Basic Concepts of Epistemology

Epistemology, the theory of knowledge, is a core component of the Western philosophical tradition. Questions about knowledge arise in Plato, presumably inspired by the career of the historical Socrates, and become the basis of a continuous historical dialogue in which virtually every Western philosopher has in some way or another been engaged right down to the present day. We will begin with the definition of knowledge as "justified true belief" as elaborated by the founders of the Western tradition, Aristotle and Plato, and basically accepted by all subsequent philosophy until the twentieth century. For these thinkers, a belief is justified as true only if it meets the criterion of being absolutely certain or "necessarily true." We will then trace the development of Foundationalistic epistemology through the Enlightenment debate between rationalists and empiricists, culminating in the critical philosophy of Immanuel Kant. Although these questions have evolved over the centuries, at a very general level they remain today the same questions with which the dialogue began. For pedagogical purposes they can be reduced to questions concerning the three "conditions" implicit in this definition of "knowledge":

If we start by asking "What is knowledge?" we have to set up criteria for identifying and distinguishing "knowledge" from what is not knowledge. We assume that if we know something we also believe whatever it that we claim to know, so the domain of "knowledge" must be a subset of the class of "beliefs." But obviously not everything people believe to be true is in fact the case; there are false beliefs, and so if what we believe is in fact not the case, then clearly we do not know it, although we may falsely believe that we know it. So knowledge is made up of only those beliefs which are in fact true beliefs. But truth cannot be the only requirement for a belief to be known, because we may believe something and what we believe may in fact be the case, so the belief is in fact true, but our believing it is just a matter of, let us say, a "lucky guess." The person who believes something just as a lucky guess cannot be said to know that thing, because just guessing cannot justify the belief. Reasoning along these lines, Plato was the first to clearly express the view that "knowledge" is "justified true belief" and this is often called the "classical" or "traditional" definition of knowledge.

We can express this definition more formally by observing that the verb "to know" is a transitive verb which takes a "subject" (the "knower") and a direct object (the "known"). Although there may be much which people know which cannot be communicated in language, we will restrict our attention to knowledge which can be expressed in language, in which case the "object" of knowledge can always be expressed as a "proposition"; this is what we will call "propositional knowledge" or knowledge that such and such is so. Thus we can analyze what is meant by saying that "S knows P." where "S" is any knowing "subject" (presumably a human subject) and "P" is anything that can be known, the "object" of knowledge, as follows. Taking its cue from Plato, the tradition has tended to identify "knowledge" with "true, justified belief"; thus to say "S knows P." reduces to three separate claims:

a) S believes P (the belief condition)
b) P is true (the truth condition)
c) S is justified in believing P (the justification condition)
Although most philosophers of the Western tradition would adhere to this classic conception of knowledge as "justified true belief," there are many rival theories with respect to each of these three conditions over which philosophers have divergent views. The following is a quick survey of the main positions.
 

a) The "belief condition," "S believes P," implies that knowledge is something that can be believed. Thus whatever is known is a subclass of what is believed, i.e. of "beliefs." Two rival accounts of what it is to have a "belief" are often debated:

i) the state/object view: to say "S believes P" means S, the subject which has knowledge, is in a certain psychological (or "docastic") state with respect to the object of belief, a proposition.  We express this by saying that the subject "holds" or "accepts" this belief.

ii) the dispositional view: to say "S believes P" means S has a "disposition" or tendency to behave in a certain way, with the stipulation that "behavior" refers to something that is empirical, or can be observed.

Although the dispositional view has been prominent in much twentieth century philosophy, especially that sympathetic to empirical psychology, the dominant tradition has essentially implicitly maintained the state/object view that speaks in terms of subject as a "mind" having "in" it certain "beliefs" or, in the language of much Enlightenment epistemology, of the "faculty of understanding" as "making" certain "judgments."

In much contemporary epistemology, especially that resulting from "analytical" philosophy, what can be believed is identified with "statements" or "propositions". Such an account of knowledge thus focuses on propositional knowledge; i.e. what can be expressed in a true or false proposition.

