliefs are lucky guesses is not logically valid. The possibility that such beliefs are the results of reliable belief-forming processes is a counter-example.
Skepticism is, however dialectically fertile and readily gives birth to a new premise to supplant the old one-Suppose, for the moment, that a belief is the product of a reliable belief-forming process, but we do not know that the process is reliable. In that case, though our beliefs may be more than lucky guesses, we will not know this. Consider a fanciful example. A young child, Alice, who is eight, is endowed with a peculiar ability. When someone asks her what the social security number of a named person is, Dane Taylor, for example, a number immediately occurs to her, 598-81-9908. She has no idea why or how the number occurs to her, but she believes it is the social security number of the person in question, Dane Taylor. Now suppose that the process is a reliable one, perhaps arranged by a Dr. O who has gained access to the braino on Alice's head and the register of social security numbers, but that Alice has no idea that the process is a reliable one. As far as she can tell, her beliefs about social security numbers are mere guesswork. But then she does not know that her beliefs about social security numbers are true!
This fanciful example is a counterexample to the externalist assumption that a true belief that is the product of a reliable belief-forming process is knowledge. That assumption is false, precisely because a person might, like Alice, be ignorant of the fact that her belief is the product of such a process. Moreover, the logical possibility that our perceptual beliefs might be in error insures that even if our belief-forming processes are connected with the external world in a reliable manner, we shall be ignorant of this fact and, therefore, be ignorant of the truth of those beliefs. We, like Alice, may have beliefs that are the products of such processes, but we, again like her, cannot tell that this is so, because of the logical possibility of the deceptions of Dr. O.
Our skeptical reply to the epistemist incorporates two new premises
10. If it is always logically possible that any of our perceptual beliefs are false, then we cannot tell whether or not a perceptual belief that turns out to be true is a lucky guess.
11. If we cannot tell whether or not a perceptual belief that turns out to be true is a lucky guess, then we never know that any of our perceptual beliefs are true.
From these two premises we deduce premise (3) of our argument
3. If it is always logically possible that any of our perceptual beliefs are false, then we never know that any of our perceptual beliefs are true.
From this premise, together with the premise that it is always possible that any of our perceptual beliefs are false, which follows from the logical possibility of the story about Dr. O, we derive our skeptical conclusion that we never know that any of our perceptual beliefs are true.
The preceding argument against the externalist defense admits of terse summary. Contrary to the externalist, we have to know that our beliefs are not lucky guesses in order to know that they are true. To know that our perceptual beliefs are the products of reliable belief-forming processes and not lucky guesses, we would have to know that those processes produce beliefs that are very frequently true and very infrequently false. So our excursion through externalism leaves the epistemist in his original quandary. He must assume that we know that our perceptual beliefs are frequently true in order to show that we know that any of them are true. To assume that is shamelessryfb beg the question.
- An Epistemist Objection: The Principle of
Charity Supports Externalism
Can an externalist justify assuming that our perceptual beliefs are frequently true, and the products of reliable belief-forming systems, without begging the question against the skeptic? There is an argument that says that a principle of clarity dictates this result. The principle of charity says that when we interpret the speech of another we should interpret it in such away as to maximize the true beliefs we ascribe to others on the basis of what we believe to be true. Suppose I see the sun shining and you say to me, "Sunny day isn't it?" It would be perverse of me to suppose that you believe it is raining and interpret your use of the word 'Sunny' to mean 'Rainy.' Why? I cannot prove you believe it is sunny, but, given what I believe about the weather, that is the most charitable interpretation for me to make of what you believe. On the basis of what I believe about the weather, your belief about the weather would be true if my belief about the weather is true. Notice that this argument appeals only to what the interpreter believes to be true. When I am charitable toward you, I should interpret what you say in such a way as to maximize the true beliefs I ascribe to you, based on what I believe to be true. It does not presuppose that I know what is true. Therefore, application of the principle of charity to perceptual beliefs of others does not presuppose that we know that our perceptual beliefs are true. So appeal to the principle does not beg the question.
The principle of charity tells us to be charitable toward others concerning the truth of their beliefs. If we follow the principle of charity, we will interpret the perceptual beliefs of others in such a way that their perceptual beliefs very frequently turn out to be true and very infrequently turn out to be false, at least in terms of what we believe. So the principle of charity yields exactly the conclusion we want concerning others, namely, that their perceptual beliefs are frequently true, indeed, almost always true. We may, moreover, apply the principle to ourselves by following the dictum that charity begins at home and conclude that our perceptual beliefs are also frequently true. Whether we apply the principle of charity to ourselves as well as to others, as it seems only equitable to allow, we obtain the desired conclusion that perceptual beliefs are almost always true.
