Lecture Notes, Lehrer's Theory of Knowledge, Chapter 9, Skepticism, Virtue, and Context

2002 Lecture Notes, Lehrer's Theory of Knowledge, second edition

Chapter 9, Skepticism, Virtue, and Context

By G. J. Mattey, Senior Lecturer, UC Davis Philosophy Department

In his concluding chapter, Lehrer deals generally with skepticism, claiming ultimately that we can know that skepticism is false. This is a reversal of the position he took in his 1971 paper, "Why Not Scepticism?" in which he argued that there is no way that skepticism can be overcome. We shall see later how it is that Lehrer eventually concluded that he could defeat skepticism. It should be noted that many epistemologists, especially externalists, do not take skepticism seriously any more. That situation is changing due to the increasing prominence of contextualism, which will be discussed below.

Since the seventeenth century, a main theme in theory of knowledge has been the attempt overcome skeptical arguments. René Descartes had set the tone with his Meditations, in which he allowed himself temporarily to accept skepticism, only to vanquish it in the end. David Hume upped the ante by arguing that certain kinds of skepticism cannot be overcome, whether by Descartes's methods or any others. Since that time, much of epistemology has been devoted to the task of combatting either Cartesian provisional skepticism or Humean skepticism. Some attention has also been devoted to more ancient forms of skepticism, including the problem of the criterion, to be described below.

Skepticism and Standards of Knowledge

There are two kinds of skeptical arguments. One type assumes a standard for knowledge and shows that this standard is not or cannot be met. The second criticizes a standard for knowledge.

If a standard of knowledge is to be assumed, which standard should it be? If the skeptic invokes his own standard for knowledge without defense, there is no reason for someone claiming knowledge to concede that standard. So, for example, if a skeptic insists on absolute certainty, it is open to the claimant to charge that the standard is too high. On the other hand, if the skeptic defends a standard and claims that it cannot be met, he transgresses against his own skepticism. A possible case of this is sort can be found in a dispute between the ancient Pyrrhonian and Academic skeptics. Writing on behalf of the Pyrrhonians, Sextus Empiricus charged that the Academics say that "all things are ungraspable. For they commit themselves to this, whereas the sceptics allow that it is possible for something to be grasped" Outlines of Pyrrhonism, Book I, Chapter 23, "In what Respect Scepticism Differs from the Academic Philosophy"). (The Pyrrhonians would have criticized what Lehrer calls an "agnoiology," or theory of ignorance, on the same grounds.)

The typical skeptical strategy is to assume provisionally the standards that are embodied in the theory of knowledge under attack. (The argument would then be ad hominem, or "against the person.") This is how the Academic skeptics attempted to undermine the theory of knowledge advanced by the Stoics, who described knowledge as unshakable conviction. The Stoics were led to modify their theory in response to the skeptics' arguments; more generally, one effect of skeptical argumentation has been to produce more solid accounts of knowledge. In fact, Descartes applied skeptical arguments to his own "preconceived opinions" in order to discover any beliefs of which he could be certain.

Skeptical criticism of a standard of knowledge may be directed against the standard itself. This is frequently done when knowledge claims are criticized by someone with competing knowledge claims. Galileo and Descartes both criticized the Scholastic physics prevailing in the early seventeenth century, and part of the criticism was that the Scholastic standards of justification were misguided. The most notorious such standard was that invoking the name of Aristotle in support of a claim was sufficient to justify it. Galileo and Descartes had other, better, standards at hand to replace those of their predecessors. Skeptics, who eschew all knowledge claims, tend to be suspicious that all standards are defective in one way or another.

The Problem of the Criterion

A general argument against the invocation of any standard for knowledge has come to be known as "the problem of the criterion." As we have just seen, there have been disputes about standards of knowledge. Some are about particular kinds of arguments that provide evidence for knowledge claims. As we will see shortly, others are about the degree of evidential support or reliability required for knowledge. The Pyrrhonian skeptics used an argument designed to instill doubt that any such standard can be established.

