Lecture Notes, Lehrer's Theory of Knowledge, Chapter 2, Truth and Acceptance

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

Chapter 2, Truth and Acceptance

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

In this chapter, Lehrer develops accounts of the first two conditions of knowledge: that p be true and that S accept that p.

Minimal Accounts of Truth

The truth condition for knowledge comes directly out of the original conception of knowledge as the recognition of correct information. In order to know that p, S must "get p right." The traditional way of expressing this is to say that p has to be true in order to be known. This is captured in the common question, "Do you know that it is true?" (There seems to be no real difference between this and the shorter question, "Do you know it?")

To get a feel for what accounts of truth look like, we will begin with one of the original definitions of truth, given by Aristotle:

To say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, and of what is not that it is not, is true. (Metaphysics IV, 7 (1011b25-26))
Here Aristotle is comparing reality (what is or what is not) with what is said about reality (that it is or that it is not). Truth is a match between the two and falsehood a mismatch. One might say that Socrates is pale and that Socrates is a father. The first is true because Socrates is pale and the second is false because he is not a father.

Notice, however, that the descriptions of "what is" and the "saying that it is" use the same sentences: "Socrates is pale," and "Socrates is not a father." Usually we use the same language to describe the reality of things and what is said about the reality. This is reflected in what Lehrer calls the "absolute theory" of truth:

(AT) It is true that p if and only if p,

which looks like an abbreviated version of Aristotle's definition. Such a theory is "minimal," in that it tells us what truth is by describing what is true using one of the components of "It is true that p. The same holds for the "disquotational theory" of truth,

"X" is true if and only if X,

for a declarative sentence "X." (It is called "disquotational" because the condition for truth is formed by dropping the quotation marks around the sentence whose truth-condition is being given.) Alfred Tarski, whose writings on truth are probably the most influential in the twentieth century, regarded such minimal conditions as only necessary, and not sufficient, for a theory of truth. But as we will see, Lehrer thinks they are enough for a theory of knowledge.

The Correspondence Theory of Truth

The minimal theories of truth just mentioned have in common that they require a "match" or "correspondence" between what is said and reality. A more substantive theory of truth can be developed by looking more closely these two elements. Let us call that which is true or false a "bearer of truth or falsehood." Typically, these are linguistic constructions, sentences. They may be words on paper, sounds in the air, or thoughts in the head. We will here consider sentences generically.

The important thing about sentences for present purposes is that they are structured. The most basic structure is a subject followed by a predicate. So,

"Socrates is pale"

is composed of the expression "Socrates" as the subject and "pale" as the predicate, joined by "is". (the coupling device or "coplua"). On Aristotle's view, objects in the world, such as Socrates, are referred to by the subject, and they have or lack something which are expressed by predicates. In this case, paleness is a quality that Socrates might or might not have. Thus, the sentence "Socrates is pale" is true just in case Socrates has the quality paleness. The relationship of subject and predicate in the sentence is reflected by the relation between object and quality in the world. This relation of correpondence between a grammatical structure and the structure of the world can be filled out in many ways, depending on one's theory of language and theory of reality. We have used Aristotle's relatively simple account just as an example here.

A feature of Aristotle's theory that may raise suspicion is that the account of reality seems to be modeled precisely on the grammar of the sentences. Perhaps reality has a different structure altogether. If so, then we might not be able to understand it in terms of our own language. In that case, we could not describe the correspondence. On the other hand, if we do understand reality in terms of the structure of our language, we are back to the problem with the minimal theory. The account of reality given by the correspondence does not tell us anything more than what we already find in the sentence whose truth is supposed to be accounted for.

Pragmatic Theories of Truth

A related kind of objection is that a correspondence theory of truth is useless. It may try to lay down conditions under which a grammatical structure and a structure of reality coincide, but how is this coincidence to be determined? Suppose I accept that Socrates is pale. I want to determine whether this acceptance is true, so I look to the structure of reality. But what kind of evidence would I have? It could be none other than the evidence that led me to accept in the first place that Socrates is pale. The objection is that we cannot get outside the structure of language to describe the structure of the world.