- A Skeptical Rejoinder: Charity without Knowledge
The principle of charity may be a useful method for interpreting the remarks of others in some cases, but it is useless, as a defense of the epistemist. What it tells us, in effect, is to assume that others have beliefs about what is true that correspond as fully as possible with our beliefs about what is true. Sometimes this is unreasonable, for example, when I know that you hold some peculiar metaphysical view, like that of Leibniz, who held that everything was made up of minds and that matter did not exist. It would be distinctly uncharitable to Leibniz to interpret him in such a way as to assume that he did not believe that everything was made up of minds. Even in those cases in which we have no reason to think that the belief system of the other is peculiar, and where application of the principle of charity does seem appropriate, it only warrants me in so interpreting the beliefs of the other that they accord as fully as possible with my own beliefs. Thus, to say that I should interpret the beliefs of others in such a way as to maximize the truth of those beliefs is just to say that I should interpret their beliefs in such a way as to maximize correspondence with my own beliefs.
The application of the principle of charity to myself is therefore redundant. My beliefs already correspond to my beliefs, and so the application of the principle to myself adds nothing whatsoever to my beliefs. It leads me to believe that my beliefs are true, but of course, I already believe that my beliefs are true, because to believe something is just to believe that it is true. To believe that I see my nose is just to believe that it is true that I see my nose. The principle of charity is, therefore, ineffective in proving that our perceptual beliefs are frequently true. In general, therefore, the epistemist has failed to show that our perceptual beliefs and our corrigible beliefs based on inductive evidence are even probable. To show that they are even probable, the epistemist must show that they are very frequently true and very infrequently false. This the epistemist cannot do, without begging the question by assuming that we know such beliefs are true in order to prove that we know precisely this.
- An Epistemist Objection: Internalism and
Complete Justification
The skeptic's reply to externalism has proven it to be an insufficient bulwark against skepticism. Fortunately, externalism is no more necessary than it is sufficient for a defense of the epistemist position. To show that our perceptual beliefs and other corrigible beliefs are probable, we do not need to show or know anything about frequencies. The skeptic lured the epistemist into her den by converting the discussion of probability into a discussion about frequency. What is required for knowledge is justification for our beliefs, and probabilities are adequate to provide such justification. The justification arises, however, not from a connection with the external world, as the skeptic alleges and her externalist opponent unwisely conceded. Justification arises, instead, from internal standards of evidence. It is, therefore, internalism not externalism that provides a defense against skepticism. It is the internal connection between our beliefs and our internal standards of evidence and justification that support the edifice of knowledge. Let us go inside for a defense of knowledge.
An internalist critique of skepticism will focus on the skeptical premise (10) above.
10. If it is always logically possible that any of our perceptual beliefs are false, then we cannot tell whether or not a perceptual belief that turns out to be true is a lucky guess.
Complete justification is a condition of knowledge. If we can tell that we are completely justified in believing something, that there is a book before me, for example, then we can also tell that the belief is not a guess and, if true, not a lucky guess. The internalist thesis is as follows: Though it is always logically possible for any perceptual or corrigible belief to be false, we can often tell that the belief is not a guess, because we can tell that it is a completely justified belief. Moreover, we can know that our corrigible beliefs are completely justified without first knowing how frequently such beliefs are true. We can know this by appeal to internal principles or standards of justification.
The very character of some beliefs stamps them as epistemically abhorrent and unjustified. A contradictory belief is one example, and a belief that flies in the face of overwhelming evidence is another. A person who refuses, in ordinary circumstances, to believe what her senses would lead her to believe, and who, moreover, believes quite the opposite instead, is an epistemically wanton and unreasonable person. On the other hand, a person who believes precisely what the evidence of her senses leads her to believe, is completely justified in terms of our internal standards of epistemic justification.
We can see that this is true by considering our predicament in the most desperate epistemic situation, namely, one in which Dr. O is constantly deceiving us. Even in this situation, we can distinguish between justified and unjustified belief by appeal to our internal standards of justification. I, being presented with expenences that duplicate those of seeing a book and no reason to doubt that there is a book before me, believe that there is a book there. Another, having exactly the same experiences, believes that the great pumpkin stands before her, even though her experiences duplicate those of seeing a book, and she has no reason to think that the great pumpkin stands before her, nor any reason to doubt that there is a book before her. However amusing her beliefs may be, it is clear that she is epistemically unjustified in her odd belief and that I am completely justified in mine. Given the deceptions of Dr. O, both of our beliefs are in error, but mine is completely justified by the standards of epistemic justification, and hers is not. Our internal standards of epistemic justification suffice to tell us that certain kinds of beliefs are completely justified no matter what the external world is like.
In summary, our internal standards of evidence and justification ensure that at least some of our corrigible beliefs, perceptual beliefs, for example, are completely justified. Thus, though such beliefs are corrigible, we can easily tell by appeal to such internal standards that such beliefs are more than guesses, and, if true, more than mere lucky guesses.
