A posteriori

2. Coherentism about empirical warrant

Coherentism about empirical justification identifies particular ‘coherence relations’ among beliefs as constituting empirical justification. Coherence relations can include logical entailment relations, relations of explanation, and various probabilistic, or inductive, relations. Whatever they include, coherentism about empirical justification must identify what makes justification empirical; it must acknowledge a role for sensory experience.

Coherentism about empirical justification faces an objection if empirical evidence can come from a kind of sensory experience that does not require corresponding beliefs descriptive of the relevant sensory experience. Suppose, for instance, that empirical evidence can come from the non-propositional contents (such as visual images) of one’s non-belief sensory awareness-states (for example, non-belief visual states). Not all sensory states are, or even require, belief states. One might believe that one has a certain visual image, but this does not entail that the image in question is a proposition one believes; nor does this entail that every visual image requires a corresponding belief descriptive of that image. Coherentism makes justification depend just on coherence relations among propositions one believes. It thus neglects the evidential role of non-belief sensory experiences in empirical justification. In particular, coherentism apparently permits that what is warranted by coherence relations among one’s beliefs can be isolated, or divorced, from the perceptual contents of one’s non-belief sensory experiences. This would be a serious defect in an account of empirical justification, justification dependent on sensory experience. Proponents of coherentism have not reached agreement on how to treat the previous objection (see Knowledge and justification, coherence theory of).

Coherentism about justification is often called ‘epistemic holism’. According to Quine and others, such holism recommends that we dispense with (1) the analytic–synthetic distinction and (2) the view that our beliefs are tested individually against the input of sensory experience (see Quine, W.V. §3). Our beliefs, given epistemic holism, are confirmed or disconfirmed as a system, collectively rather than individually. In addition, Quine holds, any of our beliefs can be held true come what may, if we make drastic enough changes elsewhere in our system of beliefs; by the same token, none of our beliefs is immune to revision. On this basis, Quine opposes any distinction between the a priori and the a posteriori due to considerations about irrevisability of beliefs. This is not, however, a general argument against the distinction between the a priori and the a posteriori, because some versions of the distinction do not appeal to considerations about irrevisability. Some versions appeal, instead, to considerations about the dependence of relevant evidence on sensory experience.