Nelson Pike’s Contribution to the Philosophy of Religion

Abstract

In this paper I attempt to capture the essence of Nelson Pike’s contribution to the philosophy of religion. My summary of his insights will revolve around three general topics: omniscience (and in particular its relation to human freedom), omnipotence (and in particular its relation to the existence of human suffering), and mysticism (with a focus on the question of whether and in what sense mystic visions can be sources of knowledge). Although the details vary in interesting ways, his work on these topics largely consists of recognizing an important challenge to the viability of the relevant doctrine or framework, sharpening that challenge by presenting it in a more forceful way, and then offering and assessing potential responses. Pike’s writings are characterized by exemplary rigor and relentless clarity, and together they constitute a rich (and under-appreciated) source of insight.

Introduction

In what follows, I have attempted to capture the essence of Nelson Pike’s contribution to the philosophy of religion. It was a daunting task, but also an enjoyable and fruitful one—in large part due to the exemplary rigor and clarity that characterized Pike’s writings. As I hope to demonstrate below, I have found his work to be a rich (and under-appreciated) source of insight.

I have divided the topics that Pike wrote about into three general categories: omniscience, omnipotence, and mysticism. I will consider each one in turn, focusing on the main claims and lines of argument. Pike’s writing was persistently technical (but never gratuitously so), and much of its philosophical value is in the technical details. Space unfortunately does not permit me to delve into these details, but I have tried to summarize them in an intuitive way.

Omniscience

Pike is perhaps best known for his work on the attribute of omniscience, and in particular on the question of whether human freedom is compatible with divine foreknowledge. He developed a forceful argument for the incompatibilist conclusion in his “Divine Omniscience and Voluntary Action” (1965), and has continued to refine that argument over the years. Although I won’t recount every twist and turn that the dialectic has taken since the publication of that seminal paper, I will highlight some of the more significant moves that have been made.

Divine Omniscience and Voluntary Action

Inspired by Boethius’ writings on divine omniscience, Pike sets out in his (1965) to show how certain traditional assumptions of Christian theism, if true, imply that no human action is ever voluntary. The preliminary assumptions are as follows. First, knowledge entails belief (and is factive): “God knows that p” entails “God believes that p” and that p is true. Second, God cannot be mistaken, which is to say that he is infallible. This follows from (perhaps because it is equivalent to) the claim that God is essentially omniscient, and what it means is that God believes all truths and believes no falsehoods: p’s truth entails and is entailed by “God believes that p.” Third, God has from eternity had foreknowledge of everything that’s ever happened. The assumption that God has foreknowledge “from eternity” can be taken to mean that God exists outside of time (as Boethius believed), or it can be taken to mean that God exists at all times. Opting for the latter interpretation (although he considers the former interpretation at length in his (1970)), Pike takes it as a priori that, for any event, God has always known that (and when) that event was going to occur.

Now suppose that God exists and that Jones decided to go to the beach last Saturday. It follows from the preliminary assumptions that

• (1) God (infallibly) believed 80 years ago that Jones would go to the beach last Saturday.

Pike’s insight is that Jones’s decision to go to the beach was not made freely unless one of the following is true of Jones with respect to that decision:

• (2) Jones was able to do something that would have brought it about that God held a false belief.

• (3) Jones was able to do something that would have brought it about that God held a different belief than the one he in fact held.

• (4) Jones was able to do something that would have brought it about that God didn’t exist.

But (given the above assumptions) none of (2)–(4) is even possible (much less true)—so it follows that Jones’s decision was not made freely. And since nothing about that particular decision is special, the conclusion generalizes to cover all decisions ever made by humans. Why are (2)–(4) not possible? As Fischer (1983, 70) points out (a summary that Pike endorses in his (1984)), (2) is ruled out by God’s essential omniscience and (3) and (4) are ruled out by the fixity of the past. (The principle of the fixity of the past tells us that nobody can do anything that would bring it about that the past would have been different.)

