AMERICAI~ PHILOSOPHICAL QUARTEIiLY

1. INTRODUCTION

The objection which I wish to consider attempts to show that Christian belief is irrational. It rests on what I shall call the evidentialist position. That position is characterized by a mutual entailment relation between one=s having adequate evidence and rational belief; i.e., that some person S has a rational belief that p iff S has adequate evidence for p. Further, the apparent failure of the project of natural theology suggests that there is no adequate evidence for Christian belief. The conclusion of the objection, then, is that Christian belief is irrational.

The purposes of this paper concern the evidentialist objection and position. I shall first attempt to explain the objection and show the pervasiveness of the evidentialist position in modern philosophy. Then I shall show how the evidentialist objection fails, and in particular I shall show that the evidentialist position is false. If this conclusion is defensible, then certain methodological constraints governing the debate about the existence of God imposed by the evidentialist objection will be altered. Finally, I wish to suggest how the debate about the rationality of believing that God exists is resolved.

2. EVIDENTIALISM

A classic statement of the evidentialist position is given by W. K. Clifford. He claims that Ait is wrong always, everywhere, and for anyone to believe anything upon insufficient evidence.@¹

The evidentialist position has played a central role since Descartes in distinguishing philosophically, acceptable beliefs from philosophically unacceptable beliefs.

Modern philosophy was founded on the doctrine, uncompromisingly formulated by Descartes, that to think philosophically is to accept as true only that which recommends itself to Reason. In England, Locke had acclimatized this Cartesian ideal. There is Aone unerring mark@ he wrote, Aby which a man may know whether he is a lover of truth for truth=s sake@: namely Athe not entertaining any proposition with greater assurance than the proofs it is built upon will warrant.@²

The evidentialist position has been held by others than those cited in the above quote. David Hume claimed that Aa wise man proportions his belief to the evidence,@³ and more recently Bertrand Russel claimed that Aperfect rationality consists. . .in attaching to every proposition a degree of belief corresponding to its degree of credibility,@ degree of credibility being a function of evidence.⁴

One of the implications of the evidentialist position with respect to the question of the rationality of belief in God concerns the constraints it puts on how the debate should go. If the evidentialist position is correct, then it would seem that Antony Flew=s methodological constraints are warranted. He says,

AIt is up to the theist: first, to introduce and defend his proposed concept of God; and, second, to provide sufficient reason for believing that this concept of his does in fact have an application.@⁵

Thus the evidentialist position serves as a crucial premiss in the evidentialist objection, and it naturally leads to certain methodological constraints on how the debate about the existence of God should proceed.

Let us be a bit clearer about the evidentialist objection. I propose that we introduce the concept Ae-reasonable,@ which is to be read Areasonable (or rational) given the evidence (alone).@ We can use this concept to define the concept of an epistemic warrant:

Dl: Person S has an epistemic warrant for proposition p=_df it is more e-reasonable for S to believe p than to believe not-p or to withhold from believing either p or not -p.

Given this definition of an epistemic warrant, we can present the evidentialist objection as:

1. For all propositions p, it is rational (reasonable) for S to believe p iff S has an epistemic warrant for p.

2. There is no S such that S has an epistemic warrant for the proposition AGod exists.@

3. Therefore, it is not rational for any S to believe that God exists.

In order for the conclusion of the above argument to be acceptable, we need good reason for accepting the premisses. Premiss (2) seems to be supported by the failure of natural theology, i.e., there does not seem to be any acceptable argument yet given that has as a conclusion that God exists. But what about the evidentialist position itself? The standard reasoning for that position seems to be something like this: Athere is a close connection between what is reasonable to believe and what is true. This connection between reasonability and truth is such that only those considerations that count for the truth of a proposition count toward the reasonability of believing a proposition. But our conception of evidence is such that, when a certain proposition e is evidence for another proposition p, p is more likely to be true given that e is true than p would be were e not true. And thus what is reasonable to believe is what is best supported by the evidence.@

