Home » Library » Modern Library » Do Mystics See God?

Do Mystics See God?


[This paper is forthcoming in Michael L. Peterson, Ed., Contemporary Debates in the Philosophy of Religion to be published by Blackwell Publishers. Electronically reprinted with Blackwell’s kind permission.]

Ex. 33:20: And [the Lord] said, Thou canst not see my face [panim]: for there shall no man see my face and live.

Gen. 32:30: And Jacob called the name of the place Peniel: for I have seen God face [panim] to face, and my life is preserved.

“There’s more than one way to skin a cat.”

1. A Cautionary Tale

Theistic philosophers have perennially cited mystical experiences—experiences of God—as evidence for God’s existence and for other truths about God. In recent years, the attractiveness of this line of thought has been reflected in its use by a significant number of philosophers.1 But both philosophers and mystics agree that not all mystical experiences can be relied upon; many are the stuff of delusion.2 So they have somehow to be checked out, their bona-fides revealed. But can they be? I will be arguing that (a) they must indeed be cross-checked to serve as good evidence; and that (b) they can’t be—or not nearly well enough to permit pressing them into service as serious support for theism. The need for cross-checking, necessary in any case, is made acute by two facts: the extreme variability of mystical experiences and the doctrines they are recruited to support, and the fact that, especially in the face of this variability, mystical experiences are much more effectively explained naturalistically. Furthermore, our ability adequately to cross-check mystical experiences (hereafter, ME’s), in a way that would reveal the hand of God, is crippled by the fact that theists offer no hypothesis concerning the causal mechanism by means of which God shows Himself to mystics.

Let’s begin with my epigraph. This insightful, if grisly, bit of folk wisdom tells much of our story. Permit me to spell out the dolorous tale. I am greeted by the sight of poor Sylvester, a heap of flayed flesh upon the lawn. I set out to reconstruct the crime. With but the denuded corpse as evidence, the possibilities are multiple. So I must locate other clues. A bloodied knife nearby might have secrets to reveal: suppose the hemoglobin tests out feline. Even better, perhaps I can find an eyewitness or two, discovering through further investigation that they are both sober and honest. I might find fingerprints on the knife. And so on.

In all this, I rely upon my senses to convey evidence. How is this managed? Why, through some causal sequence, a continuation of some of those sequences that converged upon the destruction of poor Sylvester, and that then diverged from there. Light-waves bearing news of cat skin and flesh make their way from the corpus delicti to my “sensory surfaces”, there to be processed in those still and possibly forever mysterious ways into cat-corpse-consciousness. Mysterious or not, what we do know is that cat and conscious episode are related as (partial) cause to (partial) effect. But for there being some suitable causal link between cat and experience, that experience, no matter its intrinsic characteristics, is not a perception of that cat.3

But if the intrinsic content of my experience can be caused in multiple ways (the presence of an actual cat-corpse being but one of these), then how shall I ascertain that my senses do not deceive? The short answer to this importunate and persistent problem, the problem of perception, is: I must cross-check. But we cannot explore the substance of this remark without making two antecedent observations. First, no amount of cross-checking can produce evidence that will satisfy the radical skeptic. I can decide to pinch myself to check that I’m not just dreaming of cats; but of course I might just be dreaming that I’ve pinched myself. Second: because of this, and because our brief is to examine whether putative experiences of God must be cross-checked to carry evidential weight, not to respond to radical skepticism, we shall have to frame our discussion with some care. One could, of course, accept a counsel of despair: neither ordinary sense experience nor mystical experience can form the basis of justified beliefs about external matters. In that event, mystical theistic beliefs are in no worse case, epistemically speaking, than ordinary perceptual beliefs. But that would be because neither set of beliefs could be in any worse shape, so far as justification goes. That sort of ‘pox on both your houses’ skepticism is however not a very interesting position, from the perspective of traditional debates about the warrant for theism. The interesting question is: if we suppose ordinary perceptual beliefs (and we may throw in scientific theory for good measure) to be warrantable by appeal to sense experience, then why shouldn’t theistic beliefs be similarly warrantable by appeal to perceptual experience, whether sensory or mystical?

Here, in a nutshell, is what I shall argue: The problem of perception derives largely from the general truth that any effect—hence a perceptual experience—can be caused in more ways than one. Our strategy for removing this ambiguity is cross-checking. Ultimately, cross-checking involves just collecting more data, which is subject to the same ambiguity. Our implicit reasoning is that the total amount of ambiguity can nevertheless be in this way progressively reduced. The means by which science draws a bead on postulated “unobservable” entities (like electrons) is not in principle or in practice different in kind; it is just more systematic and careful than the humdrum of everyday perceptual judgments. In everyday contexts, cross-checking is informal, and it is so automatic, continuous, and pervasive that, except under duress (e.g., as we try to catch out a magician), it is scarcely noticed. I propose to show how cross-checking works; to argue that it is a mandatory feature of any recruitment of perceptual experience to epistemic ends; and to show that, therefore, it is a requirement that must be met in theistic appeals to mystical experience as evidence for theism. Finally, I shall argue that this requirement has not and probably cannot be met. So, I shall conclude, mystical experience provides hardly any useful support for theism.