Suppose there is a dispute about a standard of knowledge. If the dispute is to be settled rationally, there must be some means for settling it. It would do no good of each side simply to assert its position without argument. So how would a standard of knowledge (or "criterion of truth," in the language of the Stoics) be defended? It could only be defended by reference to some standard or other. If the standard under dispute is invoked, then the question has been begged. If another standard is appealed to, the question arises again, to be answered either by circular reasoning or by appeal to yet another standard. So either the process of invoking standards does not terminate, or it ends in circular reasoning, and in neither case would the dispute be settled rationally.

Lehrer takes on the problem of criterion in the guise of the question whether he can justify his acceptance of his own theory of justification. He rejects the appeal to a higher-order theory of justification as well as dogmatic acceptance of the theory (p. 228). This leaves only circularity, or a "loop" of justification. Lehrer defends the loop of justification as being virtuous.

Meta-Justification and Explanation

The original question Lehrer poses is this: "When we construct a complete theory of justification, a special issue arises when we ask whether the theory itself is justified" (p. 228). The range of answers he proposes shows that he is not really concerned with justifying a theory of justification (providing an "meta-justification"), but with whether we can use the theory to explain why we are justified in accepting the theory of justification. Couched in these terms, the task is much easier, since when we explain why something is the case, we suppose that it actually is the case. I would not try to explain why there are thirty people in my office right now, because I am the only one here now. But I might try to explain why I am alone in the office, as I presently am.

Lehrer thinks it would be a defect in a theory of justification if it did not explain why one is justified in accepting the theory itself. This is because we cannot appeal to a different theory to explain it, on pain of generating a potentially infinite regress. So the choice is between explaining why the theory is correct and not explaining it. If we seek to maximize explanation, we must just put up with the fact that the only resources we have available are drawn from the very theory whose justification we are trying to explain. This is what makes the "explanatory loop" here a virtuous one. "The loop in the theory of justification is a virtuous one for the purpose of explaining as much as we can and leaving as little unexplained as we must" (p. 229).

This detour into explanation, does not, as mentioned above, give us an answer to the question whether the theory of justification actually is justified. The problem of the criterion applies here, and it appears that Lehrer should defend a loop in meta-justification as well as the loop in explaining justification. If he were to do this on the same grounds, he should have to say that we seek to justify as much as possible and leave as little as possible unjustified. But while this principle promotes the epistemic goal of accepting as many truths as possible, it conflicts with the goal of not accepting as many falsehoods as possible. The reason only one of these two goals is mirrored in explanation is that what is to be explained is already given and does not have to be established by the explanation. So the issue remains whether it is legitimate to use a theory of justification to justify acceptance of its own truth.

Lowering the Standards for Knowledge

Skeptical arguments can be criticized for presupposing too high a standard. Although Descartes took certainty to be necessary for knowledge, it is nearly universally held nowadays that a certainty-theory of knowledge is doomed to failure. As Lehrer argued in Chapter 3, there are very few beliefs we have that count as "incorrigible." His response is typical of contemporary epistemology: he embraces "fallibilism," the view that one can have knowledge when one's evidence falls short of certainty.

This is a radical break from the tradition, which considers fallible evidential connections (which would include Lehrer's "personal justification") to establish, at best, probability. Philosophers such as Plato, Aristotle, and Descartes have held that knowledge is the pinnacle of human intellectual attainment. The Stoics went so far as to question whether human beings have enough cognitive ability to attain any knowledge at all. Twentieth-century epistemologists, following the lead of G. E. Moore in the early part of the century, have been willing to lower the standard of evidence in order to be able to fend off skeptical objections to the certainty of virtually all our beliefs. In so doing, they take as their model the knowledge they would attribute to an ordinary person, so-called "common-sense knowledge.".

Of course, the dispute between fallibilists and infallibilists is just the sort of dispute over standards that triggers the problem of the criterion. Infallibilists hold that their standard allows us to avoid skepticism because there is no possibility for error when one has an infallible belief. Fallibilists such as Lehrer side with the skeptic here and claim that the standard is too strict for knowledge. But they think their fallible standard is strict enough for knowledge. This may be because many common-sense beliefs, which are generally held to be knowledge, meet the standard. But now a new standard, that what is generally accepted as knowledge really is knowledge, has been invoked. How can it be defended?

