CHAPTER 4

 

Proper Basicality and the Relevance of Reasons

 

 

 

 

 

 

 

 

            In the last chapter I articulated the two core claims about the proper basicality, immediate justification or warrant of theistic belief, what I have labeled the soft and hard proper basicality theses.[1]

Following that exposition I want now to begin to develop the plausible roles which propositional evidence might play in Plantinga’s religious epistemology, building toward a set of evidentialist requirements which are compatible with Plantinga’s soft and hard theses. The kind of propositional evidence I have in mind is primarily, though as we shall see not exclusively, the sort that is supplied by natural theology. In the course of the chapter I will present reasons for modifying the hard proper basicality thesis. After establishing a relatively uncontroversial pair of evidence functions in Part I, I will develop some less obvious ways that evidence may contribute to positive epistemic status. First, I will examine the epistemic significance of noetic structures in which theistic belief is (psychologically and epistemically) overdetermined by propositional evidence. Next I will do the same with respect to noetic structures in which theistic belief is partly based on propositional evidence (resulting in partly basic/nonbasic theistic belief). These noetic structures are then contrasted with those in which theistic belief is based solely on some immediate source(s) or ground(s). My evaluation will yield the following conclusion. A noetic structure in which theistic belief is epistemically overdetermined is epistemically superior to a noetic structure in which theistic belief is not epistemically overdetermined. Secondly, a noetic structure in which theistic belief has multiple grounds or sources of support is epistemically superior to a noetic structure in which theistic belief is based solely on experiential or propositional grounds. My argument will rest primarily on the diachronic dimension to positive epistemic status. Overdetermination and multiple sources of support each provides the appropriate cognitive resources to sustain a firm belief in God and to remain epistemically warranted in their theistic belief(s) in the face of putative defeaters to theistic belief. In Part III I draw some inferences from the analysis of the chapter for two kinds of evidentialist requirements.

 

I. Negative “Epistemic” and Positive “Apologetic” Evidentialism

 

A. Defeaters and Negative Evidentialism

 

            In the last chapter I indicated that Plantinga recognized that the conditions which confer justification on basic theistic belief confer a prima facie rather than ultima facie (or all things considered) justification. There are defeating conditions that can override the justification of basic theistic belief. Even Reformed epistemologists who do not construe justification in deontological terms (Alston 1989c, 1991c) make a distinction between prima facie and ultima facie (or all-things-considered) justification.[2] Typically, defeating conditions are spelled out in terms of two kinds of reasons which defeat the prima facie justification of a belief that p: (a) a reason for supposing that p is false (rebutting defeaters) and (b) a reason for supposing that the grounds of p are inadequate (undercutting defeaters).[3]

But the account of (undercutting and rebutting) defeaters need not be restricted to justification. More recently, Plantinga has extended his account to warrant defeaters. Unlike justification defeaters, warrant defeaters are construed in externalist terms. Justification defeat depends on what S (justifiably) believes or knows about the world. S is within his epistemic rights in believing that <it is raining outside> when S is appeared to rainly. But this justification can be defeated by good reason to believe that it is not raining outside (rebutter) or S’s justified belief that his visual apparatus is not in proper working order (undercutter). In the case of warrant defeat, what defeats warrant is some circumstance or state of affairs to which S has no special mental access. For instance, the warrant which S’s belief that <it is raining outside> has will be defeated if in fact S’s visual apparatus is not functioning properly (regardless of whether S is aware of this cognitive malfunction). Perhaps S suffers from a brain tumor that causes him to be appeared to rainly at certain times during the afternoon. Alternatively, S believes that <it is raining outside> upon being appeared to rainly, but unbeknownst to S his neighbor is spraying water over the fence on an overcast day. Things awry in one’s environment also defeat warrant, even though they are not cognitively accessible upon reflection. In these two examples, S may well be within his intellectual rights in believing that <it is raining outside>, but he does not know this proposition.

            There are also what Plantinga calls rationality defeaters. Rationality here is normative in nature, but not the normativity of epistemic dutifulness. It is the normativity of proper function that is involved in warrant. The rational person is the one whose cognitive faculties are functioning properly, subject to no dysfunction or malfunction). This conception of rationality leads to an account of defeaters in terms of a rational kinematics that spells out the proper or correct ways of changing beliefs in response to experience and the acquisition of new beliefs. On this way of looking at things what gets defeated is not justification or warrant but the belief in question, for defeaters are thought of as reasons (broadly speaking) for revising one’s noetic structure. As Plantinga explains: “. . .the basic idea is that when S acquires a defeater for B, she acquires a reason for modifying her noetic structure in a certain way. . .when she acquires a defeater for a belief, then if her cognitive faculties are functioning properly, further change in her noetic structure will occur, rationality (in the sense of proper function) requires a change in the rest of her noetic structure” (1994a, p. 31). The revision may be a matter of believing p less firmly given a defeater (partial defeat), or more radically, no longer believing p at all (complete defeat). To cast this in terms of the design plan: Given some person S’s noetic structure N in which there is some belief that p, given the conjunction of N and the p-relevant defeater, the design plan requires either the deletion of p from N or that S believe p less firmly. Moreover, the relevant defeater may be either another belief or an experience, and - according to Plantinga - in the case of beliefs need not have a lot of warrant to defeat another belief.

            A particularly important point concerns how we should understand the conditions under which a person has a defeater for some belief that p.  There is, of course, the externalist view just considered according to which the mere existence of certain objective features of a person’s cognitive situation suffice to defeat a belief, and where these conditions are not cognitively accessible items. The account of rationality defeaters, though, is broadly internalist, since defeaters here will be either experiences or other beliefs, both of which are cognitively accessible items. But, according to Plantinga, having a rationality defeater is internalist in even a stronger sense than this.  Suppose that S believes that p at time t₁, and then at t₂ S acquires two new beliefs, q and r.  Suppose further that q and r conjointly entail or make probable the negation of p, but S doesn’t see this connection between his newly acquired beliefs and his belief that p.  Does S have a defeater for his belief that p at t₂? Plantinga thinks that ordinarily S does not have a defeater for his belief under such conditions (1994a, pp. 36-37). Although we might speak loosely about the acquisition of some belief or experience as the acquisition of a defeater, the actual defeater is acquired as a result of S’s reflecting on the acquired experience or belief(s) and seeing that it sustains a negative evidential relation to S’s original belief.  In the case above, what constitutes S’s having a defeater for his belief that p is S’s having (i) the beliefs q and r and (ii) the belief that the conjunction of q and r entail (or make probable) the negation of p.[4] Must S’s belief about the connection between [q and r] and p be occurrent? Plantinga does not say. To the extent that beliefs need not be occurrent, there may be something like a distinction like something like dispositional defeaters and occurrent defeaters. I will be thinking primarily of occurrent defeaters.

