The critique of evidentialism
In recent decades some religious philosophers have welcomed the demise of traditional Western epistemology, Footnote 1 a legacy often characterized as strong foundationalism (including Greek and Enlightenment versions). Moreover, some critics have included evidentialism in that obsolete legacy.Footnote 2 As Nicholas Wolterstorff summarized the evidentialist challenge: ‘No religion is acceptable unless rational, and no religion is rational unless supported by evidence.’Footnote 3 In addition, evidentialism typically includes an epistemic duty to proportion the strength of belief to the strength of evidence (the proportionality principle).Footnote 4 The following argument denies that the demise of strong foundationalism discredits evidentialism. I conclude instead that recent criticisms of strong foundationalism strengthen the heart of evidentialism, the proportionality principle.
The most influential anti-evidentialist critics in religious epistemology include the Reformed epistemologists. A milestone in their view is Alvin Plantinga's article (Reference Plantinga and Nicholas1983), in which he asserted that ‘it is entirely right, rational, reasonable, and proper to believe in God without any evidence or argument at all’.Footnote 5 Likewise, in his defence of Reformed epistemology, Kelly James Clark echoed: ‘The religious hypothesis, as with questions of morality and personal relations, is justified without the support of evidence or arguments’ (1990, 113). Concerned for the demands of faith, Clark concludes that evidentialism is ‘not so much false as irrelevant to the discussion of the rationality of belief’ (ibid., 119). Indeed, in light of the moral demands of faith, evidentialism is worse than irrelevant: it is ‘perverse, obdurate, or improper’ (ibid., 122).
The claim that a belief can be rational without any evidence (or perverse for appealing to evidence) would strike many as counterintuitive, placing a considerable burden of proof on anti-evidentialists. According to the anti-evidentialist view, however, their burden of proof has been lessened by mainstream epistemology's rejection of traditional foundationalism, which they sometimes associate with evidentialism. But the demise of traditional foundationalism does not imply the demise of evidentialism, as the two views, while historically linked, are not identical. Their conceptual differences explain why, for example, there has been widespread consensus on the offending features of traditional foundationalism but little consensus on the concept of evidence or on the strength of evidence required for rational belief, and therefore little consensus on the offending features of evidentialism, or on the epistemic standards that religious epistemology should prefer to evidence.
In fact, since the more revolutionary pronouncements of the 1980s, even the Reformed epistemologists' view of evidentialism has been significantly reformed, including much needed distinctions between foundationalism and evidentialism (Wood (Reference Wood1998); Wolterstorff, (Reference Wolterstorff, Greco and Sosa1999)). Perhaps most significantly, Reformed epistemologists have expanded their concept of evidence to include religious experiences (see below), the effect of which is to cure evidentialism of a possibly inappropriate narrowness. To date, the implications of these developments for evidentialism have not been fully appreciated.
To begin with, the question of evidentialism suffers from different notions of evidence, leading philosophers to talk past each other. But the crucial questions for religious epistemology concern (a) the authority of the proportionality principle and (b) whether that principle renders faith too beholden to reason. In Peter Forrest's view, contemporary religious epistemology is tending toward a middle ground between the ‘the extremes of Enlightenment evidentialism [a too narrow rationalism] and fideism [a too irrational faith]’.Footnote 6 In that case, the fate of evidentialism depends on whether that happy medium will prevail, (a) by leaving evidentialism behind (as strong foundationalism is being left behind), or (b) by reforming evidentialism into a more adequate, post-Enlightenment form that more adequately defines rational religious belief. In short, it is not yet clear whether the recent revolution in epistemology spells the end of evidentialism or merely its reformation.
I argue that the future of evidentialism depends on the viability of the proportionality principle. Moreover, against earlier anti-evidentialist arguments I hold that the current drift is in the direction of an evidentialism expanded beyond its Greek and Enlightenment versions.Footnote 7 In that case, the post-strong-foundational revolution in epistemology is not the anti-evidentialist movement it was first thought to be, but rather the rebirth of evidentialism in a post-positivist, liberalized form.Footnote 8
In the following sections I re-examine the notion of evidentialism both (a) for the ambiguities that have marred previous discussions, and (b) for its prospects when expanded to include religious experience. I will focus primarily on the work of two major defenders of a post-traditional religious epistemology, Alvin Plantinga and William Alston.
