1 Pollock, J. Contemporary Theories of Knowledge (Savage, MD: Rowman and Littlefield 1986), 81Google Scholar

2 For a survey of various causal theories of the basing relation, see Korcz, K. ‘Recent Work on the Basing Relation,’ American Philosophical Quarterly 34 (1997) 171–91Google Scholar.

3 I explain the additional conditions required for a meta-belief to be appropriate below.

4 Of course, the account of the basing relation I offer does not presuppose that a deontological account of justification is correct.

5 Lehrer, K. ‘How Reasons Give Us Knowledge, Or the Case of the Gypsy Lawyer,’ The Journal of Philosophy 68 (1971) 311–13CrossRefGoogle Scholar. Pollock, J. and Cruz, J. Contemporary Theories of Knowledge, 2nd ed. (Lanham, MD: Rowman and Littlefield 1999), 79Google Scholar, offer a similar interpretation of Lehrer's example.

6 E.g., Goldman, A. ‘What is Justified Belief?’ in Pappas, G. Justification and Knowledge (Dordrecht: D. Reidel 1979)Google Scholar; and Audi, R. ‘The Causal Structure of Indirect Justification’ in Audi, R. The Structure of Justification (New York: Cambridge University Press 1993)Google Scholar and R. Audi, ‘Belief, Reason and Inference’ in The Structure of Justification.

7 Other versions of this objection appear in Pappas, G. ‘Basing Relations,’ in G. Pappas, Justification and Knowledge 51–63, at 57-8CrossRefGoogle Scholar; and Lehrer, K. Theory of Knowledge Boulder, CO: Westview 1990), at 169–70Google Scholar.

8 This is a point in which the case of the persistent lawyer differs significantly from Lehrer's original example. The point of this alteration is to help avoid the objection that the lawyer is merely using the complicated line of reasoning to rationalize his belief that his client is innocent. I discuss this point in detail in text, below.

9 My aim here is to merely provide motivation for the lawyer to search for an additional reason so as to establish the point of the example. I do not mean to suggest that there is anything epistemically wrong with the lawyer's acceptance of the card reading.

10 Audi, R. ‘Rationalization and Rationality,’ in The Structure of justification, 429Google Scholar

11 Others have also questioned causal analyses of rationalization. See, for instance, Jacobson, A.J. ‘A Problem for Causal Theories of Reasons and Rationalizations,’ The Southern journal of Philosophy 31 (1993) 307–21CrossRefGoogle Scholar.

12 Alston, W. Epistemic Justification: Essays in the Theory of Knowledge (Ithaca, NY: Cornell University Press 1989), 82–3Google Scholar

13 Swain points out that there is still a pseudo-overdetermination relation from reason to belief in such cases, and I agree. However, I do not include this as an additional condition for showing a belief to be justified because (a) I am not sure how to adequately explicate the notion of pseudo-overdetermination and (b) it seems to me that the five conditions I present are in fact sufficient for showing a belief to be justified.

14 Tolliver, J. ‘Basing Beliefs on Reasons,’ Grazer Philosophische Studien 15 (1982) 149–61Google Scholar. Audi holds a similar view with regard to beliefs based on other beliefs, but not perceptual states (‘Belief, Reason and Inference’). Richard Fumerton may seem to hold the view that the basing relation consists of (a) the reason's causing the belief and (b) having the appropriate meta-belief that the reason confirms the belief (Fumerton, R. Metaphysical and Epistemological Problems of Perception [Lincoln: University of Nebraska Press 1985], 51–2)Google Scholar. However, he merely intends to require the meta-belief as a necessary condition of the belief's being justified, not of the belief's being based on the reason (personal conversation). All make additional stipulations as to what counts as an appropriate meta-belief.

15 Note that r need not actually be a good reason to believe p. Appropriate meta-beliefs can be mistaken as to the actual quality of the reason r, and yet establish a basing relation. Thus, appropriate meta-beliefs can establish basing relations upon bad reasons and leave the belief so based unjustified.

16 Let me emphasize that (SCD) is just a sketch of the causal-doxastic theory, and that I lack the space to address other important issues relevant to the formulation of it. ·

17 r being a member of the set of reasons R, having one or more members, and consisting of all the reason(s) upon which S's belief that p is based. Thus, a belief may be based on several reasons.

