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

In the third through sixth chapters of Theory of Knowledge, Lehrer discusses various accounts of justification. Chapters 3 and 4 are devoted to a type of account of justification known as "foundationalism." We begin by taking a look at the nature of justification, which is given as the third condition for knowledge:

(iJ) If S knows that p, then S is justified in accepting that p.

In the course of stating (iJ) Lehrer does not say much in Chapter 1 about what justification is. We are told that in a case of justification, "S's acceptance is based on adequate evidence of hers" (p. 15). This implies that to be justified in accepting something, one must be in possession of evidence, and that this evidence is the basis for the acceptance. Moreover must give the acceptance strong support: it must do more than merely make it "reasonable" for S to accept that p.

An easily-overlooked feature of condition (iJ) is that what is to be justified as a condition of knowledge is something that S actually accepts. In Chapter 7, Lehrer states that "Undefeated justified acceptance obviously implies acceptance" (p. 171). And as undefeated justified acceptance is a kind of justified acceptance, justified acceptance implies acceptance as well.

As we saw in Chapter 2, the kind of acceptance relevant to knowledge is based on an evaluation of whether S's accepting that p furthers the goal of accepting all and only what is true. Evidence justifies an acceptance when it promotes the attainment of the goal of acceptance. The central issue is how evidence can promote correct acceptance. We got a clue in Chapter 1, where Lehrer claims that he is unjustified in accepting that his secretary is in her office, on the grounds that because he is not there, he cannot exclude the possibility that she had left the office. Our present task is to discover how evidence works to exclude possibilities for error in what we accept.

If we take the requirement for excluding the possibility of error in the strongest sense, we should require that justification exists only when the evidence makes error impossible, when it guarantees the truth of p. An account of justification that adopts this strong requirement can be called "infallibilist." Throughout the history of philosophy, ancient and modern, there have been numerous advocates of infallibilism. Lehrer, however, rejects infallibilism, as can be seen already in his comment in Chapter 1 that to be justified, one need not be "completely certain" (p. 14). Lehrer explicitly embraces "fallibilism" in his account of justification. "Though we are fallible in what we accept, when acceptance has the purpose of obtaining truth and avoiding error, it may achieve these objectives in a way that yields justification and knowledge" (p. 18).

Varieties of Foundationalism: Doxastic vs. Non-Doxastic

What all foundationalist theories do have in common is the view that all justification ends with evidence that justifies but is justified by nothing else. Such stopping points are the foundations of all justification, and therefore of all knowledge. On Lehrer's account of foundationalism, the stopping point is a "basic belief." Any version of foundationalism of this sort is "doxastic" foundationalism, that is, a version where the foundational evidence is a belief. (The Greek word 'doxa' signifies 'belief' in English.) Lehrer's foundationalist also holds that basic beliefs must be justified in order to justify other beliefs. Since they are not justified by anything else, the basic beliefs could only be self-justified. As we will see, a foundationalism described in the way will have a hard time getting off the ground.

There is room for some variation here, however. One account would be that there are basic beliefs which, although they justify other beliefs, are not themselves justified. This does not seem to be a very promising approach, however. Beliefs without justification of their own would seem to have no credentials for conferring justification on other beliefs.

This kind of objection is less urgent if the foundational evidence consists of non-doxastic psychological states, such as sense-perceptions or memories. As has been pointed out by many philosophers throughout the ages, there is no truth or falsehood in sense-perception, so we should not expect sense-perceptions or memories to be justified. John Pollock holds a view of this sort. (John L. Pollock and Jospeh Cruz, Contemporary Theories of Knowledge, second edition, Chapter 7. As this idea and others cited below are found in the first edition, whose sole author is Pollock, I shall attribute the quotations to Pollock.) Lehrer acknowledges this as "an alternative to the kind of foundationalism discussed in this chapter," but allows only that it "resembles foundationalism" (p. 68). He is silent about the merits of this approach, but it may well be be that it would be able to stand up to the objections Lehrer raises against foundations theories involving self-justified basic beliefs. We will have to put off these considerations until the next chapter.

The Doxastic Assumption

Pollock calls the view that the only thing that can serve to justify what one believes is something else one believes the doxastic assumption: "the justifiability of a belief is a function exclusively of what beliefs one holds of one's 'doxastic state'" (Contemporary Theories of Knowledge, second edition, p. 22). The doxastic assumption is motivated by the view that "all our information about the world is encapsulated in beliefs," so that "we cannot take account of anything except insofar as we have beliefs about it" (p. 22). Lehrer makes the doxastic assumption explicit in Chapter Six.

