CHAPTER 2

 

Theistic Evidentialism

 

 

 

 

 

 

 

The objective of this chapter is to offer a critical exposition of an epistemology of religious belief that has dominated Western philosophy since the 17th century - so-called theistic evidentialism. My account aims at explicating this view of the rationality of belief in God in the light of the critique of it presented by Alvin Plantinga. In Part I I explicate the evidentialist challenge and objection to theistic belief, as well as Plantinga's critique of evidentialism and the epistemological thesis of classical foundationalism on which it is traditionally based. Part II will involve a discussion of what I take to be some necessary, though often overlooked, variants on the evidentialist challenge. In the final part I re-examine theistic evidentialism and relocate the evidentialist challenge and Plantinga’s critique in preparation for the arguments I will develop later in the thesis. Since Plantinga's religious epistemology consists largely of a set of claims that are (at least prima facie) negations of the evidentialist position, the careful spelling out the latter (in the present chapter) will provide a useful framework for articulating Plantinga’s own positive epistemological position and developing the prospects for a compatibility theory.

 

 

                                       I. The Evidentialist Objection to Theistic Belief

 

A. Evidentialism and Classical Foundationalism

 

“Scarcely anything has been more characteristic of the modern Western intellectual,” wrote Nicholas Wolterstorff, “than the conviction that unless one has good reasons for one’s theistic beliefs, one ought to give them up” (1986, p. 38). According to Wolterstorff and Plantinga, the Enlightenment initiated what has come to be a long-standing tradition in Western philosophy.[1] This is a distinct way of looking at the rationality of theistic and religious belief according to which belief in God is rational only if it is based on or appropriately related to reasons which provide adequate evidential support for it. The view perpetuated by this tradition - the so-called tradition of evidentialism - embodies what has been labeled the evidentialist challenge to theistic belief.

In the words of Wolterstorff:

 

 

[T]his challenge can be thought of as consisting of two claims: first, if it is not rational to accept some proposition about God then one ought not accept it; and second, it is not rational to accept propositions about God unless one does so on the basis of others of one’s beliefs which provide adequate evidence for them, and with a firmness not exceeding that warranted by the strength of the evidence. (1983a, p. 136)

 

 

John Locke, perhaps the fons et origo of the evidentialist tradition, wrote:

 

 

[H]owever faith be opposed to reason, faith is nothing but a firm assent of the mind; which if it be regulated, as is our duty, cannot be afforded to anything but upon good reason, and so cannot be opposite to it. He that believes, without having any reason for believing, may be in love with his fancies; but neither seeks truth as he ought, nor pays obedience to his Maker, who would have him use these discerning faculties he has given him, to keep him out of mistake and error. . . . (1975, bk. IV, 17)

 

Living in a time of great religious diversity (even if Christian) with competing claims to divine revelation, Locke was concerned to articulate principles for the regulation of our assent to religious propositions. Locke’s stand on religious epistemology is of course a consequence of his more general epistemology according to which we should only believe what is probable for us with respect to what we know. For Locke, knowledge is not something we can regulate since we do not voluntarily give assent to objects of knowledge. What we know is certain for us (e.g., what is self-evident and truths about our own mental states, and perhaps also what may be deduced therefrom). What must be regulated is the nonepistemic cognitive state of opinion or belief. We have an intellectual duty to believe only what is likely to be true (i.e., what is probable) relative to what is certain for us, and to proportion our degree of belief to the degree to which the propositional object of belief is rendered probable by our evidence. The same is true, and most relevant, in religious belief. Locke thought that the existence of God could be known (by demonstration). Nevertheless, there are still many religious beliefs that cannot be known stricto sensu and so evidence will be needed if we are to give rational assent to them. It is the task of reason to establish the grounds for believing, for any putative proposition based on or derived from an alleged divine revelation, that this particular proposition is indeed a revelation from God. It is a kind of higher-order proposition of the form [it is probable that <p is revealed by God>] for which we must have rational grounds if we are rationally to give our assent to p.  So with respect to any noncertain religious proposition PR, a person S is rational in believing PR only if (i) it is probable that <PR is revealed by God> with respect to what is certain for us and (ii) S believes PR with a degree of firmness proportional to the degree to which <PR is revealed by God> is rendered probable by what is certain for us.[2]

After Locke (perhaps beginning with Hume), Locke’s evidentialism gets applied to theistic belief. The evidentialist requirement becomes a requirement imposed on theistic belief. Locke’s central question was “How do we know X is revealed by God?”. For his intellectual descendants the first question is “How do we know that there even is a God?” In the 19th century W.K. Clifford gave a succinct statement of the evidentialist position: “To sum up: it is wrong always, everywhere, and for anyone to believe anything upon insufficient evidence” (1879, p. 183). Similarly this century Brand Blanshard maintained: “everywhere and always belief has an ethical aspect. There is such a thing as a general ethics of the intellect. The main principle of that ethic I hold to be the same inside and outside religion. This principle is simple and sweeping: Equate your assent to the evidence” (1974, p. 401). And Bertrand Russell wrote: “Give to any hypothesis which is worth your while to consider just that degree of credence which the evidence warrants” (1945, p. 816).

Anthony Flew writes with reference to theistic belief:

 

 

It is by reference to this inescapable demand for grounds that the presumption of atheism is justified. If it is to be established that there is a God, then we have to have good grounds for believing that this is indeed so. Until or unless some such grounds are produced we have literally no reason at all for believing; and in that situation the only reasonable posture must be that of either the negative atheist or the agnostic. (1976, p. 22)

 

 

The evidentialist objection to theistic belief arises when the evidentialist challenge is conjoined with the further claim that the challenge has not been met, that there is not, or at least some individual or group does not have, the required evidence for theistic belief.

The basic evidentialist package, then, is comprised of two claims:

 

 

 (I)       It is irrational or unreasonable to accept theistic belief in the absence of sufficient evidence or reasons.

