Epistemology is the study of our right to the beliefs we have. More generally, we start from what we might call our cognitive stances, and ask whether we do well to have those stances. Cognitive stances include both our beliefs and (what we take to be) our knowings; and in another dimension they include our attitudes towards the various strategies and methods we use to get new beliefs and filter out old ones, as well as the products of those strategies and methods. Epistemology, on this showing, is explicitly normative; it is concerned with whether we have acted well or badly (responsibly or irresponsibly) in forming the beliefs we have.
In pursuing this inquiry, we do not, of course, ask only about the beliefs and strategies we find ourselves with at the beginning. We also ask whether there are not others which we would do better to have, and whether there are not others which we should have if we have these ones to start off with. The hope is to end up with a full account of how a responsible cognitive agent should behave, with some assurance that we do not fall too far short of that ideal. 1. Justification. We can distinguish between two sorts of belief: the mediated and the unmediated.
Mediated beliefs are those which we reach by some strategy which starts from other beliefs we have. Inference is such a strategy (but not the only one); we infer that will rain soon from our separate beliefs that it is mid-morning and that it is growing very dark outside. Mediated beliefs raise the question of whether the strategy we adopt is one to which we have a right—one we do well to use. Unmediated beliefs are those which we adopt without moving to them from other beliefs we already have.
These raise different problems, which concern the source of our right to believe. I open my eyes and, because of what I see, immediately believe that there is a book in front of me. If I do well in adopting that belief, it is justified (or I am justified in adopting it). This focus on justification is one way of expressing the idea that epistemology is normative. What makes it the case, then, that this belief is justified? Various answers suggest themselves. One is the reliabilist answer: that the belief is justified because it is the result of a reliable process.
Another is the coherentist answer: that this belief is justified because my world is more coherent with it than it would be without it. A third is the classic foundationalist claim that this belief is not in fact unmediated, but inferred from a belief about how things seem to me just now. If this last were true, we are thrown back to two questions. The first is whether, and how, the belief about how things seem to me just now is justified. The second is whether the inference from that belief is justified. We might ask what principle of inference is employed.
Suppose it is this: that if things seem to me that way, they probably are that way. What makes lithe case that we do well to use this principle? 2. The Structure of Justification. This brings us to one particular question about justification, which has received much attention. Suppose that we give the justification of a mediated belief A which appeals to its relation to some other belief B. This belief, B, justifies that one, A; my belief that it is Sunday justifies my belief that there will be no mail today. There is a very strong intuition that B can only transmit justification to A if it is itself justified.
So the question whether A is justified has not yet been answered, when we appealed to B, but only shelved. Whether it is justified depends on whether B is. What justified B? We might appeal to some further belief C, but then the problem will simply recur. We have here the beginnings of an infinite regress. The first belief in the series is not justified unless the last one is. But will there ever be a last belief in the series? This is the infinite regress of justification. Foundationalism takes this regress seriously, and tries to find basic beliefs that are capable of stopping it.
Promising ways of doing this include the idea that basic beliefs are justified by their source (they are the immediate products of the sense, perhaps), or by their subject-matter (they concern the nature of the believer’s current sensory states). Empiricism, in this connection, wants in some way to ground basic beliefs in experience. Foundationalism concerns itself with the structure of this empiricist programme. So a concern with the regress of justification is a concern with the structure of justification.
Coherentism tries to show that a justified set of beliefs need not have the form of a superstructure resting on a base: the idea here is that the foundationalist programme is bound to fail, so that the ‘base’ is left groundless, resting on nothing. If this were the result, and foundationalists were right about the structure of a justified belief set, the only possible conclusion would be the sceptical one that none of our beliefs are in fact justified. Coherentists reject the base-superstructure distinction; there are no beliefs which are intrinsically grounds, and none which are intrinsically superstructure.
Beliefs about experience can be supported by appeal to theory (which would be going upwards in terms of the foundationalist model), as well as vice versa (theories need the support of experience). The whole thing is much more of a mess, and cannot be sorted neatly into layers. 3. Knowledge. Epistemology, as so far explained, focuses on justification. There is a second focus, on knowledge. Someone whose belief is justified does well. But justification comes in degrees, and so does our epistemic status (determined by how well we are doing).
The top status is that of knowledge. Someone who knows that p could not be doing better (at least with respect to p). There is a natural interest in this top status. Two main questions arise: what is the most we can hope for, and in what areas do we get it? The traditional attempts to define knowledge focus on the first of these. These attempts come in two main families. The first tries to see knowledge as some clever form of belief; the best-known form of this view is the ‘tripartite definition’, which takes knowledge to be 1) belief which is both 2) justified and 3) true.
The second family of views takes knowledge to start where belief gives out. Plato’s version of this was that belief is concerned with the changing (especially the material world), and knowledge with the unchanging (e. g. mathematics). Other versions might suggest that we can have knowledge of our surroundings, but only when some physical thing is directly present to the mind. So knowledge is a direct relation, while belief is conceived as an indirect relation to the thing believed. The second question about knowledge, namely what areas we can get it in, introduces us to the distinction between global and local.
In some areas, we might say, knowledge is available, and in others it is not—or at least it is less freely available. It is common to hear people say that we have no knowledge of the future, of God, or of right and wrong, while allowing that there is at least some scientific knowledge and some knowledge of the past (in memory). Similarly, in discussing the justification of belief we might say that our beliefs about our present surroundings are on firm ground, as firm as that which supports our (rather different) central theoretical beliefs in science, while our beliefs about God and about the future are intrinsically less well supported.