18 Note that this ‘or’ is exclusive. This also follows from condition (b)(i).

19 S's meta-belief must be functionally equivalent to the belief that r is a good reason to believe that p, but need not have this exact content. For example, the meta-belief could be that if r is true, then p is, or that r entails p, or that r makes p very likely, etc.

20 Similar examples occur in Pollock, Contemporary Theories of Knowledge (37)Google Scholar and Winters, B. ‘Inferring,’ Philosophical Studies 44 (1983) 201–20, at 203CrossRefGoogle Scholar. Earlier examples occur in G.E. Moore's discussion of inference in his notebooks written in 1919, but not published until 1962: Lewy, K. ed., Commonplace Book 1919-1953 (London: George Allen and Unwin 1962), 7Google Scholar.

21 Plantinga, A. Warrant: The Current Debate (Oxford: Oxford University Press 1993), 69n8CrossRefGoogle Scholar. An evidential basis relation is a basing relation where one's reason is itself a belief (70).

22 Alston, W. ‘An Internalist Externalism,’ in Epistemic Justification, 228Google Scholar

23 I take the phrase ‘causal chains of events’ as primitive.

24 This is the same sort of reply D.M. Armstrong makes to a different objection. See Armstrong, D.M. Belief, Truth and Knowledge (Cambridge, UK: Cambridge University Press 1973), 83.CrossRefGoogle Scholar

25 Deutscher, M. ‘A Causal Account of Inferring,’ in Brown, and Rollins, eds., Contemporary Philosophy in Australia (London: George Allen 1969), 112Google Scholar

26 Davidson provides another example involving a belief causing an individual to become unnerved and his becoming unnerved then causes an action. See Davidson, D. ‘Freedom to Act,’ in Essays on Actions and Events (Oxford: Oxford University Press 1973), 79Google Scholar. Plantinga's example of Sylvia, discussed above, contains a similar element.

27 It may be objected that we do talk of justificatory reasons for emotions. But, in reply, this is not epistemic justification, it is some other kind of justification.

28 I should emphasize that this is a fictional example and is not meant to characterize dogs in general as being violent. I encourage you to support your local humane society.

29 One possible objection to this handling of the Red Santa example is to argue that seeing the red house and believing that the dog attacked the doll are a conjunctive cause of the belief that the dog is violent. If so, then wouldn't seeing the red house be a reason for which I believe that the dog is violent. One difficulty with this objection is that it seems clear that this is not what happens: the causal chain is clearly linear. Another difficulty is that attributing such a conjunctive cause to someone attributes to them an incoherent complex reason state.

30 r being a member of the set of reasons R, having one or more members, and consisting of all the reason(s) upon which S's belief that p is based. Thus, a belief may be based on several reasons.

31 r being a member of the set of reasons R, having one or more members, and consisting of all the reason(s) upon which S's belief that p is based. Thus, a belief may be based on several reasons.

32 Note that this ‘or’ is exclusive. This also follows from condition (b)(i).

33 S's meta-belief must be functionally equivalent to the belief that r is a good reason to believe that p, but need not have this exact content. For example, the meta-belief could be that if r is true, then p is, or that r entails p, or that r makes p very likely, etc.

34 Note that this leaves open the possibility that a basing relation could be established between some reason and the meta-belief described in (ii) by some second meta-belief satisfying condition (b). However, no such additional meta-belief is required by CD.

35 I plan to pursue this possibility in future work.

36 I develop this point in greater detail in my paper ‘An Objection to Process Reliabilism,’ in which I argue that counter-examples such as the case of the persistent lawyer are far more decisive than the standard objections to process reliabilism, which leave the door open for successful revisions of the process reliabilist theme. For a recent example of such a defense against the generality problem, see Alston, W. ‘How to Think About Relativity,’ Philosophical Topics 23 (1995).CrossRefGoogle Scholar

37 For a discussion of holistic coherentism and the basing relation, see Bender, J. and Davis, W. ‘Fundamental Troubles With the Coherence Theory,’ in Bender, J. The Current State of the Coherence Theory (Dordrecht: Kluwer 1989), 57–8CrossRefGoogle Scholar and Lehrer's reply in the same volume (272).

38 Currently, the most extensive bibliography of works regarding the basing relation is to be found in Korcz, K. The Epistemic Basing Relation, PhD Dissertation, Ohio State University, 1996Google Scholar.