Personal Justification and Real-World Standards

Personal justification resembles a standard used in United States courts of law in criminal cases, that of being rendered "beyond reasonable doubt" by the evidence. It is hard to make this notion precise, given the vagueness of the term "reasonable." On Lehrer's scheme, the claimant must be able to answer all objections. If we look at the objections as potential doubts, we could say that the doubt is "reasonable" in that its truth makes the claim less reasonable than does its falsehood. And being able to answer an objection would be to put it beyond doubt.

So it looks like Lehrer's standard of evidence is at least as high as that required for the conviction of a criminal in the United States. It may be higher, depending on one's interpretation of the criminal conviction standard. This might be desirable because Chisholm, in his own Theory of Knowledge suggested that being beyond reasonable doubt is not high enough.

It does appear that being beyond reasonable doubt is at least a minimal standard for justification and knowledge. The standard for conviction in civil cases is "preponderance of evidence," which means that the evidence supports the charge more than its denial. Lowering the standard of knowledge to this level would be thought scandalous today, but Descartes would have thought the same thing about fallibilism.

External Standards for Knowledge

A way of avoiding weakening the standards of justification is to abandon justification as a condition of knowledge, as externalists do. One can avoid weakening any standards at all by adopting a strong condition like Dretske's nomological condition, which originally was a way to generate certainty by tying belief to laws of nature. Another such strong condition is Dretske's and Nozick's counterfactual condition. If these strong standards are retained, then skepticism seems all the more urgent: do we really measure up to such strict conditions for knowledge?

Most externalists follow the path taken by Lehrer. They are willing to lower the standards to something like reliability. According to the reliabilist view, one has knowledge when one's true belief is produced by a reliable process or method. The method need not be 100% reliable. Most importantly, the failure of a method to produce true belief in unusual situations does not count against its overall reliability. Such situations are not "relevant alternatives."

One such situation is that in which a Cartesian demon produces a thoroughly deceptive overall experience to serve as the basis of all our beliefs. By hypothesis, no-one could distinguish the demon-induced experience from experience produced by the physical world. But, according to the reliabilist, this breakdown is not relevant to the reliability which is sufficient for knowledge. Thus, the externalist can avoid dealing with the possibility of radical deception merely by discounting its relevance. So why did Descartes think it was a relevant alternative to his present experience? The obvious answer is that unlike the externalist, he was looking for certainty.

But inasmuch as reason already persuades me that I ought no less carefully to withhold my assent from matters which are not entirely certain and indubitable than from those which appear to me manifestly to be false, if I am able to find in each one some reason to doubt, this will suffice to justify my rejecting the whole. (Meditation I

The Cost of Externalism

But there is a deeper disagreement between them. Descartes wanted to be able to determine, from his own point of view, whether his faculties (reason, sense, memory) were reliable. Unless he was able to do this, he believed, he could not trust those faculties. This is typical of internalist theories, including Lehrer's. Recall that for Lehrer, a necessary condition for personal justification is that one trust one's own ability to accept what is true and not to accept what is false. But an externalist does not impose this requirement. Externalism only requires that the faculties in question be reliable or meet some other external standard.

One criticism of this view, presented by Richard Fumerton, is that externalism has thereby made itself irrelevant to philosophical expertise. Which processes or methods are the reliable ones can be determined only by investigators steeped in psychology, physiology, physics, genetics, etc. One cannot determine from one's own point of view whether one's methods are reliable (Metaepistemology and Skepticism, Chapter 6). It must be noted, however, that reliabilists generally think of this as an advance rather than a loss.

The Argument for Skepticism

According to Lehrer, skeptical arguments have some force, in that they expose the "pretense" of those claiming certainty in what they believe (p. 209). The type of skeptical argument Lehrer has in mind here is that which shows that a standard of knowledge, in this case, infallibility or certainty, is not met.

Skeptical challenges to given standards are of two types. The "real-world" challenge confronts us with evidence of past failures of those who meet the standard to have knowledge. So, for example, one could hold that a standard for empirical knowledge is that a belief be based on sense-perception. Then a skeptic can claim that a vast number of false beliefs are based on sense-perception. For example, Descartes noted that the sun appears to our senses to be quite small relative to its real size. More sophisticated standards would be challenged with more elaborate cases of error.