If we want a theory of truth to be of any real value, on this view, we should look not to an invisible "structure of reality" but to the conditions under which we accept sentences as being true. What we want to know is how acceptable our acceptances are, so to speak. Typically this kind of theory is concerned to uncover the social practices which lie behind our judgments about what is accepted. So we call such a theory a "pragmatic" theory of truth.

A leading type of pragmatic theory is one which ties truth to the practices of some kind of authoritative individuals or groups, e.g. to scientists. What is true is, on this view, what scientists pronounce to be the case on the basis of their scientific practices. Since scientists over the millenia have constantly changed their pronouncements, it seems that on this view, the truth changes as well. The American philosopher C. S. Peirce proposed that truth is what the scientists would agree upon in an ideal state of the completion of their work - a state that may never exist (C. S. Peirce, Collected Papers, Volume V, p. 268, pp. 394-5).

Another sort of pragmatic theory regards truth as a kind of valuable state. Thus accepting certain things might be so valuable to our survival that we regard what we accept as true. But as various philosophers have noted, there is no reason that what is valuable to our survival need have anything to do with the way the world is, which would be a failing of the theory. Of course, the pragmatist could respond that it makes no sense to talk about the way the world is independently of our practices.

Fortunately, we need not decide here between correspondence and pragmatic theories of truth. Both are consistent with the minimal accounts that yield such instances as:

It is true that Socrates is male if and only if Socrates is male,


'Snow is white' is true if and only if snow is white.

Paradoxes of Truth

One would think at first blush that such minimal conditions of truth would be trouble-free, due to their trivial-seeming nature. But that would be a mistake. We can generate paradoxes when we try to apply the conditions to individual sentences, as in sentence named 'S'

S. Sentence S is false.

It is not hard to see that, S is true if and only if it is false. Any account of truth that applies to sentences such as S is untenable. So even the minimal truth-condition (AT) is too far-reaching. That is, it gives conditions for the truth of too many sentences, including S.

Tarski held that the problem is that the truth condition is stated in a language (English, but it could be any natural language) which also contains the sentences to which the truth condition applies. Sentence S is in English, and so is its truth condition. Any language which can give truth conditions for its own sentences is called "semantically closed," and Tarski concluded that truth conditions cannot be given for any semantically closed language, including English. So he confined his attention to artificial symbolic languages such as those studied in Philosophy 12 and 112.

Lehrer is not so pessimistic. He thinks that we can use truth conditions like (AT) so long as we are careful to avoid applying them to sentences, like S, which refer to themselves. (The name 'S' occurs in the sentence which 'S' names. So S says of itself that it is false.) "Most substitutions in . . . (AT) will lead to nonparadoxical equivalences, indeed, to equivalences which are necessarily true" (p. 30). So "silence" is what is dictated when paradoxical substitutions might be made.

Lehrer recognizes that not all self-referential sentences are paradoxical, so we don't have to be silent about very much. For example,

SR. SR refers to itself, and

UT. All true sentences are true,

do not cause any semantical problems. If we recognize this, we can apply the minimal account of truth to them without fear of paradox. The trick is to exercise due care.

I believe that Lehrer's prescription does not fully evade paradox, however. Consider the following sentence:

D. D is false or (AT) does not apply to D.

Since D refers to itself directly and seems to generate a paradox, (AT) should not apply to it, according to Lehrer's prescription. But then it follows that D is false or (AT) does not apply to D. (This by a rule of inference of the form: P; therefore P or Q.) But then D is true, in which case (AT) applies to it when it should not. Whether or not Lehrer can avoid the paradoxes with his policy of avoidance of certain self-referential sentences, he is still open to crticism for the minimalist nature of his account of truth. To be sure, there are philosophers who adopt minimalist theories on their own merits. We cannot settle the issue of their adequacy here. Lehrer thinks the bottom line is that we cannot define truth, but that this should not bother us. For we cannot define color-terms like 'red' either, and yet we can use them competently.

