Abstract: William Alston’s Perceiving God argues that some mystical experiences should be regarded as perceptions of God analogous to the perception of physical objects in sense experience. On Alston’s account, beliefs about God derived from mystical experiences are prima facie justified and thus rational in the absence of reasons to believe that apparent mystical perceptions are unreliable. This essay questions the soundness of the analogy between mystical and sense experience and of Alston’s epistemic commitments. I conclude that there are several reasons for doubting that mystical experience generally—or Christian mystical experience specifically—can be a form of perception, even given Alston’s epistemic commitments.
William Alston’s Perceiving God contends that some people have mystical experiences that can constitute direct, nonsensory perceptions of God. These involve a simple presentation of some object of mystical perception to a subject. The subject takes this object to be God. It is the presentational character of the mystical experience that Alston thinks qualifies it to be considered a form of perception analogous to sense perception in some ways but not in others. Beliefs formed about God on the basis of the presentational character of mystical experiences are labeled as manifestation beliefs or M-beliefs.
Alston argues that M-beliefs are prima facie justified because the objects of perception are presented to the subject as such-and-such. For example, on Alston’s ‘theory of appearing’ a subject’s perception of the color of his house (or any other property of a given object of perception) is directly determined by the way the house appears to him. This perception is irreducible in the sense that the presentational properties of perceptual objects are not due to conceptualization, that is, to Kantian categories of thought imposing structure on a structureless perceptual experience. In order to determine how we assess the reliability of perceptual beliefs, Alston asks what reasons we have for believing that sense perception—the least controversial example of (what everyone takes to be) a reliable means of forming perceptual beliefs—is reliable. He concludes that we can find no good reasons for supposing that sense perception produces reliable beliefs about the physical world because arguments for the reliability of sense perception are all infected with what he calls ‘epistemic circularity’; that is, they justify their conclusion—that sense perception is indeed reliable—by using arguments whose premises are based on what one has learned through sense perception. For example, in order to show that sense perception is reliable because it makes predictions that do come to pass one would have to rely on sense perception to determine that those predictions were indeed fulfilled. If epistemic circularity infects beliefs formed on the basis of something as uncontroversial as sense perception then it cannot be used as an argument against mystical experience qualifying as a form of perception.
Because epistemic circularity also infects other sources of belief—such as memory, introspection, rational intuition, and reasoning—one must seek a practical rather than epistemic justification for forming beliefs. According to Alston’s practical theory of justification, we form and evaluate beliefs by following certain doxastic (belief-forming) practices. A doxastic practice consists of perceptual inputs (experiences), belief outputs (such as “This is a red chair”), and input-output functions which yield particular kinds of beliefs given particular experiences. Perceiving God contends that it is prima facie rational to engage in a doxastic practice insofar as that practice is socially established, does not yield massively inconsistent outputs, and does not generate outputs which come into conflict with the outputs of more socially established (and presumably more reliable) practices. Doxastic practices are to be deemed reliable until proven unreliable. Sense perception, memory, introspection, rational intuition, reasoning, and mystical perception can all be construed as different doxastic practices. Each doxastic practice (except perhaps introspection) has an overrider system where putative (apparent) perceptions are subject to being rejected as unreliable due to reasons to think the putative perceptions are false (rebutters) or because the grounds for the perceptual beliefs are not justified under the given conditions (underminers) (79). The overrider system consists of other background beliefs putative perceptions can be checked against. Background beliefs are primarily supplied by the doxastic practice in question, although beliefs from any relevant doxastic source can serve as reasons for rejecting the reliability of a doxastic practice. Alston labels such reasons for rejection ‘overriders’.
A doxastic practice’s claims are strengthened by what Alston calls significant self-support. Our ability to use sense perception and reasoning to predict and control events in the physical world to a remarkable degree is an example of significant self-support. Determining that sense perception allows us to predict and control requires us to use sense perception to look at the world to see that, in fact, we can predict and control events. This is why attempts to establish the reliability of sense perception are circular, but it is still a significant fact that not all doxastic practices can generate such self-support within their own domain.
‘Sense perceptual practice’ (SP) is a basic doxastic practice. Alston argues that the ‘Christian mystical perceptual practice’ (CMP), a particular form of ‘mystical perceptual practice’ (MP), can be rationally engaged in because prima facie it serves as a reliable basis for forming true beliefs about God. Alston considers reasons for denying that CMP constitutes a full-fledged doxastic practice and reasons for believing that CMP has been overridden because it produces unreliable beliefs. He addresses the results of historical studies, scientific findings, and an extrapolation from scientific findings to metaphysical naturalism as potential overriders of CMP. The most difficult problem for holding that CMP generates reliable beliefs is the diversity of mystical practices from different religious traditions which, when taken together, generate massively inconsistent output beliefs. Given no external reason for accepting CMP outside of CMP itself, there is no rational basis for choosing the mystical practice of one religion over any of the others. At the end of this essay I will discuss Alston’s rather unsatisfying solution to the problem of religious diversity.
The Phenomenology of Mystical Experience
A distinctive feature of Alston’s approach is his extensive use of the actual reports of mystical experiences (mostly from the Christian tradition) as illustrations of the qualitative features of mystical perception. His aim is to show that mystical experiences are not, contrary to popular belief, purely subjective in character. By ‘purely subjective’ I mean completely internally-generated and not representative of some objective state of affairs to which mystical experiences allegedly refer. He attempts to show this by finding phenomenal commonalities between the various reports he cites. Using his own examples, it easy to see how one could conclude that such experiences are inherently purely subjective—e.g., “God was present, though invisible; he fell under no one of my senses, yet my consciousness perceived him” (13). Here we have a report that is so vague that one can conclude little about the nature of mystical experiences beyond the fact that God was not discernible to sense perception yet somehow was perceived. The report is vague because it lacks any indications of the characteristics of ‘what it is like’ to have a mystical experience rather than some other sort of experience. Nevertheless, Alston does manage to find the following basic commonalities to all the accounts of mystical experience he cites: “(A) They report an experiential awareness of God. (B) The awareness is direct. (C) The awareness is reported to be of God” (14). Alston’s aim in finding such commonalities is to show that MP, like SP, is a form of perception; the existence of such commonalities allows Alston to give mystical experience a plausible perceptual form.