            So a person S might acquire a defeater for theistic belief belief, say in the form of an evidential argument from evil. The acquisition of such a defeater calls for revision in a person’s noetic structure. A rational person who comes to believe that he has reasons for regarding his theistic belief as - all things considered - likely to be false (or who believes that it does not have an adequate ground) would not continue to hold the belief, certainly not with the same degree of firmness, unless of course he had a defeater-defeater for this initial defeater. Although any defeater will, as long as one has it, be an undefeated defeater, the defeater can be defeated, thereby allowing S rationally to hold his original belief once again. At t₀ S rationally believes that <there is a God> (Pt), and so at t₀ S’s noetic structure N contains the belief that Pt. At t₁ S acquires a (complete) defeater D for his belief that Pt, say as the result of reading one of William Rowe’s versions of the atheological argument from evil. To be more precise, S has acquired a rebutting defeater for his belief that Pt, i.e., a reason for supposing that it is unlikely that there is a God. If rational, S’s noetic structure N at t₁ will include D but not Pt. (If D had been a partial defeater, then N would have rationally included the belief Pt at t₀ and at t₁, but the belief that Pt in N at t₁ would have had a degree-of-firmness indexical with a value less than it had in N at t₀). Suppose now that at t₂ S acquires a defeater-defeater DD for the defeater D. Rationality now permits S’s holding the belief that Pt again (though perhaps with a less degree of firmness, if DD was a partial defeater-defeater). The acquisition of DD at t₂ (rationally) permits S to move from [N at t₁ including D minus Pt] to [N at t₂ including DD and Pt].

            According to Plantinga, defeater-defeaters come in two main forms. He distinguishes between intrinsic and extrinsic defeaters. In the former case, the belief itself that is the target of a potential defeater, by virtue of its degree of warrant, acts as a defeater-defeater against the potential defeater. In the latter case, the defeater-defeater would be reason to think that the potential defeater is false (or that its ground is inadequate), but where this reason is independent of the belief that has been targeted for defeat by the potential defeater. One of Plantinga’s claims, to be looked at shortly, is that basic theistic belief by virtue of its degree of nonpropositional warrant may act as an intrinsic defeater-defeater against putative defeaters of theistic belief. Whether this claim is true, and if so the range of its application, is a complicated matter I shall take up later in the chapter. Clearly, this does not necessarily alter the basicality of S’s theistic belief. Even if (a) S’s belief that Pt is properly basic, (b) some R constitutes a prima facie rebutting defeater to the belief that Pt, and (c) S’s belief that q is an undercutting defeater-defeater against R, S’s belief that Pt may still retain its status as properly basic. There need not be an alteration in the grounds of the belief that Pt simply because reasons have played a role in restoring a belief to a warranted status after an initial defeater (though, as I will argue shortly, sometimes this will happen). It may be that, though S has a reason against not-Pt, S nonetheless can continue to believe on the basis of religious experience. In fact, even if S had a rebutting defeater-defeater, which would be a reason to believe that not-Pt is false (i.e., that it is the case that God exists), S could still believe on nonpropositional grounds. Having good reasons for believing a proposition is compatible with believing it for some other reason. So the proper basicality of theistic belief is not compromised by negative epistemic evidentialism.

            So one form of evidentialism, compatible with Plantinga’s claims, would be of a negative sort.

 

[R1]     There are some people S* in some circumstance C, whose properly basic belief that Pt is defeated by reasons in the form of either rebutting or undercutting defeaters.

 

            Lastly, there is an important connection between rationality defeat and warrant defeat. A necessary condition for a belief B to have warrant is that one’s relevant cognitive faculties are functioning properly in producing or sustaining B.  This entails that one’s belief B will have warrant only if one’s defeater system is also functioning properly with respect to B.  There are (at least) two ways one’s defeater system may go wrong. S could acquire a complete defeater for B but still hold B, or S could acquire a partial defeater for B but still continue to hold B with the same degree of firmness as S did prior to acquiring the partial defeater. Recall my earlier point that a belief that “X is green” formed as a result of brain tumor which causes appeared-to-greenly states does not have warrant because there is some kind of cognitive malfunctioning going on. Such conditions defeating warrant, even if S knows nothing about his cognitive faculties malfunctioning. Similarly, if S acquires a defeater for some belief of his and there is not the appropriate noetic modification, the belief no longer has warrant for him since his defeater system is malfunctioning. In other terms, Plantinga’s externalist account of warrant and proper function suggests a necessary internalist condition for epistemic warrant: the no-defeater condition:

 

 

[ND]    S’s belief that Pt is warranted only if S does not believe that he has reasons for believing that his belief that Pt is likely to be false or that its grounds are inadequate (or the cognitive process generating and/or sustaining it is unreliable).[5]

 

B. Apologetic Evidentialism

 

