What are we to say to these arguments against (iB) and (iA)? The most direct reply is one conceded by Radford. At one point, Radford admits that his examples are
borderline cases of knowledge. 13 This is precisely the case, but what is a borderline case? To say that a case is borderline means there are considerations in favor of applying the term, and equally strong considerations in favor of not applying it. For example, if we see something that is very similar in color to many red things, so much so that this is a quite conclusive reason for saying that the object is red; but at the same time it is very similar to things that are orange and not red, so much so that this is a quite conclusive reason for saying that the thing is not red, then we have a borderline case of something red. Such cases abound.
For most terms of everyday speech, we can expect to find that the term applies without doubt or controversy in a large number of cases, and that it also clearly fails to apply in many cases. On the other hand, in between these cases there are
examples of things where it is not evident whether or not the term applies, no matter
how much we know about the example. Here we are very likely to conclude that the case is borderline. Debate on whether the term applies in such cases can produce arguments and profound speculation, but no one can win because the case is precisely one in which the application of the term is not fixed. As Stephan Körner has
suggested, such terms are inherently inexact, and, therefore, the decision to apply or not to apply the term in borderline cases is a matter of choice. 14
If we wish to defend (iB) and (iA) by maintaining that the cases of John, Alan, and Joan are all borderline for the application of the word 'know,' two tasks remain. First, we must show that the cases are genuinely borderline, and, second, we must justify our choice of applying epistemic terms in the manner required for the truth of (iA).
Some argument for the conclusion that John's case is a borderline case of knowledge has already been given in the earlier presentation of two persuasive arguments, one yielding the conclusion that John does not know that Elizabeth died in 1603, and the other the exact opposite. Similar argument may be offered in the cases of Alan and Joan. Since each sincerely reports and believes that they do not know, neither knows that the correct answer to the question is 3.1416 at the time of their
-32-report. In order to know that 3.1416 is the correct answer, however, all that either would have to know is that 3.1416 is the expansion of pi to four decimal places.
Hence, neither person knows that 3.1416 is the expansion of pi to four decimal places when they say they do not know. This argument is in direct opposition to Lemmon's, which was that, since Joan and Alan have not learned the answer to the question between the time of their first and second remarks, each must have known all along that 3.1416 was the expansion of pi to four decimal places. Once again, there are equally persuasive arguments for contradictory conclusions. Such arguments and counterarguments concerning these examples show them to be borderline cases. It is precisely like arguing about whether something is red when it is as close to red as it is to orange; and no argument can settle that.
Knowledge Implies Acceptance
Given, as we have argued, that the cases of John, Alan, and Joan are borderline, how can we justify refusing to apply the term 'know' in such cases? The appropriate
justification is theoretical, one concerning the role of acceptance and the evaluation of information in knowledge and justification. In all three cases a person may be said to possess the information that Elizabeth died in 1603 or that pi is 3.1416, but they do not, at a given time, know that the information is correct because they do not accept that it is. John possesses the information that Elizabeth died in 1603 in that it is retained in his memory. Since the information is retained in his memory, we say, when he produces an answer, that he remembered that Elizabeth died in 1603, even though he does not know that Elizabeth died in 1603. This is a case, of which there are many, where remembering that p does not logically imply knowing that p. The reason John does not know is that, though he possesses the information that p in memory, John does not know that the information is correct.
Some philosophers, Dretske, for example, have assumed that if a person receives
information that p, then they know that p. 15 Others, like Radford, have assumed that if a person receives and retains the information that p, then the person knows that p.
Both views have the same defect. There is an important distinction between receiving or possessing information and knowing that the received or possessed information is correct. Only when one knows that the information one receives or possesses is correct does one have knowledge.
Imagine, for example, that Mary had been told that Elizabeth died in 1603 by Peter, who is notoriously untrustworthy in such matters. Imagine further that Mary retained that report in memory only because
-33-she had accepted what Peter had told her, in spite of knowing him to be untrustworthy in such matters. It would be obvious that Mary did not know that Elizabeth died in 1603 when Peter told her because she did not know that the report she received was correct. It should be equally obvious that John does not know that Elizabeth died in 1603, when that information is retained in memory, because he does not know that the information is correct. John, unlike Mary, might have once known that Elizabeth died in 1603 because he once knew that the information was correct. Now, though the information is retained in memory, John no longer knows that the information is
correct. That is why John does not know that Elizabeth died in 1603. John does not know this because he does not know that the information he possesses is correct. He does not know that the information he possesses is correct because he does not even accept it.These reflections provide us with the basis for a perfectly general proof that knowledge implies acceptance. The proof is as follows:
1. If a person does not accept that p, then the person does not accept the information that p.
2. If a person does not accept the information that p, then the person does not know that the information that p is correct.