Pike then rejects three potential objections to this argument. The first—inspired by Leibniz—is that the argument fails to distinguish between absolute necessity and hypothetical necessity. Although it is necessarily true that if God foreknows that Jones will (decide to) go to the beach, then Jones will go to the beach, it doesn’t follow from that hypothetical necessity that it’s absolutely necessarily true that Jones will go to the beach. Pike points out that this objection misses the mark because the “necessity” that he imputes to Jones’s action is not the strong modal concept of necessity but rather the notion of involuntariness. In other words, we can suppose that it is contingently true that Jones decides to go to the beach, and thus contingently true that God believed that Jones would go to the beach, and moreover contingently true that Jones was not able to do otherwise.

The second objection comes from Molina, who claims that human freedom is in a sense built into God’s foreknowledge. God doesn’t merely believe that Jones will decide to go to the beach, but rather believes that Jones will freely decide to go to the beach. Pike dismisses this view as incoherent (for reasons that he fleshes out in later work, and which we will discuss below).

The third and final objection, from Schleiermacher (who is himself following Augustine), posits a parallel between divine foreknowledge and human foreknowledge. This objection involves two related claims: first, that human foreknowledge doesn’t compel the action foreknown; and, second, that divine foreknowledge is similar to human foreknowledge in this respect. But, as Pike points out, there is at least one relevant difference between the two cases. Suppose that, prior to last Saturday, Smith believed that Jones would go to the beach. By hypothesis, Smith’s belief was true. But Jones, at the time of his decision, nevertheless had the ability to bring it about that Smith held a false belief. Jones won’t exercise this ability, of course, but it still seems that he has it. The same cannot be said with respect to God’s beliefs: (1) both entails and is entailed by “Jones went to the beach last Saturday.” It’s impossible for God (as defined above) to have a false belief, and so it’s not the case that Jones has the ability to bring it about that God held a false belief.

Pike (1965, 43) draws out the following lesson from these considerations: The crucial difference between human foreknowledge and divine foreknowledge is a difference in the way that the constituent belief is connected to truth. Human beliefs, when true, are only contingently true; but divine beliefs, given essential omniscience, are necessarily true. If Smith’s belief about Jones’s future action is true, then it follows that Jones’s ability to do otherwise will remain unexercised. But since God’s belief about Jones’s future action must be true, it follows that Jones’s ability to do otherwise is unexercisable. And an unexercisable ability seems to be no ability at all (cf. Pike 1993, 153).

The most obvious way to avoid the foreknowledge problem, of course, is simply to deny one of the assumptions. If God’s knowledge—unlike human knowledge—doesn’t entail belief; or if God’s cognitions don’t take place in time; or if God is omniscient but not essentially omniscient, then the problem is dissolved. With the exception of the timelessness move (which Pike considers at length in his (1970), and which we’ll examine below), these denials have not been popular with Christian theists. Avoiding the problem is simple—but not easy.

Rejoinder to Saunders

Saunders (1966) responds to Pike’s argument by emphasizing the distinction between changing the past (i.e., causing something to happen in the past) and acting such that the past would have been different. The former is impossible, but the latter arguably is not. For there seems to be a class of past facts—“soft facts”—that are temporally related to, and counterfactually dependent on, future events. Saunders provides the following example of a soft fact:

• (5) Caesar was assassinated 2009 years before Saunders wrote his paper.

Relative to any time between Caesar’s assassination and Saunders’s writing of his paper (e.g., the year 1900), the fact in (5) is a soft fact. In addition to its temporal relation to Saunders’s writing of his paper, there is also a relation of counterfactual dependence: If Saunders had not written his paper when he did, then (5) would not have been a fact. From the perspective of 1900, Caesar’s assassination is of course firmly rooted in the past; but there is another event, namely Saunders’s writing of his paper, that is not, and it is this event on which the truth of the fact as a whole depends. Speaking informally, then, we might say that a soft fact has two “parts”—one of which is about or pertains to the future. Or we might put it this way: soft facts, while genuinely about the past, are not solely about the past. (An appeal to soft facts is characteristic of the Ockhamist response to Pike’s argument, about which more below.)