            Although it is not necessary to cite an argument to defeat a prima facie defeater to theistic belief, engaging in such an activity would be one way to do so.  This gives rise to another closely related form of evidentialism, what we can call apologetic evidentialism. There have been two general apologetic methods exemplified by Christian apologists in the history of Christian theology.  The first, negative apologetics, consists in answering objections made against the Faith.  The second, positive apologetics, consists in the giving of reasons for Christian belief.[6] This distinction can be explicated in terms of rebutting and undercutting defeaters. The apologetic task is initiated by objections to theistic (or more specifically Christian) belief(s), where these objections take the form of either rebutting or undercutting defeaters. The distinction between negative and positive apologetics, then, turns on what sort of defeater-defeater is required to answer the objections to theistic belief. Positive apologetics involves the use of rebutting defeater-defeaters, whereas negative apologetics would employ undercutting defeater-defeaters.  Plantinga (1983a) anticipates an objection to proper basicality; namely, argument becomes irrelevant. He argues that if theistic belief is prima facie justified argument is not irrelevant, for (1) there can be defeaters for theistic belief and (2) since defeaters are prima facie, there can be defeater-defeaters for defeaters to theistic belief. But Plantinga (1986e) claims that, in the face of any defeater to the prima facie justification of theistic belief, only undercutting defeater-defeaters are required for the apologetic task.  The type of argument Plantinga has in mind, then, is negative apologetics - answering objections made against theistic belief by showing that the reasons for such objections are in some way faulty (i.e., that the argument underlying the objection is either invalid or unsound).

            Plantinga makes the connection to apologetics:

 

The distinction between undercutting and rebutting defeaters is of central importance to apologetics.  If the propriety of basic belief in God is threatened by defeaters, there are two ways to respond.  First, there is negative apologetics: the attempt to refute the arguments brought against theism (the atheological argument from evil, the claim that the conception of God is incoherent, and so on).  Second, there is positive apologetics: the attempt to develop arguments for the existence of God.  These are both important disciplines; but it is only the first, clearly enough, that is required to defeat those defeaters. (1986e, p. 313n)

 

 

            So we may say that a person S who accepts the soft thesis ([P1]- [P4]) or the hard thesis ([N3]) would engage in negative apologetics if and only if

 

 [A₁]     S believes[7] that there exists a prima facie defeater D for at least one Pt_i of any set of Pt₁, . . . ,Pt_n, and where D is either a rebutting or undercutting defeater,

 

and

 

 

 [A₂]     S believes that there exists a prima facie defeater-defeater DD for the defeater D, where DD is an undercutting defeater.

 

 

            As the practice of negative apologetics follows from either the soft or hard proper basicality theses if and only if they are conjoined with [A₁] and [A₂], positive apologetics follows from

 [E]      Given any person S, S is rational in believing some theistic proposition Pt only if S’s belief that Pt is based upon evidence in the form of adequate reasons, and where these reasons are not theologically biased or loaded.

 

 

if and only if [E] is conjoined with

 

 

 [A₃]     S believes that there is evidence for at least one Pt_i of any set of Pt₁, . . . ,Pt_n, and where the evidence has some prima facie plausibility of being adequate or of contributing to a total evidence-base which is adequate.

 

 

     Although Plantinga favours negative apologetics over positive apologetics, the former is not incompatible with the latter. If one holds to [E] and [A₃], then one would engage in positive apologetics, but the converse is not true.  One may engage in positive apologetics by holding solely to [A₃] (or by holding [A₃] and rejecting [E]). Positive apologetics then is not incompatible with holding to immediately justified (or warranted) theistic beliefs. Moreover, whether positive apologetics is required will depend on the strength of the initial objections to theistic belief and the strength of the undercutting defeater-defeaters one has at one's disposal.  For instance, some versions of the atheological argument from evil may carry a high degree of evidential force. A person’s stock of undercutting defeater-defeaters may not be sufficient in force to defeat such a defeater. But it may be that a version of the cosmological argument will do the trick. Plantinga is doubtful about the degree of warrant atheological arguments from evil actually have. “No atheologian," he says, "has given a successful or cogent way of working out or developing a probabilistic atheological argument from evil; and I believe that there are good reasons why it can’t be done” (1986e, p. 309). But surely the notion of probability here can be given various relevant interpretations. It is evidential probability which I take to be at issue here, roughly the kind of probability some proposition p has given some body of evidence (in the form of other propositions).  More specifically, it is the evidential probability for beings with limited logical capacities and knowledge as opposed to the logically omniscient (see chapter 1, section II.D). But we can distinguish between the evidential force a defeater (or defeater-defeater) has for human cognizers relative to (i) inductive standards that are correct (epistemic probability) and (ii) inductive standards that S regards as correct upon some degree of reflection or investigation, but that may in fact be incorrect (subjective probability). Plantinga’s general claim, then, that only undercutting defeater-defeaters are required for the apologetic task is not correct. Although in some instances defeater-defeaters of either sort may defeat a particular defeater, in other instances rebutting defeater-defeaters will be required. So I would also emphasize under apologetic evidentialism that both negative and positive apologetics (being compatible with [P1]- [P4], and [N3]) provide another possible role for reasons which would be compatible with the soft and hard theses. This role stipulates that reasons (rebutting or undercutting) are necessary for the activity of justifying theistic belief. Whether for any specific case they ought to be rebutting will depend on whether the strength of the undercutting defeater-defeater outweighs the strength of the initial objections (be they undercutting or rebutting) given the sorts of conditions to which I have alluded.

            As explained in chapter 1, one should be careful to distinguish this notion of justifying a belief (where this involves presenting arguments) from the concept of being justified (or warranted) in some belief (where this is a state a person is in vis-à-vis his belief). These two concepts are frequently conflated in both general epistemology and in the epistemology of religious belief.[8] Plantinga claims there are nonpropositional conditions sufficient for conferring warrant (and justification) upon theistic belief, but it does not follow from this that reasons can play no role in justifying theistic belief. Indeed justifying entails having a mediate justification. Therefore, since apologetics involves the activity of justifying theistic belief, it would be completely fallacious to reject positive apologetics simply because one advocates immediately justified (or warranted) theistic belief. It is frequently the case that what one cites in defense of a proposition is not the actual ground on which one personally holds the belief. So one might advocate an apologetic evidentialism without holding to an epistemological evidentialism.[9]

            So this gives us:

 

[R2]     Given any person S, if S shows (to some audience A) that the belief that Pt is rational, then (a) there must be propositional evidence which supports the belief that Pt, (b) the propositional evidence must be adduced, and (c) the propositional evidence must come to be believed by A (as well as some kind of connecting belief linking the evidence and  the proposition Pt).