- A Skeptical Rejoinder: Satisfaction of
Internal Standards Is Not a Guide to Truth
The internalist speculations of the epistemist rest on an appeal to internal standards. Such appeals are the common refuge of all who seek to escape from the pains of inquiry and criticism. It is time to expose this form of reasoning as the intellectual protector of the status quo. Having exposed it, we can then elaborate a skeptical alternative to the dogmatic conservation of accepted opinion.
Let us lay out the internalist argument of the epistemist with greater care than he has been wont to do. It is as follows:
1. Some of our beliefs are completely justified in terms of our internal standards of epistemic evaluation.
2. If any of our beliefs are completely justified in terms of our internal standards of epistemic evaluation, then those beliefs are completely justified even if beliefs of that kind almost always turn out to be false.
Therefore
3. Some of our beliefs are completely justified even if beliefs of that kind almost always turn out to be false.
This is the argument, and surely it need only be stated this baldly to be rendered ineffective for the purpose of refuting skepticism. The internalist, in presenting an epistemist argument, has very cleverly attracted our attention to premise (1). But his strategy is doomed by the inadequacy of premise (2), which is needed to bring us to the epistemist conclusion.
The problem for the epistemist is that it does not at all follow, from the fact that something is completely justified in terms of our standards of evaluation, that is completely justified for the purpose of attaining knowledge. This is especially clear when the kind of belief in question almost always turns out to be false. What does it mean to say that some belief is completely justified in terms of our internal standards of evaluation? What it means, surely, is that we accept a principle according to which beliefs of that kind are completely justified. But the fact that some principle is accepted in no way shows that it is true. Hence, when such a principle tells us that a belief is completely justified, we still may reasonably ask whether the principle is correct. If it is not, then the beliefs it certifies as completely justified may be absolutely counterfeit; that is, they may not be completely justified at all.
To conclude that some belief is completely justified, because it is so justified in terms of some internal standard of evaluation we accept, is to offer an argument without any merit whatever. For it simply does not follow in any way that we are completely justified in some belief merely because that belief conforms to some internal standard of evaluation is intended to fulfill some purpose. What purpose are our internal epistemic standards supposed to fulfill? Obviously, they are supposed to lead us to obtain truth and avoid error. What then if conformity to our internal epistemic standard makes us fall into error? Then they are defective, a snare and delusion, and we should set them aside.
The point may be further illustrated, and usefully so, by considering a controversial epistemological claim. Imagine that some person is entirely convinced that she has some extrasensory powers, and more specifically, that she can tell what cards are drawn from a deck, even though she does not see the cards, by concentrating in a special way. She then repeatedly claims to know what card is drawn from the deck. When we challenge her claim to know, she says her claim is completely justified in terms of her internal epistemic standards. We then note that she is much more often incorrect than correct in such claims; indeed, we might even note that she is no more often correct in what she says than one would expect by chance. She then regards us with disdainful incredulity and remarks that apparently we have not understood. She informs us that her beliefs in such matters are completely justified in terms of her internal standards of epistemic evaluation, and hence her beliefs are completely justified, even if those beliefs are mostly false. Those are her standards, and there is an end to it.
No one would accept that argument as having any credibility whatever. It obviously does not follow from the fact that a belief is completely justified in terms of her internal standards that such a belief is completely justified. This point remains cogent when generalized. No matter what her internal standard might be, it does not follow from the fact that the belief is completely justified according to her internal standard, that the belief is epistemically justified. It might not be epistemically justified at all.
We are likely to overlook this point when we accept the standards in question. But the internalist argument of the epistemist is no better than the argument of our self-acclaimed mistress of extrasensory powers. The objective of epistemic justification is the attainment of truth and the avoidance of error, when an internal epistemic standard is useless for this purpose, it fails to yield the complete justification required by our epistemic enterprise. The appeal to internal epistemic standards is but one more ineffective attempt to evade skepticism. An external connection with truth is not, as the externalist claimed, sufficient for epistemic justification, but it is necessary to that end. The epistemist must show that his internal standards of justification are connected with truth, that perceptual beliefs, for example, are more frequently true than false, or candidly concede defeat.
- An Epistemist Objection: Innocent Justification
It may be conceded that our appeal to internal epistemic standards is ineffective if some argument is required to show that such standards are connected with truth. But is anysuch argument needed? The skeptic continually supposes that we have to know that certain frequencies obtain, or that certain frequencies do not obtain, if we are to know that our corrigible beliefs are true. But she has put the shoe on the wrong epistemic foot. Certain corrigible beliefs-for example, cautious perceptual beliefs and distinct memory beliefs-we need not justify by antecedently establishing that beliefs of this kind are more frequently true than false. On the contrary, unless there is some reason to believe that such beliefs are more frequently false than true, we are completely justified in such beliefs. One way of putting the matter is to say that such beliefs are epistemically innocent until proven guilty. They are justified unless they are shown to be unjustified.