The evaluation of all claims to truth, whether those of our senses, of reasoning, of memory, or of the testimony of others, must be based on our acceptance system, which includes our conception of the world and our access to it. There is no exit in evaluation from the circle of what we accept. Acceptance is the fuel of the engine of justification. (p. 124)

Basic Beliefs and Acceptance

One thing worth noting here is that Pollock understands the doxastic assumption in terms of belief while Lehrer understands it in terms of acceptance. In discussing foundationalism, however, Lehrer adopts the language of belief.

It is traditional to formulate and discuss the foundation theory in terms of belief rather than acceptance, and we shall do so as well; but the kind of state required for knowledge is acceptance directed at attaining truth and avoiding error in what is accepted. (p. 45)

Varieties of Doxastic Foundationalism: Infallibilist vs. Fallibilist

Lehrer divides doxastic foundations theories into two kinds: those in which the basic beliefs are infallible (could not be mistaken) and those in which they are fallible (could be mistaken). The latter kind of theory will be discussed in detail in Chapter 4, but for now, Lehrer notes that a fallibilist foundationalism is not a pure version of foundationalism. How could a belief be self-justifying when it runs some risk of error? If a belief by itself does not exclude all possibilities that would make it false, then why are we justified in accepting it? Lehrer thinks this is due to our assessment of our own cognitive powers. For example, one may think that memory beliefs are justified, though they are subject to error. "This means, however, that the allegedly basic beliefs in question are justified by relation to other beliefs and are not genuinely basic. Such a theory is not a pure foundation theory" (p. 47).

Motivations for Foundationalism

Traditionally, there have been three different, though related, motivational reasons that have led philosophers to embrace some form of foundationalism. The reasons can be called:

• logical
• psychological
• epistemological

The Logical Motivation for Foundationalism

Aristotle provided the classic formulation of the logical motivation for foundationalism. He was a foundationalist in the sense that in the "demonstrations" that yield "scientific knowledge," there must be "primary premises" which are also "true," "immediate" and "better known than their conclusion" (Posterior Analytics, Book II, Chapter 2).

Some hold that, owing to the necessity of knowing the primary premises, there is no scientific knowledge. Others think there is, but that all truths are demonstrable. Neither doctrine is either true or a necessary deduction from the premises. The first school, assuming that there is no way of knowing other than by demonstration, maintain that an infinite regress is involved, on the ground that if behind the prior stands no primary, we could not know the posterior through the prior (wherein they are right, for one cannot traverse an infinite series): if on the other hand—they say—the series terminates and there are primary premises, yet these are unknowable because incapable of demonstration, which according to them is the only form of knowledge. And since thus one cannot know the primary premises, knowledge of the conclusions which follow from them is not pure scientific knowledge nor properly knowing at all, but rests on the mere supposition that the premises are true. The other party agree with them as regards knowing, holding that it is only possible by demonstration, but they seen no difficulty in holding that all truths are demonstrated, on the ground that demonstration may be circular and reciprocal.

Our own doctrine is that not all knowledge is demonstrative: on the contrary, knowledge of the immediate premises is independent of demonstration. (Posterior Analytics, Book I, Chapter 3)

Lehrer discusses this kind of argument in Chapter 5, where he gives his own solution to the problem raised by Aristotle.

If a person is justified in a belief on the basis of evidence, then appeal to that evidence would constitute a correct answer to the question, How do you know? Now suppose no beliefs are basic. Then every justified belief is so justified by appeal to evidence, but evidence must itself be justified belief and therefore must also be justified by appeal to evidence. This means that every justified belief must be justified by some other, thus leading either to an infinite regress or to a justificatory circle. If both those alternatives are unacceptable, then there must be some basic beliefs. (p. 98)

The Psychological Motivation for Foundationalism

Foundationalism is often adopted by empiricists, who according to Lehrer think that justification is based on basic beliefs whose content is "empirical statements," which are "statements of observation" (p. 48). According to empiricists, we must begin with statements of observation because observation is the only way in which we can learn factual matters about the world. (Empiricists are inclined to agree that some of what we accept can be justified independently of experience, or a priori, but only about logical or other non-factual matters. See David Hume, An Enquiry Concerning Human Understanding, Section IV, for a classic statement.) Pollock succinctly describes this motivation, and the way it is combined with the logical motivation.

The simple motivation for foundations theories is the psychological observation that we have various ways of sensing the world, and all knowledge comes to us via those senses. The foundationalist takes this to mean that our senses provide us with what are then identified as epistemologically basic beliefs. We arrive at other beliefs by reasoning, construed broadly). Reasoning, it seems, can only justify us in holding a belief if we are already justified in holding the beliefs from which we reason, so reasoning cannot provide an ultimate source of justification. Only perception can do that. (Contemporary Theories of Knowledge, p. 29).