 

 

 

II. Overdetermining and Partial Sustaining Grounds

 

 

 

            [R1] and [R2] are each what I would consider some very minimal sorts of roles for reasons. Both are compatible as far as I can see with the general soft thesis:

 

[ST]     There are some people S* in some circumstance C and at some time T who believe that Pt and S*'s belief that Pt is rational, even if there exists no available propositional evidence in support of the belief that Pt, and where “rational” = either (i) “deontologically justified” or (ii) “is the result of proper function,” and if (ii), then (subject to the satisfaction of the other warrant requirements) the belief that Pt is warranted and, if true, constitutes knowledge.

 

And the hard thesis:

 

 

[N3]     Typically, a noetic structure N₁ in which theistic belief is properly basic is epistemically superior to a noetic structure N₂ in which theistic belief is nonbasic, where epistemic superiority = the degree of warrant for (basic) theistic belief in N₁ is greater than the degree of warrant for (nonbasic) theistic belief in N₂, and the degree of theistic belief in N₁  is sufficient to transform true belief into knowledge (but not so in N₂).

 

 

            The conjunction of [R1] and [R2] I shall henceforth refer to as weak evidentialism. I want now to introduce some additional functions for reasons that I think will strengthen weak evidentialism.

 

A. Psychological and Epistemic Overdetermination

 

            Plantinga’s discussion of basic and nonbasic beliefs frequently gives the impression of being an exclusive disjunctive thesis, thereby excluding the scenario of a person’s belief being both basic and nonbasic (at the same time). Foundationalism of course need not be committed to such an exclusive disjunctive thesis. Audi (1993, pp. 262-265) explains the possibility of believing some proposition for more than one psychologically sufficient reason, in contrast to believing a proposition for exactly one such reason. We can then distinguish between a belief which is basic by being based on at least one psychologically sufficient immediate source and one that is basic on exactly one psychologically sufficient immediate source. Just as one may hold a belief on multiple psychologically sufficient reasons, one may have more than one psychologically sufficient immediate source for belief.[10] A case where (at least) two sources or grounds of belief each wholly generate or sustain a belief is a case of overdetermination, specifically psychological overdetermination. If S were (at some point) to lack either ground, S would still believe that p. The more general point is that a person may have more than one causally sufficient source or basis for a belief. It follows from this, though, that a belief may be wholly basic and wholly nonbasic at the same time. Suppose (i) belief B is based on an immediate source X and on a mediate source Y and (ii) X and Y each sufficiently causally sustain B.  If S came to lack either X or Y (though not both), S would still hold B. Call this kind of overdetermination heterogeneous mode overdetermination (i.e., the mode of belief is both immediate and mediate) in contrast to the previous cases of overdetermination which were homogeneous in mode.  What is true of beliefs in general would, ceteris paribus, hold for theistic belief in particular. It could be formed or held on the basis of both an immediate source (e.g., the ground of religious experience) and reasons, where each is sufficient for the belief’s formation and/or maintenance. We can then understand the hard thesis as referring in the negative to noetic structures in which theistic belief is wholly based on exclusively psychologically sufficient propositional evidence, but which nonetheless allows theistic belief to be heterogeneously mode overdetermined.

            Summarizing then:

 

[N4]                 Typically, a noetic structure N₁ in which [theistic belief is wholly properly basic on at least one psychologically sufficient source (or ground)][11] is epistemically superior to a noetic structure N₂ in which [theistic belief is based exclusively on propositional evidence], and where epistemic superiority = the degree of warrant for (basic) theistic belief in N₁ is greater than the degree of warrant for (nonbasic) theistic belief in N₂, and the degree of warrant for theistic belief in N₁ is sufficient to transform true belief into knowledge (but not so in N₂).

 

 

            It may also be the case that  immediate and mediate grounds are each severally sufficient to confer a high degree of warrant on S’s belief that Pt (a degree sufficient for transforming true belief into knowledge), so that in the absence of one of the grounds the belief that Pt remains warranted and so still constitutes knowledge. In this way the belief may also be epistemically overdetermined.[12] Overdetermination may have reference either to the psychological source of a belief or its epistemic status. On Plantinga’s account, as with all externalist theories, the causal origin or psychological source is closely related to the epistemic status of a belief. What generates or sustains a belief is partly responsible for conferring warrant on a belief. Of course, a belief might be partly or wholly psychologically sustained on some ground where this constitutes malfunction of some sort, or perhaps the modules of the design plan governing the sustenance of the belief on that particular ground is not aimed at truth (as is required for warrant). There could be psychological overdetermination without epistemic overdetermination. Whether there could also be actual epistemic overdetermination without actual psychological overdetermination depends on whether epistemic status depends (at least in part) on the psychological source of a belief. If it does, the possibility is precluded.

            But a person’s noetic structure may be such that his belief that Pt is wholly based on experiential grounds alone, but nonetheless within his noetic structure he has other beliefs b₁, . . . ,b_n which provide some degree of evidential support for the belief that Pt (i.e., he believes some propositions which in fact probabilistically support the belief that Pt).[13]  Here we may say that S has a reason r for believing that Pt and S does believe that Pt.  But r does not figure into the actual basis of the belief (i.e., r does not play a causal role in either generating or sustaining the belief, though it has that potential). This is another case of overdetermination, what we might call potential psychological overdetermination. If S lacked these other beliefs b₁, . . . ,b_n, then S would still hold the belief that Pt. If psychological source is necessary to epistemic status, we must regard this as only potential epistemic overdetermination. Nevertheless, since it is possible to recognize a source-relevant view of knowledge (or what transforms true belief into knowledge) and also the (at least propositional or impersonal) justificatory relevance of such beliefs, we can think in terms of actual evidential overdetermination even where overdetermination is only potentially psychological (or epistemic). Ex hypothesi, if S were to believe that Pt on the basis of b₁, . . . ,b_n S would believe that Pt on evidentially good grounds and perhaps S would not only be doxastically justified but know the relevant theistic proposition. On Plantinga’s theory we could think in terms of S’s belief that Pt being epistemically warranted if S were to believe on the basis of b₁, . . . ,b_n (assuming that the other conditions of warrant were satisfied).