The collapse of strong foundationalism
Contemporary discussions of evidentialism suffer from several equivocations, including the conflation of evidentialism with strong foundationalism. Even if it were true that some forms of evidentialism have been influenced by strong foundational assumptions, it does not follow that evidentialism must share the assumptions of strong foundationalism. To see why, let us consider some widespread objections to strong foundationalism.
Pre-modern foundationalism descends from the Greeks through the scholastics.Footnote 9 Its central idea was that weaker beliefs ought to be grounded in more certain beliefs. Ideally, science would be an organized structure of beliefs in which as many beliefs as possible would be deduced from self-evident foundational principles. Epistemic authority would descend from the most certain to less certain beliefs. Three lines of criticism discredited this theory of knowledge:
The ideal of certainty: Classical epistemology held that we can have certain knowledge of necessary, self-evident truths.
Contemporary objections: The rise of empirical epistemology has challenged the ideal of certainty with the belief that most knowledge is at best probable. Though some beliefs are more certain than others, today's epistemic ideal is not foundational certainty, but rather a coherent and well-confirmed network of beliefs.
The range of basic beliefs: Since epistemic foundations should be the strongest possible beliefs, only self-evident beliefs are truly foundational, i.e., properly basic beliefs.
Contemporary objections: A number of current epistemological views are inclined to admit as properly basic not only (a) logically and sense-perceptually self-evident beliefs, but also (b) more complex beliefs that seems to be directly produced by truth-conducive cognitive processes, as well as (c) necessary presuppositions of beliefs that are already firmly believed. Thus, today's epistemologists often accept as properly basic not only self-evident truths, but also properly functional beliefs and logically presupposed beliefs.
Derivation by deduction: One ought to deduce as many beliefs as possible from certain beliefs, thereby preserving certainty as far as possible.
Contemporary objections: For epistemic holists of several stripes, the basic unit of knowledge is not the single belief, but the coherent, well-confirmed system of beliefs. In that case, the epistemic authority of all beliefs depends on their coherence with each other (notwithstanding case by case differences in degrees of justification, warrant, or logical priority). Though deductive links may guarantee some internal consistency, the justification of the whole is not deduced, but mutually and collectively implied.Footnote 10
In short, the traditional emphasis on the certainty of foundational beliefs and deductive dependence between beliefs has given way to a wider category of basic beliefs and a more complex appreciation of how epistemic authority derives from belief systems as wholes rather than from privileged beliefs.Footnote 11
For Reformed epistemologists, contemporary epistemology's critique of strong foundationalism supports their contention that faith can be justified without argument. In their view, strong foundationalism wrongly requires argumentative support for all religious beliefs that are not self-evident truths. On the contrary, they argue, it is presumptuous to demand that principles of faith be justified by reasoning. As Plantinga paraphrased Herman Bavinck, the Christian believer ‘should start from belief in God rather than from the premises of some argument whose conclusion is that God exists’ (1983, 65).
Rather than hold faith hostage to philosophical arguments, Reformed epistemologists appeal to religious perceptions (e.g. Calvin's sensus divinitatis) for properly basic beliefs, just as Thomas Reid appealed to natural modalities of knowledge (memory, intuition, perception) as rationally adequate grounds for common-sense beliefs. If persons naturally have a divine sense, their beliefs in God may be properly caused, yet not be properly basic (self-evident) by strong foundational standards. But if those beliefs are properly caused, Reformers argue, they should be properly basic, i.e. rationally foundational for religious belief. Thus, Reformers welcome the recent extension of the range of properly basic beliefs.
Granted the demise of the strong foundational paradigm, however, it does not follow that evidentialism must share a similar fate. That non sequitur becomes clear when we examine evidentialism more closely.