[If knowledge is identified with a subclass of "beliefs" and if "having a belief" is considered to be equivalent to asserting that a certain proposition is "true," since propositions are things in language, then all knowledge becomes something that can be expressed in language. Obviously the verb to "know" is often used to express other non-linguistic kinds of knowledge, such as knowledge of "how" or knowledge by "acquaintance". Whether or not such kinds of knowledge can be reducible to knowledge that certain propositions are true, remains a debatable issue.]

b) The "truth condition" implies that if "S knows P" not only does S believe P, but also that P, the proposition which is believed, is in fact true. Deciding whether a statement meets this condition means a theory of truth must be given. Three rival theories of "truth" are debated:
 

i) the "correspondence" theory of truth: to say P is true is to say there exists a "correspondence" relation between what P says is the case and what is really the case. Giving an account of what is really the case, i.e. the nature of "reality," which is normally taken as a metaphysical question. If beliefs are propositions, things in language, then the correspondence theory of truth must explicate how language somehow "corresponds" to reality, and that involves further inquiry into the philosophy of language.

ii) the "coherence" theory of truth: to say "P is true" is to say it "coheres" with an entire system of other beliefs, it has a certain place in the totality of all truths

iii) the "pragmatic" theory of truth: to say that "P is true" means that believing P leads to the satisfaction of certain expectations; the belief "works" or is successful in satisfying certain goals, aims, or "interests."


Throughout the nineteenth century the "coherence" theory enjoyed considerable allegiance and in the twentieth the "pragmatic" theory has been a recurrent theme; but overwhelmingly the dominant tradition has maintained some form or another of the correspondence theory.
 

c) The "justification condition": implies that in order for S to know P, S's believing P and P's being true is not enough. P cannot just be a "lucky guess." Being in possession of knowledge requires more than just "holding true beliefs," the subject, S, also has to have good reasons or reliable evidence for believing that P is true. Only in such a case is a true belief, P, said to be "justified." But to justify P by appeal to certain reasons is to say that we know those propositions expressing those reasons, and to say S knows them requires that those reasons in turn be justified.

Those propositions which the subject may or may not be justified in believing to be true can thus be divided into two classes: those whose justification appeals to other propositions being true, and those which are justified as being true by other means than appealing to other propositions. Those in the first class, whose justification appeals to other propositions, are said to be known mediately, i.e. through the medium of other propositions. Those in the latter class, whose justification makes no appeal to the truth of other propositions are said to be known "immediately." or "directly." Are there such directly known ultimate justifiers, those propositions which are used to justify all others? Four different views are debated:
 

i) foundationalism: by far the majority view, it holds that certain propositions are known directly and do not need to be justified by further propositions. These propositions form the "foundation" of all knowledge

ii) coherentism: a proposition is justified by fitting it into a whole system of beliefs; its justification is its part within the whole system of knowledge.

iii) infinitism: the "regress" of one proposition justifying another, which justifies another, etc., goes on infinitely.

iv) contextualism: certain propositions, the ultimate justifiers, cannot themselves be justified, thus not known; but once they are accepted, other propositions can be known by justifying them by appeal to these ultimately unjustified propositions which form the "context" for a "world-view" or system of knowledge. This gives a relativist view of knowledge; a proposition can be known only relative to a particular "framework".


The dominant tradition has defended foundationalism and basically rejected infinitism and contextualism as implying an impossibility of genuine knowledge. It is this third condition, the justification condition, that has been the central concern of traditional Western epistemological debate. Relatively little attention have been given traditionally to the questions of "belief" and "truth" in comparison to the great attention they have received in the past century. However, it is concerning the arguments for and against the different varieties of foundationalism that has been the primary concern of the tradition of western epistemology through Kant.

Obviously how a philosopher responds to one of these conditions will affect his response to others. Thus foundationalists have also tended to hold a correspondence theory of truth and a state/object view of belief. Philosophers who embrace coherence theories of truth naturally find compatibility with coherentist theories of justifications, and those of a pragmatist leaning with respect to truth are likely to embrace a dispositional view of belief and a contextualist account of justification. But the connections between these different "isms" is by no means rigid.