Truth and Knowledge

In sum, Lehrer does the best he can to avoid saying anything substantive about truth and to avoid solving the paradoxes of truth. His truth condition is minimal, and at that, not completely general. But he thinks that it is enough to serve the pruposes of a theory of knowledge. We have already noted that we could dispense with any reference to truth and reformulate condition (iT) as follows:

(iT') If S knows that p, then p.

We have seen that Lehrer ties truth to acceptance, as its goal: one accepts that p in the interests of accepting all and only what is true. He thinks that there is no problem with this goal as long as it is restricted to non-paradoxical sentences. All the paradoxes mean is that the goal of accepting all that is true cannot be met. Moreover, as Lehrer notes, one could reformulate the goal of acceptance: to accept that p if and only if p.

A final twist is Lehrer's attempt to give an account of correspondence based on the notion of true acceptance. He sets up a condition on true acceptance,

(G) What S accepts, that p, is true if and only if S accepts that p and p,

which is a kind of extension of (AT). If p is not paradoxical, we might want to say that S's acceptance that p is true just in case S's acceptance that p corresponds to the fact that p. True acceptance is what corresponds to the fact. This has the consequence that S's acceptance that p corresponds to the fact that p if and only if S accepts that p and p. "Thus, it appears that an account of what it is to accept that p, what it is for a mental state of acceptance to be an acceptance that p, to have that content rather than another, yields at least a minimal account of correspondence" (p. 31). I do not believe that we gain any understanding about the nature of correspondence given this equivalence, though in some sense it may be an "account." The engine that makes it work, (G), sweeps the basis of true acceptance under the rug. We should like to know about the correspondence between acceptance and what is accepted in order to understand why (G) holds. So (G) does not reveal anything about what acceptance is. Thus it seems to me that there is no "account" of correspondence here.


Our discussion here is in two parts. The first part concerns Lehrer's conception of "acceptance," emphasizing its difference from "belief." Once this is clarified, we can look at some counter-examples to Lehrer's acceptance condition for knowledge:

(iA) If S knows that p, then S accepts that p.

What is Acceptance?

The first thing to notice about acceptance is that it is goal-directed. In Chapter 1, Lehrer makes this explicit.

Acceptance is an attitude defined in terms of some purpose, that is the expression to accept that p is an abbreviation or shorthand for the more explicit expression to accept that p for O, where O is some specific objective or purpose. It involves an evaluation of whether the attitude fulfills the purpose. (p. 13)
There are many purposes which acceptance might fulfill. We often accept something for the sake of argument or test an hypothesis, in a way that is provisional and does not involve any lasting or real commitment to its being true. We are also told on p. 14 that we may accept something to make another person happy. Most importantly for our purposes, a person may accept something with the goal of accepting all and only what is true. Let us call this "epistemic acceptance." This is what Lehrer has in mind for his condition on knolwedge, which, in its unabbreviated form, should read,

(iA') If S knows that p, then S accepts that p with the objective of accepting all and only what is true.

Note that in all these cases, when a person accepts something, it is with a goal in mind. Acceptance is the outcome of an evaluation of whether accepting will fulfill the goal in question.

A classic example of acceptance is "Pascal's wager." Blaise Pascal was a seventeenth century philosopher who wrote about the choice of whether to accept that God exists. Pascal contended that the evidence of God's existence is inconclusive: we might as well toss a coin. "Reason cannot make you choose one way or the other, reason cannot make you defend either of the two choices" (Pensées, XLV. "Discourse Concerning the Machine"). This means immediately that the acceptance of God's existence is not epistemic in Lehrer's sense. To break the impasse, we must look at the risks and rewards of making the choice one way or the other. Pascal held that here the choice of believing that God exists is infinitely better than believing that God does not exist. Pascal's elaborate account of why this is so reveals clearly an evaluative component of this acceptance.