The ambiguities of evidentialism
Historical approaches to modern evidentialism often focus on Locke.Footnote 12 In Wolterstorff's account, Locke argued for a doxastic ideal based on three desiderata: (1) seek evidence for the truth or falsehood of belief; (2) assess the logical implications of evidence in terms of probability of truth; (3) adopt a level of epistemic commitment (‘confidence’) in one's belief in proportion to the strength or probability of the evidence.Footnote 13 Let us call these the evidence, assessment, and proportionality principles. Notably, these principles are not the offending principles of strong foundationalism: none of evidentialism's three principles requires self-evidence, certainty, or deductive necessity. So why should anyone assume the liability of evidentialism to strong foundationalism's demise?
If evidentialism has been incorrectly associated with strong foundationalism, it may be due to confusion regarding (a) what counts as evidence, and regarding (b) the strength of evidence that is required for belief to be rational. These points must be clarified before the prospects of evidentialism can be assessed. I will consider first the strength of evidence issue.
By most accounts, evidentialism ties the rationality of belief to strength of evidence. Here again is Forrest's summary of the contemporary view:
Contemporary epistemology of religion may conveniently be treated as a debate over whether evidentialism applies to the belief-component of religious faith, or whether we should instead adopt a more permissive epistemology. Here by ‘evidentialism’ I mean the initially plausible position that a belief is justified only if ‘it is proportioned to the evidence’.
Straightforward as this idea seems, however, it may be vulnerable to a strong foundational misreading. According to Forrest's explanation:
Evidentialism implies that it is not justified to have a full religious belief unless there is conclusive evidence for it. It follows that if the known arguments for there being a God, including any arguments from religious experience, are at best probable ones, no one would be justified in having full belief that there is a God. And the same holds for other religious beliefs, such as the Christian belief that Jesus was God incarnate. Likewise, it would not be justified to believe even with less than full confidence if there is not a balance of evidence for belief.Footnote 14
If not read carefully, Forrest's references to ‘conclusive evidence’ and ‘full belief’ and to the inadequacy of merely ‘probable’ belief might seem to imply the obsolete strong foundational standard of certainty (a criticism invited by some extreme evidentialists). But subsequent sentences show that the issue is proportionality, not certainty. What offends is the holding of full belief with only probable evidence. The proportionality principle is flexible enough to allow for less than full belief with less than certain evidence, but not full belief with less than certain evidence.
In short, the proportionality principle does not require a particular level of evidence or belief, but only that the two coincide. From here on, therefore, I will identify evidentialism with the proportionality principle rather than with the requirement of certainty for rational belief. But if evidentialism does not require any particular level of evidential strength, one may wonder how it became associated with the demise of strong foundationalism. That link has much to do with how one defines evidence. That point is best approached through the issue of what is properly basic.
Evidentialism, strong foundationalism, and proper basicality
The recent association of evidentialism and strong foundationalism appears to have been abetted by the debate over proper basicality. Here is one possible reconstruction:
(1) Strong foundationalism requires self-evident or incorrigible foundations.
(2) Since basic religious beliefs are not self-evident or incorrigible, they require rational justification, i.e. evidence and argument (according to strong foundationalism).
(3) But religious experience can directly warrant basic religious beliefs (God is comforting me, God disapproves of my actions), rendering them rational though neither self-evident nor incorrigible.
(4) Ergo, directly warranted beliefs are rational without evidence or argument.
Seen this way, the collapse of strong foundationalism might appear to take evidentialism (and the demand for proportionate evidence) with it.
Still, an evidentialist may rightfully object: granted that strong foundationalism's range of basic beliefs is too narrow, why would not the expansion of proper basicality to include religious perceptions constitute an expansion of the range of justifying evidences for true belief? In other words, if we are willing to revise what counts as properly basic, why not revise what counts as evidence? If we open proper basicality to religious experience, why does that not open evidentialism to religious experience? It seems that the status of post-strong-foundational evidentialism depends on what you mean by evidence.