Moreover, often our perceptual beliets are so completely justified that it would be epistemically pointless, and indeed unreasonable, to attempt to justify them by arguing that such beliefs are more frequently true than false. For example, if I see my index finger and feel it there as well, my belief that there really is such a finger is not one that could reasonably be defended by such an argument. This belief is so completely justified that any such argument mustered in defense of it would surely proceed from premises less evident than the conclusion they support. Such an argument would be epistemically useless. For, to justify a conclusion, some of the premises of the argument must be more evident than the conclusion, and none of the premises may be any less evident. Only such an argument can add epistemic weight to the conclusion. Unless some of the premises are initially more evident than the conclusion, there will be no more reason for accepting the premises than there was originally for accepting the conclusion. Thus, for an argument to justify the conclusion deduced from the premises, at least some of the premises must be more evident, more reasonable to accept, than the conclusion.
However, the belief about my index finger is so completely justified that there is no belief that is more reasonable or evident. When I see something that I also touch and feel, in the absence of evidence to the contrary, the belief that such a thing exists is so evident, so reasonable, that it would be pointless to seek a frequency argument to justify the belief. To what premises could I appeal? Surely, any premise to which I might appeal would be less evident, or at least no more evident, than the very belief I was attempting to justify. The belief is so completely justified that no such argument to justify it could reasonably be given.
- A Skeptical Rejoinder: Appeal to Innocent
Justification Does Not Rid the Epistemist
of the Burden of Proof
The primary defect of this defense of epistemism is the manner in which all equity and fairness in disputation are set neatly aside for the convenience of epistemism. We begin by questioning whether perceptual beliefs are completely justified. And what is the response to our query? It is the bold assertion that these beliefs are so evident and completely justified in themselves that no argument can even be offered to sustain them. But this reply constitutes a most immediate and obvious begging of the question against skepticism. The epistemist has simply laid it down that what seems most evident and completely justified to him must be conceded to be completely justified without argument or debate. We agree that the beliefs in question may seem completely justified to the epistemist, indeed, so completely justified that no argument could serve to render those beliefs any more evident to him. What we deny is that those beliefs are com-pletely justified, and what we require is some argument to convince us that those beliefs are so justified. The epistemist has become desperate with his dubious appeals to innocence, and we are left unsatisfied.
If matters remain at the level of denial and simple assertion, we have arrived at an impasse, and neither party to the dispute can claim victory. It is essential that we move beyond this level of argumentation. How should we proceed? With impartiality, of course. A principle of impartiality requires that until some justificatory argument is offered, we shall not assume that the claims of either party are justified or unjustified. But here is the crux: this principle of impartiality curiously favors the skeptic. If the claims of neither the epistemist nor the skeptic are assumed to be completely justified, then the perceptual beliefs of the epistemist must not be assumed to be completely justified. Therefore, they must not be assumed to constitute knowledge either. In this way, simple fairness and impartiality in discourse and disputation sustain the case for skepticism. We must assume at the outset that the beliefs in question, including those perceptual beliefs cited by the epistemist, are not completely justified until some argument is presented to justify them. If skepticism is treated fairly before the bar of evidence, the burden of proof must reat entirely with the epistemist.
- An Epistemist Objection: An Argument to the
Best Explanation Provides the Required Proof
The reply to the skeptic, ingenious as she is proving to be, is to shoulder the burden of proof and provide a justification for our corrigible beliefs, perceptual beliefs, for example, without begging the question. The burden is not a heavy one. We need only compare the fanciful hypothesis of the skeptic to our modest perceptual beliefs to notice the superiority of the latter. To avoid begging the question, however, let us consider our most modest perceptual beliefs as mere hypotheses and compare them to the skeptical hypotheses. Let us proceed scientifically, and ask which hypothesis we should accept.
Certain experiences lead me to believe that a book is before me. Here are two hypothesis. The epistemist hypothesis is that I have the experiences I do because there is a book before me which I see. The skeptical hypothesis is that I have the experiences that I do because there is a braino attached to my head, controlled by Dr. O, who is attempting to deceive me into believing that there is book before me when there is not. How should we decide which hypothesis to accept? Surely the choice is not a difficult one, if we proceed scientifically. We need only ask which hypothesis provides the best explanation, the epistemist hypothesis or the skeptical hypothesis. When confronted with competing hypotheses, a scientist asks which hypothesis provides the better explanation of data, the observed path of the planets, for example. The better explanation, the Copernican hypothesis that the planets revolve around the sun, rather than the Ptolemaic hypothesis that the sun revolves around the earth, is the one the scientist accepts.
Now compare the hypothesis that there is book before me to the skeptical hypothesis that a braino controlled by Dr. O is attached to my head. Which is the best hypothesis to explain my experiences? The epistemist hypothesis, without a doubt. It is a better explanation to suppose that there is a book before me than to accept in the fanciful story of Dr. O. Justification and knowledge arise from an inference to the best explanation of the experiences from which we began. We are completely justified in our perceptual beliefs precisely because they provide the best explanation of our perceptual experiences. I have all the familiar experiences of seeing a book before me. The explanation is that there is a book before me which I see. What could explain better? Nothing? Then the epistemist is right. We are completely justified. The conclusion easily bears the burden of proof against the skeptic.