2. Cross-checking Explained

So, what is cross-checking, why is it needed, and how does it work? Let ‘cross-checking’ denote all those procedures and strategies we use to settle questions about the causes of something. These include, in particular, (1) using Mill’s methods to pick out causally relevant antecedent conditions, (2) exploiting the fact that events have multiple effects, to “triangulate” the event in question, on the principle that qualitatively different causes will have some differences in their (potential) effects,4 and (3) confirming the existence of causal mechanisms allegedly connecting a cause to its effects (when it is not a proximate cause). These strategies depend upon putting forward hypotheses and testing them by means of diagnostic experiments. I shall discuss mainly tests of type (3), but invoke strategy (2) when considering prophetic revelations as a test of ME’s.

There are various ways in which cross-checking principles can be formally stated. One way to approach type-3 cross-checks is to consider the problem posed by Duhem’s Thesis. We have a hypothesis H (e.g. that the cat was skinned with a knife), on the basis of which we can, with the assistance of auxiliary hypotheses A, infer some observable effect E0. In general, the occurrence of E0 should confirm H, and its non-occurrence disconfirm H. How that goes depends upon how strongly H & A probabilifies E0 (or its negation), and how strongly it or its negation is probabilified by competing hypotheses-cum-auxiliaries.

But, as we know, even when E0 fails to materialize and H & A is thereby disconfirmed, the opprobrium need not fall on H: the falsehood of A may be to blame. Here is where type-3 cross-checking comes in. It comes in two varieties. First, we can run further tests on H, pitting it against its rivals either in repeat performances of the first experiment or, often more tellingly, in different experiments which call upon different auxiliaries and predict different observations. Second, we can check A, now employing it as a hypothesis to be tested.5 Thus, a defender of H in the face of not-E0 might insist that the relevant auxiliary is not A, but A*, where H & A* entails not-E0. Now A and A* are competitors, and we can require an independent “crucial experiment”6 in which they make conflicting predictions, E1 and E1*. But of course, those predictions cannot be made without invoking further auxiliaries—call them A1 and A1*. Clearly, if the experimental outcome is E1, the defender of A* (and hence H) can protect A* by insisting upon modifying his A1*. And then we can play another round. Can this testing-game go on forever, or will the regress eventually run the quarry (the truth-value of H) to ground? One way to capture the radical skeptic’s intuition is by arguing that the cycle of modifications to save the appearances can go on forever. (This is one form of the so-called underdetermination of theory by data.) The other side of the coin is that this way of formulating the problem of skepticism helps us see what sorts of minimal assumptions might head off an infinite regress, thereby making the evidential issue an interesting one. And this is just what we need to see, where the observations in question are the mystic’s experiences, and the hypothesis is theism.

I do not regard it a settled question whether adjustment of auxiliaries in the face of recalcitrant data can go on forever. But even if after-the-fact revising can proceed indefinitely, there is a strong intuition that a system of beliefs which must constantly be revised as new evidence comes in loses plausibility in relation to one that does not. Let us make this anti-skeptical assumption. Evidence, in the face of which a hypothesis can be rescued only by revision of auxiliary assumptions, works to the disadvantage of that hypothesis—though perhaps not decisively so—in comparison to competitors which accommodate that evidence without revisions.7

An obvious objection to all this will be that, plausible as it may be as a rational reconstruction of scientific reasoning, it does not at all capture the process by which we acquire warranted perceptual beliefs. Perceptual knowledge seems much more direct than this, even to those who concede the obvious fact that it is causally mediated. So now want I to argue that this is an illusion, that in fact warrant accrues to perceptual beliefs only insofar as, rationally reconstructed, their acquisition, too, requires inference to the best explanation.

3. The Pervasive Need for Cross-checking

What, then, is it about cross-checking that establishes its essential and fundamental place as an epistemic method, even in the case of sense perception? Surely this is a consequence of the fact that we are physical beings, situated within a spatio-temporal world in an environment with which we communicate via physical—that is, causal—processes. But the centrality of cross-checking is still more fundamental than this. It is demanded for knowledge of any causal process, in which causes are known via their effects. Thus, it is demanded in connection with any claim to have perceptual access to an extra-mental reality. It would be demanded, for example, if we were bodiless minds claiming perceptual contact with disembodied demons, evil or benign, with angels, or with a god. That is because the contact is perceptual, and because of the principle that:

(P) If S perceives (has a perceptual experience of) X, then X is a suitable cause of S‘s experience.