 

(II)       We have no evidence or at any rate not sufficient evidence for the proposition that God exists. (Plantinga 1983a, p. 27)

 

The central principle of theistic evidentialism involves the imposition of what we might call an evidentialist requirement for theistic belief. That requirement may be stated in the following theistic, evidentialist principle:

 

 [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[3] in the form of adequate reasons, and where these reasons are not theologically biased or loaded.[4]

 

 

Although one might hold to [E] because one holds that there is an evidentialist requirement for all beliefs (and so advocate a coherentist-based evidentialism), Alvin Plantinga takes the evidentialist challenge to be based on foundationalism, specifically classical foundationalism (hereafter CF).[5] A crucial part of the foundationalist position is the specification of what beliefs are appropriate as basic or foundational in a rational noetic structure. Which beliefs will be rational for a person to hold other than on the basis of other (rational) beliefs? In short, which beliefs are properly basic - basic and yet rational? As noted in chapter 1, CF is a species of foundationalism that holds to rather narrow criteria of proper basicality. As Plantinga explains it, CF maintains that a basic belief is properly basic if and only if it is either (a) evident to the senses (e.g., it is raining outside), (b) self-evident (e.g., 2 + 2 = 4), or (c) about one’s immediate experience (e.g. I am being appeared to redly.)[6]

Thus:

 

 

[CF]     A proposition p is properly basic for a person S if and only if p is either self-evident to S or about S’s immediate experience or evident to the senses.

 

 

To be more precise, Plantinga has distinguished between two forms of CF: ancient/medieval and modern CF. According to the former (e.g., Aristotle and Aquinas) a properly basic belief will be either self-evident or evident to the senses. Modern CF accepts the self-evidence of a proposition as a sufficient condition for proper basicality, but typically replaces what is evident to the senses with what appears to the senses (so-called self-presenting states or truths appropriately about one's immediate introspective experience, such as being appeared to greenly). In other terms, a modern classical foundationalist takes something like Cartesian certainty as the controlling criterion for proper basicality and accordingly delimits the scope of foundational beliefs. A rational belief is either a properly basic belief or a nonbasic belief (remotely or proximately) based upon properly basic beliefs, where the based-upon relation ranges over three possible modes of support: deductive (Descartes), inductive (Locke), and abductive (Peirce).[7]

If we integrate the epistemological framework of evidentialism, then we may restate the evidentialist requirement as follows:

 

 

[ECF]    Given any person S, S is rational in believing that Pt only if S’s belief that Pt is based upon evidence in the form of adequate reason(s), where these reasons are (a) either properly basic beliefs or nonbasic beliefs that ultimately terminate in beliefs that are themselves properly basic and (b) are adequate just if they provide either deductive or good inductive support for the belief that Pt.

 

 

Over against the background of modern CF and [ECF], the evidentialist objection naturally develops in the following argument:

 

 

(P1)     In a rational noetic structure, every belief is either basic or nonbasic.

 

(P2)     A properly basic belief is either self-evident or immediately about one’s experience.

 

(P3)     A rational nonbasic belief is adequately supported (deductively or inductively) by other beliefs that are either properly basic or nonbasic, and if nonbasic which ultimately terminate (via the based-upon relation) in beliefs that are properly basic.

 

(P4)     Theistic belief is neither self-evident nor immediately about one’s experience.

==================

 

(C1)     Therefore, theistic belief is not a properly basic belief. (from (P2) and (P4))

==================

 

(C2)     Therefore, theistic belief is a rational belief only if it is a rational nonbasic belief. (from (P1) and (C1))

 

(P5)     Theistic belief is not adequately supported (deductively or inductively) by properly basic beliefs or by nonbasic beliefs that ultimately terminate (via the is-based-upon relation) in beliefs that are properly basic.

==================

 

(C3)     Therefore, theistic belief is not rational. (from (C2), (P3), and (P5))

 

The evidentialist challenge to theistic belief, then, claims that if belief in God is to be found in a rational noetic structure, then it will be a nonbasic belief in that structure. The evidentialist objection raised by the atheologist maintains the further claim that there is no sufficient evidence for theistic belief. Therefore, it is unreasonable or irrational to believe in God. But what is the up-shot of the objector's claim that the theist is irrational? What does this amount to?

Plantinga makes it clear that he understands the evidentialist challenge (as well as its foundationalist framework) to be a normative thesis.

 

 

The first thing to see is that this [evidentialist] objection is rooted in a normative view. It lays down conditions that must be met by anyone whose system of beliefs is rational, and here “rational” is to be taken as a normative or evaluative term. According to the objector there is a right way and a wrong way with respect to belief. People have responsibilities, duties, and obligations with respect to their believings just as with respect to their actions. . . .(1983a, p. 30)

 

 

The type of normativity here is deontological. This deontological view of rationality may be traced back to Locke[8] and has had a steady line of advocates since. Plantinga says:

 

 

The Cliffordian idea is that there is a sort of intellectual duty or obligation not to believe in God without having evidence, or sufficient evidence. If there is no evidence, or insufficient evidence, the believer is unjustified; she is flouting her epistemic duties. . . . . . . . .

Contemporary evidentialist objectors (for example, Brand Blanshard, Anthony Flew, John Mackie, Bertrand Russell, Michael Scriven). . .join Clifford in putting their objection in terms of obligations, permission and rights. . . .The problem with the believer in God, they say, is that she holds her beliefs without having sufficient evidence; and the problem with that is that it goes contrary to our intellectual duties and obligations. Evidentialist objectors to theistic belief argue that there is insufficient evidence for theistic belief, and to believe something for which you have insufficient evidence is to go contrary to your epistemic duties. (1991, pp. 290-291)

 

 

On this way of looking at things, then, the sense in which it is necessary for a theist to base his belief in God on reasons is along deontological lines. By doing so he avoids violating his epistemic or intellectual duties.

 

 

[ED]     Given any person S, S violates no intellectual obligations in believing that Pt only if S’s belief that Pt is based on evidence in the form of adequate reasons.