- A Skeptical Reply: Explanation without
Truth Is Worthless
Is the epistemist hypothesis a better explanation than the skeptical hypothesis, if the skeptical hypothesis is true and the epistemist hypothesis is false? On the contrary, if our perceptual experiences are caused by Dr. O, then that is the better explanation of those experiences. When the epistemist says that his explanation is best, he is not providing a justification for our perceptual beliefs; he is begging the question by appealing to his epistemist assumption that such beliefs are true. Of course, if a perceptual belief is true, if really I see a book, for example, as I believe I do, then the hypothesis that I see a book is a better explanation of my perceptual experiences than the hypothesis that my experiences are generated by Dr. O. If, however, my experiences are caused by Dr. O, then the best explanation of my experiences is that they were caused by Dr. O. It all depends on which explanation is true. The best explanation is the true one.
Sometimes we are not completely justified in accepting that one rather than the other of two conflicting hypotheses is true, when we judge that one explains some phenomena better than the other. For example, two conflicting hypotheses might explain the motions of some planet, say the irregular orbit of Mercury, and we might note that one is simpler than the other. If we do not know which of the hypotheses is true, we may regard the simpler hypothesis as the better explaination. In general, when we are unable to decide which of two hyptheses is true, we may appeal to some explanitory feature of the two hypotheses, such as simplicity, to decide which is better.
That one hypothesis is better than another in terms of some explanatory feature other than truth fails to provide a complete justification for accepting that the better explaination is true, however. One explaination may be better than another without our being completely justified in accepting that either of them is true, because neither of them is the best explaination of the phenomena in question. The best explaination must be a true explaination, simply because a true explaination is better than a false one. Moreover, this assumption is essential to the argument of the epistemist. If a best explaination could be false, then it would be useless for the epistemist to argue that his hypothesis provides the best explaination. Whatever explainatory merits other than truth his hypothesis might have, they would fail to show that we are completely justified in accepting his hypothesis as true.
It is truth, not other explanatory merits, that is the objective of knowledge and complete justification. What the epistemist must prove, therefore, is that his hypothesis is true, and he cannot do this by appealing to other explanatory merits of his hypothesis. Explanation without truth is epistemically worthtess. The epistemist appeal to the best explanation must beg the question by assuming the truth of such explanation or else become irrelevant to the reality of knowledge and complete justification. Thus, the appeal proves impotent against the skeptic.
4) Another Argument for the Skeptical Conclusion:
Complete Justification As Excluding the
Chance of Error: The Lottery Paradox
In fact, as we shall now prove, a completely justified belief must be one that leads to truth without any risk for error. We shall argue that a person is not completely justified in what she believes unless there is no chance whatever that she is mistaken. If there is some chance that she is mistaken, however small, then she is not completely justified in her belief, and therefore, she lacks knowledge. This thesis is highly contentious. People often say that they know, when there is obviously some chance that they are in error. So ordinary speech suggests that we should not require that all chance of error be excluded before a person may be said to know. However, as we shall prove by appeal to the lottery paradox, a contradiction results from assuming that a belief may be completely justified when some chance of error remains. Hence we shall conclude that this assumption must be rejected to save our conception of knowledge from inconsistency.
Suppose, for the sake of argument, that a belief could be completely justified without all chance of error being excluded. How great a chance of error is to be allowed? One chance in ten? One chance in a million? It won't matter. If there is one chance n, whatever number n may be, we shall be led into contradiction. Imagine we say one chance in a million is acceptable. Now, suppose we set up a fair lottery with a million tickets numbered consecutively from 1, and that a ticket has been drawn but not inspected. Of course, there is only one chance in a million that the number 1 ticket has been drawn. So by the current proposal, we would be completely justified in believing that the number 1 ticket was not picked. There is only one chance in a million of error. Hence we would be completely justified in claiming to know that the number 1 ticket was not picked.
Moreover, people really do speak this way about lotteries; they do say they know that the ticket they hold was not drawn because there is so little chance of it. However, a similar claim can be made concerning the number 2 ticket, for there is equally little chance that it was picked. So we can say that we know that the number 2 ticket was not picked. But then the same reasoning applies to each ticket in the lottery. Of each ticket in the lottery, we would be completely justified in believing, and, hence, in claiming to know, that the ticket has not been drawn. But the set of things we would thus claim to know is inconsistent. It is contradictory to claim that each of the tickets in a fair lottery with one winning ticket is net the winner. For if each is not the winner, then the lottery with one winning ticket has no winning ticket. Of course, requiring the chance of error to be less than one in a million will not help. However small the chance, we can find a large enough lottery to create the paradox. Since the assumption that a belief may be completely justified though there is some chance of error leads to contradiction, we must reject it. To analyze knowledge in terms of complete jusification that allows for some chance of error is to render knowledge logically inconsistent.