First, three words about (P); and then, more on the connection between (P) and cross-checking:

(1) When I say X is a cause of S‘s experience, I mean just that it plays a role as one of the causal antecedents of S‘s experience.

(2) Strictly speaking, it is events or states of affairs that are causes. If X is a particular, then it is not X per se, but X‘s having some property or undergoing some change which constitutes the cause in question.

(3) When I say that this is a suitable cause of S‘s experience, I mean that it must cause the experience in the right sort of way, for the experience to count as perception of X. Obviously, not all of the causal conditions of my now perceiving this pen are conditions I now perceive (those conditions include my eyes and brain’s working properly, the pen’s being illuminated, and even God, if God caused the pen to exist and sustains me in existence). We cannot say in general what criteria distinguish the “right” sort of causal ancestry from the wrong sorts; but cross-checking has everything to do with how we justifiably identify the right items in particular cases.

Knowing what we are perceiving is a matter of knowing what is causing our experience in the right sort of way. But that is a matter of narrowing down the candidate causes of an experience so that—ideally—just one cause, situated in the right way, can explain our data. It is precisely here that cross-checking plays the crucial role, by enabling us to eliminate alternative possibilities, and to form a sufficiently precise conception of our environment and the causal processes that occur in it, to “zero in” on the (or a) “suitable” cause.8

William Alston misses the mark when he insists that a demand for similarly cross-checking the claims of mystics amounts to a kind of epistemic imperialism.9 Alston insists that each epistemic practice, including mystical practice, gets to dictate its own standards and cross-checking criteria. But as we shall see, those invoked by mystics are characteristically vacuous. Obviously, the sorts of evidence relevant to checking a perceptual claim will depend upon its modality and content. But determining what makes something count as evidence and justification is dictated by the causal structure of perception and cannot be commandeered by epistemic practices, so-called.

Many philosophers will reject this conception of perception and perceptual knowledge. They do so partly for dialectical reasons—i.e., because they believe that the road so paved leads straight to skeptical perdition. They do so, further, for broadly phenomenological reasons—that is, because we do not ordinarily make perceptual knowledge claims on the basis of anything more than having the right sort of experience. We don’t indulge in any cross-checking or inference in judging, e.g. that there is someone in the seat next to us.

But these objections are, in the present context, misdirected. The phenomenological objection ignores what we might call subliminal information-processing, both past and occurrent, and the vital role that cross-checking plays in this processing.10 What sorts of perceptual seemings a given environment can produce in one is a function, not only of recent sensory stimulation, but of much else: of attention and motivational factors, of past experience and concepts thereby acquired, of expectations for which an inductive rationale could be supplied if required (but which ordinarily does not—and need not—enter into perceptual engagement with the world). We can look and just “see” that the refrigerator in the kitchen is white, in part because we have acquired an understanding of what refrigerators are, and what they look like, readily expect such items to appear in kitchens, and know that white things look a certain way under the apparent conditions of illumination. An ability to “just see” directly that this refrigerator is white is a hard-won skill. Learning endows us with unconscious cognitive mechanisms that operate to apply concepts in forming a percept as if on the basis of various inductions.

Moreover, past learning and also the present cognitive processing incorporate cross-checking in fundamental ways. What our cognitive systems have learned is how “automatically” to make judgments that, were we rationally to reconstruct them, would involve causal reasoning to the best explanation for the multitude of sensory inputs with which we are provided. For example, the supposition that light travels in more or less straight lines, together with the hypotheses that there is a bulky, stationary, solid white object before us, and that we are in motion in a certain way relative to it, can help explain the sequence of our visual/tactile inputs. But any such reasoning (or unconscious surrogate for it) must invoke, implicitly, cross-checking. It is as if, e.g., the various visual and tactile inputs serve to corroborate the judgment that there is a refrigerator, by eliminating alternative possibilities.

This kind of implicit cross-checking is absolutely pervasive; it comes to permeate all of our perceptual “takings” as we mature and piece together our world.11 This feature of sense-perceptual processes explains a fundamental phenomenological feature of perceptual judgments, namely how we can directly take ourselves to be en rapport with our physical surroundings, even though no single bit of sensory information could form an adequate basis for such a judgment (or even, I would add, for the formation of the concepts required to envision a three-dimensional space inhabited by physical continuants). It explains how it is that we do this without seeming to engage in any processes of inference from representations—of inference from effects to causes. That is why direct realist theories of perception can seem so plausible, even though in a causal sense, we are obviously not in direct contact with our physical surroundings.