Contrary to the broad view of evidence, anti-evidentialist arguments often take a narrowly propositional view of evidence as exemplified by Plantinga's definitions of classical foundationalism and evidentialism. In his view, classical foundationalism involves ‘believing one proposition on the evidential basis of others’ (2000, 82, original italics). Similarly, his definition of evidentialism involves ‘the view that belief in God is rationally justifiable or acceptable only if there is good evidence for it, where good evidence would be arguments from other propositions that one knows’ (2000, 70, original italics). For Plantinga, evidentialism and traditional foundationalism are both guilty of holding rational belief hostage to propositions or arguments.Footnote 15
Evidence vs grounds
If the Reformed epistemologists' dismissal of evidence strikes the uninitiated reader as a mystery, the Reformed definition of evidence is the key to that mystery. The common reader is mystified because she probably assumes that ‘belief without evidence is rational’ means ‘belief without grounds is rational’ (i.e. rational without support of any sort). But anti-evidentialists who define evidence propositionally do not count as evidence non-propositional grounds that many would count as evidence. In particular, if evidence is propositional and if religious experience is perceptual rather than propositional, religious experience would not count as evidence by the anti-evidentialist's propositional definition. According to the evidence-is-propositional view, to believe that one has sensed the presence of God is to believe in God without evidence, in which case it can be ‘entirely right, rational, reasonable, and proper to believe in God without any evidence or argument at all’.
Clearly, the distinction between experiential grounds and propositional evidence is crucial to the anti-evidentialist's case. Rather than seeing experience as a kind of evidence, the Reformed view contrasts experiential grounds and evidential grounds. As Plantinga explains, beliefs such as ‘I see a tree’, ‘I had breakfast this morning’, and ‘[t]hat person is in pain’ are grounded in experience, not in inferences or arguments from evidence, and therefore are not grounded in evidence by the definition of evidence-as-argument. Likewise, ‘belief in God resembles belief in the past, in the existence of other persons, and in the existence of material objects’ (1983, 17), insofar as many religious beliefs are not grounded in arguments, but in mental states that directly cause theistic beliefs.Footnote 16
How much weight should be put on the experience/argument distinction? To be sure, the distinction has both philosophical and theological substance. In particular, the grounds/evidence distinction not only instantiates the causes/reasons distinction dear to philosophers, but also the grace/effort distinction dear to theologians. In the former case, the causes/reasons distinction explains how naïve believers can know God without knowing how or why they know – they may sense the presence of God without understanding its underlying mechanisms, just as we perceive objects without understanding the underlying mechanisms. In the latter case, the sensus divinitatis is a gift of grace rather a product of effort, being caused by divine action rather than by argument. Whereas traditional epistemological theories of justification require reflective knowledge of the rational basis of belief, Plantinga's warrant model warrants any belief produced by any proper mode of epistemic access (perception, memory, logical inference, etc.) that is functioning in a healthy way under propitious circumstances regardless of the knower's epistemic self-consciousness or lack of it.Footnote 17 For anyone who believes that a divine sense is a rational ground for religious belief, the distinction between divine causation and rational argument would be fundamental. In order to avoid a too narrow notion of rational religious belief, one must distinguish caused rational beliefs from argued rational beliefs.
But even granting the difference between caused and argued beliefs, if the warrant model reforms the notion of basicality, why should it not reform the notion of evidence to coincide with common usage? Why not say that the sensus divinitatis opens the believer to divine evidence? This question is crucial for the fate of evidentialism since the adequacy of the proportionality principle depends on how one draws the boundaries of evidence.
What is evidence?
Resolving the meaning of ‘evidence’ would be simple if there were a standard conception of evidence. Unfortunately, there is not. According to Thomas Kelly,Footnote 18 there is a wide difference between the quasi-legal common sense idea of evidence as objects (exhibit A) and the plethora of philosophers' special meanings of evidence, including sense data as mental items of one's present consciousness (Russell); the stimulation of one's sensory receptors (Quine); observation statements or protocol sentences (logical positivists); the totality of propositions that one knows (Williamson); the occurrent thoughts that one is having at a given time (Conee and Feldman); those beliefs of which one is psychologically certain (Baynesianism).