“Hard facts,” on the other hand, are solely about the past; they only have one part. As Pike puts it (1966b, 370), they are “fully accomplished,” “over-and-done-with,” etc. Thus the fact that

• (6) Caesar died on the steps of the Senate

is a hard fact about the past. This distinction between hard and soft facts has been remarkably resistant to analysis, but it remains reasonably clear at the intuitive level. So there are some facts which are counterfactually dependent on our actions: what this means (among other things) is that we sometimes have the ability to act in such a way that a fact about the past would not have been a fact. The question, of course, is whether facts about God’s beliefs exhibit this special quality of softness. What exactly is the status of (1)?

• (1) God believed 80 years ago that Jones would go to the beach last Saturday.

Is (1) a soft fact, like (5), or is it a hard fact, like (6)? Pike insists that (1) is a hard fact. He acknowledges that there may be a difference between his original (1965) claim that

• (8) Nobody can bring it about that someone who held a belief didn’t hold that belief

and the revised (1966b) claim that

• (9) Nobody can act such that someone who held a belief didn’t hold that belief,

but he maintains that they are nonetheless both a priori truths. So a fact such as (5) may be a counterexample to an unrestricted fixity of the past principle, such as

• (10) Nobody can act such that a fact about the past wouldn’t have been a fact,

but it is not a counterexample to Pike’s (9). (Fischer (2011) examines the interesting question of why Pike’s (9) is so drastically restricted in scope.) And (9) is enough to deliver the result that Jones was not able to do something such that God would have held a different belief (i.e., to deliver the result that (3) is false). So it would seem that Saunders has failed to refute Pike’s argument.

Despite his dismissal of Saunders’s objection, Pike is quick to point out that reflection on Saunders’s piece does raise some interesting questions about the concepts of belief and person. In particular, Pike recognizes that he is susceptible to the complaint that he is trying to have it both ways with the logic of those two concepts. God’s beliefs are different than human beliefs because they have a necessary connection to truth: that Jones went to the beach last Saturday both entails and is entailed by (1). Given this unique feature of God’s beliefs (i.e., given that they aren’t governed by the logic of ordinary beliefs), we might be inclined to think that (1) behaves like (5) rather than (6), and thus that (9) doesn’t apply to God’s beliefs—in which case

• (3) Jones was able to do something that would have brought it about that God held a different belief than the one he in fact held

might be true after all.

A similar (albeit arguably less plausible) line of argument can be advanced in favor of the truth of (4):

• (4) Jones was able to do something that would have brought it about that God didn’t exist.

Our initial conclusion was that (4) is impossible. This conclusion is supported by something like the following principle (cf. Pike 1966b, 378):

• (11) Nobody can act such that a person who existed at an earlier time didn’t exist at that time.

This principle seems unassailable when we’re talking about the ordinary concept of personhood. But—to foreshadow a point that will come up later—does it apply to the sort of person God is assumed to be (within the context of Pike’s argument)? Recall that in this context, God is essentially omniscient. Not only are his beliefs necessarily connected to truth, but his very existence is necessarily connected to all true beliefs (and necessarily disconnected from all false beliefs). Perhaps the ordinary logic of personhood—in particular, the principle in (11)—doesn’t apply to persons who are essentially omniscient, and perhaps (4) is true after all.

Pike takes these considerations to supplement, rather than refute his argument (1966b, 378–9). The principles in (9) and (11) about beliefs and persons, respectively, may not apply to God. But if they don’t, then that suggests that God—at least as traditionally conceived—doesn’t have beliefs and perhaps isn’t even a person. In other words, if the connection between God’s beliefs and the truth means that those beliefs aren’t subject to the constraint in (9), then perhaps they shouldn’t count as beliefs in the first place. Likewise, if God’s essential omniscience means that (11) does not apply to him, then perhaps he shouldn’t count as a person in the first place. (Perhaps essential omniscience and personhood are in tension.) In any case, these considerations, while far from decisive, do point toward the conclusion that the problem lies with the assumptions that generate the argument—a conclusion that Pike endorses from the beginning.