Lehrer does not deny the psychological observation that our knowledge of the world begins with sensory stimulation. But he does explicitly reject the notion that the way in which we come to have a belief makes it justified. What he calls in Chapter 8 the "causal fallacy" is "to confuse the reason a person has for believing something with the cause of his believing it" (p. 195). What matters for justification is not where our information comes from, but how we evaluate it (which might include consideration of its origins). This reflects Lehrer's "doxastic assumption" that justification is always in the context of "our conception of the world and our access to it" (p. 124).

The Epistemological Motivation for Foundationalism

Most contemporary discussion of foundationalism has focused on a perceived need for an "unshakeable basis" for human knowledge. We want our evidence for the truth of p to be good evidence: if it is already tainted, then how can we trust it as a way to get at the truth of something else? So philosophers have searched for a basic stock of evidence that would not, in itself, lead us astray, though we may make bad use of it.

The empiricist Epicurus (3rd century BCE), for example, held that "clear facts are the foundation and cornerstone" of opinions (Sextus Empiricus, Against the Mathematicians, 7.216). Clear facts are a worthy foundation because they are sense presentations which are always true. In the passage from Sextus just mentioned, Epicurus's foundationalism is developed by showing how different kinds of opinion can be supported or refuted by clear facts. This is the kind of motivation which is the main target of Lehrer's attacks in the present chapter.

Incorrigibility

Like Epicurus, the contemporary infalliblist foundationalist holds that basic beliefs are guaranteed to be true. Such beliefs are often said to be "incorrigible." Lehrer tries to give a precise analysis of the concept of incorrigibility. His preliminary attempt is this:

The belief that p is incorrigible for S if and only if it is logically impossible that S believes that p and p is false.

And it is logically necessary that if it is true that p, then S believes that p.

Pollock adds a different second clause, which seems to work better (Contemporary Theories of Knowledge, p. 35). Reformulated to conform to Lehrer's proposal, it looks like this:

And it is logically necessary that if it is true that p, then S does not believe that not-p,

It is logically necessary that if S believes that not-p then p is false.

It should be noted here that the reference to logical necessity is not critical to Lehrer's argument. He shows on p, 59 that "nomological" necessity, the necessity had by laws of nature, do not change the analysis in any way relevant to his arguments against foundationalism.

It is questionable, though, whether incorrigibility analyzed in this way is a sufficient condition for basic belief. Traditionally, philosophers who have held that we have knowledge based on incorrigible beliefs have tied their incorrigibility to the way in which they are held. The most famous example of this is Descartes, who claimed that "whatever I perceive very clearly and very distinctly is true" (Meditation Three). Empiricists tie the incorrigibility of beliefs about one's psychological state to a kind of direct observation or introspection of the state by the mind. We shall see below that the manner in which a belief is held is important determining whether it is basic.

Are Beliefs About Psychological States Incorrigible?

We want to find out whether a belief can justify itself, in the sense of guaranteeing its own truth. There are (at least) two ways in which this might occur, and each will be discussed in turn.

Every belief has two components, an "act" of believing, and a content, what is believed. In some very limited cases, the very act of believing brings it about that the belief is true. The act of believing that I exist requires my existence in order to take place at all. So the act of believing that I exist guarantees the truth of the content, that I exist.

Similarly, the act of believing that I believe something guarantees the truth of the content, that I believe something. [But note that the act of believing that I believe, say, that Nome is in Alaska, does not guarantee the truth of the content, that I believe that Nome is in Alaska. It seems possible that I may be mistaken about whether I believe that very specific proposition.]

There are not many beliefs whose truth is guaranteed by the very act of believing. Most foundationalists of the sort we are considering look elsewhere to find enough basic beliefs to justify everything else we believe. The obvious place to look is for beliefs that have contents about which one cannot be mistaken because of some special characteristic of those contents.

Descartes, for example, appealed to the simplicity of contents involving low-level truths of mathematics, such as that 2+3=5. The contents are so simple that one just cannot make a mistake about them provided that one is in full possession of one's faculties and in an alert and attentive state of mind. In the unpublished Rules for the Direction of our Native Intelligence, Descartes described "intuition" as a source of knowledge.