Granted that acceptance can have different kinds of objectives, what do all cases of "accepting that p for O" have in common? There is an attitude that results from evaluation of how well it would meet an objective, but what is the attitude? In his first book, Knowlege (1974), Lehrer described the attitude required for knowledge as one of "conviction," which is reflected in one's readiness to assert that p (pp 63-69). One need not be so convinced as to feel certain that p, however. By 1980, in a survey of his own work, Lehrer called the attitude "acceptance," rather than "belief." He added that to accept something is to perform an action: by saying "I accept," I actually accept, just as when I say "I promise," I actually make a promise. ("Self-Profile" in Bogdan, ed. Keith Lehrer, p. 80). So acceptance is optional.

In the 1990 first edition of Theory of Knowledge, Lehrer dropped the description of acceptance as conviction and as a performative in favor of the notion of a "functional role." He uses the same language in the second edition.

Acceptance is the sort of mental state that has a specific sort of role, a functional role, in thought, inference, and action. When a person accepts that p, he or she will reason in a certain manner and perform certain actions assuming the truth of p. Thus, if a person accepts that p, then the person will be ready to affirm that p or to concede that p in the appropriate crcumstances and use p to justify other conclusions. (pp. 39-40).
The key point in this account of the role of acceptance is that of "assuming the truth" of what is accepted. We might describe the attitude in acceptance as one of "taking-to-be-true." (This generic notion was utilized extensively by Immanuel Kant in his Critique of Pure Reason, Transcendental Doctrine of Method, Chapter 2, Section 3, "Opining, Knowing, and Believing.") Epistemic acceptance is the pure form of taking-to-be-true, in which one takes p to be true solely for the purpose of taking to be true exactly what is true.

Acceptance and Belief

In both editions of Theory of Knowledge, Lehrer emphasizes the difference between his condition (iA) and a related condition,

(iB) If S knows that p, then S believes that p.

He is willing to use (iB) at times because this is how the condition has been formulated traditionally, but "when precision is needed" (p. 14), he reverts to (iA). We must not think that belief is simply a broader category than acceptance, as if acceptance were a belief which is held for some purpose or other. Instead, both belief and acceptance are ways in which we take p to be true. The essential difference between the belief and epistemic acceptance is that "acceptance involves evaluation in terms of the epistemic purpose." Presumaby, belief would differ from other kinds of acceptance in that it does not involve evaluation in terms of their purposes, either. If belief is related to a purpose at all, is just a by-product. "Belief may result from the pursuit of some purpose, but it is not defined in terms of any purpose" (p. 13).

If belief is not the outcome of a purposive evaluative process, where does it come from? What Lehrer calls "simple belief," is said to "arise in us naturally without our bidding and often against our will" (40). Lehrer makes the bold claim that humans literally are of two minds. A lower-level mind has beliefs independently of a higher-level "metamind" which accepts things on the basis of evaluation of how well a goal is served by adopting a certain attitude. (It is "meta" because one of its objects is the mind itself, i.e., the attitudes such as acceptance that the mind might take on.) Lehrer's example of a conflict between the two levels is one in which a person believes against all the evidence available to her that her loved one is safe.

Metamind can positively evaluate simple belief, and on this basis accept what is believed (which it would not do in the safety case). On the other hand, it may negatively evaluate simple belief and accept the opposite of what is believed (as it might do in the case of the belief that one's loved one is safe). There is no necessary connection between belief and acceptance. Lehrer states that "acceptance that p involving evaluation in terms of the epistemic purpose may coincide with belief that p. We may, therefore, expect the appropriate kind of acceptance ot be accompanied by a kind of belief, but we should not assume that belief is a kind of acceptance" (p. 14).

To motivate his view further, Lehrer makes an analogy between simple belief and desire. Both arise involuntarily, and both can be positively or negatively evaluated. Corresponding to acceptance is "preference," which is an evaluative notion. What we prefer may conform to desire or conflict with it. My powerful desire to drink heavily is balanced against my preference not to, so as to avoid a hangover. Or I may prefer to drink a lot and suffer the consequences.

An example from ancient epistemology can be adapted to illustrate the difference between the two levels of mind as well as the process of negative evaluation.