At first, this embarrassment of riches might suggest that choosing a definition of evidence is just a matter of convenience. But even if the definition of evidence were somewhat discretionary, the logic of one's choice ought to be consistent with all of one's epistemological commitments. On that point, some anti-evidentialist arguments are inconsistent. On the one hand, anti-evidentialism uses a narrow definition of evidence to make the case for rational belief ‘without evidence or argument’ while on the other hand appealing to a reliabilist warrant for religious experience although reliabilism typically takes a wider view of evidence.
The anti-evidentialists' inconsistency on evidence (and its pragmatic equivalents) can be highlighted through the contrast between internalist and externalist theories of rational belief. Internalist theories of knowledge ground rational belief in the contents of consciousness (e.g. sense data, the coherence of available beliefs, propositions, etc.). Externalist theories include as rational grounds transpersonal processes of belief-production whose mechanisms the knower may not understand or be wholly conscious of (memory, perception, peer-reviewed research, etc.). Proponents of each approach tend to define evidence to fit their model: whereas some internalists treat evidence as propositions (contents of consciousness),Footnote 19 externalists typically take the more common approach that includes as evidence external objects or processes (causes of beliefs).Footnote 20 Given that the warrant model supplies externalist reasons for warranting beliefs – i.e. a sense of divinity – it would seem arbitrary, curious, or perhaps self-contradictory for Reformers to restrict evidence to the internalist's definition of evidence-as-argument. Likewise, the evidentialist would be prompted to assert: if propositional standards of rationality are to be faulted for excluding reliabilist standards of reasonable belief, then Reformers can be faulted for excluding reliabilist standards of evidence.
Significantly, between 1983 to 1993 Alston, Feldman and Connee defended a notion of evidence inclusive of warranted epistemic functions, prompting Plantinga to concede a wider construal of evidence, i.e. allowing that warranting grounds of belief could be counted as evidence.Footnote 21 In that case, however, defenders of the warrant theory cannot fault evidentialism for tying rationality to evidence if the warrant model does the same thing in another key. When taken to its logical conclusion, in other words, the warrant theory of rational belief does not render evidentialism's principles of evidence and assessment dispensable; rather, it expands the notion of evidence and assessment to include the sort of unreflective but truth-conducive grounds typified by healthy epistemic functions.
In short, the Reformist's argument against a narrow notion of rationality also works against a narrow notion of evidence. Likewise, evidentialism's appeal to evidence is expanded rather than mitigated by the warrant model. Today Reformers enjoy a wider notion of basicality, but evidentialism too enjoys a wider notion of evidence. In that case, the reformed Reformed position cannot be that belief can be rational without evidence, but only that belief can be rational without argument.
Thus, the semantics of evidence helps clarify the distinction between traditional reflective justification and experiential warrants by proper functions. But these definitional issues do not clarify the authority of the assessment and proportionality principles in the wake of strong foundationalism. Granted that religious experience warrants religious belief, does rationality require assessing such experiences and adjusting our beliefs accordingly? If so, then the proportionality principle has not been superseded by, but assimilated to the warrant model of rational belief.
Warrant and proportionality
Having discovered an expanded definition of evidence, it is worth noting the perhaps unexpected point that the definition of evidence is less essential to the spirit of evidentialism than the proportionality principle. Even if one were to maintain a distinction between evidence (arguments) and grounds (reliable belief-producing causes), evidentialists could preserve the essence of their view of rational belief simply by revising their proportionality ideal to read:
A belief is rational for a person iff the strength of that belief is proportionate to its evidence or to the reliability of the process that produced it.
Likewise, whether or not reliable cognitive processes are called evidence, the warrant model preserves the spirit of the proportionality principle insofar as the degree of warrant depends on the degree of reliability of epistemic access. The epistemic link between reliability and warrant is the rational equivalent to the epistemic link between evidence and justification. Notably, Reformed epistemology has variously acknowledged this extension of the proportionality principle to cover warranted religious belief. Alston (Reference Alston1993), for instance, rationalizes belief on the basis of the reliability of spiritual perception. Like Plantinga's notion of rational belief, Alston's also appeals to religious experience as a reliable mode of epistemic access.Footnote 22 On the other hand, though Plantinga and Alston imply the relevance of the proportionality principle under the rubric of warranted religious belief, their approaches are quite different.