By 'intuition' I do not mean the fluctuating testimony of the senses or the deceptive judgment of the imagination as it botches things together, but the conception of a clear and attentive mind, which is so easy and distinct that there can be no room for doubt about what we are understanding. (Rule Three)

Since I sometimes consider that others go astray in cases where they think they have the most perfect knowledge, may I not similarly go wrong every time I add two and three or count the sides of a square, or in some even simpler matter, if that is imaginable? (Meditations, One)

The cases discussed by Lehrer are more appropriate to empiricism. Some beliefs may be self-justified because of their very limited content. Information about how things appear to me is of this sort. How could I go wrong in accepting information that desribes the mere look and feel of things? It may be that the thing which looks and feels this way is not really the way it looks and feels, but the properties of appearances themselves are a subjective matter.

Lehrer develops several cases to refute the incorrigibility of beliefs with appearance-content. His basic contention is our beliefs about those contents can go wrong because the way we describe these contents depends on inference.

Whatever one can believe as a result of introspection, one can instead believe as a result of inference, and the inference can be based on a false premise. If a woman believes she is in pain as a result of feeling pain, then, of course, she will be correct in believing that she is in pain. If, however, a man believes that his is in pain because some scientific or religious authority figure tells him that he is in pain, then what he believes may well be false. In such a case, the person has accepted a premise, namely, that what the scientific or religious authority says is true which, together with the premise that the authority says the person in sin pain, leads the person to infer, and tehrefore, to believe that he or she is in pain and to believe this falsely when the authority is untruthful. (p. 58)

One might argue for this version of foundationalism on semantical grounds. The idea is that if one feels pain and knows what the word 'pain' means, one cannot be wrong in believing that he is in pain. The word 'pain' just means that kind of feeling that one is having at the time. Lehrer maintains that it is possible to have false beliefs about one's psychological state without deviating from the standard meanings of the terms involved. I might have an itch and believe it not to be an itch because I have been convinced by my doctor that it is a pain. He regards this as an odd kind of belief but holds that it is not semantically deviant.

Opacity

The second edition of Theory of Knowledge presents a new argument against infallible basic beliefs as foundations for justification which can be used against the more restricted version just described. Lehrer notes that infallibility may be "opaque," in the sense that the infallibility of one's belief may not be apparent to the believer, in which case there is no justification.

A belief may be infallible even though the person has no idea that this is so. When the infallibility of a belief is concealed from the subject of the belief, it is powerless to justify the belief for the person. The infallibility of the belief may not reveal itself to the subject of the belief; it may remain concealed and opaque rather than revealed and transparent. Opaque infallibility is too dark a feature to yield to the light of evidence. (p. 61).

This kind of objection will be deployed later, in Chapter 8, against externalism. The reader may wish to reflect now on its cogency. I think that the best rebuttal to this kind of objection is to claim that the "transparency" requirement reflects an overly intellectualized account of knowledge. People in general may not be sophisticated enough in their self-evaluation to be familiar with all the required evidential connections involved in a satisfacory answer to the question, "How do you know?" Moreover, Pollock has argued that justification operates at a much less self-conscious level then the intellectualized way described by Lehrer. It is more like what it takes to ride a bicycle than like looking up directions in a book (Contemporary Theories of Knowledge, Chapter Five).

How Basic Beliefs Infallibly Support Other Beliefs

If there are very few or no basic beliefs, then foundationalism seems to have little point. But Lehrer recognizes that he might be wrong about this. And there may be many fallible basic beliefs that could support other beliefs (a view that will be examined in Chapter 4). In the spirit of his examination of infallibilist foundations theories, he discusses in Chapter 5 the possibility that basic beliefs guarantee the truth of the other beliefs that they support. It must be emphasized that the epistemological motivation for infallibilism need not apply to non-foundational justification. It may be that all justification must be based on something secure, but that the secure foundation may support beliefs that are less secure.

The most important case is the one where basic beliefs are about psychological states such as sense-perception and memory, and the beliefs they are said to support are about the world "external" to those states, which allegedly are represented by them. Can the way something looks to me, for example, guarantee that it is that way?

Typically we say a belief or set of beliefs infallibly supports a belief that p when p is logically deducible from it. Given the truth of the supporting beliefs, the supported belief must be true. But this infallible support comes at a price. There is no more information in the conclusion of a logical deduction than there is in the premises. So unless the information contained in our beliefs about our psychological states already contains the requisite information about physical objects, deduction from those beliefs will not yield truths about physical object statements.

And the prospects look pretty bleak. In order to justify accepting things about the world outside us, we need to have information about the connection between our subjective states and the outside world. For example, I need the information that how things look to me correspond to the sizes, shapes, colors, etc. they really have. But this information is hardly basic. My acceptance of my own reliability in seeing colors, for example, depends on a wealth of experience in identifying colors correctly.