            This yields something like:

 

[N5]                 A noetic structure N₁ in which [theistic belief is wholly properly basic and potentially or actually heterogeneously mode overdetermined] is epistemically equal to a noetic structure N₂ in which [theistic belief is based solely on immediate sources].[14]

 

 

B. From Overdetermination to Partial Sustaining Grounds

 

            One of the advantages which arises from potential psychological overdetermination is that the reasons one has which evidentially support theistic belief might be cited by the person in defending his belief in God, thereby putting the person in the position of carrying out the project of positive apologetics. Assuming of course that the reasons are the sort of thing the person has cognitive access to, they could in principle be cited in defense of belief in God (even though the person does not himself believe on the basis of those reasons). I do not mean to suggest that the activity of citing r in support of some proposition p one happens to believe confers justification on S’s belief that p, nor would this imply that S is justified in his belief that p.  In chapter 1 I noted that a belief’s being justified should sharply be distinguished from the activity of justifying a belief by citing reasons for it, specifically that the conditions for the latter are both too strong and too weak to be conditions for the former. Nonetheless, this apologetic advantage suggests a more directly epistemically relevant point.

            One’s having (good) reasons for belief in God might make a contribution to one’s own epistemic state. How this contribution is made will depend on whether one takes a belief's actual psychological source (i.e., what causally sustains it) to be necessary for justification, for it is possible to cite as reasons for a belief what does not factor causally in a person's actual ground of belief. I may cite as my reason for believing that the San Francisco 49ers will win the 1996 NFC American Football Championship against the Dallas Cowboys a set of relevant statistics about the offensive (and defensive) strengths of the 49ers which considerably outweigh Dallas’s. Say I cite this evidence at mid-season when the 49ers are doing well. Then they take a turn for the worst late in the season - they lose their quarterback due to injury and their offense significantly suffers. By the season’s end Dallas’s offense ranks first in the league. Suppose, though, that I still believe that the 49ers will win the NFC championship. Apart from the cited reasons being merely ad hominem in force, a very likely explanation for this very common sports-fan phenomenon is that my belief that the 49ers will win the Championship game was (and is) based on reasons other than what I cited in defense of my belief. After all, when the evidence is no longer relevant, in fact collapses totally, I still hold the belief.[15] It may be that my belief is being sustained by an adequate ground (e.g., my belief that defense wins games), and in that case there would seem to be no problem in regarding the belief as justified. Doubts arise though when the sustaining ground is inadequate (e.g., where it is my own bias, a hunch, or wishful thinking), in which case many would deny that the belief is justified or constitutes knowledge.[16]

            What this case exemplifies is how one may express a reason for believing some proposition by citing (adequate) evidence in support of the proposition, though this reason is not causally operative as a reason for which a person actually believes the proposition. The case exemplifies what Audi  (1993, pp. 227-232) calls belief rationalization.  As suggested above, belief rationalization does not entail that S’s belief that p is justified, nor that it is not justified. What it does imply is that there is a kind of justification available for some proposition or belief-type. And for this reason we would want to say that epistemically it is a good thing to have good reasons (or potential grounds) for a belief. This is a desirable epistemic state, but it is one which we would at least want to distinguish from the sort of justification which has a causal component.  So we might distinguish between the justification a person may have for a proposition (or belief-type) and the justification a person has for their actual belief (and where the latter involves a causal component). Call the former propositional (or impersonal) justification; the latter doxastic (or personal) justification. In the case where someone believes that Pt on the basis of some experiential ground, but cites,[17] say the cosmological argument, as a reason for believing in God, they have and are offering a propositional justification.[18]  If the experiential ground is adequate, then they also have a personal justification. This way of cutting the cake approaches a more pluralistic account of justification. Here possessing (good) reasons is recognized as a valuable epistemic state which can, in some sense, imply justification of a sort (even if it falls short of the source-relevant concept thought to be necessary for knowledge on externalist accounts).

            One objection to this line of reasoning is as follows. Having good reasons for a belief is a good thing epistemically only if one recognizes that the reasons are good ones. But if one recognizes that some reasons are good reasons for believing that p, those reasons will operate causally in one’s holding the belief that p.[19]   But I do not find this line of reasoning very satisfying for a few reasons.  First, on an externalist position what makes any reason or set of reasons good from the epistemic point of view is that they are reliable or truth-conducive sources of belief, not simply (if at all) that we believe that such reasons are reliable (as we may be wrong). Beliefs formed on the basis of such good reasons are likely to be true (where “likelihood” is understood in any one of the externalist senses of probability discussed in chapter 1, section II.C and D) and for that reason are epistemically relevant and a good thing. Secondly, suppose we drop externalism. As explained in chapter 2 (section II.A) we can understand having good reasons in two senses. One’s having good reasons for believing that p may mean that S holds other beliefs that, together with S’s inductive standards, make p probable. But for whatever reason S might not see the connection between these other beliefs, his inductive standards, and the belief that p. Alternatively, S’s having good reasons could be analyzed like “having defeaters” was earlier in the chapter. S has good reasons for believing p just if S has evidence that S recognizes makes p probable given S’s own inductive standards. I think that usually our second case will entail that S holds p at least in part because of the evidence in question. But I don’t see why this must always be the case, for the same reason that I can recognize a good reason for doing some action but am personally motivated to do it for reasons other than (not merely in addition to) such reasons. More importantly, I think that having good reasons in the first sense is a good thing since it likely creates a disposition realized upon reflection (i) to come to see that such reasons are good (at least on non-externalist assumptions), (ii) to hold the belief that p (if one doesn’t already), and (iii) to hold the belief (at least in part) on the grounds of such reasons when upon reflection one sees that such reasons are good ones. Here we have both one’s coming to believe that one has good reasons for a belief and the potential of such good reasons playing a causal role in sustaining the belief. These points are, I take it, very important and bear significantly on why potential psychological overdetermination or merely having good reasons for belief in God (and so having a propositional justification) is an epistemically valuable state.