This argument for the evidentialist position rests on unacceptable epistemological presuppositions. I shall label these presuppositions Aobjectivist epistemology.@ The central claim of objectivist epistemology is that there are certain objective (subject-independent) connections between propositions which, in relation to cognizers, determine what is justifiably believed by these cognizers.⁶I agree that there are certain objective connections between propositions; however, I wish to claim that these connections do not have much to do with justified beliefs. Any such objectivist epistemology runs afoul of Cartesian-evil-demon considerations. In an evil-demon world, although there still are objective connections between propositions, these connections have nothing to do with what the cognizers in that world justifiably believe since the evil demon deceives them about the truth of most if not all of what they believe. But then any objectivist epistemology entails that persons in such an evil-demon world almost never have justified beliefs, since their evidence fails to bear any objective connection to (fails to make more likely true) the propositional content of their beliefs. For example, take any proposition p believed by some inhabitant S of an evil-demon-world. If S believes p, then p is false given standard operating procedures for evil demons. Now suppose that our objectivist epistemology cites e₁, Y e_n, as evidence for p, and assume S has evidence e₁, ..., e_n. In this (or any) evil-demon-world, e₁, ..., e_n, fails to make p probable at all; and thus S fails to have a justified belief that p since S=s evidence for p fails to make p more likely to be true. It seems to me that such a view is fundamentally wrong, since there is nothing about our epistemic situation that can distinguish their world from ours. Thus, if we allow that we have any justified beliefs, we ought also to hold that inhabitants of a Cartesian-evil-demon world have roughly the same justified beliefs, since, after all, their world could be our world and they could be us.⁷

The relation of this denial of objectivist epistemology to the evidentialist position is that the denial of objectivist epistemology entails that there is no longer any direct connection between rationality and truth. At best, there is only a subjective connection, i.e., that what is justifiably believed is justifiably believed to be true. But that sort of connection between rationality and truth is not one that could support such a grandiose claim as that what is reasonable to believe is determined by what is supported by the evidence, since what is reasonable to believe fails to be more probably true (in any non-subjective sense of Aprobable@). The evidentialist position will have to look elsewhere for support. And I am not sure where to look, since the evidentialist position (on the subjectivist epistemology that I defended above) claims that what is reasonable to believe on the whole is what is determined as reasonable to believe within the proper domain of epistemology. But it is not at all clear that pragmatic concerns cannot outweigh the results of the epistemologist. I shall argue in the next section that at least this much is true: it is possible that the results of the epistemologist be outweighed by pragmatic concerns.

3. A RESPONSE TO THE EVIDENTIALIST OBJECTION

In this section I wish to argue that even if the proposition AGod exists@ is such that no one has an epistemic warrant for it, it is still possible that someone rationally believe that God exists. Thus, in this section, I shall assume that premiss (2) of the evidentialist objection is true, but I shall argue that premiss (1) is false.

Generally we distinguish between what it is rational to believe and how it is rational to act. One way that we determine how it is rational to act is to roughly calculate what risks we face and what benefits can be obtained were a certain proposition true.

Let us introduce the concept Au-reasonable (rational),@ which is to be read Areasonable (rational) in terms of utility.@ Then we can make a first attempt at defining what it is for there to be a prudential warrant for a proposition:

D2: S has a prudential warrant for p = _df it is more u-reasonable for S to act as if p than to act as if not-p.

There is a problem with D2 as it stands. Suppose that we have a prudential warrant for both p and q by D2, and p and q are competitors, defined as:

D3: q is a competitor of p for S=_df (a) both p and q are prudentially warranted on D2, and (b) it is not possible for S to act as if both p and q.

In such a case we ought to restrict the notion of a prudential warrant to that proposition which is most reasonable to act on. But even that won=t avoid all of the problems, for it is possible that two propositions are tied in terms of u-rationality. Let us incorporate into our definition the seemingly plausible intuition that when two propositions are tied in terms of u-rationality, one ought to act on that proposition which, roughly, is most likely true. With these alterations, we get:

D2=: S has a prudential warrant for p = _df (a) it is more u-reasonable to act as if p than to act as if not-p; (b) if there were a q such that q would be a competitor of p, then it would be more u-reasonable to act as if p than to act as if q, or (c) if there were a q such that q would be a competitor of p, and it would not be more u-reasonable to act as if p than q nor q than p, then it would be at least as e-reasonable to believe p as q.⁸

There is one other problem that we must address concerning the notion of a prudential warrant. We have used the notion of acting as if a proposition is true in both D2 and D2=. The problem is that there are many ways of acting as if a proposition is true, and not all of them can be done at the same time. Thus, in order to understand the notion of a prudential warrant, we need some way of identifying which ways of acting on a proposition are intended. Let us introduce the notion of an efficacious act to handle this problem:

D4: Act a is an efficacious act for a prudentially warranted proposition p for S = _df were p true, a would enable S to avoid the risks and/or receive the benefits by which p was prudentially warranted for S.