We now have established a critical premise in our final argument for skepticism, namely, that if a person is completely justified in what he believes, then his justification must exclude all chance of error. There is some chance, however slight, that the braino hypothesis, or some other hypothesis of the same skeptical cut, is true; and therefore some chance that our ordinary perceptual beliefs are in error. Therefore those beliefs are not completely justified. Because they are not completely justified, we do not know that they are true.
- Summary of the Argument
The argument just advanced may be laid out as follows:
1. If anyone knows that any perceptual belief of his is true, then he is completely justified in his perceptual belief.
2. If anyone is completely justified in his perceptual belief, then his justification for his perceptual belief excludes all chance of error.
From these two premises we conclude Therefore
3. If anyone knows that any perceptual belief of his is true, then his justification for his perceptual belief excludes all chance of error.
Having reached this conclusion, we appeal to material from an earlier argument to reach our skeptical conclusion.
4. If there is some chance that the braino hypothesis is true, then the justification anyone has for his perceptual belief does not exclude all chance of error.
5. There is some chance that the braino hypothesis is true.
Therefore
6. The justification anyone has for his perceptual belief does not exclude all chance of error.
From conclusions (3) and (6) we obtain our further skeptical conclusion
7. No one knows that any perceptual belief of his is true.
- An Epistemist Objection: The Skeptic's Braino
Hypothesis Is a Mere Possibility
We may concede most of this argument without conceding the conclusion. For we may deny there is some chance the braino hypothesis is true. We have conceded that the braino hypothesis is logically possible. But the logical possibility of truth does not show there is any chance whatever that the hypothesis is true. To argue that a belief is not completely justified because some conflicting hypothesis is logically possible is to argue fallaciously. We have shown this earlier. So, if the skeptic attempts to argue from the logical possibility of the braino hypothesis to the conclusion that there is some chance that it is true, her argument will be ill-founded. There is no chance that the braino hypothesis is true; it is simply preposterous.
- A Skeptical Rejoinder: The Survival Hypothesis:
A Realistic Skeptical Hypothesis
We do not by any means concede that there is no chance that the braino hypothesis is true. After all, how do you know that it is false? Note that any alleged evidence of the falsity of the hypothesis could be explained in terms of the attempts of Dr. O to mislead us. However, again for the sake of argument, it will be useful to present a skeptical hypothesis that is more apt to obtain agreement from an impartial consideration. So we shall construct, ultimately, a hypothesis which obviously has some chance of being correct.
Since our capacity for scientific discovery is not matched by our capacity to use scientific discoveries wisely, our survival as a species may depend on our ignorance of the true nature of the world. There is at least some chance that if our perceptual beliefs were not slightly incorrect, then we would indeed destroy ourselves, as a result of scientific discovery and mishandled technology based upon it. Thus, we propose that there is at least some chance that erroneous beliefs have survival value, and, moreover, that the erroneousness of our perceptual beliefs has saved us from destroying ourselves long ago. There is some chance, however small, that the erroneousness of our perceptual beliefs has survival value.
Suppose, to illustrate, that there is some particular theory that is especially dangerous to humankind. Imagine that some discovery in physics would enable us to understand how to release vast amounts of energy in a simple way with common materials. If we imagine, moreover, that such devices might have the power of thermonuclear bombs and that anyone could easily learn how to construct them from materials to which we all have access, then we can see that such devices would place us all at the mercy of demented and desperate people willing to destroy themselves to destroy others. And then the holocaust would result from the madness of the few.
The foregoing is but one way in which the discovery of some principle might serve to destroy us. One might imagine countless others. If the discovery of such a principle would obliterate life, then the failure to discover it would be necessary for our continued existence. Now one way in which we might be prevented from discovering such a principle is by virtue of perceptual error. If we are misled at the perceptual level, our attempts to check those theories that might lead to the fatal one will be subverted at the level of observation. We shall, by dint of our defective observational beliefs, be encouraged to accept some slightly incorrect theories that are more probable, in terms of our slightly erroneous perceptual beliefs, than those correct but fatal theories we reject. In short, theory is based, either directly or indirectly, on observation, and, therefore, faulty observation, erroneous perceptual beliefs, can prevent us from arriving at correct theories. We propose that there is some chance, however small you might think it is, that our very survival at this moment depends on our failure to accept some correct theory, because of our erroneous perceptual beliefs. In this way, then, there is some chance that our perceptual beliefs are erroneous, and, indeed, that our survival has depended on it. If you ask how the error could have gone undetected, the answer is that, had it been detected, no one would be here now to report the result.