The "possible-world" challenge invokes only possible failure. Here the most famous example is the "invincible deception" that might be inflicted on us by a powerful deceiver, be it a malign demon or an evil scientist. The depth of the deception is such that the victim has no means whatsoever to discover the truth. Since there are no known cases of invincible deception, the problem is, as Descartes put it, exaggerated and metaphysical (Meditation III). Lehrer's response to this kind of case is to concede it but claim that it only shows that infallibilism sets too high a standard.

Lehrer thinks that the chief skeptical argument lies not in fanciful examples but in a commonplace observation.

It matters little what the source of error may be. What is critical is most obvious. People often accept what is false and, when what they accept is true, there is some chance that they might have erred. This is the fundamental skeptical premise. (p. 206)
The argument proceeds by claiming that it is a standard of knowledge that error be precluded. So the standard of knowledge cannot be met, and we are ignorant. Since Lehrer accepts the premise, he needs to find a way to avoid its conclusion.

Before turning to the task of answering the skeptic, Lehrer gives an example of how the premise might be supported. He describes how human conception is subject to change. Humans have often rejected previously entrenched concepts because they play no further role in our view of the world. This is especially true of theoretical concepts used to explain the natural world. Philosophical concepts are especially vulnerable. The medieval Scholastic concept of "substantial form" is seldom, if ever, used any more to explain why things are as they are. Lehrer notes that the concept of belief is under attack by some materialists. On the other hand, the medieval concept of a "haecceity" or individual essence has been "recycled" by recent philosophers.

One might think that Lehrer goes too far, however, in his generalization of this phenomenon: "No concept or belief is sacrosanct in the quest for truth, and there is always some chance that any one may be cast off as misleading and erroneous" (p. 207). The concept of water has been with us since before the birth of Western philosophy with Thales, and it shows no sign of being on the way out. Lehrer's sweeping pronouncement may be the result of the influence of Quine, who made a similar claim in his influential 1951 article "Two Dogmas of Empiricism" (Philosophical Review, widely reprinted). "No statement is immune to revision," even statements of logic. Quine's argument for this claim is that some "recalcitrant" experience simply might not fit in with our current body of beliefs unless we make massive re-adjustments in our world-view to accommodate it. Given his claim that statements have meaning only by reference to an entire body of beliefs, nothing is sacrosanct--anything can go.

Lehrer's Response to the Skeptical Argument

Recall that the skeptical argument is that there is always a chance of error, that a chance of error is not permitted by the standard of knowledge, so there is no knowledge. Lehrer admits the first premise, so he must show show that his standard of knowledge permits the chance of error. He can do this within the framework of personal justification. That there is a chance of error is an objection to a claim to be justified. The objection cannot be answered, but it can be neutralized. In terms of the justification game, the claimant can state, "But I am not in error in the present case," and show (on the basis of the evaluation system) that it is just as reasonable to accept that there is some chance of error but no present error as it is to accept that there is some chance of error.

It might be worth exploring a bit further Lehrer's confident response to the skeptic. What we need to understand is why the acknowledgement of the chance of error can be neutralized. This depends a great deal on his reasons for accepting that there is always a chance of error, which is "our fallibility as we conceptualize what we experience and change how we do it, discarding concepts that formed our discard beliefs, replacing or recycling old concepts and beliefs with new or recycled ones" (p. 208). The claim seems to be based on the history of human conceptualizing, but what does that history show? That "the beliefs that have been most cherished and in which people have placed their greatest confidence . . . have been demoted from literal truths to figures of speech" (p. 207).

If the lesson of history is that the beliefs in which we have placed the greatest confidence have turned out to be false, why should we place confidence in our present beliefs? Why should we confidently trump the skeptic by saying, "I am not in error in the present case?" More generally, why should we, in this light, accept that we are trustworthy? Perhaps exactly the opposite conclusion is called for. There is danger in conceding too much to the skeptic, as when a strategic retreat inadvertently brings the enemy to the gates of the capital.

The Argument from Ignorance

An advantage of Lehrer's response to skepticism is that it works particularly well on a specific skeptical argument known in the literature as the "argument from ignorance." The argument proceeds from a claim about my ignorance of whether an extreme skeptical hypothesis holds to a conclusion that I am ignorant about matters of common sense, such that I have two hands. The skeptical hypothesis is that I am deceived in such a way that I could not possibly detect the deception. This may be due to the machinations of a Cartesian evil demon or a scientist who has removed my brain and kept it alive in a vat. In both cases, I have no hands and no perception at all of a physical world. And I have absolutely no way of telling that my perception of the physical world is entirely induced by the deceiver.