These commonalities are so basic, however, that they (save the third one) are also applicable to SP—SP similarly reports a direct experiential awareness of physical objects. Thus, while Alston has indeed found phenomenal commonalities in mystical experience, he has not offered any phenomenal qualities that are specific to MP. This is an important point because for SP we can find phenomenal qualities that are not only specific to it, but are specific to various modalities within SP (e.g., vision, hearing, touch, taste, and smell). A falling tree causes a completely different phenomenal content to be present to one’s visual experience (e.g., colors, shapes, depth, motion, etc.) compared to the auditory phenomenal content it evokes (e.g., volume, pitch, etc.). Thus we are faced with a significant difference between MP and SP which provides us with reasonable grounds for suspecting that mystical experience does not qualify as a form of perception.
Alston does not think that the inability to list the phenomenal qualities of MP constitutes a significant problem for regarding mystical experience as a form of perception. He contends that such an inability is merely a reflection of the state of our knowledge, that we are largely ignorant of the specific causes of mystical experiences (49). We can say generally that God causes mystical perceptions just as we can say that physical objects cause sense perceptions, but we lack the detailed knowledge of the means of causation in the case of MP that is abundant in the case of SP. For example, we can describe the causal processes by which light (a stimulus) strikes the retina (a receptor) and causes a visual experience in SP. But we have no idea what kinds of mystical stimuli might exist or how human beings could detect such stimuli if they exist. Thus our ignorance of the MP phenomenal qualities that may exist is not a fundamental difference for Alston because he thinks it merely reflects the different ways in which knowledge is acquired in the two doxastic practices:
We have discovered quite a bit about the stimulus conditions of various sensory qualities, and we have been able to subject the experience of those qualities to a considerable degree of stimulus control… But nothing like this has happened with respect to the perception of God, nor is it at all likely to. We know nothing of the mechanisms of such perception… Thus we lack the most elementary prerequisites for analyzing divine appearances into the phenomenal elements, cataloging them, associating them intersubjectively with names… and so on (49).
Does Alston’s response really do justice to the issues involved? Certainly phenomenal qualities may exist that we have not extensively categorized on the basis of stimulus-response mechanisms, but it is likely that we could distinguish phenomenal qualities purely on the basis of their phenomenal character rather than on the causal mechanisms which evoke them in the absence of knowledge of such mechanisms.
Alston implicitly assumes that sensory phenomenal qualities are always distinguished on the basis of their causal mechanisms. In the case of distinguishing the various sense modalities, for example, one could point out that long before a physiological understanding of sensory modalities had been developed people were still able to distinguish the traditional ‘five senses’ on a crude causal basis such as knowing that taste sensations are only present when some substance is placed in the mouth or that visual sensations are only present when the eyes are not damaged. The inability to completely eliminate such causal influences does complicate any argument to the effect that individuation of sense modalities can be based solely on phenomenal character, but there are examples of individuation of phenomenal qualities within a sense modality that can be made in the absence of any knowledge of their causes. Within visual perception, the ability to distinguish colors constitutes such an example. People had been able to distinguish between different colors purely on the basis of their phenomenal character long before it was known that colors are caused by different wavelengths of light striking the retina. The fact that Alston cannot provide any examples of phenomenal qualities that are specific to MP gives us strong grounds for believing that such qualities do not exist and thus mystical experience does not qualify as a form of perception.
In response to this criticism, Alston could claim that mystical phenomenal qualities do exist but we lack the vocabulary to describe them or we can only vaguely characterize them in terms of our common sensory language. If that were the case, then nonmystics would not be expected to have a vocabulary that captures mystical phenomenal qualities, but we would expect mystics to have developed some sort of basic vocabulary for mystical phenomenal qualities amongst themselves since they have all had the same kind of experience. This distinguishing mark between nonmystics and mystics is found in other cases of perception: The congenitally blind do not have a common vocabulary for colors but those who see do share a common vocabulary for distinguishing colors. Since Alston presents no evidence for even a basic common vocabulary for mystical phenomenal qualities amongst mystics he provides no grounds for believing that mystical phenomenal qualities exist.
Alston’s Theory of Justification
On Alston’s account there is a fundamental relationship between the justification of a belief and its reliability. For a perceptual belief to be justified, that belief must at least be likely to be true, that is, truth-conducive (69). Similarly, in order to be justified in trusting a source of perceptual beliefs, that source must also be truth-conducive. In general, for a source of perceptual beliefs to be reliable (and thus justified) it must yield true beliefs most of the time. Thus, although one can sometimes be justified in holding certain perceptual beliefs even if those beliefs are in fact false, these cases will constitute exceptions to the rule. Justified putative perceptions will usually be genuine perceptions of objects that exist outside of one’s experience (69).
In order to be justified in holding a belief, according to Alston, it is necessary for that belief to be based on an adequate ground—a ground that maximizes true belief and minimizes false belief (74). A ground that is adequate renders a belief very probably true if the belief was formed on that ground (74). Alston plausibly argues that a belief is justified if there is an adequate ground for that belief even if the person who holds the belief does not know what that ground is. All that is required for justification, then, is that the ground for a belief be adequate in the objective sense that it renders the belief likely to be true (75). Since the grounds for many of our beliefs, such as ‘Nothing can travel faster than the speed of light’, are not clearly spelled out for us, our knowledge of those grounds (or lack of it) is irrelevant to the objective probability of such beliefs being true (75). Alston allows for the possibility that different beliefs will have different degrees of justification, but that in his sense of the term justified belief means “a degree of justification sufficient for rational acceptance” (81).
Alston distinguishes immediate (direct) justification from mediate (indirect) justification. To be mediately justified in a belief is to be justified by reasons; to be immediately justified is to be justified in the absence of reasons. Beliefs formed on the basis of experience, the self-evidence of the truth a proposition, and beliefs about one’s own conscious mental states are candidates for immediately justified beliefs (71). A belief can be simultaneously justified mediately by reasons and immediately by experience rather than falling exclusively under one or the other category. Immediately justified beliefs are to be understood as being prima facie justified only. Perceptual beliefs are at least partly immediately justified and thus perceptual beliefs can be construed as prima facie justified in the absence of sufficient reasons (overriders) for considering the perceptual beliefs to be unreliable. Such reasons can include rebutters, reasons for thinking a perceptual belief is false, and underminers, reasons for thinking that under the special circumstances of the case at hand the grounds for the perceptual belief do not actually justify it, although they would under more usual circumstances (79). Alston’s thesis is that M-beliefs formed in the standard way described above, which includes cases in which the basis is purely experiential and cases where the basis also includes background beliefs the subject brings to the experience, are often justified (81).