 

 

Plantinga notes that in addition to taking “rationality” in terms of “duty” it could also be taken in the sense of “epistemic excellence” or “the avoidance of epistemic defect” (Plantinga 1983a, p. 52). In 1986a (pp. 110-111) Plantinga carefully notes a distinction between “deontological evidentialists” (Blanshard, Clifford, and Scriven) and “axiological evidentialists.” Here Plantinga raises the problem of doxastic voluntarism to cast doubt on the plausibility of the deontological evidentialist objection to theistic belief. Because of the involuntary nature of belief (see chapter 1, section II.B), Plantinga suggests that the evidentialist objector desert the deontological position and take up axiological evidentialism. Instead of arguing that the theist who believes in God without basing that belief on evidence has violated some intellectual duty, he should argue that theistic belief without evidence is an indication of a defective or flawed noetic structure. The theist, who believes in God, but not on the basis of adequate reasons, is like the person who suffers from a spastic colon.  The person has a deficiency or flaw that places him in a deplorable condition. In the case of belief it is not an unhappy physical condition but an unhappy epistemic condition.

On the axiological view, then, the relevant evidentialist requirement would be something like:

 

 

[EA]     Given any person S, S possesses epistemic excellence (or is epistemically nondefective) in believing that Pt only if S’s belief that Pt is based on evidence in the form of adequate reasons.

 

 

B. The Plantingian Objection to Evidentialism

 

 

 

Traditionally, theists have sought to answer the evidentialist objection articulated in the atheological evidentialist argument by showing that

 

 

(P5)     Theistic belief is not adequately supported (deductively or inductively) by properly basic beliefs or by nonbasic beliefs that ultimately terminate (via the based-upon relation) in beliefs that are properly basic

 

 

is false. Or, to detach the objection from foundationalism,

 

 

(II)       We have no evidence, or at any rate not sufficient evidence, for the proposition that God exists

 

 

is false.

 

 

Theistic philosophers have taken on claims like (P5) and (II) by presenting arguments for the existence of God, what is traditionally called the project of evidentialist (or positive) apologetics. Plantinga certainly agrees that (P5) is a strong statement. Is it really true that theistic belief lacks evidence, sufficient or otherwise? The criteria of self-evidence and evident to the senses (as in ancient/medieval CF) allow a rather broad properly basic evidence-base from which theistic arguments could be constructed. In fact, it is precisely these kinds of starting-points that characterize Thomas Aquinas’s Five Ways in the Summa Theologiae. Prospects are even further strengthened by extending support relations beyond purely deductive modes of support. Even if there are no sound deductive arguments for the existence of God (say, no cosmological argument starting from self-evident principles and what is evident to the senses and proceeding by necessary inference to the existence of God), why should this preclude there being adequate or sufficient evidence for the claim that God exists? How many reasonable philosophical or broadly scientific theses could be supported in this fashion? Very few it would seem. Moreover, narrowing the scope of properly basic beliefs (as in modern CF) may easily be compensated for by broadening the kinds of evidential support relations between the foundations and superstructure, so as to allow the generation of a broad scope of nonbasic beliefs.  In the face of the many inductive or probabilistic theistic arguments put forth this century by F.R. Tennant, C.S. Lewis, E.L. Mascall, Basil Mitchell, Richard Swinburne, and others, it seems that (P5) and (II) are philosophically vulnerable claims.

Plantinga, though, rather than drawing on the arguments which constitute the stock and trade of the natural theologian to launch an evidentialist apologetic, sets out to dismantle the evidentialist objection by challenging the evidentialist challenge. Since this challenge is rooted in CF, focus is taken on the fundamental principle of CF:

 

[CF]     A proposition p is properly basic for a person S if and only if p is either self-evident to S or about S’s immediate experience or evident to the senses.

 

 

More specifically, since the evidentialist challenge to theistic belief is rooted in modern CF, Plantinga’s critique focuses more directly on modern CF. According to this principle, two things follow.  First, the principle claims that a proposition is basic if it is either self-evident or about one’s immediate experience. Secondly, the principle claims that a proposition is properly basic only if it meets one of these conditions.  Plantinga’s critique of modern CF focuses on this second element. Why suppose this is true?

Plantinga’s first argument against modern CF is what we might call the counter-example argument (1983a, p. 60). It purports to show that the basic principle of modern CF is incongruous with the stock in trade of our ordinary everyday beliefs that we take to be rational beliefs. Let us suppose for the moment that the modern CF criterion of proper basicality is true. How many of our everyday beliefs - which we regard as rational - are either self-evident or about our immediate experience are either deducible from or rendered more probable than not on the basis of propositions that are?  Very few of them. Indeed every individual finds himself with an array of beliefs to the effect that he is appeared to such and such. And the same can be said about self-evident propositions. They certainly make up a portion of our noetic structures. However, the problem is that there are a whole lot of everyday beliefs which do not satisfy the modern classical foundationalist’s requirements for proper basicality. Belief in the external world, other minds, the occurrence of past events, and the rest of our so-called common-sense beliefs or knowledge are not properly basic on the model of modern CF. Consequently, these beliefs are rational only if they can receive adequate support from beliefs which are either self-evident or appropriately about one’s immediate experience. Plantinga finds it hard to see how such beliefs can be rendered more probable than not, let alone deducible, on the basis of some properly basic beliefs. This does seem to be a genuine problem if we restrict the class of properly basic beliefs to those that possess Cartesian certainty, unless we allow a rather liberal range of probabilistic relations between the foundations and nonbasic beliefs.[9] It would seem then that on the view of modern CF, many of our everyday beliefs are not rational. Under a deontological construal of rationality, individuals are violating some intellectual duty when they hold such beliefs. If one prefers a nondeontological account of rationality, humans are epistemically defective, or exemplifying cognitive malfunction of some sort, in holding to most of their ordinary, everyday beliefs about the world. But this seems absurd.