  1. I do not know that I am not indetectably deceived
  2. If I do not know that I am not indetectably deceived, then I do not know that I have two hands.
  3. Therefore, I do not know that I have two hands.
The argument is formally valid, so long as the meanings of the premises and conclusions are the same. That is, if there is no equivocation in the argument, the conclusion must be true given that the premises are.

The first premise applies equally to internalist and externalist theories of knowledge. On the side of internalism, the claim is that one's epistemic perspective is necessarily limited: the possible deception cannot be detected, so that one is not in a position to rule it out when evaluating whether one is connected to the world in the right way. On the side of externalism, most accounts of the truth connection needed for knowledge are such that the connection would not be appropriate. For example, I cannot reliably detect that I am not being deceived. Or, if I were indetectably deceived, I would not believe that I am.

The second premise may not at first seem plausible. Why does my ignorance about being indetectably deceived imply my ignorance about something so commonplace as my having two hands? Descartes had demanded this because he had deliberately decided to withhold belief in statements about his body until he had determined whether or not he was indetectably deceived (Meditation I). But this is an extreme policy that he imposed on himself, and most philosophers think that it sets a standard that cannot be met, thus playing into the hands of the skeptic.


Consider first that under the assumptions of the argument from ignorance, if I am indetectably deceived, then I do not have two hands.

If I have two hands, I am not indetectably deceived.
It may be that knowledge is "closed" under conditionals of this kind. That is, they endorse the following "strong closure" principle.
If I know that p, and if p then q, then I know that q
This principle would yield the result that if I know I have two hands, then I know that I am not indetectably deceived, which is equivalent to step 2 in the argument from ignorance.

Strong closure is too strong a principle for knowledge. It may be that the conditional is one that I have never conceived, so I should not be said to know its conclusion. Or it may be that I have conceived the conditional and am not justified in accepting it.

These objections can be addressed by weakening the closure principle.

If I know that p, and I know that if p then q, then I know that q
To know that if p then q, I would have to have conceived of the conditional and be justified in accepting it. It might still be objected that I might never have drawn the conclusion that q, even though I know the premises. In that case, we could add a further conjunct to the antecedent of the conditional and require that I accept that q. Since I know that p and that if p then q, both those statements are true, in which case q is true as well. And q should be justified, indeed "irrefutably" justified (Chapter 7) if they are.

Rejecting Closure

The defense of closure just given is from an internalist perspective. Things look much different from the side of externalism. It may be the case that I can reliably tell that I have two hands but not that I am not a victim of undetectable deception. Or that my belief that I have two hands tracks the truth, but my belief that I am not indetectably deceived does not. Basically, the reason is that with the ordinary statement, extreme alternatives like that of undetectable deception are not considered relevant. An ordinary person can reliably believe that he has two hands without being able to detect reliably an undetectable deception, but it does not matter to whether he knows.

Lehrer gives an indirect defense of closure by appealing to an unpublished article of Saul Kripke. The upshot is that if we allow that closure fails in some cases, we would then have to show how to delineate permissible from impermissible cases of failure. Kripke shows how a tracking version of externalism can lead to an impermissible failure of closure. He takes it as intuitively clear that one cannot know that he sees a blue barn without knowing that he sees a barn. To know that one sees a barn, it must be the case that if he did not see a barn, he would not believe that he sees one. This condition would fail to be met in a case where the environment is full of barn facades which closely resemble actual barns.

But one might be said to know that he sees a blue barn despite the presence of many realistic barn facades. The reason is that it might be a law of nature that no blue barn facades can exist in the area. Maybe a rare soil condition prevents the construction of a blue barn: pick whatever outlandish explanation you like. In that case, if the person does not see a blue barn, he would not believe that he sees a blue barn. This is an impermissible failure of closure. Lehrer notes that an externalist could try to circumvent this kind of failure but adds that it is not necessary if one adopts his solution to the argument from ignorance.