It is generally held that beliefs formed on the basis of mystical experience are not justified, whereas beliefs based on sense experience are. Thus mystical experience is often denied to be a source of knowledge about God for lacking features that are evident in sense perception. In order to shore up the status of mystical experience, Alston proceeds to tear down the (apparent) reliability of beliefs formed on the basis of sense perception. His aim is to show that sense perception is no better off than mystical perception is as a source of knowledge. He labels the practice of forming beliefs about the physical world on the basis of sense experience ‘sense-perceptual practice’ (SP) and the practice of forming beliefs about God on the basis of mystical experience ‘mystical perceptual practice’ (MP) (103). Alston argues that although a high degree of reliability is necessary for a belief to be justified, just how high that reliability should be will differ for different doxastic practices. Thus one could expect that MP-based beliefs would be less reliable than SP-based beliefs due to the inherently different natures of the practices (105). Alston concedes that a belief can misrepresent reality in various ways and yet still be a reliable guide to reality; for example, SP-based beliefs attach color to physical objects when in fact colors are really qualities of the light reflected from the objects and not of the objects themselves (105). Thus we should not be too concerned about exact truth in these matters; a doxastic practice should be considered reliable insofar as “it would yield mostly true beliefs in a sufficiently large and varied run of employments in situations of the sorts we typically encounter” (104-105).
In order to learn anything about the physical world we must ultimately rely on sense perception. If we wish to reason to a conclusion on the basis of observed facts we must rely on sense perception to determine what those facts are; if we want to use an instrument to make observations, the ‘calibration’ of those instruments will require us to use sense perception to determine that they are functioning properly; and if a theory about the physical world is to be informative it must make successful predictions that we confirm by using sense perception to determine that those predictions were in fact correct (107). Since sense perception provides us with some of the premises of our arguments for the reliability of sense perception, our arguments already assume the reliability of sense perception by using sense perception to determine those premises (107-8). Arguments for the reliability of sense perception are all infected with epistemic circularity and thus have no force. It is often held that the fact that SP allows us to successfully predict and control events in the natural world establishes the reliability of SP, but the only basis we have for believing that our predictions have been successful or that we do control nature must be established by using SP itself. There is also the evolutionary argument that we are inclined to take SP to be reliable because such an inclination was selected by evolution. That would not have occurred unless trusting SP tended to increase our chances of survival; if our chances of survival were increased then SP must be reliable (109). But this is also epistemically circular because our knowledge of evolutionary processes is based on observation and thus ultimately based on SP. Such track-record and evolutionary arguments can only establish the reliability of SP if SP is in fact reliable. We need independent grounds for determining whether SP is reliable (108).
Since none of our attempts to establish the reliability of SP escape epistemic circularity and none of the other basic doxastic practices can be shown to be non-circularly reliable (e.g., the reliability of memory is established based on premises about memory derived from memory), there can be no epistemic justification for forming beliefs and we must seek a practical theory of justification. Because all doxastic practices are infected with epistemic circularity, “there is no appeal beyond the practices we find firmly established, psychologically and socially… Hence what alternative is there to employing the practices… to which we find ourselves firmly committed”? (149-150). Alston argues that we cannot get out of our doxastic practices and criticize them from some external, objective position. Thus the only basis we have for holding that doxastic practices are reliable is that they are socially established (150-151). Given that sole external criterion for the justification of doxastic practices, we should regard socially established doxastic practices to be rationally engaged in the absence of overriders that disqualify them as unreliable (153). Insofar as a doxastic practice is prima facie justified, its outputs are prima facie justified and thus it is reasonable to take those outputs (generally) to be reliable (183).
Doxastic practices present means of assessing and correcting beliefs formed on the basis of those practices. The input-output function is a principle of prima facie justification for practice-based beliefs. Thus if a person is incapable of determining when such beliefs are overridden, the practice cannot yield prima facie justified beliefs (158). To know that beliefs formed on the basis of doxastic practices are not overridden we must have an extensive knowledge of the relevant subject matter that the beliefs are about—the objective probability of the belief being true, the means by which we can further investigate the issues involved, and the psychological characteristics of the observer who formed the beliefs (158-9).
Doxastic practices depend on other practices (e.g., reason depends on SP for its premises and SP depends on memory) and on the outputs of other practices (159-61). Each doxastic practice has a distinctive mode of justification and distinctive principles describing sufficient conditions for being justified in belief outputs and in the unqualified beliefs that constitute the overrider system (162). Such conditions will differ for different doxastic practices. For example, SP-based beliefs will be justified according to different standards than beliefs formed on the basis of a set of premises (162). Doxastic practices are also subject to change. The overrider system of SP has changed as we have learned more about SP and change is obvious in religious doxastic practices (163-64). A potential overrider of a doxastic practice is the generation of contradictory belief outputs from the practice—i.e., internal inconsistency (170). To some extent all doxastic practices are going to yield contradictory outputs, but these will not override a practice so long as they constitute a small proportion of the output of the practice (171). Another overrider of a doxastic practice is interpractice inconsistency: if two doxastic practices yield contradictory outputs, one of them must be unreliable, and we must assume that the more reliable one we should trust is the one that is more socially established (171).
If a doxastic practice produces significant self-support, as in the case of the ability of SP to predict and control events, this strengthens our justification in taking that practice to be reliable. Because self-support does not escape epistemic circularity it cannot be used to show that a doxastic practice is reliable, but it does increase the prima facie rationality of the practice (174). Thus the practical rationality of a practice is no guarantee of its reliability, its likelihood to produce mostly true beliefs (178). What Alston has concluded is that it is rational to take a doxastic practice like SP and its outputs to be reliable even though he cannot show that such basic doxastic practices or their outputs actually are reliable (178). Alston makes it astonishingly explicit how weak this approach is when it comes to establishing the likelihood that beliefs formed on the basis of a doxastic practice are really true:
Any doxastic practice that is not internally inconsistent can be strongly supported if epistemic circularity is allowed. Since it will agree with itself, it will turn out that it is invariably correct. Hence if the reliability of SP can be strongly supported only with circularity, it cannot be claimed to be superior to any other practice, including MP in this respect (143).