Someone seeing a coiled rope in a dark room jumps over it, thinking at first that it is a snake, but afterwards he turns around and searches for the truth, and finding that it is motionless, his intellect is already inclined to thinking it is not a snake, but still reckoning that sometimes snakes do not move owing to winter frost, he prods the coil with a stick, and having thus tested the presentation that occurs, he assents to its being false that the body presented to him is a snake. (Sextus Empiricus, Against the Mathematicians, 7.187-8.)
The thought that the object is a snake might be a belief, one that arises involuntarily upon seeing the object. The intellect's "search for the truth" is the process of evaluation, and the "assent" could be called acceptance that the object is not a snake.

We are not given any examples of belief being positively evaluated with acceptance resulting. We might adapt the case of the snake to this end. Suppose the object really is a snake, but the dark room is in Ireland, and the person accepts that there are no wild snakes there. Although the person believes it is a snake (as evidenced by his jumping over it), he does not accept it. But after seeing it move in the way snakes do, he decides that it really is a snake, a fact that might be explained after he makes his getaway!

One thing that might be puzzling in this account is Lehrer's assertion that "belief may fulfill an epistemic purpose" (p. 14) even without being evaluated. Apparently the idea is that when we successfully take p to be true, perhaps automatically and even against our will, we have added to our stock of truths, which is the goal of acceptance. "My belief that p may fulfill the epistemic purpose if p is true, even though I have not evaluated the matter in terms of that objective but have only sought to conform to authority to obtain peace of mind" (p. 14). It would seem that this kind of case would most naturally be illustrated by the initial reaction to seeing the real snake.

But Lehrer's illustration is tricky, because it seems to be a case of non-epistemic acceptance, like the outcome of Pascal's wager. (Lehrer also describes the safety case as one where the belief is "for the sake of felicity," which makes it look like an acceptance, p. 14.) One could say that even though the belief is the outcome of a goal-oriented process, there is no evaluation of the suitability of the belief for the purposes of meeting the goal. This may explain the safety case, where the belief would naturally arise involuntarily and hence without evaluation.

In cases of conforming to authority, however, people usually are well aware of what is going on, and they often "accept" things (in the performative sense) that they do not believe. There is plenty of evaluation going on! One way to make the example work better is to say that the belief is the outcome of a kind of habituation, whereby one learns to take as true what one is told as a kind of reflex action. No doubt many people have been consciously or unconsciously adopted this kind of habit in the face of oppressive authority.

Because evaluation is always involved in acceptance, one's knowledge is restricted to the products of "metamind," to what has been evaluated. This is an important consequence of Lehrer's account of justification, because it is only a matter of speculation as to how much we evaluate the information that comes before us. If it turns out that metamind is not so very active, it thereby turns out that we do not know so very much.

Objections to the Acceptance Condition

Objections to the acceptance conditions are of two types. One type is based on general features of linguistic usage. The second type brings forward a specific counter-example purporting to be a case of knowledge without acceptance. Lehrer notes that the objections are directed at a belief condition for knowledge. He thinks that his distinction between belief and acceptance is not relevant to the force of the objections, and he couches them in the traditional language of belief, though he answers them in terms of acceptance.

Linguistic Objections to the Acceptance Condition

The first objection is dealt with rather easily. It is correct to say, "S does not believe that p, S knows that p." This may be understood as indicating the possibility of knowledge without belief. But as Lehrer notes, this way of speaking does not deny thatS believes what he knows, but only thatS merely believes what he knows.

The second objection is based on J. L. Austin's account of claims to knowledge as performative utterances. When a person claims to know, perhaps by asserting "I know that" before a declarative sentence, the person is putting his personal authority behind the truth of the sentence. The act of making a knowledge claim creates a new reality, a public commitment, for which the claimant can be held accountable. Suppose my wife and I leave our house for a long trip, and she asks me whether the oven is off. I tell her that I know that it is. When we return, the house has burned down, and I will be held accountable. In this way, claiming to know is like making a promise. (Both are called "performatives," in that performing the ritual of assertion brings about a new state of affairs, a commitment.)