            To elaborate this further, it is important to distinguish between generative and sustaining grounds of belief. The is-based-on locution is time relative and so may refer to a person’s noetic structure at the time of belief formation t₁ or some later time t_n+1 when the belief is being sustained. It is incorrect to assume the rather implausible thesis that what justifies or confers warrant on a person’s belief at the time of the belief’s acquisition will be what justifies or confers warrant on the belief at all times during the belief’s maintenance. This is especially true if we focus solely on question of the causal source of a belief (thought by many to be necessary to its epistemic status), for different factors may be involved in the causal generation and sustenance of a belief over time. As Moser points out: “It might be that one causal factor is responsible for one’s coming to believe a proposition, and second causal factor is responsible for one’s maintaining the belief, because without the second factor, a third causal factor would extinguish the state of believing” (1989, p. 17).  Equally, the second causal factor may be necessary for sustaining the state of believing some proposition with a high degree of firmness, if in the absence of this factor some third causal factor might, even if not extinguish the state of believing, reduce the degree to which a person believes some proposition. Now the second causal factor could be a reason, an experience, or some conjunction of the two. I only want to consider a case in which the reasons one may have for theistic belief at time t₁ are combined with experiential grounds at time t_n+1 and so contribute to the psychological sustenance and epistemic warrant of theistic belief at that later time.

            Let’s apply this to theistic belief in the context of Plantinga’s theory of warrant and proper function. Suppose John firmly believes at t₁ that <God is present> on the basis of some experential ground e. At t₂ John reads one of the versions of the evidential atheological argument from evil and is strongly persuaded by the force of the argument. John now has a rebutting defeater for his theistic belief <God is present>, namely a reason (which he regards with some significant degree of plausibility) to believe that God does not exist, and so that <God is present> is false. The situation is such that an experiential ground is responsible for John’s coming to hold a belief in God’s presence, but another causal factor (the relevant rebutting defeater) provides John with a reason for revising his noetic structure. It could do this in two ways, depending on the actual force of the atheological argument. It could be a complete defeater, and so extinguish John’s state of believing that God exists and so extinguish the state of John’s believing that <God is present>, where these beliefs are held on the experiential ground e. Alternatively, the defeater could be a partial defeater by merely reducing John’s degree of belief. On this latter construal, even if John’s belief remains warranted, it may lack the degree of warrant necessary for knoweldge. Now John might acquire an undercutting defeater-defeater for the original rebutting defeater, namely a reason to regard the atheological argument from evil as unsound or invalid. Alternatively, if John had reasons to believe that God exists (independent of the experiential ground) he would have a rebutting defeater-defeater. Whether one or the other is required in the actual situation would, as I pointed out earlier, be determined by the strength of the particular atheological argument and John’s intellectual resources and view of the weight of the evidence. Theistic arguments might happen to strike John as stronger than arguments that undercut the atheological argument. Having a defeater-defeater reason is an example of a second causal condition essential to the maintenance of John’s belief at some point beyond the experiential grounding of John’s belief at the time of its formation, given the atheological argument as a third causal factor that would otherwise extinguish or reduce John’s state of believing. In this particular situation, I suggest one of two possible sustaining roles for a rebutting defeater-defeater. (1) At t₃ (after the acquisition of the initial defeater) John’s belief that <God is present> is partly sustained by reasons to believe that God exists; for not only will the latter function as a rebutting defeater-defeater, but it also lends support to the veridical nature of the experiential ground.[20] Alternatively, (2) although the conjunction of the experiential ground and the defeater does not entail that S does not believe that <God is present>, it may entail that S no longer believes that <God is present> with the same degree of firmness. On Plantinga’s account this latter possibility has implications for the degree of warrant John’s belief has and his knowing that <God is present>.  John’s propositional evidence for <God exists> would be necessary for John’s remaining sufficiently warranted for knowing that <God is present> if (a) the degree of warrant the belief <God is present> has for John is significantly reduced because John holds the belief with a significantly less degree of firmness after acquiring a defeater for the belief <God is present> and (b) the degree of John’s belief that <God is present> is significantly raised by propositional evidence for <God exists>.

            Cases of potential psychological overdetermination, then, are epistemically relevant because they provide one with resources for remaining epistemically warranted in one’s theistic belief in the face of defeaters. In the case of actual psychological overdetermination this is very clear, but it is just as relevant when overdetermination is, in the first instance, merely potential. Upon reflection a person may come to believe in God (or the specific theistic proposition) on the partial basis of both experiential and propositional grounds, and where the degree of warrant on both grounds is greater than on either of the grounds alone. In such a case it may be that reasons for holding a theistic proposition which were up to that point merely potential grounds become part of the actual grounds for the person's theistic belief.  Here the belief is partly immediately and partly mediately warranted, but its status as knowledge may depend on the conjunction of both sources. In other words, at time t₁ S’s noetic structure is propositionally overdetermined with respect to theistic belief. At t₂ S acquires a partial defeater for theistic belief, with the consequence that S now holds his theistic belief with a less degree of firmness.[21] Consequently, his theistic belief now has little by way of warrant for him. He no longer knows the relevant theistic proposition. I am claiming that the firmness of S’s theistic belief may be shored up to the requisite level for epistemic warrant by his at holding theistic belief at t₃ on the basis of the original experiential ground and those reasons which were at t₁ merely overdetermining reasons. Those reasons become necessary at t₃ for the causal sustenance of S’s firm belief that Pt. Indeed, owing to the different degree of influence multiple sources of belief may have, S might even believe the theistic proposition in question more for the reasons than on account of the experiential grounds.