A second argument (1983a, pp. 60-61) raises the question of the coherence of modern CF. Plantinga claims that modern CF appears to be self-referentially incoherent, as it does not fulfil the conditions of proper basicality it lays down, and it is difficult to see how it could be adequately supported by beliefs which do meet that narrow criterion.  According to modern CF, a person S is rational in accepting the principle of modern CF only if that principle is either a properly basic belief or a nonbasic belief appropriately based upon a properly basic one. But if the principle of modern CF is a properly basic belief, then it is either self-evident or about one’s immediate experience. Clearly, though, the proposition in question - “a belief is properly basic if and only if it is either self-evident or appropriately about one’s immediate experience” - meets neither of these conditions of proper basicality. This is especially clear if we recall that properly basic beliefs possess Cartesian certainty and are therefore characterized by epistemic immunities (from error, doubt, etc.). As for self-evidence in particular, the principle of modern CF is not expressed by a sentence that a person believes merely by understanding the terms involved. Therefore, it is not a properly basic belief. If a person is rationally to believe it, then, it must be a nonbasic belief based on beliefs which are either themselves properly basic or which ultimately terminate in beliefs which are properly basic, and where these beliefs provide adequate evidential support for the principle of modern CF. However, if we take our set of self-evident beliefs and beliefs about one’s immediate experience, it doesn’t appear that they entail or render probable the proposition in question. Equally, it is hard to see how they entail or render probable other beliefs that in turn (directly or indirectly) entail or render probable the principle of modern CF. Certainly no modern classical foundationalist has ever produced an argument showing this. Therefore, belief in the proposition that asserts the modern classical foundationalist criteria of proper basicality is irrational. Anyone who accepts it is either flouting his intellectual duties or is in some way epistemically defective.

Plantinga’s argument from self-referential incoherence was criticized by Alston (1985, pp. 296-299). Alston agrees that Plantinga has established a fundamental challenge to modern foundationalists, but he does not believe that he has substantially supported the charge of “self-referential incoherence.” Alston argued that, “It often happens that people believe that p on the basis of q without ever citing q in support of their belief that p” (p. 297). Therefore, simply because no argument has been given for p, it does not follow that p is not based upon premises which adequately support p. Plantinga (1985a, p.386) concedes as much. In 1983a (pp. 61-62) he had presented an alternative, “more rigorous” version of his argument. In this argument the requirement that modern CF’s principle be supported by properly basic beliefs (or by beliefs which terminate in beliefs which are properly basic) is substituted by a weaker requirement to the effect that “there are paths in S's noetic structure from [that principle] to some propositions q1, . . . ,qn which are properly basic and support [it]”. He agrees that there could be such paths, even if a person cannot say what these paths and their corresponding propositions are and even if others cannot think of any viable candidates. Plantinga, though, concludes that “This seems unlikely, however, and in the absence of reason to think that there are propositions of that sort, the better part of valour is to reject. . .[it]” (p. 62). On the basis of his critique Plantinga concludes: “It is evident. . .that classical foundationalism is bankrupt, and in so far as the evidentialist objection is rooted in classical foundationalism, it is poorly rooted indeed” (1983a, p. 62).[10]

 

                                             II. Critical Epistemological Elucidations

 

 

 

A. Evidence Possession and Evidence Basis

 

 

 

Thus far I have stated the evidentialist challenge as requiring that the theist’s belief in God be based on evidence in the form of adequate reasons. There is, however, a variant on this formulation frequently encountered in the literature. Instead of [E], the evidentialist requirement is sometimes given as something to the effect of

 

[Ep]      Given any person S, S is rational in believing that Pt only if S has evidence in the form of adequate reason(s) for the belief that Pt.

 

 

I will call this the evidence possession form of the evidentialist requirement, in distinction from the evidence basis form expressed by [E].  These two accounts of the evidentialist requirement are found throughout the literature on Reformed epistemology. I cite just a few examples (italics mine). Wolterstorff writes: “If one is rationally to believe some theistic proposition, one must believe it on the basis of others of one’s beliefs that constitute good evidence for it” (1986, p. 39). And yet we also find “Scarcely anything has been more characteristic of the modern Western intellectual than the conviction that unless one has good reasons for one’s theistic beliefs, one ought to give them up” (1986, p. 38).[11] Plantinga states the challenge both ways as well. “The question is whether. . .someone might be rationally justified in believing in the existence of God on the basis of the alleged evidence offered by [theistic arguments]” (1983a, p. 30). The evidentialist “holds that one who accepts the belief that there is such a person as God but has no evidence for that belief is so far forth, unjustified, intellectually irresponsible, or in some way unreasonable” (1986a, p. 110). 

Admittedly there is an ambiguity in phrases like “has evidence” or “has reason” for believing p. One may have other beliefs b1, . . . ,bn such that, given one’s inductive standards, b1, . . . ,bn make the belief that p probable, but failing to see the evidential connection between b1, . . . ,bn and the belief that p one does not also hold (even nonoccurrently) the higher-level belief that <b1, . . . ,bn make the belief that p probable>, though one will typically have the disposition to form such higher-level beliefs given such conditions. Alternatively, one may think of S’s having evidence for p as involving such higher-order beliefs. I will shortly come to the matter of evidence with such higher-order elements. Here I am simply concerned with having evidence in the first sense above, where this does not involve occurrent or nonoccurrent higher-order beliefs about the evidence b1, . . . ,bn counting as evidence for the belief that p. Now given the first sense of “having evidence,” it might be thought that nothing of great epistemological importance rides on the evidence-basis and evidence-possession distinction. After all, a rational belief that satisfies an evidence-basis requirement will satisfy an evidence-possession requirement.  Nevertheless, it is not the case that if a rational belief satisfies an evidence-possession evidentialist requirement it also satisfies an evidence-basis requirement. This is where the difference is found, and the difference here is an important one.[12]

Having adequate reasons (or “grounds” generally speaking) for believing that p (and hence having a justification, so to speak) is independent of believing that p.  I may have a set of adequate reasons to believe that I will not pass my Latin mid-term exam (e.g., I haven’t studied all week, my last two exams were D’s, and I was absent the day the instructor went over the material to be covered on the exam), yet I remain unmoved in my belief that I will pass it (perhaps as a case of wishful thinking). Here I have adequate reasons to believe not-p, but in fact I believe p. In other cases one may have some belief b and b may in fact provide evidential support for the proposition p, but S may fail to believe p (or its negation). It is also quite possible to have adequate reasons for some belief and yet believe it on the basis of reasons that are inadequate. It is quite possible for me to believe that I will pass my Latin mid-term on Monday, and I may have adequate reasons to believe this (e.g., I have passed all previous tests with high marks, my homework has had few mistakes, and I have not missed a single class all term). But suppose that, even though I have these adequate reasons, I actually believe that I will pass the exam on the basis of the belief that I am a logically omniscient being.  Here is a case where a belief is generated or causally sustained in a way that is intellectually disreputable, and this is not epistemically desirable, if we bear in mind the epistemic point of view. For generally, beliefs formed on such a basis are not true. Even if having good reasons is sufficient for having a justification for the proposition p, many would think this insufficient for being doxastically justified in one’s belief that p (or in knowing p).

