Epistemology explores the possibilities and limits of knowledge as applied to different cognitive domains (knowledge of the external worlds, knowing one’s own mind, understanding language, etc.). The classic definition of knowledge is that a person knows that p iff s/he has the justified, true belief of that p (see Dancy 1985 and Dancy & Sosa 1992 for details). A central issue is the correct characterization of the justification required for knowledge. Inaugurated by Descartes, classical foundationalism is perhaps the most influencial position in epistemology. Foundationalism divides our beliefs into two groups: those which need support from others, and those which need no support themselves (truths we know without relying on perceptual evidence). The latter constitute our epistemological foundations. Foundationalism need not be associated with cartesian rationalism, however. Empiricism identified states which do not need justification with sensory states, our own immediate experiences of the external world. They are infallibly true and stand on their own feet. Other beliefs are supported by them. A central objection to both positions is that no belief is infallible — there is no region in our thoughts which is entirely immune from the possibility of error. Recent epistemology has therefore drawn a distinction between the possibility that all beliefs could be false (the sceptical option) and the global claim that each particular belief could be false without the global system of cognitive attitudes being false. The latter claim is weaker and leads to the well-known fallibilist position in epistemology and the philosophy of science. Classical epistemology (the position we have just characterized) describes itself as a philosophical discipline that explores the limits of knowledge. Epistemology is thus seen as a part of philosophy, not science. Naturalized epistemology, a concept inaugurated by W. V. O. Quine (1960), rejects the assumption that epistemology is not a branch of science and that epistemology, as a ‘first philosophy’, is not able to independently justify claims of sciences like physics or, more mundanely, sense perception. Naturalized epistemology has it that epistemology is continuous with and part of scientific inquiry, not a separate inquiry into the meanings of crucial words such as ‘to know’ or ‘to justify’, or a reach for infallible knowledge. The Quinean approach to epistemological issues is based on holism: there is a difference of degree between sentences to whose truth we are firmly wedded on and those which we are more prone to abandon in the light of recalcitrant evidence. All sentences are synthetic, one might say, but some are more synthetic than others. This epistemological holism reflect the kind of semantic holism that is present in contemporary philosophy of language. The consequences of Quine’s naturalized epistemology for contemporary theories of meaning are wide-ranging, to say the least. A number of these consequences have been drawn by Donald Davidson (1984), but one could also count Wittgenstein’s work On certainty as having a profound influence on our current thinking about what it means to know the meaning of a sentence, or to understand other persons. The crucial connection between epistemology and philosophy of language is that the evidence of one’s senses is not just what we appeal to in the justification and verification of statements; it is also what we start from when learning a language. The notion of evidence of one’s senses is held by empiricists to be basic in the theory of meaning. This has led to the thesis that a statement has empirical meaning if and only if its truth would make a difference to the evidence of our senses, a statement later relaxed by Quine, who held that a non-observation sentence does not have its own observational consequences, which in turn shows that there is nothing that the sentence means if taken all by itself. To the extent that we can say something about the meaning of a sentence, our answer is dependent upon the nature of the theory surrounding that sentence. Sentences have no determinate meaning. It is the pattern in which they are woven that confers meaning upon them. Sentential meaning is indeterminate, according to Quine. Recent epistemological discussions which are relevant for semantics and pragmatics include discussions about what it means to follow a linguistic rule and what it means to know that one follows a rule (of language) correctly. The connection between epistemological and semantic issues is obvious as soon as one thinks of understanding as knowing the meaning of the words and sentences a speaker uses. Interpreting Wittgenstein’s arguments on rule-following in the Philosophical investigations, a number of contemporary analytical philosophers (most notably Kripke 1982) have written extensively on the so-called private language argument and connected it with general considerations about rules and objectivity. The discussion centers around two different and to a certain extent incompatible views of language. The first one connects objectivity to a community of rule-followers. The basis for objectivity and correctly understanding other persons lies in the present behavior of the linguistic community. What makes deviant behavior incorrect is that it is not in step with other persons’ behavior. In line with this communal view on (correctly) understanding is the view that knowledge of conventions is of crucial importance in speaking and understanding. The alternative view is based on the Quinean insight that a theory of meaning (a theory that enables us to speak and understand sentences of a natural language) is a constantly evolving theory based on interpreting other persons’ behavior in the light of the overall rational character of their actions. Languages are, from this point of view, idiolects. No person speaks the same language, and what is central in a theory of meaning is not shared conventions or a communal set of rules one is supposed to know or to follow; what confers meaning upon one’s utterances is the desire to be understood and the evidence speakers create so as to achieve that goal. It should not be a surprise that the idiolect view of language is defended by followers of Quine, such as Davidson. The relation between Quine’s holism (to know the meaning of a sentence is to know the language that sentence belongs to) and current anti-holistic tendencies in the philosophy of language and the philosophy of mind is discussed in J. Fodor & E. LePore (1992). Their central anti-holistic argument is that if holism is true, one cannot understand any sentence unless one can understand all sentences. But how could one then be in a position to learn a language? The countless reactions against this view show that discussions on this issue are far from settled.