Some philosophers want to uphold both premises of the argument from ignorance but reject its conclusion. I mentioned in introducing the argument that it is formally valid so long as the statements in the premises and conclusion have the same meaning. Contextualists claim that the argument is paradoxical because the conditions under which one would be willing to assert the premises are different from those under which one would be willing to assert the conclusion.

Both the first premise and the conclusion are attributions of knowledge to myself. I attribute knowledge to myself in ordinary contexts, so in those contexts I would deny the conclusion. But in the same contexts I would not assert the first premise. This is because ordinary contexts are ones in which I am not considering the exotic possibility of undetectable deception.

If I do consider that possibility, the context shifts. In that case, a contextualist is willing to assert the first premise. And given the contextualist's embrace of closure, he is willing to assert the conclusion as well. Let us call this context-relativity of knowledge attribution "weak contextualism." This may explain why it is that at times I think the argument applies to me and at times I do not think it does. (Notice that I cannot say, "and at times I think it does not," since in that case I would be in a context when I am considering the possibility of an undetectable deception!) But we would like to know whether the premises are true or not. Does their truth-value shift from context to context?

Strong contextualism is the view that whether one knows or not depends on the context of attribution. David Lewis is one philosopher who recognized this paradoxical result. When we do epistemology, bringing up skeptical hypotheses, we lose knowledge that we once had. On Lewis's view, epistemology should proceed by discovering rules according to which possibilities that destroy knowledge can properly be ignored.

What is epistemology all about? The epistemology we've just been doing, at any rate, soon became an investigation of the ignoring of possibilities. But to investigate the ignoring of them was ipso facto not to ignore them. Unless this investigation was an altogether atypical sample of epistemology, it will be inevitable that epistemology must destroy knowledge. That is how knowledge is elusive. Examine it, and straightaway it vanishes. ("Elusive Knowledge," Australasian Journal of Philosophy 74, 1996, widely reprinted.)
It is not hard to see why Lehrer thinks that his own account of knowledge, which does not appeal to contextual shifts of knowledge attribution, is to be preferred.

Lehrer's Response to the Argument from Ignorance

On Lehrer's analysis of knowledge, in order for S to know that he is not undetectably deceived, the following conditions must hold:

Lehrer correctly states that if there is no demon deception, a person who trusts his acceptance system not to be the product of deception knows that he is not undetectably deceived. In this way, such a person knows that this kind of skepticism is false, given that it is false. On the other hand, if there is demon deception, the person does not know that skepticism is false. (A further, related, consequence of Lehrer's theory is that in general, if you know something, you know that you know it.)

Thus Lehrer's answer to skepticism is conditional. He acknowledges that his answer is close to being skepticism itself, but he settles for it as the best we can do. Of course, such an answer will not be satisfying to many. As Barry Stroud points out, this kind of conditional response to skepticism will leave you wondering whether you really do know that skepticism is false, even if (technically speaking), you know that you know that skepticism is false. For knowledge, in Lehrer's system, does not require certainty. Like so many anti-skeptics who reach the end of the line without arriving at the promised land, Lehrer is happy to have gotten as far as he has. Stroud, on the other hand, thinks that the power of skepticism shows that the whole project of epistemology needs to be re-thought from the ground up. (See The Significance of Philosophical Skepticism.)

Intellectual Virtue

After presenting his fallibilist solution to skepticism, Lehrer aims to elucidate the point of his anti-skeptical argument by appealing to the notion of intellectual virtue. What he wants to illustrate is that undefeated justification yields knowledge despite the possibility of error because an essential element of justification is the acceptance of our own "intellectual trustworthiness and virtue" (p. 212). If our justification is undefeated, then we are intellectually trustworthy and virtuous.

The critical point is that falsity of the skeptical hypotheses of deception is sufficient to convert justification into knowledge. Most critically, it is that justification for accepting [the claimant's] intellectual trustworthiness and virtue must be successfully connected with truth in a reliable manner. (p. 212)
If my justification is undefeated, then I really am trustworthy, and my reasons for accepting that I am trustworthy are sound reasons. This supports the intuition that one knows in this case. But it raises the further question as to whether the notion of virtue adds anything beyond what is already in the notion of trustworthiness.