For the above reasons it might plausibly be thought that [Ep] is subject to some significant epistemic liabilities. That it is a weaker requirement is clear, but one’s concept of rationality will probably determine whether [Ep] lays down an inadequate requirement. Its inadequacy would have to be an epistemic inadequacy, and that could be judged from at least two different perspectives, one deontological and another nondeontological. Perhaps one relevant epistemic obligation is simply to have evidence for one’s beliefs. Then one may be rational in that sense of the term if one does have the evidence. But even on a deontological view of justification or rationality, we want to have obligations which are suitably connected to being in a good position to acquire true beliefs, and it’s not clear that simply having reasons (even adequate ones) helps us sufficiently in that regard. Moreover, since [Ep] is compatible with a person’s noetic structure being defective or deficient with reference to the belief in question (e.g., when p is formed on inadequate grounds or one has adequate grounds for p but either does not believe that p or believes its negation), it could plausibly be held that [Ep] doesn’t capture what we want on even a nondeontological view of rationality. On the other hand, if rationality is non-source-relevant, the liabilities of [Ep] are not as clear. Perhaps having occurrent evidence that renders p epistemically (or subjectively) probable is sufficient for rationality, though S’s seeing that his evidence makes p probable may also be necessary.

Even on a source-relevant view, though, merely having an adequate ground for some belief that p can be viewed as “a” desirable epistemic status (whether or not one actually believes that p). In most cases, it is better to believe something for which one has an adequate ground than to believe something for which one does not have an adequate ground, for presumably it at least puts one in the epistemic ballpark. A person suitably reflective might eventually become cognitive of the fact that he has adequate reasons for believing that p and then believe that p on the basis of those reasons. This will be true if, for instance, having evidence creates the disposition to believe that p and to believe it on the basis of that evidence. In later chapters I will exploit these moves to develop the positive justificatory role of having good reasons. But, given the points above, it also seems true that it is a better thing (from the epistemic point of view) to believe something on the basis of an adequate ground than merely to have an adequate ground for the belief in question.[13] For the moment I will rest content with understanding the evidentialist principle to require that the basis of one’s belief in God is another rational belief or knowledge.

 

B. Dialectical and Structural Evidentialism

 

There is a related distinction I wish to draw between dialectical and structural evidentialism. The distinction hearkens back to the discussion of foundationalism and the formulation of the infinite regress problem.  In chapter 1 I noted that the infinite regress problem can be stated in a skeptical context according to which the question “What justifies you in believing that p?” is roughly equivalent to “Show me that you are justified in believing that p.” On the other hand, one may simply issue the original question in an informative tone as equivalent to “What is the source of your justification for believing that p?” Whereas the latter is a purely structural question, the former is dialectical or discursive. Similarly, the emphasis in the evidentialist challenge may fall on showing (to oneself or others) that Pt, where this is the upshot of producing some argument or engaging in the process of inference.

 

 

[ED]     S is rational in believing that Pt only if S has (consciously) inferred the proposition Pt from a set of other propositions that constitute adequate evidence for Pt.

 

 

This is a common understanding of the evidentialist challenge. Wolterstorff states that the evidentialist's claim “is that theistic conviction, to be rational, must be arrived at, or at least reinforced, by the process of inference” (1983a, p. 158). Plantinga, commenting on Flew’s evidentialism (see Flew quote in Part I), writes: “Flew is claiming that it is irrational or unreasonable to accept theistic belief in the absence of arguments or evidence for the existence of God” (1983a, p. 26). Not to anticipate too much of Plantinga’s own positive position, but when he discusses the so-called “Reformed Objection to Natural Theology” (1983a, pp. 73) he states that objection in such a way that it presupposes a dialectical evidentialist requirement. John Calvin is said to maintain that “a Christian ought not to believe [in God] on the basis of argument” (p. 67) and “belief in God ought not to be based on arguments” (p. 71). “In rejecting natural theology. . .these Reformed thinkers [Calvin, Bavinck, and Barth] mean to say that the propriety or rightness of belief in God in no way depends upon the success or availability of the sort of theistic arguments that form the natural theologian’s stock in trade” (p. 72). All this suggests that the original evidentialist challenge holds that a person must base their belief in God on arguments (or at least have such arguments at their disposal).

The first thing to be said here is that there is a subtle ambiguity and epistemic confusion which may largely be responsible for leading philosophers to adopt a form of dialectical evidentialism. The ambiguity and confusion surrounds the idea of inferential belief and inferential justification.  There is one sense in which a person's belief that p may be inferentially justified, even though S has not gone through any conscious process of inferring p from q. The ground of a belief (some belief or experience) may be implicitly possessed and its being utilized in the formation of a belief need not be something of which the subject is aware. As long as other propositional attitudes are the features of the input taken account of, S may believe p on the basis of some other belief even though S is not conscious of the fact that she believes on the grounds of some other beliefs she possesses. This can be contrasted with an inferential justification that is the result of the subject inferring some proposition p from some other proposition q that constitutes an adequate reason for p.  We should not confuse a belief’s being based on other beliefs (and thereby being mediated by them) with the activity of inferring a belief from other beliefs (and thereby being mediated by them in this sense).