Perhaps it could be objected that while the belief that my sensory states reliably represent the outer world is not basic, it is justified by other basic beliefs. But how could this be, since we always start with the purely subjective? We cannot get outside our sense experience to compare how things look with the thing itself. So there is no justification for what we accept about how things outside our mental states really are, at least if the infallible foundationalist's view of justification is correct.

Phenomenalism

The eighteenth-century philosopher George Berkeley pushed this kind of argument to the limit. His solution was to deny that there is a real world of physical objects existing independently of our sense-perception. A physical object is real only insofar as it is perceived or at least perceivable. Thus he used an epistemological argument to motivate a metaphysical conclusion. The doctrine that physical things are no more than collections of sense perceptions is known now as phenomenalism (from the Greek ''phaenomenon ,' which means ''appearance'). For Berkeley, the appearance is the reality.

Twentieth century versions of phenomenalism are tied to language. They claim that sentences about physical objects are logically equivalent to sentences which are about sense-perception alone. The physical-object sentence can be translated into sense-perception-sentences, and vice versa. We will here examine phenomenalism in its original Berkeleyan version. But the "analytical" version is useful to illustrate an important point. The information contained in beliefs about sense-perception need not logically imply information about physical objects in order to guarantee the truth of the latter. If statements about physical objects are equivalent by virtue of their meaning to statments about sense-perception, then beliefs about sense-perception can guarantee the truth of statements about physical objects.

Phenomenalism seems to be faced with a grave difficulty from the very start. If things are just as they appear to be, then how can error be possible? Suppose I see my golf ball in the water, then reach out to the spot where it looks to be, missing the ball. I would say that the ball was not really where it appeared to be. Berkeley would say that in one sense the ball really was where it looked to be, that is, that its visual location is where my eyes placed it. But its tactile location is where my hand finds it (eventually). Visual and tactile location are completely different from each other. He claimed that error consists in making bad predictions about how tactile experiences are correlated with visual experiences.

Another objection to phenomenalism is that what is unperceived, what does not appear to anyone at a time, does not exist given that its being is to be perceived. Berkeley answered this objection by enlarging his criterion of reality to include how things would look or feel under given conditions. Thus the table in his study exists insofar as he would see it and feel it if he were to enter the room. This kind of contrary-to-fact conditional is required to uphold the existence of most of what we regard as real, given the very limited range of actual sense experience.

But now we are faced with a further problem which is less easy to deal with. How could we be justified in accepting the required contrary-to-fact conditionals? Although they do not purport to connect mental states with an external reality, but only connect mental states to mental states, they are still too general to be basic. If Berkeley accepts that the table in his study exists when he is not in the study, he must have justified information about what would happen if he were to enter the room. But what individual mental states could justify this information?

This problem with phenomenalism is a problem which holds in general when it is demanded that the truth of non-basic beliefs be guaranteed by the truth of the basic beliefs. What can be justified on the basis of our meager mental states are only beliefs about what we have experienced. But we think there is a far-wider world out there.

Yet another problem with foundationalism is the way in which it purports to distinguish appearance from reality. Even though physical objects are ultimately composed of sense experiences, we need to be able to distinguish those experiences which have to do with actual physical things from those which do not. My dreaming experience of there being a whale in my bathtub does not establish the existence of such a thing. Berkeley proposed that "objective" sense experiences can be distinguished from "purely subjective" experiences like dreams insofar as our sense experiences form a coherent whole, while our dreams, hallucinations and imaginings do not.

But how can we justify the claim that our experiences cohere with one another? It would seem that this claim is not basic, nor can it be justified by appeal to beliefs about what mental state we are in at a time. Descartes had faced this problem by appealing to God as having made us in such a way that we are not deceived about what we carefully judge to be real. That is, God would not allow us systematically to mistake dreams for reality. This is hardly a basic belief. Descartes recognized this, and to dispel this possibility he thought he had to prove that God exists and is no deceiver. Berkeley would have to provide some reason to think that my subjective states are not a coherent dream.

Conclusion

The upshot of Lehrer's discussion of infallible foundations theories is that they fail to provide an adequate account of the justification of what we think we know. One reaction to this result is to wave the white flag of surrender to the skeptic. If we insist on infallible foundations or infallible support from any foundations, we will fail to know. Lehrer prefers to retreat in order to fight another day. He embraces fallibilism. "The lesson is that we are fallible in what we believe and must proceed without any guarantee of success" (p. 68). The skeptic will, of course, disagree. Lehrer postpones their showdown until the final chapter of his book.

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