Divine Timelessness

Pike’s (1966b), then, can be viewed as an initial attempt to fend off a fledgling Ockhamist solution to the foreknowledge problem. (He responds to a more fully-developed Ockhamist response in his (1977), as we will see below.) He next considers the Boethian solution (or, as he styles it, the Boethian dissolution (1970, 75)), which consists of a denial of (1). The ground for this denial is the claim that God exists outside of time—that God is timeless. What this means is that God has neither temporal location nor temporal duration. (Pike (1970, 8) points out that a lack of temporal location implies a lack of temporal duration.) And if God does not have temporal location, then we cannot say that he believed anything at a time. Thus, (1) is false if God is timeless and the argument for theological incompatibilism fails.

According to Pike, this is one of the two viable solutions to the foreknowledge problem. (I will briefly discuss the second below.) If God is timeless, the problem simply dissolves. Unfortunately, however (for proponents of the Boethian solution), this appears to be one of those cases in which the cure is worse than the disease. Pike’s (1970) is a rich and rigorous treatment of the doctrine of divine timelessness, which I unfortunately won’t be able to present in the detail that it deserves. But I will draw out what I take to be the main points of interest. The first point is that timelessness is problematic because it appears to be in tension with two qualities that are often ascribed to God: omnipotence and personhood.

God is typically understood to be the omnipotent creator of the universe. But if God is timeless, then it’s not clear that he could have intervened in or sustained—much less created—the universe (cf. Pike 1970, 111–18). For creation involves a temporal relation (since a created thing has a beginning in time), and it doesn’t seem as though a temporal relation could have a timeless being as one of its terms. Pike illustrates this lesson using several examples of the kind of activity that might give us an idea of what creation looks like for a timeless God. The first example is of a mental action: the conjuring up of a mental image (1970, 101). An act of conjuring does share some features with the doctrine of creation (a conjured mental image is in a sense created out of nothing, and doesn’t require tools, instruments, or even a body), but there is one crucial difference: the product of a conjuring (i.e., the image) is not independent of the mental activity in the way that the universe is supposed to be independent of God. In other words, it’s difficult to make a distinction between the mental act of conjuring and the product of that act. To describe the product is merely to describe the action in a different way. And this is not typically how the doctrine of creation is interpreted. Typically the created world is considered to be a genuine product, rather than simply an alternative description of God’s mental activity. Of course, the world (despite being a genuine product of God’s creative act) is not completely independent of God; it remains dependent in the sense that it is sustained, or preserved, by God—which leads us to Pike’s second picture.

The second picture is of an utterance. When someone utters a sound, she produces a sound “out of nothing” in much the same way that a mental image is conjured out of nothing. But the product of an uttering (i.e., the utterance itself) is distinct from the action in a way that the conjured image is not. And if we consider temporally extended utterance (perhaps an act of humming or singing), we can get a feel for the sense in which God is supposed to sustain creation: in much the same way that someone might sustain a note that she is singing. But of course this picture isn’t perfect either, because an utterance requires a body (and thus a temporal location). Moreover, an utterance, like any product of creative activity, has a beginning. So the utterance picture helps us understand preservation, and has an independence that the conjuring illustration doesn’t, but it remains inadequate.

These considerations suggest that any proposition of the form “x creates y” implies that there is a temporal relation between x and y (Pike 1970, 109). Recognition of this implication has led some (e.g., Schleiermacher) to reduce the doctrine of creation to the doctrine of preservation—which doesn’t appear to have the same temporal implications. In order to see why, consider a third picture: that of the sea supporting a ship. The way the sea supports a ship is not the same way that vocal chords (diaphragm, etc.) support a note. The way that the sea supports a ship is more foundational than it is productive, and as a result we might be tempted to view God’s preservation in light of this last picture. Unfortunately, this picture is too passive (Pike 1970, 115). Even if creation reduces to preservation (itself a bitter pill to swallow), we still want to be able to say that preservation involves activity on God’s part. And there’s really no sense in which a support or foundation is active.

What we need, in order for a coherent picture of an omnipotent timeless being, is a situation in which someone is actively producing or supporting something, where the product is genuinely distinct from the productive act and yet no temporal relation is implied by the details of the situation. And this, unfortunately, does not seem to be a conceivable picture. (We can come up with examples in which two of those three features are present, but none in which all three are present.) For all that it may still be possible for a timeless being to create or sustain in a way that exhibits each of those features, but to leave it at that seems no better than an appeal to mystery.