4. Skepticism Bracketed

I have dwelt upon this point because I take it to be crucial to an assessment of the epistemic status of mystical experience, interpreted as perceptual contact with supernatural realities. But it also permits a response to the objection that conceding perception to involve an “indirect” (causal) contact with extra-mental reality, and perceptual judgment to require reasoning from effects to causes (or surrogates for that), gives the skeptic all he needs to undermine claims to have knowledge or justification.

Alston is particularly forceful and insightful in making this case with respect to sense perception (but of course it applies to mystical perception equally).12 He argues that any attempt to justify a perceptual practice must fail either on grounds of unsoundness or circularity. Though Alston’s argument is complex, we have seen why this result is to be expected, and can consequently specify the way in which I believe the issue concerning mystical perception ought to be framed.

So as not to beg any questions, I shall adopt Alston’s view that there are distinguishable belief-forming practices, including different perceptual practices 13. Two such practices take as their inputs sense perception and mystical perception. If the possibility of mystical evidence for God is not automatically to be ruled out, we must find some way of deflecting skeptical objections, as they apply to perceptual judgments generally. Seeing how this goes for sense perception will enable us to generalize to other perceptual practices, for the relevant similarities between them are more important than the differences. Alston, in spite of his insistence that each perceptual practice is beholden only to its own epistemic standards, recognizes this when he invokes, for all perceptual practices, what amounts to a kind of Principle of Credulity.14 Alston takes it that, provided a perceptual practice meets certain conditions,15 perceptual judgments formed in the normal ways provided for in that practice are prima facie justified. (They are only prima facie justified: every such practice must include what Alston calls an overrider system, and so a judgment can be overridden. Indeed, Alston’s overrider systems reflect the importance of cross-checking, without properly recognizing its fundamentality.)

Any appeal to prima facie warrant—warrant occurring in the absence of even implicit or preconscious processes that could be rationally reconstructed in terms of inductive inference and cross-checking—is just the wrong way to bracket (radical) skepticism and frame our question. It is wrong because it short-circuits precisely the crucial justificatory procedures (or at least a crucial stage in their application), thereby begging, or at least certainly obscuring the bearing of, critical questions that the mystical theist must confront. They include the question whether cross-checking procedures must be, but are not, appropriately “built into,” and cannot retrospectively be applied to, mystical experiences and the judgments they deliver. I shall argue that they are not, and that this flaw is fatal to mystical justifications of theism.

Cross-checking and cross-checkability must be integral parts of any perceptual epistemic practice because what a perceiver takes to be present on the basis of her experiences might not be what is in fact causally responsible for those experiences. Cross-checking ‘pins down’ stages of the causal process, thereby eliminating alternative hypotheses as to how the input is produced.16

What goes for sense perception goes for mystical experience as well. Theists who invoke such experiences as evidence may help themselves to the same inductive principles that our sensory practices evidently presuppose—in particular, those that vindicate cross-checking. However, if, granting those principles, mystical experiences fail to supply significant evidence for theism, an appeal to them will be of little help to theists.

I have been insisting that what we need to frame the debate productively is not some principle of credulity, but more general and fundamental inductive principles that will not short-circuit the issues. But even if I were to grant some form of credulity principle,17 it would avail the theist little. For the warrant it confers is only prima facie warrant, and, as it happens, there are good reasons to question that warrant, in the mystical case. Since that is so, cross-checking can’t be avoided, and its demands are made acute in proportion to the cogency of the cognitive challenges that mystical practices (MP’s) confront.

5. Christian Mysticism: Challenges and Checks

There are a number of such challenges, in the form of alternative explanations for mystical experiences (ME’s). One of these, which I shall not pursue, comes from within many MP’s. It is the possibility that an ME is demonically caused.18 There are also naturalistic explanations. Here I shall mention two which complement one another and are jointly strong enough to outdistance any theistic explanation.19 Fortunately (and pace Alston, 1991, 230-233), patterns of mystical encounter are so predictable and overtly manifested, in religious traditions ranging from Pentecostal worship to the ritual seances of Dinka and Tungus shamans, that it has been possible for anthropologists and psychologists to study the phenomenon in great detail in its natural settings.20

The first naturalistic explanation is due to the anthropologist I. M. Lewis, and derives from world-wide comparative studies which reveal certain general patterns among MP’s. In brief, Lewis (1989) shows that, at least where mystics “go public” and appeal to their experiences in the social arena, mysticism serves mundane interests either of the mystic him- or herself, or of some group with which he/she identifies. Lewis discerns two types of mystics: socially marginalized mystics whose mysticism is a weapon in the struggle to achieve social justice for themselves and their group, and upwardly-mobile-mystics who use their mystical experiences as credentials to legitimate their claim on positions of social leadership. Lewis shows how the descriptions mystics give of their experiences, and the behaviors they exhibit prior to, during, and after mystical episodes, serve these social ends in quite precise and predictable ways.