To illustrate how our definitions function, consider an example of a stranger saying that there is someone down the street giving away money. The proposition Athere is someone down the street giving away money@ is prudentially warranted for S because of the benefits available to S were that proposition true. Now, there are many ways that S could act as if that proposition were true: he might tell the next person he sees that there is someone down the street giving away money; he might write in his diary that there is someone down the street giving away money; he might even believe that there is someone down the street giving away money. However, none of these acts are efficacious on D4. The only act that S ought to perform is the act of walking down the street and getting the money from the person giving it away. Thus, D4 enables us to pick out just which way of acting as if p one can rationally perform.

We are now in a position to see why premiss (1) of the evidentialist objection is false. One way to act as if AGod exists@ is true is to believe that God exists. Since believing a proposition can be the act picked out by D4 as the efficacious act in some cases, it is possible that one ought to believe some prudentially warranted propositions. Further, since it is reasonable to act on prudentially warranted propositions, it is also reasonable to perform the particular acts which are efficacious in a given circumstance. Thus, when the efficacious act is the act of believing the proposition, it is reasonable to believe the proposition. And thus, it is possible that there are some propositions that are reasonable to believe which are not best supported by the evidence, i.e., premiss (1) of the evidentialist objection is false.⁹

One possible world in which believing that God exists is rational on prudential grounds is a Pascalian world. In a Pascalian world, persons have such limited conceptual abilities that they are able to conceive of only two possibilities: either God exists as Christians think of him, or there is no God. In such a world, Pascal=s wager argument¹⁰ yields that there is a prudential warrant for the proposition AGod exists,@ and the efficacious act with respect to this proposition is the act of believing that God exists.

It might be thought that the evidentialist could alter his objection since the above response employs a possible world which is surely not the actual world. The evidentialist might consider altering his first premiss to read: it is reasonable for S to believe p when there is no prudential warrant for p only if p is best supported by the evidence. However, once the evidentialist grants that premiss (1) of his argument is false since prudential considerations can in some cases yield rational belief, it seems to me that it is not at all clear that a revised acceptable premiss can be offered. The one just offered is not acceptable since I can construct other notions of warrant other that D2= that can also entail rational belief in some cases. Consider:

DS: S has a pragmatic warrant for p = _dfamong all the propositions that meet minimal evidential requirements E, (i) it is more u-reasonable to act as if p than not-p; and (ii) (b) or (c) of D2= is true.

When we develop our epistemic logic, suppose that we have ten levels of epistemic acceptability for propositions. Then D5 is really a schema for ten separate definitions, instantiated by replacing E in D5 with each of these levels. Similarly, we could offer another warrant principle schema whereby we took some level of utility to circumscribe the available class of propositions, and then picked one as superior on evidential grounds. Again, we would have as many different warrant principles here as we have distinguishable levels of utility.

The point of all of this is that an acceptable replacement for premiss (1) for the evidentialist is going to be very difficult to construct. The evidentialist will have to argue that either these various warrant principles are to be ordered in a certain fashion so that evidential considerations are usually more important, or he will have to argue that none of these warrant principles yield in the actual world that belief that God exists is reasonable. Of course, he could substitute this premiss in place of (1): it is reasonable for S to believe p when there are no warrant principles other than strictly evidential ones that yield that it is reasonable to believe p only if there is sufficient evidence for p. That premiss is obviously true, but then the altered second premiss would claim that there are no warrant principles other than strictly evidential ones that entail that it is reasonable to believe that God exists. As far as I know, no one defends that this altered second premiss is true; and a defense of it would be imposing at best.

