It has been argued that a person's epistemological theory about factual propositions can be expressed as a belief-policy covering the degree and kind of evidence for such propositions, and the importance that he attaches to both. There are different possible belief-policies embodying differing (though no doubt overlapping) standards of rationality.

As I shall use the expressions, to justify a belief involves providing grounds for that belief which are not person-relative, but which are generally acceptable and convincing to anyone no matter what his belief-policy; a strong requirement. To render a belief rational to a person it is sufficient to show that it is consistent with the belief-policies held by that person at a time, and that the belief-policy or policies in question do not violate certain necessary conditions for rationality as such. It is not necessary that the belief be consciously derived from the belief-policy, since a belief may, as a matter of fact, antedate the formulation of any policy. In this chapter we shall be pursuing the question of which belief-policy or policies it is rational to accept.

In chapter four a number of broad classifications of belief-policies was suggested. Restricting our attention to beliefs about contingent matters of fact it was claimed that it is possible to classify belief-policies into three broad categories: belief-policies which people find themselves holding at an early age, evidential belief-policies, and non-evidential belief-policies. In addition a distinction was drawn between permissive and mandatory policies, and between verificationist and falsificationist policies. Each class of policy is not exclusive of each of the others, and so various combinations are possible.