One of the great strengths of Lewis’ theory is that it cuts across the entire spectrum of MP’s, providing a unity of explanation that the theist cannot hope to match.21 Lewis’ theory has, however, a significant lacuna. It says little about how the occurrence of favorable social circumstances gets translated into the incidence of mystical phenomenology. Moreover, Lewis gives no very adequate explanation for the apparent frequency of ME’s which remain private. Many people, it seems, have occasional mystical experiences, but almost never disclose them.

But it looks now as if these gaps can be closed by the second naturalistic approach, which has begun to indicate the details of the neurophysiological mechanisms by means of which mystical experience is mediated. Such experiences, it turns out, are associated with microseizures of the temporal lobes of the brain. When these seizures are severe, they result in temporal-lobe epilepsy. But mild seizures, which can even be artificially induced during brain surgery, can result in powerful mystical experiences.22 A substantial portion of the general population has a disposition for such mild seizures, and there is some circumstantial evidence that they can be provoked by techniques traditionally used to induce mystical trance-states.23

A theist may wish to reply here that God may well have a hand in these mechanisms, indeed employ them as His means for appearing to His worshippers.24 But this is implausible on a number of counts. For one thing, it is extraordinarily hard to explain why God would appear through the figure of Jesus to a Christian, as Allah to a Muslim, Brahman to a Hindu, the god Flesh to a Dinka, and as a variety of loa spirits to voodoo practitioners. And, if a purely naturalistic explanation can be given for the non-theistic experiences, then why not also for the theistic ones?25

There are other problems. Suppose we take a naturalistic explanation of ME’s, and tack on the hypothesis that God was involved in some way. This is a God-of-the-gaps strategy. Given the lacunae in our understanding of even simple physical processes—to say nothing of the neurophysiology of the brain—this strategy is one a theist can deploy with some ease.

Indeed, it incurs the danger of being too easy. A theist could invoke divine intervention to explain why the radiator of my car cracked overnight. Our natural explanation is full of holes: we may not know exactly how cold the engine got last night, nor exactly how strong the walls of the radiator were at the rupture point, nor how to apply the known laws of nature to such a complex system. So in principle, all the theist need do is find some gap in the posited causal etiology, and tack on the hypothesis that here the finger of God helped the process along—no doubt, to punish my sins.

Why do we (most of us!) not credit such an “explanation”? First, of course, because a long history of experience teaches us that such gaps often are eventually filled by natural causes. But second, because the theistic explanation comes too cheaply: there are no constraints on when, how, and where God is likely to act, no attendant procedures for cross-checking or ferreting out the precise mode and locus of divine intervention, no positive suggestions about how the theistic account of theophysical interaction might be investigated, fleshed out, ramified—and virtually no concomitant predictive power. This theoretical poverty cripples cross-checking for divine influence.

Still, the presence of naturalistic competitors makes it imperative that we examine what sorts of cross-checking MP admits, and how successful such cross-checks have been. We run here into a number of obvious difficulties. Most prominent among them is the fact that mystical experiences are not public.26 Moreover, the sorts of checks typically invoked, by Christian mystics at least, are either epistemically irrelevant or question-begging, absent quite strong auxiliary assumptions.

It is not that mystics are unconcerned about the veridicality of ME’s. On the contrary: they often display a lively concern with this, and offer multiple tests. But let us look at some of these tests, using Teresa of Avila as a guide. Teresa exhibits a strong interest in the question how veridical experiences are to be distinguished from those produced by what she calls “melancholy,” and by Satan. (This interest is hardly surprising, given the regularity with which the Inquisition accused mystics—especially women—of nefarious motives, fraud, or demonic possession.) Teresa’s list of tests includes: (1) the fruits of an experience—both in the actions and personality of the mystic, and as producing an inner peace rather than a troubled state of mind; (2) the vividness of the memory of the experience; (3) conformity to Scripture, and (4) validation by the mystic’s confessor.

It is not hard to see how these criteria might be designed to secure for the mystic immunity from Inquisitional prosecution, but not easy to see what epistemic force they could have.27 (3) looks straight-forwardly question-begging, inasmuch as the authority of Scripture rests largely on the supposed authority of the revelations upon which it is based.28 (1), (2), and (4) have no epistemic force except on the assumption that only God, and neither Satan’s best deceptive efforts nor natural causes, can produce experiences that are memorable, convincing to confessors, or have good fruits. But what independent evidence is there of that? What cross-checks for these claims can theists supply? On this, Teresa is silent.