3. If a person does not know that the information that p is correct, then the person does not know that p.
Therefore,
4. If a person does not accept that p, then the person does not know that p.
This is equivalent to the acceptance condition, (iA) If S knows that p, then S accepts that p with which we began.
The premises of the argument, once made explicit, may seem so obvious as to require no proof. They do, however, reflect our concern with a kind of acceptance and
knowledge consisting of the recognition of information. Acceptance of information is not a sufficient condition for knowing that the information received is correct, but it is necessary. Consequently, the failure to accept information results in failure to know that the information is correct. The problem in the cases of John, Alan, and Joan was the same. Each retained some information that p in memory, but, at a specified point in time, none accepted the information that p
-34-and, as a result, did not know that p. In the case of Alan and Joan, subsequent
reflection produced the acceptance of the information that pi is 3.1416, and, therefore, knowledge replaced ignorance. In short, it does not suffice for knowledge that some information is stored in memory as a result of communication or perception. One must, in addition, have the appropriate sort of access to the information and acceptance of it.
Knowledge and the Functional Role of Acceptance
The foregoing remarks may be construed as further elucidation of the kind of knowledge that is the object of our study. There may be living beings, such as a
gullible child, a well as machines, such as an answering device, that receive and retain information. If we know enough about those beings and machines, they may be a source of information and knowledge for us. They, however, do not know that the information they possess, retain, and transmit is correct. We, on the contrary, not only receive and transmit information, we process it. We accept some but not all of what we receive and know that some but not all of the information is correct.
The nature and role of acceptance in knowledge requires some clarification.
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 draw certain inferences 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 circumstances. They will also be ready to justify the claim that p. If they accept information received from the senses or retained in memory, they will regard such information as correct and proceed accordingly in thought and action. The reluctance of John to affirm that Elizabeth died in 1603 and the initial reluctance of Alan and Joan to affirm that pi is 3.1416 reveal an initial ignorance and lack of acceptance of the information. As a result of the appropriate processing of information retained in memory, this ignorance is replaced by acceptance and
knowledge. Acceptance of p sometimes arises from considered judgment that p, but a functionally similar state of judgment may arise in other ways. To accept that p is to accept the information that p. To accept the information that p implies a readiness in the appropriate circumstances to think, infer, and act on the assumption that the information is correct.
The result of our argument is that we shall resolve the borderline cases of John, Alan, and Joan by saying they lack knowledge. For some purposes of everyday speech, we can afford the semantic imperfection that yields the sort of contradictory conclusions we have derived from
-35-the study of John and company, but an acceptable -35-theory must eliminate such
imperfection for the sake of consistency. We shall require that the epistemic terms in question carry the implication of acceptance, for this will enable us to extricate our employment of the terms from the contradictions noted earlier. Such a requirement
leads directly to the conclusion that John, Alan, and Joan do not know that p when they say they do not know, because they then lack acceptance that p. By so doing, we are not dogmatically ruling out the possibility that some theory of knowledge might be constructed which would rule in the opposite direction. We should welcome the
development of such a theory. However, as we affirmed earlier, our concern is to present a theory of knowledge and justification to explain how we are justified in accepting information as correct and claiming to know that it is. Examples of alleged knowledge in which a person does not know that the information they accept is correct may be of philosophical interest, but they fall outside the concern of knowledge that is characteristically human.
In defence of the foregoing restriction, it should be added that our decision to require the implication of acceptance of the information that p, and accompanying readiness to affirm that p in the appropriate circumstances when a person knows that p, is nothing arbitrary or idiosyncratic. It is warranted by the fact that our edifice of scientific knowledge and practical wisdom depends upon the social context in which criticism and defence determine which claims are to be employed as the postulates of scientific systems and the information for practical decisions. In such contexts, when a person admits ignorance, one is taken at one's word; for such a person is not willing to make the sort of epistemic commitments that would enable us to check their cognitive credentials. Of course, we may well be interested in their reasons, if they have any, for conjecturing what they do, but this is quite different from asking whether they are completely justified in accepting something as correct information and claiming to have knowledge. An affirmative answer to that question not only shows that the person has knowledge, it also tranfers that knowledge to those who understand the justification and apprehend its merits. Our theoretical concern with critical reasoning, and our attempt to explain how such reasoning succeeds, warrants our decision to rule that knowledge must involve the forms of acceptance cited above.