According to Robert Audi (1993, pp. 237-239) the preceding distinction may be stated in terms of a distinction between structurally inferential and episodically inferential beliefs. S may believe that the Jones’s are home for the reason that their living-room lights are on. S does not believe that the Jones’s are home because he has gone through a process of inferring this as a conclusion from a premise (or set of premises). S takes note of the fact that the lights are turned on in the Jones’s living-room and, already possessing the belief that when the Jones’s living-room lights are turned on in the evening the Jones’s are home, the belief that they are home is formed. S’s belief that when the Jones’s living room lights are turned on the Jones’s are home is taken account of (along with the belief that the lights are now turned on) in the formation of the belief the Jones’s are home. What we have here is a belief due to a reason (or reasons), but not due to a reasoning process. According to Audi, there is an argument that underlies this case, in the sense of “an abstract argumental structure,” and one that is probably enthymematic, from the grounding reason (or reasons) r to the target belief b. Believing for a reason in such an instance is structurally inferential. On the other hand, b may arise from a tokening or internal recitation of that structure, the actual process of inferring b from r. Such a belief would be episodically inferential. Although every belief for a reason that is episodically inferential is structurally inferential, the converse does not hold.[14] [ED] of course only follows from an episodically inferential belief requirement.

 [ED] doesn’t seem very plausible as a general epistemological account. To require that a person’s nonbasic belief that p be based on arguments or a process of inference is generally too strong of a requirement for the justification of nonbasic beliefs. Such a process is a moderately sophisticated one which many individuals (e.g., children, idiots, and some simple-minded adults) are either not capable of performing, or in fact do not perform. Yet, they nonetheless frequently form justified nonbasic beliefs (and even possess knowledge) that p.  For example, a young boy forms a belief that the stove is hot because he sees that the burner is red, recalls that in the past he touched it when it was red and was burned. This often takes place even when a person has not engaged in a process of inference, and yet the belief is not merely the result of a perceptual experience. This would be an example of a structurally inferential belief. Does the boy violate some epistemic duty in believing as he does? Is his noetic structure in some way defective? Such accusations appear most implausible. On many accounts of epistemic justification, the boy would have a mediately justified belief, but he has not performed any conscious inference.  One might amend the epistemological theory to account for this counter-argument by accepting two forms of inferential justification and yet impose the episodically inferential evidentialist requirement on theistic belief (excluding many of the paradigmatic cases adduced in favour of structural inferential justification). Unfortunately, until some independent reason is given for applying this requirement to theistic belief in particular, such a move would smack of atheological adhocness and exemplify a type of epistemic double-standard which many would take to be philosophically unacceptable.[15] Most fundamentally, as noted in chapter 1, the property theory of justification is to be preferred over the process theory of justification. The former better captures the general desideratum of epistemic justification. For this reason structural evidentialism is to be preferred over dialectical evidentialism. Perhaps dialectical evidentialism has some other significance (as I will argue later in the thesis). But if it is epistemic justification we are targeting, I believe we need to focus on structural evidentialism.

 

C. Internalist and Externalist Considerations

 

A more interesting variety of options arise in relation to the internalist/externalist debate.  The evidentialist requirement for theistic belief maintains:

 

 

  [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.

 

 

Inasmuch as [E] restricts the scope of justifiers to reasons (i.e., other rational beliefs or knowledge), the evidentialist requirement can be seen as imposing a minimal internalist constraint on the conditions of the rationality of theistic belief, a PI constraint. I also think that there is a case for naturally construing the evidentialist requirement as involving some kind of second-order internalist condition. Recall the association of classical evidentialism with deontologism, as highlighted by “rationality” taken as “the fulfilment of epistemic duty.” Recall also my point in chapter 1 that a duty to conduct my cognitive life in such a way that it aligns itself to the alethic goal of believing leads rather straightforwardly to requiring adequacy to be cognitively accessible. Intellectual dutifulness would seem to require, not only that a person believe p on the basis of q, but that he (justifiably) believe that <q is an adequate ground for p>, or at least that this is something which he could come to believe fairly readily upon reflection. Only if I believe on the basis of a ground that I take to be an adequate truth-indicator, or could come to see this upon reflection, will I have done all that I could to see to it that I am believing what is true and not believing what is false. Very easily, then, can we move from an evidence-base e to an evidence-base e*, where the latter includes a rational judgement about the efficacy of evidence e (or where such a judgement could be reached upon reflection).

The notion that “adequacy” must be a cognitively accessible item has important implications for how adequacy is construed. Externalism, owing to its taking probability in an objective sense aligned with statistical or physical probabilities, will naturally take “adequacy” as a kind of truth-conducivity which is not reflectively accessible (at least not in any obvious sense). What is crucial to externalist accounts are objective empirical features of the subject’s cognitive situation (e.g., the reliability of belief-forming processes and the possession of reliability-characteristics), and we do not come to grasp these a priori. True enough, the internalist’s evidence may and usually does consist of contingent truths discovered a posteriori but what constitutes the adequacy of that evidence for the internalist will be an evidential probability relation grasped (or capable of being grasped) a priori on the basis of inductive or deductive criteria. For the externalist “adequacy” is measured empirically (e.g., do beliefs of x-sort usually produce true beliefs, do criteria x and y usually yield a statistically high probability of true beliefs), and that kind of adequacy measurement is grasped a posteriori (by relying on what we know from cognitive psychology and other empirical sciences).