Alston’s concession above weakens his whole doxastic practice approach. It shows that, aside from the single and highly questionable external criterion of social establishment, we have no basis for believing that engaging in a particular doxastic practice (as long as one of them does not contradict itself) is more rational than engaging in any other. Alston’s notion of practical rationality is so weak that it fails to provide us with any means for deciding whether a belief formed on the basis of a doxastic practice is likely to be true in the objective sense he was aiming for in his discussion of the relationship between being justified in holding a belief and there being an adequate ground for that belief.
The Christian Mystical Perceptual Practice (CMP)
There is a fundamental difference between SP and MP. In the case of SP the physical environment is portrayed in basically the same way for everyone, even in widely divergent cultural traditions, thus there is only one form of SP. However, for MP there are as many ‘maps’ of the spiritual realm as there are different religions (188-89). Each of the different forms of MP will employ extremely different overrider systems (189). We cannot create a universal overrider system for all MPs because the doctrines of different religions are mutually exclusive; any universal MP overrider system we tried to create would be wrought with massively inconsistent belief outputs. Alston concentrates his focus on one particular form of MP, CMP, and says that the main purpose of Perceiving God is to show that it is rational to regard CMP as a reliable source of beliefs about God insofar as CMP “is a socially established doxastic practice that is not demonstrably unreliable or otherwise disqualified for rational acceptance” (194). The overrider system of CMP is constituted by the Bible, the ecumenical councils, and Christian tradition in general (193). Our task here is to determine whether CMP qualifies as a full-fledged doxastic practice.
What reasons could we have for doubting that CMP constitutes a doxastic practice? It is often complained that the overrider system of CMP is based on the MP of others; Christian doctrine is justified only if the MP it is based on is justified—but what served as an overrider for that MP? (211). Alston claims the same sort of problem arises for the overrider system of SP, for the SP we engage in is based on the SP of others. Thus, such an objection invokes a double standard by disqualifying CMP for features it shares with a basic doxastic practice like SP that is not disqualified on the same grounds. I will give grounds for believing that such an objection does not actually apply to SP in what follows.
CMP is often considered to be disqualified because there are more abundant overriders for SP than there are for CMP (212). Other people can serve as reliable checks for SP and we can test the perceptual systems of those people to see if they are functioning properly, but in CMP other people cannot serve as reliable checks for beliefs formed on the basis of CMP because we have no means to test the ‘mystical receptivity’ of practitioners (we can say mystics have certain personality traits in common, for example, but not all people who share those traits will engage in CMP or any other MP) (213). In the case of SP, if many other witnesses on a scene say they saw something that contradicts what you believe you saw on the basis of SP, your SP-based belief can be overridden; but MP-based beliefs cannot be overridden by the mystical experiences of others in the same way (214). Alston takes this to amount to the claim that
[S]ince CMP lacks this kind of check by other observers it does not have the kind of overrider system that a doxastic practice must possess if it is to confer prima facie justification on its outputs. Experientially based beliefs must be subject to test by the experience of other observers in the SP way if they are to count as prima facie justified by that experience… Objective epistemic worth requires intersubjective validation (215-16).
This is a very plausible objection because in the absence of intersubjective confirmation of a belief we have no means of distinguishing genuine from delusory perceptions (216). Alston contends, however, that this is simply a case of epistemic imperialism, inappropriately applying the standard of one doxastic practice (SP) to another (CMP) (216). He claims that we shouldn’t expect CMP to use intersubjective confirmation as an overrider of its outputs in order to regard CMP-based beliefs as reliable objective claims.
Is the requirement that beliefs based on experience need to be intersubjectively confirmed in order to be prima facie justified really nothing more than epistemic imperialism? Alston says that it must be because introspective reports of my experience cannot be checked by other people’s experience (216). But introspection is the sole example of a doxastic practice that may not have an overrider system and thus yields unqualifiedly justified beliefs, not prima facie justified beliefs (160). In all the other doxastic practices where an overrider system must be present, there is only prima facie justification insofar as beliefs formed on the basis of those practices can be overridden. My perception of my own mental states is expected to be accessible only to myself through introspection, so in this case a public test appealing to the experience of others would be inappropriate. But God is allegedly accessible to all those who engage in CMP; thus it is completely appropriate to ask for intersubjective confirmation among those who engage in CMP. Without such confirmation what could possibly constitute a sufficient overrider for CMP?
Recall the requirement for practical justification that one must have an extensive knowledge of the subject matter of a doxastic practice. Without such knowledge, one has no ability to assess when beliefs formed on the basis of a doxastic practice have been overridden, and thus beliefs so formed lose their prima facie justified status. For such beliefs to be justified, we need to know objective probabilities and objective means of investigation (158-9). The nature of doxastic practices opens up Alston’s larger project of justifying M-beliefs to serious criticism, for one can ask: How do we acquire knowledge of the relevant metaphysical subject matter of beliefs about God? For example, how can we assess the objective probability of God being a certain way rather than another?
We can suppose that God presents his characteristics to people in mystical perception, but then we can have no background knowledge to constitute the overrider system to check such M-beliefs against. This leaves M-beliefs incapable of being overridden and thus provides no basis for rejecting some M-beliefs as unreliable, with the undesirable result that anything could be supported on the basis of CMP. It is likely that Alston would appeal to revelation or natural theology as the basis of the overrider system for CMP-based beliefs, but as sources of knowledge about God these sources are extremely problematic. Ultimately revelation simply reduces to a form of mystical perceptual belief—a catalogue of other people’s mystical perceptual beliefs. Thus to say that one’s own beliefs formed on the basis of CMP are prima facie justified because they have not been overridden by the revelation component of the background overrider system is simply to say that one’s own MP-based beliefs have not been overridden by others’ MP-based beliefs. But why should we regard others’ MP-based beliefs as prima facie justified? If the MP-based beliefs that constitute the overrider system of CMP are not justified, then neither are the CMP-based beliefs that depend for their prima facie justified status on the justification of that overrider system. Natural theology is no help to us here because a strong case can be made that beliefs about the nature of God are read into the conclusions of arguments for the existence of God rather than being derived from them.