My purpose in this section has been to show how complicated the debate about the rationality of believing that God exists is. The methodological constraints imposed by Flew are no longer acceptable once we have seen the unacceptability of the evidentialist=s premiss (1). The proper conclusion is that there is a burden of proof on both sides C neither side can sit back and wait for proposals from the other. With this in mind, I wish to defend in the last section of this paper that the proposition AGod exists@ is epistemically justified.

4. EPISTEMICALLY JUSTIFIED BELIEF THAT GOD EXISTS

In this section, I will propose a condition for justification of a somewhat Reidian character. The Reidian insight into the nature of confirmational support is that the connection between evidence and supported propositions is marked in a great number of cases by a disposition to believe that supported proposition in certain circumstances. For instance, Reid=s explanation of induction is characteristically dispositional.

It is undeniable, and indeed is acknowledged by all, that when we have found two things to have been constantly conjoined in the course of nature, the appearance of one of them is immediately followed by the conception and belief of the other.@¹¹

The point of this account of induction is that beliefs formed as a result of such general human tendencies are reasonable beliefs in the absence of specific grounds for doubt. With one additional clause, I think we can present this insight as a sufficient condition for justified belief. That condition is that it must be acceptable for the person in question to believe that the circumstances which naturally prompt belief in a certain proposition count as evidence for that proposition. Thus I propose:

RA: S is justified in believing p if (i) there is a general human tendency to believe p when e; (ii) it is acceptable for S that e is evidence for p~, (iii) S believes p on e without ground for doubt; and (iv) e.

In order to understand RA, we need some more definitions:

D6: S believes p on e without ground for doubt = _df (i) S bases his belief that p on e; and either (ii) nothing counts against p for S, or (iii) what counts against p for S is overridden for S.

D7: Nothing counts against p for S= _df no conjunction i of propositions which S believes or are acceptable for S is such that if i and the trees of justification for each conjunct of i were all that were evident for S, then not-p would have some presumption in its favor for S.

D8: What counts against p for S is overridden for S = _df (i) there is a conjunction i of propositions which (a) counts against p for S and (b) is larger than any other conjunction of propositions that counts against p for S; and (ii) there is a conjunction i +j of propositions not including p, e or anything that entails e which is such that (a) i+j and the trees of justification for the conjuncts of i+j all that were evident for S, p would be no worse than counterbalanced for S.

D9: T is a tree of justification for proposition p for S= _df, (i) if p is a foundational belief for S, then T=p; and (ii) if p is not a foundational belief for S, then T is composed of levels L₁, . . .L_n such that (a) L1 contains proposition e₁, . . ., e_m which are S=s epistemic grounds for p, (b) for any Lm, where m is less than n, L_m+1 contains propositions f₁, Y, _fmwhich are S=s epistemic grounds for the propositions g₁Y, g_n of L_m.¹²

I propose that RA is a sufficient condition for D2=, i.e., that if S has a justified belief that p by meeting the conditions stated in RA, then S has an epistemic warrant for p. This ought to be rather uncontroversial since, for the sense of justification that we are considering, S can be justified in believing p only if S has an epistemic warrant for p. Thus, the only question is whether RA is an adequate principle of justification. Although I do not have space here to defend that it is, I submit for inspection that RA is such an adequate principle.

We need then to examine the implications of RA for belief in God. First, are there conditions in which there is a natural human tendency to respond with belief in God? I think there are, and the pervasiveness of belief in a Diety supports such a claim. Even Kant and Hume, after criticizing versions of the teleological argument, agreed that there is something natural about such considerations prompting belief in God. The failure of certain arguments fails to cast doubt on the claim that there is a natural human tendency to respond with belief in God upon contemplating the majesty of the mountains, the starry heavens, or the beauty of a flower. Further, the intuitive force of such questions such as Aif God does not exist then where did the universe come from?@ although possibly inadequate as proofs for the existence of God, certainly reveal circumstances in which it is natural to believe that there is a God.