Nor is this to deny that the internalist package has objective features. We can, on the internalist view, distinguish between (at least) two senses of evidential probability, and so two senses of  “adequate reasons.” Inasmuch as we can think of the evidential relation between some proposition h and contingent evidence e as an objective logical relation, there is an objective feature of the situation. More specifically, we can distinguish between a proposition h being more probable than not on evidence e given correct inductive standards (epistemic probability) and h’s being probable relative to the particular inductive standards employed by some person which may in fact not be correct (subjective probability). It is also clear that we can think of P(h/e) relative to S’s own inductive standards but where S’s own inductive criteria have been subject to critical scrutiny by S and which S judges to be adequate. There is also the question of how much S looked into the evidence, whether he thinks that the evidence has resulted from investigation and reflection which was in fact adequate. And so naturally we can see a very complex set of evidential adequacy relations, with distinctions between S’s inductive standards, S’s investigation of and reflection on those criteria and the evidence, and whether these are merely believed by S to be adequate or whether are in fact adequate.[16]

It would appear that the theistic evidentialist requirement is, on the internalist position, susceptible to several kinds of permutations.[17] Moreover, although space constraints do not allow a development of this, I think that these permutations of the core evidentialist requirement have nearly all been represented in the tradition of classical evidentialism.  For instance, Locke writes: “The mind if it will proceed rationally, ought to examine all the grounds of probability, and see how they make more or less for or against any proposition, before it assents or dissents from it; and upon a due balancing of the whole, reject it or receive it, with more or less firm assent, proportionably to the preponderancy of the greater grounds of probability on one side or the other” (Essays, iv. 15. 5).  What Locke seems to be saying here is that the rationality of our beliefs (in what is not self-evident or certain) requires that we have higher-order beliefs (presumably justified) about the extent to which the grounds of our beliefs render them probable. If the mind proceeds rationally with respect to some belief that p, then there should be a determination about (a) what the ground of the belief that p is and (b) how probable G makes the belief that p.  He further assumes that we have control over our doxastic states and so can believe or withhold belief or believe to some degree dependent on how probable G makes p.  This locus classicus of traditional evidentialism suggests that the evidentialist requirement is fundamentally a demand that a person not only have good reasons for belief but have among those reasons a justified belief about the evidential adequacy of those reasons and perhaps reasons for supposing that the reasons are in fact adequate.[18] What is not clear is how Locke sits on the distinction between epistemic and subjective evidential probability (as distinguished from the logical evidential probability, see chapter 1, section II.D).

We can, I think, lay down two versions of the evidentialist requirement.

First, there is an epistemic evidential probability requirement (for the less than logically omniscient person):

[E1]     Given any human person S, S’s belief that Pt is rational only if (a) the belief that Pt is based on propositional evidence e that renders Pt probable relative to correct inductive standards and (b) S rationally believes (or is capable of believing just on reflection) that <e provides adequate evidential support for Pt>.

 

Secondly, there is a subjective evidential probability requirement.

 

[E2]     Given any human person S, S’s belief that Pt is rational only if (a) S’s belief that Pt is based on propositional evidence e that renders Pt probable relative to S’s own inductive standards and (b) S rationally believes (or is capable of believing just on reflection) that <e provides adequate evidential support for Pt>.

 

 

 

III. Re-examining Theistic Evidentialism

 

 

A. Assessing the Challenge of Theistic Evidentialism

 

 

What can be said of the basic evidentialist package then?

Let’s begin with the second-order internalist requirement. In chapter 1 I developed some of the difficulties with higher-level requirements for justification, so I will only mention those here and how they bear on theistic evidentialism. If the theistic evidentialist does not build a PI2 requirement into his concept of justification as such, there is no threat of infinite regress (which would otherwise be generated by requiring for every belief B S be justified in believing some correlated higher-level proposition). But PI2 and AI2 seem to flounder given the sorts of conceptual requirements involved in higher-order beliefs. Beliefs about “probability” or “the adequacy of evidential support relations” require the possession of the requisite concepts, even where such beliefs are not occurrent or explicit. Even if adequacy is spelled out in internalist terms so that it is in principle cognitively accessible, this accessibility will in fact only be for those people who have the conceptual prerequisites for forming such beliefs on reflection. Perhaps this doesn’t rule out the majority of human subjects, but it does rule out the possession of rational theistic beliefs by some humans (e.g., little children and some idiots). Secondly, I have also noted what seems to be the close relation between higher-level requirements and deontologism. And there has already been a consideration (in chapter 1) of the problematic nature of deontological rationality given the involuntary nature of belief, a difficulty which can only be overcome it would seem by applying obligations to practices which influence belief formation (or by distinguishing between belief and acceptance). So one of the main motivating factors for higher-level requirements diminishes the plausibility of such requirements. Thirdly, the higher-level requirement seems to involve a conflation of being justified and showing justification. If one is to perform the latter, higher-order beliefs would be required. But why suppose that being justified in a theistic belief requires the activity of justifying that belief? If this is not true of beliefs in general, we would need an explanation as to why it is a reasonable demand for belief in God. Theistic evidentialism it seems stands in need of modifications.

The case for theistic evidentialism may be strengthened by laying down the distinction between what Paul Moser (1989, chapter 5) calls static epistemic rationality and procedural epistemic rationality. In chapter 1 the idea of epistemic justification was associated with the need to locate the surplus value of knowledge over true belief. This surplus value seems best construed in terms of a ground (or process) that in some sense counts toward the truth of the belief or is a truth-indicator. But the question of whether we have such truth-indicators is a different question from assessments of whether I have done my best to make sure that I possess them. There is an important distinction between one’s being in a favourable epistemic state and one’s conducting or regulating one’s intellectual life to see to it that one is in such an epistemic state. The conditions imposed on static epistemic rationality correspond to something on the order of the possession of (actual) truth-indicators or probability makers. The conditions imposed on procedural epistemic rationality are different. They have to do with our engaging in certain belief influencing activities in order to maximize truth and minimize falsity, and perhaps a notion of intellectual obligations which fall on the voluntary actions which influence belief formation and sustenance. What the conditions of the latter tell us is whether a person has been a responsible truth seeker. What is required for this sort of epistemic rationality is something like reflective rationality. This is a process of critical reflection on the evidence for one’s belief that p that results in a judgement about the (static epistemic) rationality of the belief that p.  I have already noted that deontologism generates second-order internalism, and the natural tendency to see cognitive obligations tied to procedural epistemic rationality yields a second-order internalist requirement for this regulative view of rationality.