The final—and in principle the best—hope for cross-checking ME’s lies with successful prophecy. Perhaps a theistic account does after all yield checkable predictions in a way that bears directly upon the evidential force of ME’s. For, often enough, one of the fruits of a mystical encounter with God has been the revelation of a prophecy. Not only that, but prophecy has figured as a central component of Christian mystical practice (CMP) and many other MP’s, and of the apologetical strategies associated with them. This is because prophecies permit, when certain conditions are satisfied, type-2 cross-checks of a fairly powerful and peculiarly direct sort. When the content of a ME contains some message, putatively from God or some supernatural source assumed to be in the know, concerning future events, the claim of genuineness can in principle be checked; ordinarily, the prophesied events will be of such a sort that it is within the purview of ordinary sense perception to determine their occurrence or non-occurrence.

Yet, Alston tries to downplay the prophetic dimensions of MP.29 Why? After all, the plain fact is that prophecy is a major and central feature of the MP’s of many religious traditions; moreover, putatively successful prophecy is regularly appealed to precisely by way of confirming the genuineness of the prophet, the veridicality of his or her ecstatic visions, and the uniquely truth-connected status of the tradition that claims him or her as its own. Ecstatics who develop prophetic practice into a vocation are familiar figures in religious traditions—witness the Oracle at Delphi, the Hebrew prophets, John of Patmos—and Jesus of Nazareth. Nor is this an aspect only of ancient MP’s, long since superseded (within the Jewish and Christian MP’s). Far from it, as anyone who considers the claims of contemporary televangelists can confirm.

Prophecy, therefore, is a feature intrinsic to CMP, a feature by means of which the truth-claims produced by that practice can be quite directly checked. However, no such check will be very informative unless certain conditions are satisfied. Briefly, these include:

  1. The prophecy must be of some event not intrinsically likely (not, e.g., “wars and rumors of wars” – Mk. 24:6)
  2. The prophecy must not be self-fulfilling, or of events the prophet or his/her followers can themselves bring about.
  3. The prophecy must demonstrably have been made prior to the events which count as its fulfillment.
  4. The prophecy must be sufficiently specific and unambiguous to preclude ex post facto reinterpretation to fit any of a wide range of possible ‘fulfillments’.
  5. The fulfillment of the prophecy must be verified independently of the say-so of the prophet or his/her partisans or tradition.30

Here we have, at last, a cross-check which really does offer a test of mystical experience. The reasoning is straightforward: given 1-5, only the mystic’s having received a message from a superhumanly prescient being (or, improbably, wild luck) can explain his or her prophetic success. (There are, to be sure, some added complications: for example, we must be careful to avoid the Jean Dixon fallacy. A clever prophet can issue hundreds of risky prophecies, in the hopes of scoring a few memorable “hits,” and calculating that the “misses” will be forgotten. Our reasoning to the best explanation must take into account the prophet’s entire track record.)

Now, just what is the record of Jewish and Christian MP’s on this score? Rather than pursue this question at length, let me observe that I know of no recorded prophecy, either within the Jewish/Christian canon or outside it, that clearly satisfies criteria 1-5. (There are, however, a number of demonstrably false prophecies. Of these, perhaps the most decisive and poignant occurs at Mt. 16:27.)31

6. Conclusion

Like any perceptual practice, CMP requires an elaborated system of cross-checks and cross-checking procedures. But, because of its theoretical poverty with respect to the causes of mystical experiences, no such system has been or is likely to be forthcoming. With respect to the one relatively strong cross-checking strategy that CMP has available (and has purported to use), its record is one of failure. Until these defects are remedied, mystical experience cannot hope to provide significant evidential support for theism.


Alston, William P. (1964). “Psychoanalytic Theory and Theistic Belief,” in John Hick, ed., Faith and the Philosophers. New York: St. Martin’s Press, 63-102.

________. (1991). Perceiving God: The Epistemology of Religious Experience. Ithaca, NY: Cornell University Press.

________. (1993). The Reliability of Sense Perception. Ithaca, NY: Cornell University Press.

Blackmore, Susan (1993). Dying to Live: Near-Death Experiences. Buffalo, NY: Prometheus Books.

Fales, Evan (1990). Causation and Universals. London: Routledge.

________ (1996a). A Defense of the Given. Lantham, MD: Rowman & Littlefield.

________ (1996b). “Scientific Explanations of Mystical Experiences. Part I: The Case of St. Teresa,” Religious Studies 32: 143-163.

________ (1996c). “Scientific Explanations of Mystical Experiences. Part II: The Challenge to Theism,” Religious Studies 32: 297-313.

Davis, Caroline Franks (1989). The Evidential Force of Religious Experience. Oxford: Clarendon Press.