Thus, I think it is clear that there is a natural tendency to believe that God exists in certain circumstances. But is it acceptable to believe that the propositions which express these circumstances are evidence for the proposition AGod exists@? It seems to me that these propositions do provide evidence for the following reasons. If we give up an objectivist epistemology, as I argued earlier we should, then there is nothing non-subjective to appeal to define what counts as evidence for something else. But, if this is so, then the following argument seems to make acceptable that e is evidence for p: e seems to give good grounds for p, and since I have no ground for doubting that e gives good grounds for p, e is evidence for p. That the circumstances that naturally prompt belief that God exists seem to provide good grounds for the existence of God can be seen by noting that we cannot explain this natural prompting on pragmatic or prudential grounds. Persons do not believe that God exists in certain circumstances because it is advantageous in some sense for them to believe so. But that just means that it seems to them that those circumstances provide good grounds for the existence of God.

Are there grounds for doubting that these circumstances give good grounds for God=s existence? The sorts of considerations that evidentialists normally appeal to here are irrelevant C it is irrelevant here whether or not good deductive or inductive arguments can be constructed with the propositional expressions of these circumstances as premisses and the proposition AGod exists@ as the conclusion. But what else could count as grounds for doubt here? If we had a specification of some objective notion of evidence, we could easily specify what would count as grounds for doubt. But we have seen the irremediable problems with objectivist epistemology, and thus this option is not available to cast doubt on these circumstances providing evidence that God exists. The only available option is that there is some group of propositions believed by the theist or which are acceptable for the theist that gives grounds for doubting the proposition Athe circumstances that normally prompt belief that God exists are evidence that God exists.@ Some propositions which might provide grounds for doubt here are ones like: ANo one I know and trust thinks that these circumstances do not provide evidence that God exists, and they are not wrong in this case.@ But neither of these propositions are acceptable for most theists, and thus I conclude

that it is acceptable for most theists that the circumstances that naturally prompt belief that God exists are evidence for the existence of God.@¹³

Thus, the only clause of RA that we have not yet shown to be met (since (iv) is obviously true) is clause (iii) C that belief that God exists on this evidence is not subject to grounds for doubt. First, it is clear that clause (i) of D6 is satisfied since, however the basing relation is clarified,¹⁴ theists do base their belief that God exists on the circumstances that prompt such beliefs. Thus, we need only to show that either (a) nothing counts against the existence of God for theists, or (b) what counts against the existence of God for theists is overridden for theists. What I wish to do in the rest of this paper is to consider two suggestions that might be considered to violate both (a) and (b).

The problem of evil, considered as generating a contradiction between God=s nature and the existence of evil, I find an implausible candidate for generating grounds for doubt. First, it does seem that there are adequate responses to that problem,¹⁵ and thus the only way that the deductive problem of evil could be a ground for doubt is if theists believed that God=s nature was logically incompatible with evil. But certainly most theists do not believe this. Thus, I find that if the problem of evil is to cause ground for doubt, the only plausible form of the problem that might cause such problems is the inductive problem of evil.

It does seem that the conjunction of all the evil in the world does count against the existence of GodCat least I shall grant that this is so. Thus, we cannot dismiss the problem of evil by D7 since we are granting that something does count against God=s existence for the theist. But in order for the inductive problem of evil to be a ground for doubt for the theist, it has to be the case that what counts against the existence of God is not overridden for the theist. But there are such overriding considerations for the normal theist. Most theists would say something like Aevil works for good in the long run,@ or Aour perspective is too limited to judge.@ Now, of course, one may doubt the adequacy of such responses since one may have things that could be said to such theists that would make him see that these responses were not adequate. But I am not trying to solve the problem of evil at this point; I am only pointing out that if these further considerations were added to the fact of evil in the world, and these propositions and what justifies them were all that were evident for the theist, then the existence of God would be no worse than counterbalanced for the theist. To see that this is so, suppose that no one had any response to make to the considerations of limited perspective or evil as necessary to further good. Surely then such considerations would be adequate for answering the inductive problem of evil. But the point is that most theists are precisely in that epistemic situation C they believe that such considerations answer the problem of evil and are not aware of any responses which cast doubt on the adequacy of such responses. Thus, I conclude that the inductive problem of evil need not be and is not a ground for doubt for most theists.

Of course, this conclusion is perfectly compatible with the inductive problem of evil being a ground for doubt for theists that have read and understood the philosophical literature on the subject and do not know how to answer the problem. But here I am not trying to answer the problem of evil C I am only trying to show that the normal theist has a justified belief that God exists. To answer the problem of evil, and hence show that the philosophically sophisticated theist has a justified belief that God exists, would require answering any and all objections that such a theist is aware of; and certainly one cannot do all that in one paper. I shall settle then for showing that the normal theist has a justified belief that God exists; and I conclude that the inductive problem of evil does not cast doubt on this conclusion.