This kind of objection—that the overrider system of CMP in turn depends on some form of MP—doesn’t apply to the overrider system of SP because SP contains within itself the means of obtaining further knowledge (that becomes part of SP’s overrider system) by employing investigative techniques that open up SP-based beliefs to tests of intersubjective confirmation. It is the lack of such methods of investigation in CMP that disqualifies Christian tradition from serving as a sufficient overrider system for CMP. SP-based beliefs can be checked against each other due to the allowance for intersubjective confirmation within SP; but the lack of intersubjective confirmation in CMP prevents M-beliefs from being checked against each other. This gives us strong grounds for doubting that CMP qualifies as a rationally engaged in doxastic practice, for prima facie justification of CMP cannot occur if CMP cannot be overridden. If CMP is to be rationally engaged in, it must be capable of being overridden on other grounds, grounds that come from outside of the insufficient overrider system of CMP itself.
I propose that Alston’s weak criterion of social establishment for the rationality of taking a doxastic practice to be reliable should be replaced with a criterion of intersubjective confirmation. Insofar as a doxastic practice has an overrider system, degree of reliability that can rationally be afforded to that practice should be proportional to the amount of opportunities for intersubjective confirmation within that practice. Only an unqualifiedly justified doxastic practice like introspection is exempt from this criterion because it lacks an overrider system. Thus SP, memory, and reasoning are to be regarded as reliable, but CMP (or any other MP), insofar as it offers no opportunities for such confirmation, should not be deemed as a reliable doxastic practice.
One ground for rejecting a doxastic practice as unreliable is that it yields massively inconsistent belief outputs or outputs that are massively inconsistent with the overrider system of the practice. A doxastic practice can be also be regarded as unreliable if the beliefs that constitute the overrider system for that practice are massively inconsistent with each other (234). Alston confesses that in this regard CMP clearly yields more inconsistencies than the other basic doxastic practices:
Especially when it comes to perceptions of communications from God, we get a wild diversity, many mutually inconsistent. The reports people give as to what God told them about His plans… could not possibly all be true unless God is incredibly confused, vacillating, or inconsistent Himself. Moreover, in many cases pairs of alleged communications are such as to envince radically incompatible divine attributes (236).
Given such inconsistencies, can one rationally engage in CMP? Well, the amount of ‘acceptable’ inconsistency will vary with the doxastic practice in question because in some practices we will have a better understanding of the subject matter than in others (237). Alston concludes that the amount (and importance) of the inconsistencies generated by CMP is not great enough to override the prima facie rational status of CMP. However, he does admit that the degree of reliability we can rationally attach to CMP is less than the degree of reliability we can attach to SP and other basic doxastic practices (238).
SP exhibits significant self-support insofar as: (1) we can make successful predictions and thereby control events by using SP; and (2) we can establish facts about how SP works that indicate that SP is generally reliable, why it is reliable, and the special circumstances under which SP is less reliable (250). These features are significant because not all doxastic practices can exhibit such self-support; this increases our confidence in the reliability of SP (250). Perceiving God maintains that CMP also exhibits significant self-support in that CMP provides spiritual guidance by sanctification, the transformation of an individual into a spiritual person. For example, CMP is said to cause people to be less concerned with material possessions and make them more likely to be charitable with their time and money (250-51). However, CMP does not exhibit self-support in any way that is analogous to the self-support SP exhibits. This is perhaps best illustrated by the fact that other MPs incompatible with CMP also exhibit sanctification; therefore sanctification cannot increase our confidence in the reliability of any MP since all of them cannot be equally likely to be true given their incompatible outputs and overrider systems.
Conflict Between the Christian Mystical Practice and Science
CMP can also be deemed unreliable if its output or overrider system conflicts with the output or overrider systems of other more socially established (and presumably more reliable) doxastic practices like SP, science, or historical studies. By ‘more socially established’ I do not mean understood and accepted by a greater number of people—clearly religious traditions that have existed for thousands of years have more social influence than the centuries old sciences. The results of scientific and historical studies are more socially established in the sense (that I believe Alston uses) that more rational people would accept those findings as ideologically-neutral grounds for establishing reliable beliefs about the world than the number of rational people who would accept the belief outputs of any particular MP as grounds for reliable belief.
In order for any such conflict to disqualify CMP from being rationally engaged in there must be massive inconsistency between CMP and the other practices (238). Conflicts between fundamentalist Christian doctrines such as creationism and science will only provide reasons for regarding fundamentalist forms of Christianity as unreliable (239). In order to determine that CMP is unreliable we will need to find cases where central Christian doctrines come into conflict with scientific findings (240). Thus it is appropriate to ask which doctrines we should regard as central to Christianity in the sense that denying these doctrines would result in a religion that doesn’t look like Christianity to most Christians. Michael Martin has listed the central doctrines of Christianity as follows: (1) God exists; (2) God is three persons in one; (3) Jesus is the Son of God; (4) Jesus existed as a historical person; (5) Jesus was born of a virgin; (6) Jesus was crucified by the order of Pontius Pilate; (7) Jesus was resurrected from the dead three days after his death; (8) the only way to salvation is through faith in Jesus; (9) Jesus will return to this world and there will be a resurrection of the dead (Martin 10-12).
Now we can proceed to ask whether science conflicts with the central tenets of Christianity. It seems that none of the tenets listed above are amenable to scientific analysis. We could, in theory, produce Jesus’ corpse to determine if he was crucified—a belief capable of being overridden if he was hanged or died of natural causes. Or we could perform a DNA test on Jesus to determine if Joseph was his father, thus overriding the belief in his virgin birth. But of course none of these options are really available to us this late in history. Thus we must agree with Alston that science does not come into conflict with any of the central tenets of the Christian faith (although science clearly does conflict with some Christian beliefs) and thus CMP is not overridden by any scientific findings. Even conceding the possibility that such central beliefs could come into conflict with science, Christianity could always modify its central tenets (as it has done numerous times in the past) to fit the results of scientific research and thus remove any source of conflict. There is always the possibility of methodological conflict insofar as scientific reasoning requires reliance on publicly-available evidence when forming beliefs and Christianity requires faith in revelation regardless of the evidence, but Alston maintains that any such apparent conflict is illusory because science and Christianity deal with different subject matters, the physical world and the spiritual realm respectively (242). The situation is not as clear-cut as Alston would have us believe because the subject matter of Christianity includes historical events and thus can come into conflict with historical evidence derived from the physical world.
