From his study of earlier work in epistemology Quine comes to the conclusion that it is impossible to construct a definition of knowledge, or justify knowledge, from sense experience alone, and that even if we could it wouldn’t reflect how we actually come to have knowledge. Quine proposes then that we simply study how people actually come to have knowledge. To do this we employ the science of psychology, we examine what kinds of experiences lead to knowledge. Truths about knowledge then are derived from the truths of psychology. However psychology itself is derived from knowledge about human behavior. This then is reciprocal containment: knowledge explains our belief in psychology, and psychology explains knowledge.
Epistemology, in Quine’s view, should be a study of how we come to have knowledge, especially scientific knowledge. Unlike rational reconstruction, Quine is not interested in an abstract account of knowledge, or a justification to trust our knowledge, but an explanation of the actual mental processes that results in people coming to have the beliefs we call knowledge. Thus, Quine holds that epistemology should be considered a subset of psychology. In general psychology is able tell us how a person’s experience creates their beliefs, at least in theory, and thus it seems reasonable to suppose that psychology could tell us how experience leads to knowledge in particular. Ideally when provided with a set of experiences we could deduce using a Quinean epistemology all the knowledge beliefs that a person will have as a result (but not all the knowledge beliefs they could have).
Obviously this may seem to put knowledge on a foundation that is at best circular. After all we only feel that our science of psychology is correct because it is knowledge, and at the same time, under Quine’s account, we should think that a belief is knowledge because of psychology. Quine claims that this circularity is not a problem for his proposal. “This interplay is reminiscent again of the old threat of circularity, but it is all right now that we have stopped dreaming of deducing science from sense data. We are after an understanding of science as an institution or process in the world …” Here Quine is saying, in a roundabout fashion, that circularity is not a problem when you have given up looking for a validation for science and are instead interested in only an explanation. If you understand knowledge to some extent, and psychology to some extent, then there is the possibility that by defining each in terms of the other you will understand both more fully.
There is, however, still a serious problem with reciprocal containment. First consider that knowledge is supposed to be true, or at least reliable or probable. Whatever knowledge is it is assumed to give us a foundation from which we can have information about the world with more certainty than a wild guess. This is why we feel comfortable deducing principles of psychology from our knowledge of human behavior, and in turn calling them knowledge. The problem then is that what many people consider to be knowledge is neither true, reliable, nor probable. For example there were ancient peoples who believed that visions from the gods were knowledge just as firmly as we believe science today. A psychological study of knowledge can’t help but conclude that both of these beliefs are knowledge, since psychology has no way to distinguish between them, as it can’t rely on objective standards of truth or justification. Knowledge then, as defined under Quine’s method, isn’t necessarily true or reliable, but in that case why should we trust psychology? And if the principles of psychology aren’t true or reliable why should we expect them to reveal what experiences and mental processes actually result in “knowledge”? Quine’s proposal then seems self-defeating, because it seems to indicate that we have no reason to trust the results of such a project.
Perhaps Quine may be able to defend himself by claiming that he is simply creating a descriptive account of what we say knowledge is, and not a normative account. Such a defense however seems hollow, because to say that something is knowledge is to recommend it, and to claim that it is reliable. If Quine defines knowledge as something that is neither of these things then perhaps he is really defining something else, such as “acceptance by society as true”, but I think that such “knowledge” is a poor basis for science, or anything else. For example we can imagine a primitive society that believes knowledge comes only from divine revelation. One of these revelations gives them a theory of psychology, which states that knowledge only comes from divine revelation. A Quine from this society could defend their account of knowledge in the same way our Quine does for his, and I think this reveals that it is a poor defense, when you can justify anything with it. What Quine’s account needs is some extra criterion of justification or truth that lies outside psychology in order to distinguish between knowledge and simply a strongly held belief. With such a criterion he can argue that psychology really is reliable. However, requiring such a criterion puts us exactly back where we started.
Thus Quine leaves the real task of epistemology, distinguishing knowledge from superstition, unfinished. Even if our actual process of knowledge formation doesn’t include such a justification that doesn’t mean we should stop looking for one. Before the rational reconstruction of mathematics mathematicians went directly from their intuitions to statements in their branch of mathematics. However, now that we have a justification of math based in logic and set theory, we can say with confidence that the branches of mathematics are all correct and consistent, and we can use the rational reconstruction as an additional tool for doing better mathematics. Likewise, a justification of knowledge would allow us to eliminate any beliefs that we had misclassified as knowledge, as well as allow us to pursue knowledge with more accuracy in the future. This is reason enough to keep looking for a justification, and not to be satisfied with an explanation. Would we have been happy if the explanation of mathematics was a theory about how mathematicians have and apply their intuitions?