A bit of science fiction may help you to imagine how our survival might depend on perceptual error. Imagine that a group of very intelligent beings, living in a remote galaxy, have the means to observe the earth from a scientific spaceship concealed from our view. Let us call these beings googols after the number googol, which is ten to the hundredth power, Because their intelligence is ten to the hundredth power greater than that of the most intelligent earthling. One googol scientist, John, reaches an alarming conclusion about the humans, namely, that their rate of scientific discovery will soon enable them to create very inexpensive weapons with sufficient power to destroy themselves. His computation leads him, moreover, to conclude that the weapons will actually be used within fifty years, due to the slow rate of moral development of humans and their inadequate understanding of how to control their aggressiveness. The humans are clearly an endangered species, and John places them on the urgent concern list, offering a prize to any googol who can find a method for saving the humans.
One googol genius, Mary, who has special fondness for the foolish humans, comes up with an elegant solution. It is possible to introduce a very small microbe that will live as a parasite on the brain tissue of humans, being passed from one to the other, which will produce a minor distortion in the way in which humans perceive the world so that their perceptual beliefs will all be slightly inaccurate. This will introduce errors into their scientific observations that will prevent them from making the scientific discoveries that would lead to the creation of the endangering weapons. There is a special elegance to the solution resulting from the fact that errors caused by the microbe will also prevent the humans from making the scientific discovery that would reveal the microbe itself. Mary points out that a minor modification of the microbe would reduce human fertility to a level that would be more compatible with resources of their habitat and proposes the modification be added.
After careful study as to whether the microbe would have harmful side effects, or whether the humans might discover the weapons by other means, it is concluded that Mary's solution will be effective and benign. Mary, revealing typical googol integrity, points out that, strictly speaking, all perceptual beliefs of the humans will be false once the microbe is introduced. Indeed, she insists that it is improper to describe the post-microbe-infestation perceptual beliefs of humans as merely inaccurate, for they will be erroneous, even if only very slightly so. Her point is thoughtfully considered by the referees, but her solution is deemed much less invasive than other alternatives, which range from turning the earth into a kind of human park for googol visitation to preserving a sample of living human brains in vats, and Mary wins the prize. The microbe is introduced undetected onto earth, it functions as predicted, and the humans are saved by their microbe-caused errors of observation. The googols extend Mary the prize of their esteem and celebrate the survival of the humans.
5) Summary of the Final Skeptical Argument
We now present a summary of the preceding argument. In our summary, we refer to the hypothesis that the erroneousness of our perceptual beliefs has survival value as the survival hypothesis. The argument then is as follows:
1. If anyone knows that any perceptual belief of hers is true, then her perceptual belief is completely justified.
2. If anyone is completely justified in her perceptual belief, then her justification for her perceptual belief excludes all chance of error.
Therefore
3. If anyone knows that any perceptual belief of hers is true, then her justification for her perceptual belief excludes all chance of error.
We continue our argument:
4. If no one knows that the survival hypothesis is false, then no one's justification for her perceptual belief excludes all chance of error.
5. No one knows that the survival hypothesis is false. Therefore
6. No one's justification for her perceptual belief excludes all chance of error.
Finally, from statements (3) and (6) we conclude
7. No one knows that any perceptual belief of hers is true.
We have, by this argument, formulated our argument for skepticism. We shall rest the matter here. However, it is important to understand the implications of the doctrine. Standards of evidence and epistemic evaluation telling us that some beliefs are completely justified, beyond all risk and chance of error, must be laid aside in favor of a more skeptical and flexible theory of reasonable belief. We are not repudiating reason; instead, we are claiming that nothing is beyond its reach. Nothing is so secure or sacrosanct as to be beyond rational criticism. Hence we may always ask whether it is reasonable to accept some statement as evidence for a hypothesis. The question is not resolved by appeal to some infallible beliefs beyond all chance of error. It is settled instead, tentatively and subject to subsequent reconsideration, by appeal to the probabilities, to the very genuine risk of error we admit. The reasonable statement to accept is the one that is more probable than those with which it competes for the status of hypothesis or evidence. In this way, and by induction, we may proceed to reason in terms of evidence and hypothesis without dogmatism or pseudocertainties of knowledge.
In conclusion, it is important to point out that we are not disputing here over the mere use of the word 'know.' Our objection to the dogmatic contention that people know certain of their beliefs to be true is the roadblock to inquiry indigenous to such claims. If a person says he knows that something is true, then he intends his listener to take what he says as true on his authority. It is not a matter to be questioned. The word 'know' functions this way in ordinary discourse, and we consider this a defect of such discourse. We affirm the right and the need to submit any statement or belief to criticism and requisite justification. None are allowed exemption from this ordeal of reason. We concede, of course, that a person might succeed in using the word 'know' without such dogmatic implications. And if someone uses the word in some weaker sense, allowing for fallibility, the chance of error, and the appropriateness of criticism, then we wish him well.