Gellman, Jerome I. (1997). Experience of God and the Rationality of Theistic Belief. Ithaca, NY: Cornell University Press.

Lewis, I. M. (1989). Ecstatic Religion, 2nd ed. London: Routledge.

Payne, Steven (1990). John of the Cross and the Cognitive Value of Mysticism: An Analysis of Sanjuanist Teaching and its Philosophical Implications for Contemporary Discussions of Mystical Experience. Synthese Historical Library, v. 37. Dordrecht: Kluwer Academic Publishers.

Plantinga, Alvin (1983). “Reason and Belief in God,” in Faith and Rationality: Reason and Belief in God, Alvin Plantinga and Nicholas Wolterstorff, eds. Notre Dame, IN: University of Notre Dame Press.

Poulain, Augustin (1978) [1901]. The Graces of Interior Prayer: A Treatise on Mystical Theology, 6th ed., Leonora York Smith, transl. Westminster, VT: Celtic Cross Books.

Russell, Robert J., Nancey Murphy, and Arthur R. Peacocke, eds. (1995). Chaos and Complexity: Scientific Perspectives on Divine Action. Vatican City: Vatican Observatory Publications.

Sargant, William (1974). The Mind Possessed: A Physiology of Possession, Mysticism, and Faith Healing. Philadelphia: J.E. Lippincott.

Swinburne, Richard (1991). The Existence of God, rev. ed. Oxford: Clarendon Press.

Wainwright, William (1981). Mysticism. Madison, WI: University of Wisconsin Press.

Yandell, Keith E. (1993). The Epistemology of Religious Experience. NY: Cambridge University Press.


1 E.g., William Alston (1991), William Wainwright (1981), Keith Yandell (1993), Richard Swinburne (1991), Jerome Gellman (1997), Alvin Plantinga (1983), Steven Payne (1990), and Carolyn Franks-Davis (1989).

2 Contrary to what is sometimes claimed, mystical beliefs are surely not self-certifying, no matter how much certainty mystical experience may generate in the mystic. On this point most philosophers—and mystics themselves—are agreed. The reason testing is needed is, as I shall show, that mystical claims, when they are about an extra-subjective reality, aren’t of the right sort to be self-certifying. It doesn’t help the mystic’s case, of course, if her mystical beliefs are contradicted by those of another mystic who displays equal certitude.

3 There are direct realists who deny that ‘S perceives C‘ entails any causal claim about C. That is not something that someone who rejects the direct realist’s theory of perception need be concerned to deny. Nor, for present purposes, need I deny the view that, when I really do perceive a cat, I do so “directly”, that is, not “in virtue of” perceiving something else more directly. So I shall here concede both these points. It suffices for my present purpose that we do not allow, where C is an ‘external’ entity or state of affairs, that S perceives C unless C in fact plays a causal role in production of S‘s experience. I should say more: for external C, it is a metaphysical necessity that C be so involved in the production of S‘s experience. The notion of externality can be sufficiently captured by saying that C is external to S‘s experience just in case C‘s existence doesn’t entail the existence of S.

4 See Fales (1990), Chapter 8.

5 In practice, A will be a long conjunction of hypotheses which describe the antecedent conditions and the laws governing the causal mechanisms upon which the making of any measurement or observation depends upon. In the case of perception, H will articulate the causal pathways which mediate the transmission of sensory information. The component hypotheses of A will typically be independently (of each other) testable. For simplicity, I shall treat A as if it were a single hypothesis.

6 That is, one that does not make use of H.

7 Our goal is to vindicate the inverse-probability reasoning we use to infer causes from their effects as the best explanations of those effects. If we employ Bayes’ Theorem (or some qualitative analogue) for that purpose, we shall also need to assume some rough way to assign credences to competing hypotheses, antecedently to considering any of the empirical data. Let us assume this can be done. For present purposes, these anti-skeptical assumptions are enough to be getting on with.

8 There is often more than one. Even though I could not, ordinarily, be said to observe an image on my retina, I could be said, when watching a presidential press conference, alteratively to see either the TV or Bill Clinton. An elementary particle physicist could rightly say both that she is observing tracks in a bubble chamber, and that she is observing electrons.

9 See Alston (1991), 209-222.

10 It also ignores the difference between our just perceptually taking there to be someone in the adjoining seat, perhaps in part because of the operation of hard-wired belief-forming mechanisms that operate “automatically,” and our being justified in that belief.

11 In young infants these cognitive processes are observably in the process of formation.

12 See Alston (1991), Chapter 3, and Alston (1993).

13 For details, see Alston (1991), Chapter 4.

14 The term, and the principle itself, are due to Swinburne, though the idea can be traced back at least to Reid.

15 These conditions include being socially established, incorporating an overrider system, and being free of massive contradiction from within and from beliefs generated by other doxastic practices (see Alston (1991), Chapter 4).