One might be unsympathetic to the above argument because of recent work done in epistemology which seems to suggest that there is a social aspect to knowledge, i.e., that one person=s knowledge depends not only on that person=s subjective epistemic situation, but also on what his or her general community happens to know.¹⁶ This question leads directly into the last possible ground for doubt that I wish to consider. That ground for doubt is the fact of disagreement about the claim that God exists.

The problem of disagreement is essentially this: for any person among a group of epistemic peers that disagree with him concerning the truth of p, doesn=t this disagreement give that person a reason to doubt p?¹⁷ I wish to claim is that it does not, i.e., I wish to claim that disagreement among epistemic peers does not count against God=s existence for the theist. Consider the proposition AS disagrees with me about the truth of p, and S is an epistemic peer of mine.@ If that proposition and whatever evidence I have for it were all that were evident for me, it would not follow that not-p had some presumption in its favor for me. But then, on D7, that proposition does not count against p for me. Further it is easy to see that a specification of what constitutes a ground for doubt ought not allow such a proposition alone to constitute grounds for doubt. The intuitive idea of a ground for doubt is something that counts against the truth of a proposition. But surely the simple fact that an epistemic peer of mine disagrees with me about a proposition does not provide evidence against that proposition. If it did, then if I assembled enough epistemic peers of mine, and all of them agreed that not-p, and I had no other information at all, the fact that they all agreed could give not-p a quite high epistemic status since each one=s belief that not-p counted in favor of the truth of not-p. But surely that is wrong. All we need to do is to imagine a member of a community that acquires justified beliefs by methods that are unavailable to other members of the community C say, for example, my grandmother trusting her rheumatism as a reliable indicator that it will rain. If some disagreement principle concerning epistemic peers were adequate, a community large enough could justify my grandmother in believing that it would not rain, even though she knew her rheumatism to be a perfectly reliable indicator concerning whether it would rain or not. But certainly the disagreement of these peers alone cannot defeat her justification. No matter how large the community is my grandmother still has a justified belief that it will rain because she knows her rheumatism is a reliable indicator of whether it will rain or not.

The lesson here is that if disagreement is important to justification, it cannot be disagreement among epistemic peers. Rather, the fact that you disagree with me about p is only relevant to the justification of my belief is I justifiably believe that you and I are in the same epistemic situation with respect to p (that you and I share roughly the same information about the truth of p), or that you are in a better epistemic situation than me with respect to p. But not even that is sufficient for your disagreement giving me grounds for doubt, for we know that people in general indulge in unjustified beliefs fairly regularly. Thus, the theist may justifiably believe that the non-theist shares his epistemic position with respect to God=s existence. The fact that the non-theist fails to believe that God exists does not give the theist a ground for doubting that God exists unless it is acceptable for the theist that the non-theist justifiably fails to believe that God exists. But for this to be acceptable for the theist, the non-theist would have to share his reasons with the theist for not believing that God exists. And the theist would have to be convinced that these reasons were good ones. But certainly most theists are not in that sort of situation.

The point of this discussion of disagreement then is this: disagreement does not give grounds for doubt unless one justifiably believes that the one who disagrees is in the same or better epistemic situation with respect to the proposition in question, and that it is acceptable for one that this person is justified in disagreeing. But clearly that does not obtain for most theists. Thus, I conclude that the problem of disagreement fails to cast doubt on the existence of God for most theists.

5. CONCLUSION

The conclusion of this paper is two-fold: first, the normal theist has a justified belief that God exists; and second, even if my argument for this first conclusion fails, it does not follow that belief in God is irrational. My evaluation of the evidentialist position has shown that there is no special burden of proof that rests on the defender of the existence of God as opposed to the agnostic. There would be such a burden of proof if the evidentialist position were true. However, the evidentialist position is false; and therefore it is still an open question, apart from my defense that beliefs in the existence of God are generally justified, whether it is rational to believe that God exists.

On the other hand, I think it is perfectly clear that it is reasonable to believe that God exists.

NOTES