C) New Knowledge as Undefeated Justification:
A Revisionist Alternative to the Skeptic
and the Epistemist
Let us reflect on the dispute between the skeptic and the epistemist. The skeptic has proven that our perceptual beliefs and corrigible beliefs generally are not completely justified in any way that guarantees the truth of those beliefs and excludes all chance of error. Must we concede the day to the skeptic? The arguments_of the skeptic are formidable. What have we learned from her? We have learned that all justification runs some risk of error. Any justification for what we believe is fallible. When we seek a justification for what we believe, the best we can find will inevitably fall short of guaranteeing the truth of what we believe. Justification can aim at truth but cannot eliminate the risk of error. If our search for knowledge is the quest for complete justification and a guarantee of truth, we must admit our ignorance and concede the day to the skeptic. There is another way, however.
We can revise our conception of knowledge. We may redefine knowledge without committing the redefinist fallacy by admitting that our new conception is a revision. We can construct a new conception of knowledge and make this new knowledge the object of our philosophical quest. How can we do this? We begin by admitting that our justification for what we believe remains fallibIe and falls short of a complete justification. We continue by noting that the fallible justification we do have tor our beliefs, the sort appealed to by the internalist, for example, may prove a trustworthy and reliable guide to truth. Such justification may lead us to truth without being based on any false premise or assumption. These reflections show us how to revise our conception of knowledge. The revisionist takes fallible justification rather than complete justification as the basis of knowledge, and affirms that when fallible justification for our beliefs does not depend on error and leads us to truth, we attain a new kind of knowledge. This kind of knowledge based on fallible justification becomes the legitimate object of philosophical and scientific inquiry. In this way, revisionism transcends epistemism and skepticism, combining the insights of both. We have not been able to prove the skeptical hypotheses to be false. We believe, however, that those hypotheses are fanciful, false constructions of the imagination, rather than a truthful account of our world. We believe that our perceptual beliefs about the objects we see, hear, and touch inform us in a trustworthy way about the truth of those objects. We believe, therefore, that beliefs that are justified by our internal standards of justification, though those standards be fallible guides to truth, are also externally connected with truth in a trustworthy and reliable manner. We believe all this.
Suppose, in fact, that our fallible internal justification for our perceptual beliefs and other corrigible beliefs does not rest on error but instead leads us to truth in some trustworthy and reliable manner, as the externalist maintains. Then a revised conception of knowledge lies shining before us. One component is fallibilism, which we take from the skeptic. Another component is internal justification, which we take from the epistemist and the internalist. The final component, which we take from the epistemist and the externalist, is that of justification that is undefeated by error and that connects us with truth in a trustworthy and reliable manner. It is easy to assemble the components, as we have seen, to obtain a revised conception of knowledge. Undefeated fallible justification is the new knowledge that we seek.
It is the object of our inquiry. We cannot prove, as the skeptic has taught us, that our justification is undefeated by error. We have learned from her that some forms of error are invincible and beyond detection. If some skeptical hypothesis of invincible deception is true, then our justification is defeated and our perceptual beliefs are errors. In that case, our situation is epistemically desperate, and we must remain ignorant. If, however, we are right in thinking that our perceptual beliefs will lead us to truth in a trustworthy manner, as our internal standards of justification tell us, then our fallible justification is undefeated, and we have new knowledge, If there is an appropriate match between our beliefs about ourselves and our perceptual relation to the external world, then internal justification matches external justification, fallible justification goes undefeated, and we obtain a new kind of knowledge.
We must, in conclusion, thank the skeptic for undermining our dogmatism and our arrogance. She has shown us our fallibility. We may, nevertheless, seek reasoning and justification that lead us to truth in a reliable manner. The nobility of the goal of truth sustains the undertaking. We enoble ourselves in seeking truth, even when we realize that we may fail to obtain that noble objective. If the justification we find does not rest on error and enables us to reach the truth, we shall have attained our revised kind of knowledge. This new knowledge is based on a fallible quest for truth without any guarantee of sucess; we may attain it, though we cannot prove that we will. To the skeptic who asks for proof that we shall succeed, we must put our hands over our mouths in silence. We have no proof. We may, however, invite her to join our quest for truth and the new kind of knowledge we seek. Once we admit to the skeptic that she is right and we have no guarantee of success, she, being a woman of insight and character, who has, moreover, freed us of our dogmatism and arrogance, may join as a sympathetic friend in our noble undertaking. We may say to her, "Let us reason further with one another to find some fallible justificafion to lead us to the truth in what interests us, concerning freedom, mind, God and morals," and she, our brilliant adversary, will become a friend to our philosophical undertaking. The modesty resulting from a recognition of our own fallibility becomes us, opens the road to inquiry and removes the roadblocks to understanding. Revisionism combines the insights of skepticism and epistemism in harmony.
Philosophical Problems and Arguments: An Introduction
by James W. Cornman, Keith Lehrer, George Sotiros Pappas
https://www.amazon.com/exec/obidos/tg/detail/-/0872201244/
======
No comments:
Post a Comment