It is not clear fom the present text why Austin's account of knowledge claims should be taken as an objection to the acceptance condition. This is explained in the original 1974 version of the text, Knowledge.

Thus, the reasoning is that if we accept condition (iB) then we must admit that when a man says 'I know', he is describing this state as being one of belief. Since he is not describing himself at all, but is performing an epistemic ritual, according to Austin, he is not implying that he is in such a state of belief. (p. 53)
The problem with this line of reasoning, as Lehrer points out, is that performing a ritual does not necessarily conflict with one's giving a description at the same time.
If I were not claiming to be justified in accepting that S is P when I say 'I know that S is P,' then why in the world would should my saying 'I know that S is P be taken as giving my authority for saying that S is P? It might be more reasonably taken as an expression of opinionless agnosticism. (pp. 33-34)
Note that this comment goes further than required to answer an objection to the acceptance condition, by holding that knowledge claims involve justified acceptance. The reference to justification is a clue that Austin's account of knowledge claims should be seen as directed against the entire project of giving an analysis of knowledge in the sense "in which 'to know' means to recognize something as true" (p. 5), rather than just against the belief condition. In any case, Lehrer's response is the same, that there is no confusion here between performing a ritual and describing one's epistemic situation.

Radford's Counter-Example

The final challenge to the acceptance condition, due to Colin Radford, is based on concrete examples which are said to be cases of knowing without believing. In these examples, a person, when quizzed, utters a correct answer when asked. He does so because he was taught the correct answer but refuses to endorse the answer as correct. Suppose it is John who learned long ago that Queen Elizabeth died in 1603 but now thinks he knows nothing of British history. Nonetheless, when asked when Elizabeth died, he answers correctly on the basis of his memory. It is clear that John does not accept (in the interest of accepting all and only what is true) that Elizabeth died in 1603. For the counter-example to be successful, it must be established that John knows that Elizabeth died in 1603.

There are reasons to think this is so. The more obvious reason is that John remembers the date accurately. We will turn to this reason shortly. The less obvious reason lies in this argument:

  1. John knows the correct answer.
  2. The correct answer is that Elizabeth died in 1603.
  3. If John knows the correct answer and the correct answer is that Elizabeth died in 1603, then John knows that Elizabeth died in 1603.
  4. Therefore, John knows that Elizabeth died in 1603.
This argument is formally valid, so if the conclusion is false, one of the premises must be. It is agreed that premise 2. is true, which leaves 1 and 3 up for grabs.

Surprisingly, Lehrer is willing to grant that John knows the correct answer. One would think this is because John has correctly remembered that Elizabeth died in 1603, but this is not the reason. Lehrer asserts that in the context of a quiz, someone who guessed the correct answer should be said to know the correct answer. It may be that people speak this way at times. If we grant that John knows the correct answer in this sense of 'know', there seems to be no reason not to grant that John knows that Elizabeth died in 1603 in this same sense. After all, he said that she did when quizzed. Of course, this is not the sense of knowledge Lehrer has in mind when he denies the conclusion. Thus it would seem better for him to deny premise 1 given that 'knows' is used in the same sense as in the conclusion.

At this point, Lehrer agrees with Radford that this is a "borderline case," one which is close enough to either side (affirming or denying that John knows that Elizabeth died in 1603) that it is a "matter of choice" (p. 37) that cannot be settled by argument. To argue for one side would be like arguing about a color which resembles both red and orange. The choice Lehrer makes is to apply the epistemic term 'know' in a way that requires acceptance. Once the choice is made, Lehrer can deny premise 3, on the grounds that the kind of knowing whose possibility he grants (knowing the correct answer without accepting it) does not imply the kind of knowing he chooses to emphasize (knowing that the answer is correct).