Though Plantinga's work on other minds and on basicality invites a reliabilist interpretation of religious experience, he does not emphasize the probability of such beliefs, but only the possibility that such beliefs are rational.Footnote 23 Without assessments of probability, Plantinga's defence consists mainly of showing that religious beliefs of a certain type have not been discredited in principle by various sceptical arguments. For that reason, some charge Plantinga's Christian apologetic with having side-stepped the problems of assessment and proportionality by establishing only the absence of defeaters.Footnote 24
Still, even if Plantinga's work is often a negative apologetic, the warrant model is not neutral regarding the proportionality ideal; rather, it translates the logic of proportionality into the terms of reliability, as when Plantinga acknowledges degrees of warrant. On that point, however, I hold that Plantinga misstates his model's implications when he says: ‘the more firmly S believes B the more warrant B has for S’ (1993b 18; 2000, 456). As with his older definition of evidence, this subjective principle opts for an internalist standard rather than the externalist standard that reliabilism is usually taken to imply. Under the typical view of reliabilism as a transpersonal process, degree of warrant should not depend on the subjective standard of firmness of belief – which could be caused by unreliable promptings, wishful thinking, or perhaps by acts of will – but on the transpersonal credibility of the mode of epistemic access.Footnote 25
By contrast to the firmness of belief standard, Alston (Reference Alston1993) defends the transpersonal reliability of Christian spiritual perceptions based on established practices. For instance, by analogy with our learned ability to recognize well-formed and ill-formed language uses without yet knowing the rules, Alston notes: ‘In the same way one picks up ways of recognizing God and His activities, and criteria for separating veridical perception of God from counterfeits, without any of this every being explicitly formulated’ (1993, 200). That is, we may pre-reflectively develop a sense of warranted (i.e. reliable) religious beliefs without being conscious of their justifications.
Furthermore, there are traditions of spiritual expertise with criteria for assessing spiritual experiences.Footnote 26 These criteria include intellectual content (sound judgments, common moral insights, profound themes) and effects on the experiencer (e.g. serenity, charity, patience, simplicity, piety, etc.; ibid., 203). In the interests of full disclosure, Alston also acknowledges the limited reliability of spiritual perception relative to sensory perception: he concedes to doubters that by comparison with sensory perception, ‘mystical experience is usually but dim, meager, and obscure. The net effect of these differences is to render [spiritual perceptions] much less useful as a source of information, even if its epistemic credentials are in order’ (ibid., 208). Also, divine behaviour lacks the predictable regularity of physical processes, even if Providence implies a divine order (ibid., 218). Thus, he concludes, Christian mystical perception/practice ‘undoubtedly generates more inconsistencies than sensory perception, or rational intuition, or memory, or any of the other basic secular practices’ (ibid., 236). So modes of knowing are not all-or-nothing warrants, but vary by degree according to the strength and consistency of evidence.
Notwithstanding the inferior reliability of religious experience, Alston does not discard the ideal of proportionality in favour of blind faith or merely subjective feelings. Having acknowledged the need for sorting the good evidence from the bad, Alston opts for a satisficing defence (i.e. arguing for sufficiency rather than conclusiveness) of spiritual experience as evidence rather than attempting a full-bore proportional approach that would quantify the typical probability of traditional spiritual practices:
Taking all this into account, I think we can say in good conscience that the incidence of inconsistencies in perceptual beliefs stemming from [Christian mystical practices] that are not resolved by [our critical standards] are not so numerous or so central as to override the prima facie claim to rational acceptance that [Christian mystical practice] enjoys by virtue of being a socially established practice …. Again, I can't say anything very definitive, both because we have no usable metric for degrees of reliability … and because there is no determinate answer to how much reliability is required for rational participation … If we had sufficient reason to judge that the degree of reliability is quite low, say 50–50 or less, that would show that the practice is not rationally engaged in. But I see no grounds for any such judgment. (238)
Though somewhat muted, Alston's assessment acknowledges the link between the rationality of belief and the reliability of its mode of access, thereby honouring evidentialism's proportionality principle under the rubric of warranted cognitive practices.