16 The trouble with this story is, as we saw, two-fold. First, the only means we have for “pinning down” the facts about a given causal process are perceptual means; and if there is a skeptical question to be raised about the original process—the one generating the perceptual experience upon which a perceptual judgment is based—then entirely similar doubts will apply to the perceptual processes upon which cross-checking procedures depend. Second, our problem arises in the first place—and hence in the second place—because effects underdetermine their causes. (This is just a special case—undoubtedly the most central case—of the problem that theory is underdetermined by data: any given data can be explained by any number of incompatible theories.) It is to evade the skeptic here that we invoke the anti-skeptical principles.

17 Whether it be that adopted by Swinburne (1991), Alston (1991), or Gellman (1997).

18 I shall also largely ignore the major challenge which derives from the enormous variety and conflicting content of ME’s worldwide. That is a severe problem in its own right.

19 So I argue with respect to Lewis’ theory in Fales (1996b, c). That can now be supplemented with the neurophysiological findings.

20 I have the report (private communication) of a Christian mystic trained in neurophysiology who has been able to record her own brain waves, and those of a colleague, during trance, and who confirms the temporal lobe finding (see below). For a more detailed summary of the evidence and references, see Fales (1996a, b, c).

21 See Fales (1996c).

22 The literature is substantial and growing. For a good bibliography, see Blackmore (1993), especially the citations for Chapter 10.

23 See Sargant (1974).

24 Alston has suggested this possibility on a number of occasions – e.g., in his (1964) and (1991), 230-233.

25 This argument is fleshed out in Fales (1996c). It is moreover very unclear just how, in principle, God would be able to communicate with human beings. If this is to occur via divine influence upon a person’s brain states, and those states are macroscopic physical states, then any divine intervention will involve local violations of the highly confirmed laws of conservation of momentum and energy. If, one the other hand, we suppose that God intervenes at the quantum level, acting as a kind of “hidden variable” in determining the outcomes of indeterministic processes, as Nancey Murphy has recently proposed, then we can avoid the violation of physical laws, but only at the price of making in principle unknowable (since hidden by quantum uncertainties) the presence of divine intervention. On these issues see the articles by Murphy and Tracy in Russell et al. (1995).

26 There are occasional reports of sense-perceptual supernatural apparitions witnessed by many—e.g. at Fatima and Zeitoun. Also, some mystics do report perceiving God via several sensory modalities—e.g., vision, hearing, and smell. I cannot pursue these matters here; and in any case, many theists—e.g. Alston and Wainwright—de-emphasize this sort of experience.

27 Indeed, most such tests aim at social acceptance within the religious community. These, and all the other tests of which I know, are such that passing them is largely under the control of the mystic or of her religious community. Thus, unlike proper cross-checks, they do not risk invalidation of the tested hypothesis by an uncooperative tester-independent world.

28 It is all too likely that the content of Teresa’s experiences, as she describes them, is conditioned by her (and her superiors’) prior acceptance of Scripture. And she gives no independent evidence that Scripture is authoritative. That authority could be independently confirmed by miracles and successful prophecy, however. Concerning the latter, see below.

29 See Alston (1991), 222-225. Alston is there concerned with the general predictive power of CMP and does not mention religious prophecy at all. On p. 291, he mentions fulfillment of prophecy in passing as a test, not of ME’s, but of divine inspiration not associated with ME’s. Yet on p. 298, he expresses skepticism concerning the record of Christian miracles generally.

30 It might be protested that this last condition reflects an improperly imperialistic imposition on CMP of criteria indigenous to SP. But to excuse CMP from this requirement on such grounds is to abandon good sense. First, the fulfilling events are typically ones which would be observed by ordinary sense perception; and second, as Hume correctly observed, the temptation to prevarication is too great here to rely upon the say-so of those whose interests are directly at stake. We have ample demonstration of the perennial creative reconstruction of the historical record by those who have a religious agenda; and the N.T. is certainly no exception.

31 Others occur at Josh. 4:7, Ez. 26:14-21, and Is. 60:1-62:12, esp. 62:8. Augustin Poulain (1978, Ch. XXI) reports with considerable embarrassment the false prophetic utterances of a number of mystics canonized by the Roman Church. As to Mt. 16:27, see parallels at Mk. 8:38 – 9:1 and Lk. 9:26-7. Jn., writing later, discretely omits this prophecy; see Jn. 6 and 12:25 which in other respects parallel the synoptic pericopes. It is clear that the parousia and final judgment are intended: cp. Mt. 25:31f and Rev. 20:11-21.

all rights reserved