 

III. Objections and Further Refinements

 

A. Defeaters and the Causal Grounds of Theistic Belief

 

            Plantinga has a number of salient points that might be thought to constitute objections to the argument I have just presented. First, according to Plantinga, since basic theistic belief can by virtue of its own degree of warrant function as an intrinsic defeater-defeater (and thus have more by way of warrant than a putative defeater), there is no need to fall back on the kind of propositional evidence that characterizes natural theology. The idea of properly basic theistic beliefs being intrinsic defeater-defeaters seems plausible enough. It would seem, to use Plantinga’s example, that Moses’ belief that God was speaking to him from the Burning Bush would have more by way of warrant than a defeater in the form of either an atheological argument from evil or the projective explanation of theistic belief advocated by Feuerbach or Freud.  But Moses’ situation and others like it are clearly extraordinary. Few of us would seem to have theistic beliefs that enjoy that degree of warrant. The range of degree of warrant which may plausibly be attributed to many (or most) cases of properly basic theistic belief (even if sufficient for knowledge) will preclude the possibility of that belief being an intrinsic defeater-defeater for at least some people under many circumstances. If we use degree of belief to measure the degree of warrant (even in part), the fact that the firmness with which some people hold theistic beliefs is often (significantly) reduced by putative defeaters shows that on these occasions even if basic theistic belief is an intrinsic defeater-defeater it is so only for a limited range of potential defeaters.  The point is that a properly basic theistic belief will be an intrinsic defeater-defeater for some belief B only if B has less by way of warrant than theistic belief, but as the present account suggests it is rather dubious to generalize unqualifiedly about properly basic theistic beliefs being intrinsic defeater-defeaters. In fact, the notion of intrinsic defeater-defeater is not even confined to theistic belief, but every belief can serve as an intrinsic defeater-defeater against some (weak enough) defeater. So the notion that theistic belief can be an intrinsic defeater-defeater is trivially true. The determinants of whether theistic belief is, on some particular occasion, an intrinsic defeater-defeater against some putative defeater are the degree of belief and warrant possessed by basic theistic belief on that particular occasion and the amount of warrant which the defeater has on that occasion. In other terms, the determinants turn out to involve matters like S’s view of the prior probability of the theistic proposition and S’s view of how improbable the defeater makes the theistic proposition given S’s own inductive standards.[22] None of these are so fixed as to permit a generalization that would exclude the necessity of extrinsic defeater-defeaters for theistic belief to be epistemically warranted on many occasions.[23]

            Moreover, Plantinga’s emphasis on undercutting defeater-defeaters suggests that there is more to be said for defeater-defeaters than is captured by basic theistic belief as an intrinsic defeater-defeater. But Plantinga has argued that (i) given any defeater (rebutting or undercutting) for theistic belief, only an undercutting defeater-defeater is required to defeat the original defeater (1986e, p. 313n), (ii) the presence of an undercutting defeater-defeater need not, perhaps should not, be the basis for theistic belief (1983a, pp. 83-86), and (iii) even if defeater-defeater beliefs are required for a person to remain rational and sufficiently warranted in their theistic belief, that belief remains basic. The defeater-defeater may be part of the warrant conferring circumstances, but it need not factor into the causal basis of the belief (1983a, p. 86; 1993b, p. 185). I have already argued that (i) is false. What kind of defeater-defeater is required on a particular occasion depends not only on the sort of original defeater but how strong that defeater is for S, to what extent it reduces his degree of belief. As for (ii), yes; but we should not preclude even an undercutting defeater-defeater from being a partial basis for theistic belief. Suppose John locates a flaw Feuerbach’s projective theory of theistic belief (and so has an undercutting defeater-defeater). Even if it is odd (and perhaps irrational) for John to believe that <God is present> solely on the grounds that Feuerbach’s argument is unsound, it may be that John’s belief that <God is present> is partly causally sustained by the experiential conditions and partly by his firm belief that Feuerbach’s argument is flawed. Perhaps John concludes (as Quinn 1993 suggests) that theistic projection mechanisms are secondary causes which God has intended to serve as generators of theistic belief, and so Feuerbach’s argument is not incompatible with the existence of God. Perhaps John holds his theistic belief more on the basis of experiential conditions than because he believes that the atheological argument is flawed, but it does seem that some degree of partial causal sustenance is involved. Just as John’s coming to hold his theistic belief with a less degree of firmness at t₂ was explained by John’s acquisition of a defeater, John’s acquisition of an undercutting defeater-defeater (and so John’s losing his belief in the force of the original defeater) now provides a partial explanation for his holding theistic belief with the degree of firmness which he does at t₃. Moreover, if (i) is false and rebutting defeater-defeaters are sometimes required, (ii) will simply not be relevant in such instances.  The unsoundness of the projective arguments of Feuerbach and Freud may not be a sufficient reason to believe that God exists, but a reason to believe that the conclusion of their arguments is false would be a rational basis for belief in God, for that would be equivalent to a reason to believe that God exists. As for (iii), I think we can only draw the conclusion that a person’s theistic belief may remain wholly basic even if it is required that a person have a defeater-defeater (given a putative defeater to theistic belief). In some instances, it will only remain partly basic.

            I have reached a conclusion here has been suggested by Plantinga himself. In 1991 he accepts the possibility that reasons may contribute to the warrant of theistic belief in an epistemically relevant manner.

 

. . .an essential feature of the degree of warrant a belief has for me is the strength with which I hold the belief in question. . . .Perhaps my belief in God, while accepted in a basic way, isn’t firm and unwavering; perhaps it isn’t as firm as my belief in other minds. Then perhaps good theistic arguments could play the role of confirming and strengthening my belief in God, and in that way they might increase the degree of warrant belief in God has for me. Indeed, such arguments might increase the degree of warrant of that belief in such a way as to nudge it over the boundary separating knowledge from mere true belief. . . .(1991, pp. 311-12)

 

           

            So in addition to noetic structures in which properly basic theistic beliefs is overdetermined, there are those in which theistic belief is partly basic and partly nonbasic - a possibility which in chapter 1 I argued was quite consistent with foundationalism.  So perhaps the best way to think of the Reformed position, and the hard proper basicality thesis in particular, is as drawing a distinction between noetic structures in which theistic belief is at least partly causally sustained by some immediate source(s) and one in which theistic belief is wholly causally sustained by some mediate source(s).

            So we actually end up with something on the order of:

 

[N6]     Typically, a noetic structure N₁ in which theistic belief is based partly on immediate and mediate sources (and so is partly basic and partially nonbasic) is epistemically equal to a noetic structure N₂ in which theistic belief is wholly based on at least one psychologically sufficient immediate source.