CMP clearly does come into conflict with an extrapolation from the findings of science to naturalism, the metaphysical thesis that everything that exists is natural and thus does not violate invariant laws of nature. Naturalism, after all, is committed to denying the existence of supernatural phenomena that Christianity embraces, such as the existence of a supernatural god that intervenes in the natural order by producing miracles. CMP can be regarded as overridden by naturalism only if we have good grounds for holding that naturalism is true. What are such grounds? The plausibility of naturalism lies in the impressive achievements of science in successfully explaining processes in the physical world with a remarkable degree of accuracy that is unprecedented in history and which competing metaphysical views cannot even come close to providing (247). Alston’s complaint is that the success of naturalistic explanations can only claim hold to the relevant subject matter of science, which is the physical world; spiritual phenomena fall outside the purview of science (247). Thus naturalism is nothing more than a working assumption of science (243). We have good reasons for holding that naturalism is true only if we have no basis for believing that there is a spiritual realm, and Alston believes that CMP gives us a basis for such belief (247). However else naturalism may be challenged, I think the points made in this essay show that CMP does not constitute a significant challenge to naturalism. Nevertheless, we must concede that CMP is not disqualified for rational acceptance by naturalism because naturalism is not demonstrably true, but is a metaphysical extrapolation from science. It is an extrapolation since scientific explanations assume for methodological purposes that only natural causes are operative in the physical world; naturalism extrapolates from this methodological assumption the metaphysical thesis that the only causes influencing the physical world are natural causes. Since such a broad metaphysical thesis cannot be proven, it cannot be used to disqualify alternative metaphysical views for rational acceptance.
Conflict Between the Christian Mystical Practice and Historical Studies
Alston argues that, although it is possible for the central tenets of Christianity to come into conflict with the results of historical research, he does not think that such conflict has actually occurred. He admits that there is no independent confirmation from other historical sources, for example, for the account of the resurrection of Jesus given in the Gospels, but this does not show that the Gospel account is false (244). He claims that those historians who reject the resurrection of Jesus as a historical event primarily do so only because they hold a naturalistic presupposition against the possibility of supernatural events (245).
Alston does not provide any historical evidence to support these contentions. He ignores historical findings that put the Christian faith in grave doubt. I will restrict my discussion of these findings primarily to perhaps the most central doctrine of the Christian faith, the Gospel account of the resurrection of Jesus. Because the Gospel account is the only source of information about Jesus’ resurrection that is available to us, we will have to treat the Gospel account as a historical document. Since the claim that Jesus was resurrected from the dead is a historical claim, we cannot argue that belief in his resurrection is simply a matter of faith, and the Gospel account is opened up to an objective historical treatment. How does the account of Jesus’ resurrection fair when subject to such a treatment?
Take the account of the empty tomb, from which it is inferred that Jesus rose from the dead. After Jesus’ death he is placed in a tomb. According to Matthew, Mary and Mary Magdalene noticed a rock in front of the tomb, felt a violent earthquake, and saw an angel come down from the sky and roll back the stone when they arrived just before dawn. In Mark and Luke, the women arrive at dawn and the stone is already rolled back. In Matthew, an angel is sitting on the rock outside the tomb and in Mark a youth is already inside the tomb. In Luke two men are inside. In Matthew the two Marys rush away from the tomb to tell the disciples that the tomb is empty and they meet Jesus on the way. In Mark they rush out from the tomb and say nothing. In Luke, the women report the story to the skeptical disciples but make no mention of meeting Jesus. According to John, only Mary Magdalene came to the tomb. She sees that the stone that had been placed in front of it was moved and informs Simon Peter and the other disciples of her discovery, who then rush to the tomb, implying that they do believe her story. The account in John has no mention of her seeing any angels or youth. In John she does not enter the tomb after her discovery that it is empty, contrary to the account given in Luke and Mark, but does enter it after she returns with the disciples. After leaving the tomb, when the others had already inspected it with her, she then sees two angels and Jesus, though she does not recognize Jesus (Martin 78-80).
There are many other massively inconsistent reports of the resurrection of Jesus in the Gospels. If we take the Gospels as historical evidence for the resurrection of Jesus we are given contradictory evidence and the reliability of the Gospels themselves becomes highly suspect. Since the Gospels are historically unreliable there is no basis for regarding the resurrection of Jesus as an actual historical event. Note that there are no eyewitnesses to the resurrection according to the Gospels. That Christ was resurrected from the dead is inferred from the presence of an empty tomb; but even here the accounts of the circumstances surrounding the empty tomb all contradict each other (Martin 81). Faced with such contradictory evidence, we have no basis for even believing that an empty tomb was present. Eyewitnesses to a car accident can completely disagree about the circumstances of the accident and yet no one doubts that an accident has occurred, but in such a case we have the evidence—a damaged car—right before our eyes that justifies our belief that a car accident has occurred. In the case of the Gospel account, all we have to go on is second or third hand reports of an empty tomb.
Six years of historical analysis of the New Testament by seventy-nine biblical scholars has created intense controversy. These historians—collectively known as the Jesus Seminar—attempted to use the best available methods of critical scholarship to determine which parts of the Gospels are historically accurate and which parts are not. The influence of theological biases on the Seminar’s historical conclusions was minimized by the use of weighted averages when tallying the votes of the historians for each Gospel claim (Funk 36-7). A certain Gospel passage (or part of a passage) would be voted on in terms of four levels of authenticity: the passage would be colored red if the best evidence indicated it was historically reliable; pink if it was deemed probably reliable because it fitted well with other verifiable historical evidence; gray if it was possible but considered unreliable because there was no historical evidence to back it up; and black if it was improbable because it was contrary to other verifiable historical evidence (Funk 36-7). The use of weighted votes in determining the authenticity of a Gospel passage ensures that the collective judgment of the Seminar is a true representation of a consensus among its members.