To explain his choice, Lehrer defends his conception of knowledge against the alternative favored by Radford. (Lehrer actually proposes to "justify refusing to apply the term 'know'" in Radford's case (p. 37). But this would seem to amount to a denial that it is truly borderline.) As has been mentioned, Radford thinks that John's remembering that Elizabeth died in 1603, along with his assertion of it, is sufficient for knowledge. Lehrer maintains that this kind of knowledge is impoverished in a fundamental way. If a person is not committed to the information as the result of an evaluation (and John clearly is not), then that information cannot play the kinds of roles that Lehrer's robust sense of knowledge does. In particular, "Examples of alleged knowledge in which a person does not know that the information he accepts is correct may be of some philosophical interest but such knowledge falls outside the concern of konwledge used in a way that is characteristically human in critical reasoning and the life of reason" (p. 41). Lehrer amplifies this claim by stating that "our ediface of scientific knowledge and practical wisdom depends upon the social context in which criticism and defense determine which claims are to be employed as postulates of scientific systems and the information for practical decisions" (p. 41).

After making his case that Radford's case is borderline and justifying his choice of one side of the border, Lehrer's final task is to explain why "philosophers of merit" would choose the wrong side. His diagnosis is that attributing knowledge is a common-sensical way of explaining why someone is correct. It is not a matter of luck that John gave the correct answer: he remembered it. Lehrer concedes that John once knew that Elizabeth died in 1603, but he denies that he still knows it. The "knowledge has vanished" (p. 42). What has vanished is knowledge in the robust sense described so eloquently in the last paragraph. All that is needed is a proper theory of memory, i.e., one that explains how one can remember information that he cannot tell that he remembers.

To reinforce his point, Lehrer adapts an example of David Armstrong, where someone says, and maintains under questioning, that Elizabeth died in 1306. It is not a matter of luck that he gives this answer, but we certainly would not explain it by saying that the man knows that Elizabeth died in 1306. "Just as it would be incorrect to explain why this man gives the answer he does by affirming that he knows the date of Elizabeth's death, so explaining why John and the others are able to give a correct answer by affirming that they know, would be otiose" (p. 42). The idea here seems to be that we should explain this man's answer and John's answer in the same way, by saying that they recalled this date, although one did so defectively and the other accurately. So neither answer was a matter of luck. But the fact that John got it right is not enough of an addition for us to switch the explanation and say that John got it right because he knew. "A theory of memory that explains how memory produces the results it does, whether correct or incorrect, does not require us to assume that one knows one's answer is correct whenver memory enables one to produce a correct answer" (pp. 42-43).

Knowledge and Competence

A final comment is in order. It may be that the explanation for Radford's choice lies elsewhere. It is quite plausible to think that the key feature of John's remembering the correct date is that it exemplifies John's competence. What distinguishes John and the man who thinks the date is 1306 is that John competently exercises his memory and the other man does not. It does seem to be a feature of ordinary language that we are willing to ascribe knowledge to people who do things competently, and that we are unwilling to do so in cases, like that of the mis-rememberer, when their faculties are not competently exercised.

Lehrer likes to draw the line between his conception of knowledge and opposing conceptions as between the sense of knowing the information is correct and the sense of mere possession of information.

Correct information is necessary to human knowledge and is useful in picking out the sense of the word 'know' that concerns us, but the possession of correct information is not sufficient for human knowledge in that sense. This sort of knowledge is something beyond the mere possession of information, since one must know that the information is correct in order to attain knowledge that supplies one with premises for reasoning and the other endeavors described above. (p. 6)
A person who is competent to give the correct answer as is John, more than merely possesses it. And a competent person could have a stock of premises for use in reasoning. Competence without acceptance may well fail to yield the kind of knowledge that is implicit of the great achievements of society. But it is a respectable concept, one worthy of "philosophers of merit," that does not yield to cases like Armstrong's.

We have now completed our discussion of the further analysis of the truth and acceptance conditions of knowledge. Truth gets a minimal account which even so must be used with caution. Acceptance is the outcome of the evaluation of information against the goal of accepting all and only what is true. True acceptance is not sufficient for knowledge, for one may stumble onto the truth as a matter of luck. For there to be knowledge, the acceptance must be justified. It is to justification that we turn our attention in the next three chapters.

[Previous Lecture | Next Lecture | Lecture Notes Menu ]

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