            The conjunction of [N6] and [N4] entails:

 

 

            [N7]                 Typically, a noetic structure N₁ in which theistic belief is based partly on immediate and mediate sources (and so is partly basic and partially nonbasic) is epistemically superior to a noetic structure N₂ in which theistic belief is wholly based soley on some mediate source(s).

 

B. Diachronic and Synchronic Positive Epistemic Status

 

            But there is more to the story.  Given that we can evaluate a noetic structure in terms of the cognitive aim of remaining epistemically warranted in some theistic belief, it could also be argued that a noetic structure that is propositionally overdetermined with respect to theistic belief is actually epistemically superior to a noetic structure that lacks overdetermination.[24]   Since there is a distinction between the conditions required for S’s coming to believe that Pt with epistemic warrant and the conditions required for S’s remaining warranted in the belief that Pt, we can I propose distinguish between synchronic and diachronic epistemic status. The former will be the warrant a belief has at some specific time (when it is generated or being sustained). The latter will be a kind of warrant-history for a belief, its epistemic status charted through time. Given that distinction, it seems that overdetermined noetic structures will possess a diachronic epistemic superiority over their nonoverdetermined counterparts.  More specifically, given the function which propositional evidence may play in a person’s remaining epistemically warranted in the face of defeaters, we get something like:

 

            [N8]                 A noetic strucure N₁ in which [properly basic theistic belief is heterogeneously mode overdetermined] is diachronically epistemically superior to a noetic structure N₂ in which [properly basic theistic belief is not heterogeneously mode overdetermined][25] in all worlds W* in which (a) the history of both N₁ and N₂ each includes defeating conditions for theistic belief and (b) properly basic theistic belief is not an intrinsic defeater-defeater in W* at any time.

 

 

            The account also suggests a way to think of the synchronic epistemic superiority of noetic structures in which theistic belief is overdetermined or partly sustained by propositional evidence.

 

[N9]     A noetic structure N₁ in which [properly basic theistic belief is heterogeneously mode overdetermined] is synchronically epistemically superior to a noetic structure N₂ in which [theistic belief is based solely on some immediate sources and lacks all heterogeneous mode overdetermination] if and only if (a) S believes that there exists a defeater D at t_n for theistic belief in N₁ and N₂, (b) basic theistic belief in N₂ (and N₁) is not an intrinsic defeater-defeater for D, and (c) D is such that theistic belief in N₂ is either extinguished or has a less degree of firmness than theistic belief in N₁, and where the degree of firmness of theistic belief in N₁ is epistemically sufficient.

 

 

            The idea that properly basic theistic belief is typically epistemically superior to nonbasic theistic belief must be substantially qualified. Perhaps the claim rings true in a world minus defeating conditions for theistic belief, but such a world, though logically possible, is not the actual one. (I shall have more to say about this in relation to the design plan in the next chapter). It would seem that defeating conditions provide us with many scenarios in which a merely properly basic theistic belief is epistemically inferior to its overdetermined and partly nonbasic counterparts. Such seems to be the case given (i) Plantinga’s contention that degree of warrant is - roughly speaking - proportional to degree of belief and (ii) some people find themselves in circumstances in which belief in God is sustained or strengthened by propositional evidence.  This is not to say that solely properly basic theistic belief is always epistemically inferior, only that [N9] gives us a statement of the synchronic epistemic superiority of partly nonbasic theistic belief under certain conditions.

            I propose then two additional evidential requirements to be added to [R1] and [R2].

            There is first what I want to call the sustaining evidentialist requirement:

 

[R3]     For some people S* at certain times t and under certain circumstances C, S* remain epistemically warranted in believing that Pt only if their belief that Pt is based at least partly on propositional evidence and where (a) C includes (i)  defeating conditions D and (ii) S*s belief that Pt is not an intrinsic defeater-defeater for D and (b) propositional evidence = either undercutting or rebutting defeater-defeaters.

 

            A parallel requirement holds for the conditions required for some people to come to hold an epistemically warranted theistic belief, what I will call the formative evidentialist requirement.

 

[R4]     For some people S* at a certain time t₁ and under certain circumstances C,  S* belief that Pt is epistemically warranted only if their belief that Pt is based at least partly on propositional evidence and where (a) t₁ = the time of belief formation, (b) C includes (i) defeating conditions D at t_n-1 and (ii) S*'s belief that Pt would not be an intrinsic defeater-defeater against D, and (c) propositional evidence = either undercutting or rebutting defeater-defeaters.

 

 

IV. Conclusion

 

 

 

 

            In this chapter I have sought to lay out the first planks in a case for an epistemically adequate version of evidentialism which is compatible with Plantinga’s religious epistemology. Matters of overdetermination and multiple partial supporting grounds, as well as the distinction between belief-formation and sustenance, have significant epistemic implications for accounts of human noetic structures and the extent to which reasons may play a hand in the game of warranted theistic belief. My conclusion at this stage is that the soft and hard Plantinga theses are compatible with overdetermination and partial basicality as the result of multiple sources of belief. Secondly, the soft and hard Plantinga theses are also compatible with (1) the diachronic epistemic superiority of noetic structures in which theistic belief is propositionally overdetermined and (2) the restricted synchronic epistemic superiority of noetic structures in which theistic belief is at least partly based on propositional evidence (in the form of either undercutting or rebutting defeater-defeaters).  Therefore, the evidentialist requirements [R3] and [R4] deduced from the account of noetic structures and defeaters are compatible with Plantinga’s soft and hard theses. The conjunction of [R1]-[R4] makes a substantial contribution toward what I will be calling modest evidentialism.

            Further questions remain. One of these questions involves the extent to which propositional evidence can be a sufficient (generative or sustaining) ground for epistemically warranted theistic beliefs. Answering this requires developing the concepts of warrant, a theistic design plan, and the noetic effects of sin. Such a development will also provide confirmation of the sorts of conclusions I have reached in this chapter and thus contribute to a more thorough case for the compatibility of modest evidentialism and Plantinga’s religious epistemology.