Given the apparent impartiality of such a procedure, why has the Jesus Seminar caused such an uproar among Christian intellectuals? It has done so because most Christians deem the conclusions drawn by the Seminar as very damaging to the Christian faith. Consider some of the Seminar’s conclusions about the resurrection of Jesus: “Belief in Jesus’ resurrection did not depend on what happened to his body”; “The body of Jesus decayed as do other corpses”; the resurrection “was not an event that could have been recorded by a video camera” (Funk 461-2). These conclusions contradict what the Gospels claim and what many Christians take to be fundamental to Christianity—yet these New Testament scholars have come to a consensus on these points. If the resurrection of Jesus could not have involved a physical resurrection of his body, as these scholars maintain, then his ‘resurrection’ must have involved the appearance of an apparition of him. Furthermore, the alleged apparition would only have been seen by a select few if a video camera could not record it. This suggests that the post-resurrection appearances of Jesus, if they occurred, were likely hallucinatory since they could not be corroborated by any physical evidence or seen by others present when the apparition allegedly appeared.
As if these conclusions were not damaging enough, the Seminar additionally concluded that because the earliest parts of the New Testament lack any accounts of the appearances of the resurrected Jesus after his death, these too are not likely to have really occurred (since these accounts were later added to the original New Testament scripts) (Funk 462). The New Testament sources for the stories of the physical resurrection, the accounts of the empty tomb, and the post-resurrection appearances are not historically reliable. 96% of the Seminar voted that the virgin birth was a later invention and appropriately colored it black (Funk 516). While the crucifixion, death, and burial of Jesus are considered to be real historical events, all the Gospel references to the details of what occurred are placed in black (Funk 562). When all of the Gospel claims the Seminar considered in The Acts of Jesus are taken as a whole, 84% of them are deemed historically unreliable (Funk 1). The Seminar concludes that this is the case because
[T]he New Testament gospels were composed during the last quarter of the first century by third-generation authors on the basis of folk memories preserved in stories that circulated by word of mouth for decades. The oral stories the four evangelists recorded had been shaped, reshaped, augmented, and edited by numerous storytellers for a half century or more before achieving their final written forms (Funk 2).
While the Seminar’s methodological procedures for determining the historical reliability of the Gospels may be questioned, it is unlikely that any methodological flaws critics of the Seminar may uncover could be great enough to account for an 84% historical unreliability assessment. Potential flaws could reduce that figure somewhat, but it is highly unlikely that they could change the assessment enough to render the Gospels historically reliable. I find it incredible that New Testament scholars could be so incompetent that their assessment of historical evidence would be off to such a degree as to result in a complete reversal of their conclusions. Since the best evidence indicates that the Gospels are historically unreliable, the historical aspects of Christianity are unreliable. Because the metaphysical aspects of CMP are not testable, the only means of determining the reliability of CMP—determining the historicity of the Gospel accounts—has indicated that CMP is likely to often produce false beliefs.
The Problem of Religious Diversity
The major world religions contain MPs that are massively incompatible with each other, in terms of both the output beliefs generated by them and their background overrider systems. Alston provides a good characterization of why this constitutes a major problem for the rationality of taking CMP to be a reliable doxastic practice:
Since each form of MP is, to a considerable extent, incompatible with all the others, not more than one such form can be (sufficiently) reliable as a way of forming beliefs about the Ultimate. For if one is reliable, then most of the beliefs that issue from it are true; and hence, because of the incompatibility, a large proportion of beliefs issuing from each of the others will be false; and so none of the others is a reliable practice. Now why should I suppose that CMP is the one that is reliable (if any are)?… [I]f it to be rational for me to take CMP to be reliable, I will have to have sufficient independent reasons for supposing that CMP is reliable, or… more likely to be reliable, than its alternatives. But no such reasons are forthcoming. Hence, it cannot be rational to engage in CMP… [or] in any other particular form of MP (268-9).
The point here is that even if there is a reliable form of MP, we have no means of determining that this MP is the reliable one (269). Alston’s social establishment criterion will not resolve the issue because the major world religions pretty much enjoy the same degree of social establishment (269). (As a side note, I believe this also gives us good grounds for rejecting social establishment as the sole means of determining the reliability of a doxastic practice.)
The naturalistic view is that the best explanation for religious diversity is that the mystical experiences of the practitioners of these religions are solely the products of their minds. For if these mystical experiences were truly representing some objective reality outside of the experiences themselves, there would be more agreement as to what that reality is like (267). Alston disagrees with this assessment, contending instead that religious disagreements arise from our difficulty in understanding the subject matter of mystical experiences (267). This might explain differences in the details of the beliefs generated by different religious doxastic practices, but if Alston’s explanation were correct, wouldn’t we expect at least some agreement beyond the most general belief that there is something out there that transcends nature? The major world religions cannot even agree on the most basic issues, such as what characteristics that something has or what our relationship to it is. Alston couldn’t have made this point any more explicit:
The ways in which [believers in different religions]… think of the objects of their worship (and of what they take to be Ultimate Reality) differ enormously. The differences are as great as that between personal and impersonal, positive and negative, concerned and unconcerned about morality, [etc.] (188-89).
The best explanation for the lack of any agreement between mystical experiences in different religions on even the most basic issues is that mystical experiences do not refer to any real subject matter outside themselves but reduce to nothing more than a specific form of introspection where subjects confuse their own internally-generated mental states with something they take to exist independently of their minds.
Unlike witness accounts of a car accident where all the witnesses engage in the same practice (SP), there are no universal criteria for settling disputes on religious matters for those engaged in competing MPs (271). Alston believes that this fact, rather than showing that it is irrational to take CMP to be reliable, gives us a reason for why we can’t show that CMP is reliable in the face of religious diversity even if CMP is reliable (272). But even if this is so, this fact does not justify engaging in CMP rather than any of the others. Alston cites an example of rationally engaging in a practice without any independent reasons for choosing one practice over another—the unresolved dispute over how best to treat neurosis between behaviorists and psychoanalysts (272-3). Since both exist within the framework of science, there are independent scientific means for determining which approach is more likely to be successful—e.g., statistical analysis of the effectiveness of treatment where both groups could find areas of agreement as to what constitutes effective treatment. Despite Alston’s claim to the contrary, there probably are neutral scientific grounds for assessing therapeutic success.
Alston asks us to suppose that there were cross-cultural diversity in the forms of SP where the various SPs had incompatible conceptual frameworks yet each exhibited significant self-support. Would it be irrational to engage in my culture’s form of SP in the absence of any reason for preferring my form over the others? (273-4). No, because all forms exhibit significant self-support so my form of SP is just as good as any of the others. Thus “the rational thing for a practitioner of [CMP] to do is continue to form Christian M-beliefs… [and] continue to accept… the system of Christian belief” (274).