Pike (1970, 116) briefly considers the objection that our inability to come up with a complete picture does not imply that the doctrine of timelessness is a failure. Pike’s response is two-fold, and, I think, instructive. His first point is that when we’re theorizing about God, all we have is the pictures. If we’re trying to understand a doctrine about God, we have no choice but to start with what we can picture and then, in Thomistic fashion, refine it by removing the imperfections. His second point is that pictures give us insight into the concept of the individual or relation that we’re trying to explicate. If we can’t come up with a coherent picture, then that suggests that the concept itself might not be coherent. In the present context, these considerations lead Pike (1970, 117) to the following conclusion: “St. Thomas says that God’s preservation activity is ‘without either motion or time.’ I wonder if this isn’t a little like saying that when Gabriel blows his horn, he does it while holding his breath.”

So whatever its merits in light of the argument for theological incompatibilism, the timelessness solution is in tension with the doctrine of omnipotence. Unfortunately, that’s not the only traditional doctrine that timelessness runs afoul of; for it’s difficult to see how a timeless being could be a person. As Pike explains in his (1970, Chapter 7), there are several properties that seem to be quintessentially personal and yet seem not to apply to a timeless being. Recall that a timeless being does not have temporal location—from which it follows that a timeless being does not have temporal extension either. Now consider our capacity, as persons, to anticipate things and to form intentions. Both of these capacities appear to require temporal position and thus are not properly ascribed to a timeless being. Similarly, the capacities of reflection and deliberation both require temporal extension, and thus are not properly ascribed to a timeless being either. As it turns out, one of the few personal attributes—if not the only personal attribute—that applies to a timeless being is the capacity for knowledge. Unfortunately, there are problems even here. For, given previous results, a timeless being would never be able to demonstrate knowledge in a way that would be detectable by temporal beings such as ourselves. In other words, apart from the bare stipulation that he has knowledge, it’s not clear how we can differentiate, even conceptually, between a timeless being who has knowledge and a timeless being who doesn’t. (Pike 1970, 127)

We have seen that the doctrine of timelessness, while dissolving the tension between human freedom and divine omniscience, is not without its problems. I would now like to briefly consider the Augustinian response to the problem, which is the second of what Pike takes to be the only two viable options.

Augustine’s Solution

The Augustinian solution to the foreknowledge problem, at least as Pike presents it (1970, 76–82), consists of two claims: the first is that God’s foreknowledge is relevantly analogous to human foreknowledge, and the second is that God has foreknowledge not merely of what we’ll do but of what we’ll freely do. The second claim (which has been assimilated by the Molinists) is superfluous if the first can be established—and may not provide additional help in any case (cf. Pike 1993, §4)—so I will set it aside for now. As for the first claim, we saw above that God’s foreknowledge differs from human foreknowledge in virtue of its infallibility. Pike (1970, 79) helpfully explicates this difference as follows.

Given a case of human foreknowledge (e.g., a case in which Smith knows that Jones will go to the beach on Saturday), we can identify two (relevant) contingent facts: the fact that Smith held the belief that Jones would go to the beach, and the fact that Smith’s belief was true. According to principle (9)—

• (9) Nobody can act such that someone who held a belief didn’t hold that belief—

Jones is not able to do anything that would result in Smith’s not holding the belief that he in fact held. But he remains free with respect to his decision to go to the beach because he is able to act such that the belief Smith had would have been false. He won’t act in such a way, because we have stipulated that Smith’s belief is true, but it doesn’t follow that he can’t. The situation is different, however, when it comes to God’s belief that Jones will go to the beach (given the constraints within which we’ve been working). When we’re dealing with God’s beliefs, there is only one contingency—namely that God held the belief that he held. Because God is essentially omniscient, there’s no wiggle room: his belief must be true. And this is what rules out Jones’s ability to do otherwise. Principle (9) tells us that Jones can’t act such that God would have held a different belief, and the doctrine of essential omniscience tells us that Jones can’t act such that God would have held a false belief.