As Lehrer presents it, intellectual virtue is virtue that aims at the intellectual goal of obtaining truth and avoiding error in one's intellectual endeavors (p. 210).

How might one do this? One must seek to reason validly in deduction and cogently in any argument as an exercise of intellectual virtue. One must consider the objections to the candidates for acceptance and only consider oneself justified in accepting things when one can meet the objections. One must be ready to change what one accepts when the objections cannot be met in order to avoid accepting what is false. In short, therefore, to proceed in an intellectually virtuous manner in what one accepts is to proceed in an intellectually trustworthy manner in what one accepts. (p. 210)
But this does not mean the the notion of virtue and trustworthiness are conceptually equivalent. Instead, they are "materially equivalent," applying to the same actual cases, though perhaps not to the same cases in some non-actual but conceivable circumstances. To show this, Lehrer gives an abstract case where, "If someone who is trustworthy instructs me to accept something that it is not at all intellectually virtuous for me to accept, I may be trustworthy but not virtuous in accepting what that person instructs me to accept" (p. 210).

To evaluate this kind of case, we need to get clearer about what is different about the notions of trustworthiness and virtue. To be trustworthy is to be "worthy of trust." This is a matter of "intellectual integrity." In particular, one must accept what one does in an appropriate way and change what one accepts when change is called for. Moreover, if the correction of what is accepted requires a change in methods of acceptance, the trustworthy person will change them (p. 140). There are only a few minor differences between this and the description of virtue. The cogent reasoning required for intellectual virtue is surely what will determine what one accepts in a trustworthy way. The changes in reasoning and methods would presumably be in response to objections. In view of the purely formal nature of these differences, it seems that Lehrer's first description of the relation of trustworthiness to virtue, that to proceed in one way is to proceed in the other, is correct.

It must be noted that Lehrer's conception of virtue, like that of trustworthiness, is a necessary condition for personal justification. The resulting justification may be defeated if a skeptical hypothesis is true, but any victims of deception remain "as faultless in the way they proceed in pursuit of the goal of truth as we are. They are as intellectually trustworthy or intellectually virtuous in their world as they would be in our world or as we are in our world" (p. 211). Standard virtue theorists, such as Sosa and Greco, did not understand intellectual virtue in this way. As Greco has pointed out, virtue epistemology is externalist. An intellectual virtue is a power or competence to attain truth and avoid error with respect to a given field of inquiry and a given context. This is more than mere reliability. Greco puts it succinctly:

Whereas generic reliabilism maintains that justified belief is belief which results from a reliable cognitive process, virtue epistemology makes a restriction on the kind of process which is allowed. Namely, the cognitive processes which are important for justification and knowledge are those which have their basis in an intellectual virtue. (A Companion to Epistemology, ed. Jonathan Dancy and Ernest Sosa, entry, "Virtue Epistemology.")
This is clearly not how Lehrer understands intellectual virtue. For example, he regards the virtue to lie in seeking to reason cogently, while the other virtue theorists would say that the virtue is a faculty of reasoning that really produces cogent arguments. Importantly, they extend the virtues to non-evaluative faculties such as perception and memory.

Externalist virtue theorists invoke intellectual virtue in the analysis of knowledge. Sosa in particular was looking for a condition of justification which would (in Lehrer's language) convert belief into knowledge. For reasons I will not discuss here, he thought that reliability does not quite do the job (See "Reliabilism and Intellectual Virtue," in Knowledge in Perspective: Selected Essays in Epistemology). Greco also presents virtue as part of an analysis of justification.


It is my view that both Lehrer and the original externalist virtue theorists have ignored what traditionally has been regarded as intellectual virtue, namely, wisdom. (Its corresponding intellectual vice is folly.) Wisdom is not a virtue in the sense that it is a faculty like memory. It is a characteristic of persons that they be intellectually virtuous (wise) or intellectually vicious (foolish). We can understand these notions, I think, not simply as being conditions of knowledge.