The relation between procedural epistemic rationality and reflective rationality is somewhat complex and requires a few elucidations. Voluntary control goes with regulating one’s cognitive life to ensure that one is holding true beliefs and avoiding false ones, for one has voluntary control over pursuing activities which will better align oneself to the epistemic goal. The second-order beliefs that I am associating with reflective rationality are not subject to voluntary control any more than first-order beliefs that p.  The connection is this: when one satisfies conditions of procedural epistemic rationality with respect to some belief that p, one is pursuing epistemic policies or seeking to ensure that one’s belief that p is likely to be true. This (voluntary) activity is what leads one to a consideration of evidence about whether one’s belief is likely to be true and thus to what extent it is a rational or justified belief. As a consequence, one comes to hold a second-order belief that one’s belief that p is rational. Ideally, S’s belief that <S’s belief that p is (statically) epistemically rational> will itself be statically epistemically rational (and one’s satisfying the conditions of static epistemic rationality for this higher-order belief will be distinct from the conditions required for procedural epistemic rationality toward this higher-order belief). What must be understood is that there are diachronic and synchronic aspects to reflective rationality. The latter refers to S’s cognitive state of holding a belief about the rationality of S’s belief that p (clearly involuntary); the former refers to the process through time of S’s looking at evidence and evaluating that evidence (matters of voluntary control). Procedural epistemic rationality via-à-vis some belief that p entails diachronic reflective rationality toward the belief that p, and the latter culminates in a state of synchronic reflective rationality toward the belief that p.

Given the above distinctions, a strengthened case for theistic evidentialism is roughly as follows: propositional evidence is necessary for procedural epistemic rationality. More specifically, it might plausibly be thought (as I shall argue in chapters 6 and 7) that higher-order beliefs which are entailed by reflective rationality require propositional evidence if they are to be (statically) epistemically rational), be that evidence discovered upon lengthy investigation or reflection. In viewing the tradition of classical evidentialism, with its strong commitment to deontologism and corresponding internalist demands, it seems quite plausible to read that tradition as developing evidential requirements for belief as a result of a two-fold conflation:  conflating (1) static and procedural epistemic rationality and (2) unreflective and reflective rationality. The classical evidentialist position faces difficulties, then, not merely for reasons adduced by Plantinga, but also because of the apparent confusion of epistemic levels tied to conflating distinct desiderata of rationality. That weakness, though, suggests an alternate construing of evidentialism that may not be susceptible to the Plantingian critique. If what I have said is at least approximately correct, then it may be that what propositional evidence is required for is reflective rationality. This relocates the impropriety of theistic belief without evidence on another level of discourse, thereby freeing up the prospects for justified or rational theistic belief, even theistic knowledge, without evidence. There is a growing recognition that there are various senses of rationality, different doxastic merits that contribute in different ways toward the larger epistemic picture. That seems to lend some credibility to the idea that evidence may be more appropriate for some epistemic merits than others, and if some of these evidence-essential epistemic merits are necessary for being justified in certain circumstances, we have a starting-point for a new way of thinking about evidentialism. My suggestion (to be developed later in the thesis) is that there are actually multiple ways of thinking about the epistemic point of view, and so multiple ways of thinking about evidential requirements. Casting one’s lot with the classical package with reference to some epistemic merits need not commit one to buying into the original evidentialist package. The plausibility of strong evidential requirements at these higher-levels and for desiderata distinct from justification, or what transforms true belief into knowledge, forms the content of chapters 6 and 7.

 

B. Modest Foundationalism and the Evidentialist Objection to Theistic Belief

 

The rise of modest foundationalism in the 20th century presents some interesting consequences for the evidentialist objection and challenges to the prospects for any version of theistic foundationalism. First, by broadening the base of properly basic beliefs the prospects for the inclusion of theistic belief in the foundations enjoys a better status than it did under the classical foundationalist model. The extension of “experience” has resulted in a form of empiricism more amiable to religious belief.  Alston’s recent thesis (1991c) regarding religious experience as a nonsensory perceptual practice of forming beliefs about God in a way analogous to the sensory perceptual practice of forming beliefs about our physical environment has illuminated the prospects for treating theistic belief as properly basic. Secondly, the rise of externalist theories of justification and knowledge has contributed to thinking of theistic belief as immediately justified or as immediate knowledge, since such theories are not saddled with the kinds of internalist demandsthat have generated classical evidentialism. On the other hand, the prospects for rational nonbasic theistic belief are strengthened by broadening the scope of properly basic beliefs and the mode of acceptable support relations between the foundations and the superstructure. This allows more both by way of data in the foundations and by way of inferential possibilities within the superstructure as to how nonbasic beliefs may appropriately be generated and sustained.[19]

But with these developments within the framework of modest foundationalism also come new ways for thinking about evidentialist requirements for theistic belief. Those philosophers of religion critical of evidentialism all admit the prima facie character of justification. And if reasons can override the justification of theistic belief, it may be that for some people their remaining justified in holding theistic belief requires something in the way of propositional evidence, even if their belief in God was properly basic at the time of its acquisition. The conditions required for the justification of a belief at the time of its formation may not be the same as the conditions necessary for justification during the belief’s maintenance. The application of a rather generous principle of credulity (according to which a belief, or doxastic practice, is innocent until proven guilty) is compatible with a heavy-handed evidentialist requirement when there are substantial reasons to doubt the truth of a proposition. Perhaps such defeating circumstances generate the need for reflective rationality. Or, as some have suggested, maybe the more appropriate evidentialist requirements under such circumstances need not require that the individual possess the relevant evidence, only that the evidence be available within the subject’s community (Wykstra 1989, Kenny 1992). Perhaps the satisfaction of such a requirement is actually a condition for proper basicality. In any event, the person and time relative nature of rationality certainly leaves open these analogical extensions of the classical evidentialist package. The extent to which Plantinga’s anti-evidentialism stretches as far as these variations on the basic evidentialist position is yet to been seen. For the moment, though, I think it is crucial to note the following by way of reflection on the material thus far presented. Even if Plantinga’s criticisms present a challenge to Enlightenment evidentialism rooted in classical foundationalism, the potential analogical extensions of evidentialism may possess some degree of immunity from those criticisms and some degree of plausibility, even if they should stand in need of further development. In short, the refutation of classical foundationalism does not destroy (foundationalist-based) theistic evidentialism, even if it reduces to ashes the aspirations of Locke and his most devout descendants.