Sanctifying evidentialism
The foregoing considerations interpret Forrest's via media between fideism and Enlightenment evidentialism in favour of a reformed evidentialism. Whether the term ‘evidence’ is taken narrowly or broadly, the warrant model translates the proportionality ideal into reliabilist terms, rendering it as relevant to religious perceptions as to sound arguments. Since the proportionality ideal is the crux of evidentialism, evidentialism is expanded rather than discredited.
Still, we should exaggerate the triumph of evidentialism if we overlooked both the flexibility and contingency of the proportionality ideal. Rationality comes in degrees, just as warrant and justification do.Footnote 27 The gradation of rationality proves crucial for questions of epistemic duty that have flavoured anti-evidentialist critiques, especially in regard to the spontaneous rationality of lay believers.
Typically, the question of rationality has too often been posed as a true/false question. On the one hand, extreme evidentialists have said that lay believers without conclusive evidence are irrational;Footnote 28 on the other hand, defenders of lay believers have countered by saying that unreflective beliefs may be fully rational, even when uninformed by philosophical standards.Footnote 29 Both views turn a gradient into a dichotomy. To the contrary, the rationality gradient suggests that spontaneous beliefs, though somewhat rational, are not ideally so. This gradation complicates the orientation of epistemic duties toward the proportionality ideal by implying the need for an epistemological division of labour. If the warrant model rationalizes lay believers in the absence of elaborate analytic justifications, that exoneration does not apply to the project of epistemology, which seeks a more ideal rationality.
The rational priority of the epistemic point of view qualifies the rationality of lay belief in a twofold way.
First, the rationality of spontaneous beliefs is ultimately contingent upon the reliability of the religious practices. The qualified rationality of believing without reflective knowledge means that such beliefs may be rational enough, given limited information. But that is hardly an ideal rationality. Far from idealizing spontaneous belief as the paradigm of rationality, reliabilism subordinates the rationality of spontaneous belief to the higher, justified knowledge of reliable practices, a knowledge that recognizes the warrants for spontaneous beliefs, a recognition that is needed to avoid error.
Second, it is the duty of religious epistemologists to discover and describe ideal rationality. Reliabilism rationalizes unreflective basic beliefs only for truth-conducive practices or processes, and we depend on epistemologists (or experts in religious practice) to assess those practices. But such a warranting of spontaneous beliefs constitutes evidential justification in the old, reflective sense, if only for epistemologists. If we can rationalize lay believers for trusting their instincts rather than theoretical analysis, we cannot rationalize instinctive epistemology as a final goal, as epistemology seeks truth and avoids error according to a higher, reflective standard.
This graded view of rationality acknowledges both the pragmatic mitigation of the proportionality ideal for lay believers and the priority of the proportionality ideal for epistemology. If lay believers are somewhat exempt from the proportionality ideal, it is because ‘should implies can’: the lay believers' obligation to the proportionality ideal is mitigated insofar as their limited knowledge of justification and warrants does not allow for fine-tuned proportionality. This is not to exempt lay believers from epistemic humility, but rather to doubt their capacity for an ultimately informed proportionality. By contrast, the validity of the proportionality ideal is presupposed by the very project of religious epistemology, which assumes the duty of knowing which beliefs are credible and which are not in order, (a) to find truth and avoid error, and (b) adjust belief accordingly. Far from being an alternative to naïve belief, proportionality is the truth-seeking ideal of rational belief at which all belief implicitly aims, and which epistemology aims to empower.
In light of these considerations, it seems that post-strong-foundationalism's call for a new conception of rationality is answered by a reformed evidentialism. Just as a post-Enlightenment foundationalism is reformed by more modest, holistic views of basic beliefs, so evidentialism survives its Enlightenment narrowness in a form congenial to religious experience. But while this reformed evidentialism avoids the Scylla of Enlightenment evidentialism, it only avoids the Charybdis of fideism if religious epistemology acknowledges its duty to justify all the grounds of faith.