Aristotle distinguished two kinds of intellectual virtue: philosophical and practical wisdom. Both kinds of wisdom are the result of training and not traits of character. (Nicomachean Ethics, Book I, Chapter 13). Philosophical wisdom specifically is "intuitive reason combined with scientific knowledge," which is knowledge of the best things, as opposed to what is to one's own advantage (Nichomachean Ethics, Book IV, Chapter 7). The ancient Stoics held up the "wise man" or "sage" as the ideal human. One of his characteristics is that he has no opinions, only knowledge. For both Aristotle and the Stoics, knowledge is infallible.

For a fallibilist, wisdom can be understood in a way that seems more characteristic of what Aristotle thought of as moral virtues. That is, wisdom can be seen as a mean between the two extremes of believing without sufficient evidence and not believing with sufficient evidence. As Hume put it, "A wise man, therefore, proportions his belief to the evidence" (An Enquiry Concerning Human Understanding, Section X). Hume thought that belief comes in degrees, so that the strength of belief would be proportional to the strength of evidence. But his basic idea can be adapted easily to the notion of acceptance: a wise person is one who accepts only on the basis of sufficient evidence.

This raises a problem that Hume did not have. Given that acceptance does not come in degrees, how much evidence does the wise person require before accepting something flat-out? Here, I believe, there is room for a good deal of variation. But the doctrine of the mean can be helpful. One epistemic goal is to accept as much as possible that is true, which by itself promotes promiscuity, while the other is to accept as little as possible that is fault, which by itself promotes abstinence. A wise person will, first, pursue strategies of acceptance accommodating both goals to the extent possible. This is the "internalist" side of intellectual virtue. Moreover, wisdom requires that a person use means which will best implement these strategies. Someone who reasons badly, perhaps through no fault of his own, is lacking in intellectual virtue. This is the "externalist" side of intellectual virtue.

Getting back to Lehrer, we can say that his view is one way of understanding the internalist side of intellectual virtue. To complete the picture, we need the connection between the internal condition for virtue with its external condition. This would be the connection between trustworthiness and reliability that Lehrer has already claimed to be a condition for knowledge. A person whose trustworthy pursuit of the truth is at the same time reliable would be intellectually virtuous in the full sense. (These remarks on wisdom were prompted by a remark by Michael DePaul at a 1999 conference on intellectual virtue held at Santa Barbara, California.)

Trustworthiness and the Truth Connection

Let us grant that we accept that our trustworthiness is reliably connected to the truth. Lehrer wants to be able to convert this acceptance into knowledge that the skeptic is wrong. The first step is to marshal the evidence for the claim that our trustworthiness is reliably truth-connected: "I have found it to be so connected with great reliability" (p. 228). But what does it mean to say that he has "found" his trustworthiness to be reliably connected to the truth? It could only mean that his evaluation system makes it reasonable to accept this. Yet it is the reliability of the trustworthiness of his evaluations which is at issue. So it looks like he is presupposing what he is trying to establish. It looks like he is begging the question against the skeptic.

Lehrer does not think he has a problem here. As usual, he makes an appeal to explanation. We can use the reliable trustworthiness of the evaluation system to explain the reasonableness of accepting that the evaluation system is one of reliable trustworthiness. "The truth of the hypothesis that my trustworthiness is successfully truth connected explains why my trustworthiness in accepting it leads me to the truth" (p. 228). This is the same maneuver Lehrer had used in defending the partial self-justification of the principle that I am trustworthy in what I accept. My trustworthiness explains why it is reasonable to accept that I am trustworthy. In the present case, Lehrer just "widens the loop" by including the truth connection along with trustworthiness. "The loop of trustworthiness and reasonableness widens to explain the reasonableness and truth of what I accept, including that my trustworthiness in what I accept is truth connected" (p. 228).

I believe that this version of the maneuver is no more successful than the other ones we have considered. The skeptic is not concerned with explaining why it is reasonable for me to accept that my trustworthiness is reliably connected to the truth. What the skeptic wants is a reason to think in the first place that it is reasonable. How can you justify an appeal to the evaluation system to show the reasonableness of accepting, in effect, that the evaluation system itself produces truth-connected acceptance? It is up to the defender of Lehrer's theory to show how the loop of justification is virtuous. Any skeptic worth his salt will not be seduced into shifting the question of circularity away from justification to the friendlier ground of explanation.

Previous Lecture | Lecture Notes Menu

[ Philosophy 102 Home Page | G. J. Mattey's Home Page ]