The difference between the case of multiple SPs and multiple MPs is, aside from the fact that CMP does not really exhibit any form of significant self-support, the various forms of SP would differ only in conceptual framework. That is, it is likely that my SP-based beliefs would be ‘translatable’ into another form of SP using that form’s framework. Output beliefs from each would be (roughly) the same, only construed in different terms. This clearly cannot be the case for multiple MPs given the massive incompatibility between them on the most basic issues. We can see an example of beliefs of this kind in the history of physics: Newtonian physics can be ‘translated’ into relativistic physics at fractions of a percent of the speed of light under conditions of low gravity with the same results despite the fundamental differences in the concepts being employed. Note that I am not saying that the two systems of physics are exactly translatable into one another, or that both are equally reliable in their explanations of physical processes; clearly relativity provides a more reliable explanation of such processes. The point is that some phenomena are just as reliably explained on both accounts despite their fundamentally different conceptual frameworks. Nothing even remotely analogous appears in the comparison of beliefs formed in competing MPs.
There are several good reasons for doubting that William Alston’s Perceiving God has shown that mystical experience in general or Christian mystical experience in particular can be a form of perception: (1) MP doesn’t qualify as a form of perception because mystical experiences lack any phenomenal qualities specific to MP; (2) Alston’s theory of practical rationality leaves us with no basis for holding that engaging in a particular doxastic practice is more rational than engaging in any other as long as the practices in question do not contradict themselves and are socially established; (3) Since there is no way to establish objective facts about metaphysical entities such as God, MPs are not prima facie justified because there is no metaphysical knowledge base that can serve as an overrider system for them; (4) The lack of investigative methods in CMP disqualify Christian tradition from serving as a sufficient overrider system for CMP. This gives us strong grounds for doubting that CMP qualifies as a rationally engaged in doxastic practice, since prima facie justification of CMP cannot occur if CMP cannot be overridden; (5) A better indication than Alston’s inadequate social establishment criterion that a doxastic practice actually is reliable is that it offers a large number of opportunities for intersubjective confirmation. Thus degree of reliability that can rationally be attributed to a practice that has an overrider system should be proportional to the amount of opportunities for intersubjective confirmation within that practice. Because CMP (or any MP) cannot offer any opportunities for intersubjective confirmation, CMP (or any MP) cannot rationally be regarded as a reliable doxastic practice; (6) CMP does not exhibit significant self-support. Sanctification cannot be an indicator of the reliability of any MP since all MPs exhibit sanctification and yet produce incompatible beliefs. Thus not all of them can be regarded as likely to generate true beliefs simply because they exhibit sanctification; (7) CMP has been shown to be unreliable insofar as the central historical claims of Christianity are likely to be false in light of the findings of historical studies; (8) The existence of massively incompatible MPs and the lack of any independent reasons for regarding any one MP as being more likely to be reliable than any of the others presents strong grounds for holding that CMP (or any given MP) is unreliable; (9) The absence of any basic agreement between mystical experiences in different religions indicates that mystical experiences have no subject matter and thus subjects confuse their own internally-generated mental states with something they take to exist independently of their minds.
 Clearly different cultures conceptualize physical objects in different ways. A bushman uninfluenced by Western culture would not ‘see’ the same thing a Westerner ‘sees’ when he looks at a compact disc because the bushman does not have the concept of a CD, although both perceive the same physical object. But it does seem like the perceptual characteristics of the CD—its feel to the touch, its shape and color, the sound it makes when dropped, etc.—would indeed be universal across cultures. I am unaware of any anthropological evidence suggesting any basic difference of this kind in the way physical objects are perceived by people in different cultures. Thus for our purposes it is safe to assume (as Alston does) that there are no multiple forms of SP pending further evidence which might suggest otherwise.
 In his final chapter, Alston includes inference to the best explanation as a source of revelation; e.g., Moses inferred that the sensory presence of a burning bush during his mystical experience was best explained by divine intervention and Jesus’ disciples felt that the presence of Jesus’ empty tomb was best explained by Jesus’ resurrection. Clearly most revelation claims, however, are MP-based claims; thus the small proportion of inference-based claims in revelation does not affect the force of this argument.
 One could argue that scientific and historical evidence are essentially the same since both are empirical. Because Alston demarcates science and history as two different subject areas, I retain his individuation of them.
 The centers of black holes and the the primordial ‘fireball’ from which the Big Bang emerged are thought to be examples of singularities, theoretical entities in cosmology where the known laws of physics break down. Singularities do not constitute examples of supernatural phenomena as I define ‘supernatural’ because they either obey unknown laws of physics or obey no laws of physics at all; but they do not violate any laws of nature as supernatural events purportedly do.
 At this point one could ask what evidence I have for the claim that multiple forms of SP would generate the same beliefs construed in different terms. As there are no examples of multiple forms of SP between cultures today (at least that I am aware of), one cannot compare any actual multiple forms of SP and see how they diverge. So long as Alston is speculating about the hypothetical (rather than actual) possibility of multiple forms of SP, I am free to speculate about how such multiple forms of SP would compare to each other. It is plausible that the hypothetical multiple SPs would yield the same beliefs since they all enjoy the same amount of predictive success. In other words, it is implausible that different belief systems would generate both massively incompatible beliefs and equally successful predictions since predictive success implies that a belief system is reflecting some aspect of the world as it (more or less) actually is.
 Note that there is an important disanalogy between the argument for the rationality of accepting your culture’s form of SP in the face of multiple, equally successful SPs and the rationality of accepting CMP in the face of multiple MPs: Alston wants to deny that competing MPs all exhibit significant self-support.
 In considering the existence of successful theories in physics with different conceptual frameworks I have provided the closest analogy to multiple SPs that does exist. In that case, different theories can yield the same belief outputs despite different conceptual frameworks.
Alston, William P. Perceiving God: The Epistemology of Religious Experience. Cornell University Press, 1991.
Funk, Robert W. and the Jesus Seminar. The Acts of Jesus: The Search for the Authentic Deeds of Jesus. San Francisco: HarperSanFrancisco, 1998.
Martin, Michael. The Case Against Christianity. Philadelphia, PA: Temple University Press, 1991.