The Augustinian solution, then, is simply to deny that God is essentially omniscient. This denial removes the barrier to Jones’s ability to do otherwise, which is to say that divine foreknowledge is no more threatening to human freedom than is human foreknowledge. If Pike is right about the Boethian and Augustinian responses being the only two viable options, then it would seem that the Augustinian denial of essential omniscience is the best hope for the theological compatibilist. It purports to solve to problem without any of the metaphysical baggage that comes with the doctrine of divine timelessness.

Additional Reflections

Thus ends a summary of Pike’s main lines of thought regarding the foreknowledge problem. In what remains of this section I would like to highlight some of the points that have emerged from Pike’s thoughtful responses to various objections that have been raised against his argument.

Plantinga (1974) objects to Pike’s argument by arguing for a concept of ability according to which Jones is able to refrain from going to the beach because even though God believed in the actual world that he would go to the beach, there is a possible world in which he refrains from going. (And in that possible world, God believed that he would refrain from going—this move is what makes Plantinga’s response to Pike’s argument an Ockhamist response.) Pike (1977) criticizes this move by pointing out that we cannot establish some agent S’s ability to perform some action A merely by appealing to a possible world in which S does A. What’s required instead is much stronger: there must be a possible world in which S does A and that world must have a history that is indistinguishable from the history of the actual world (Pike 1977, 216). And given that requirement, Jones is indeed unable to refrain from going to the beach. Plantinga responds (1986, 263ff.) by denying that he is committed to the logical possibility of Jones’s refraining being sufficient for his ability to refrain. He then adduces various examples intended to show that Pike’s requirement is too strong.

Without wading too deeply into the fray, I would like to highlight one additional facet of this dispute between Pike and Plantinga. Alston (1985) argues that Plantinga’s version of Ockhamism commits the Ockhamist to a compatibilist analysis of free will according to which it is within S’s power to perform A even if A is causally or logically necessitated. Since God’s belief that Jones will go to the beach entails that Jones will go to the beach, that action of Jones’s is logically necessitated. And yet, according to the Plantingian Ockhamist, Jones can refrain from going to the beach (in which case God would have held a different belief than the one he actually held.) Thus the Ockhamist is committed to compatibilism about free will. This is a compelling criticism that has, to my knowledge, not been addressed in the literature—by Plantinga or by anyone else. Pike, however, presses the point even further in his (1990), arguing that Plantinga’s response to the argument for theological compatibilism commits him not only to a compatibilist account of ability but to the absurdly permissive account (mentioned above) that collapses ability into possibility. Whether or not Pike’s stronger criticism hits the mark, it does seem as though we have good reason to doubt that Plantinga’s Ockhamist solution to the foreknowledge problem can preserve the libertarian freedom that theological compatibilists typically want to affirm.

Despite Pike’s (1977, 1990) rejection of Plantinga’s Ockhamism, he appears in his (1984) to be more sanguine about Marilyn Adams’s (1967) version of the Ockhamist response. Adams focuses, like (Pike’s interpretation of) Augustine, on one of the key assumptions of Pike’s argument: the de re notion of essential omniscience. Rather than denying the assumption, however, Adams points out that one of the implications of this assumption—namely, that the person who is in fact God would not be God were he to hold a false belief—is not an implication of the normal concept of personhood. Given this non-standard notion of personhood, perhaps

• (4) Jones was able to do something that would have brought it about that God didn’t exist

is, contrary to appearances, not impossible after all. Perhaps, to put it in Ockhamist terms, God’s existence (qua essentially omniscient being) is a soft fact. Adams’s (1967) analysis of the distinction between hard and soft facts turns everything into a soft fact (as Fischer (1983) demonstrated), and thus can’t establish the softness of facts about God’s existence—but the claim that the existence of a being who is essentially omniscient is a soft fact is at least worth exploring.

Omnipotence

The bulk of Pike’s early work focused on the apparent tension between human free will and divine omniscience—i.e., the tension between a fundamental element of human existence and one of the attributes that belongs to God as traditionally conceived. I would now like to consider Pike’s contribution to the discussion arising from the tension between another fundamental element of human existence (the presence of evil in the world) and another of God’s traditional attributes (omnipotence). The problem here—the problem of evil—is that it initially seems as though God, being omnipotent and perfectly good, both could and would create a world without any evil. But of course there is evil in the actual world, which suggests that there is no such being as an omnipotent, perfectly good God.

God and Evil

Pike (1958, 116) presents the problem of evil as an apparent contradiction between two theses. The first thesis is theological:

• (12) There exists an almighty and omniscient being who is a perfectly good person and who is God.

And the second is ethical:

• (13) There is something in the finite world, created by that being, which is evil.

If (12) and (13) are contradictory, then we are faced with a dilemma in which we are forced to deny that God exists or deny that evil exists (which for all practical purposes forces a denial of God’s existence). This dilemma presents a problem for any religious adherent or theologian who doesn’t want to deny that God exists. Of course, one way of solving the problem is to qualify the claim that God exists so that it becomes compatible with the falsity of (12). (For example, one could deny that God is perfectly good.) But this “theoretical” solution creates a practical problem for most theologians, who are typically committed to a conception of God that is not compatible with the falsity of (12). Given the nature of this commitment, Pike (1958, 117) identifies a desideratum on solutions to the practical problem: they must “involve not only the removal of the formal contradiction in question but the preservation of the essential religious and moral attitudes of the believer.” The practical problem of evil is thus the problem of eliminating a contradiction that strikes at the heart of religious commitment.

Pike’s insight is that the practical problem of evil, thus described, is an illusion. The problem is illusory first of all because there is in fact no contradiction between (12) and (13), but also because the only framework within which it arises is also a framework within which a solution is presupposed. I will consider both of these points in turn.

Pike (1958, 119) introduces two possibilities, either of which by itself is enough to demonstrate that (12) is consistent with (13). The first possibility is that evil exists in every possible world—i.e., that the existence of evil is logically necessary. The second possibility is that an almighty, perfectly good God could prevent evil but has a good reason for not doing so. What we need, then, in order to generate a contradiction, are two auxiliary premises:

• (14) The existence of evil isn’t logically necessary.

• (15) There exist no good reasons for God to allow evil.

Before we take a look at the second reason why the problem of evil (as described above) is illusory, I would like to pause briefly to consider an example that Pike discusses in defense of the need for (15). He points out that a perfectly good being could allow the occurrence of evil—or even do evil—as long as there is some motive or condition that renders that being blameless for that occurrence of evil. Pike then makes the following point about the connection between blameworthiness and moral responsibility (1958, 119):

Such a motive or condition could, of course, eliminate the property of moral blame without at the same time alleviating moral responsibility for the act in question. The father is morally responsible for the pain of the punished child, yet we would not feel it appropriate to blame him for his action if it were done with the child’s ultimate good in mind.

Although Pike skates over this point, I think it’s worth dwelling on for a bit. For there is a common, intuitively plausible belief that there is a necessary connection between moral responsibility and blameworthiness. Those who deny the connection (e.g., Fischer and Ravizza 1998, 83) often do so on the basis of examples such as that of a sea captain who decides to jettison his cargo in order to save his ship. Supposing that the captain’s decision was the result of a rational deliberative process, it seems that he is morally responsible for throwing the cargo overboard. But we should be reluctant to blame him for dumping the cargo, as (we can suppose) it was the right decision given the circumstances. Another example involves a woman who has been the victim of repeated physical and mental abuse (Fischer 2007, 186). If the victim eventually retaliates against her abuser, it seems we should say roughly the same thing about her as we said about the sea captain: she may be responsible for the retaliation, but we should be reluctant to blame her.

I have been convinced by these examples, but I do think it’s worth noting that Pike’s own example has the virtue (shared by much of his work) of being straightforward and compelling. Moreover, Pike’s example forestalls one common response to some of the other slightly more exotic examples in the literature. In response to the physical abuse scenario, for instance, some will deny that it’s a case of responsibility without blameworthiness. They will insist that if the woman is not blameworthy for retaliating, then she’s not morally responsible for doing so; and if, on the other hand, she is responsible, then she’s also blameworthy. But the parallel response does not seem available when it comes to the example of a father punishing his child. There is no reason to hesitate to ascribe moral responsibility to the father, and absolutely no reason to suggest that he is blameworthy for the pain that accompanies the punishment.

Let us now return to the main line of argument. Pike first demonstrated that (12) and (13) are not contradictory unless they are supplemented with (14) and (15). His second claim is that whatever problem remains is only of minor concern to the practicing theologian. We can see this by approaching the problem from the theologian’s perspective. She will take both God’s existence and the existence of evil (i.e., both (12) and (13)) as premises or axioms on which to build her theological system. It follows from these axioms either that evil exists in every possible world (in which case (14) is false) or that there is some good reason for the existence of evil (in which case (15) is false). If we consider evil’s being necessary as a “reason” for God’s allowing it to exist, then we can simply say that the axioms of this theological system dictate that there be some good reason(s) for the existence of evil. Within the theological framework, then, the only problem is ascertaining what those good reasons are. A system that has no proposals for what such reasons are might be incomplete, but it isn’t rationally deficient. Pike (1958, 121) applies this point to the study of Job:

This, so it seems to me, is the principal theological insight to be gained from a study of Job. Though particular solutions fail, Job’s theistic commitment undergoes no crisis-threat in the face of hardship. His faith is tested, but not threatened. When the theological thesis is affirmed, Job teaches, the crisis is past; some reason for evil is assured and the failure of specific proposals will be of relatively little consequence.

But even if we set aside the theological framework, the presence of evil in the world is not a decisive reason to deny God’s existence. Pike compares the situation to a scientific prediction: perhaps a scientific theory leads to a prediction that a planet will be observed at certain coordinates at a certain time. If the planet is not observed when and where it’s expected to be, then that’s a problem for the scientific theory—but only insofar as the independent support for the theory is weak. If the theory is extremely well-supported, then one failed prediction doesn’t force its rejection. There are other reasons why the prediction might have failed. The current situation is similar: The theological thesis might predict that there would be no evil in the world, but the failure of that prediction is not in itself reason enough to reject the thesis. The failed prediction (i.e., the existence of evil) is only a problem insofar as the support for God’s existence is weak.

And so the tables have been turned: rather than the theologian facing a dilemma, it is now the proponent of the problem of evil who faces a dilemma. The dilemma is that the problem is either philosophical or theological. If it’s philosophical, then it’s not a crisis—because “philosophy knows no crises” (cf. Pike 1970, 122). If the problem is theological, then it’s also not a crisis—because the framework within which it arises is one that already presupposes a solution.

Hume on Evil

Pike’s most well-known and oft-cited contribution to the problem of evil is his “Hume on Evil” (1963). He reiterates his challenge to the overly facile presentation of the problem that we saw above, and presents a sharpened version of the argument (1963, 183):

• (12) There exists an almighty and omniscient being who is a perfectly good person and who is God.

• (16) The world contains instances of suffering.

• (17) An omnipotent and omniscient being would have no morally sufficient reason for allowing instances of suffering.

This triad is clearly inconsistent, and if (17) is a necessary truth then (12) and (16) are contradictory (1963, 184). (If (17) is only contingently true, then there is a world in which (17) is false and in that world both (12) and (16) could be true; thus if (17) is contingently true then (12) and (16) aren’t contradictory.)

It’s worth noting here, as a scholarly aside, that the standard story about the problem of evil fails to do justice to Pike’s work on the issue in the 50s and 60s, and in particular fails to do justice to his (1958) and (1963). The standard story (cf. Peterson 1999, Gelinas 2009, Tooley 2010) is that Mackie and others pressed the “logical” problem of evil, which was refuted by Plantinga (1965, 1974, 1977) and subsequently reborn as the “evidential” problem of evil. And a key detail of the standard story is that Plantinga was the first to point out that all the theist needs, in order to defeat the argument, is a possible scenario that contains a morally sufficient reason for allowing suffering. In other words, the argument does not go through unless (17) (or something that plays the role of (17)) is a necessary truth. But here we see that Pike made this observation in 1963. (Moreover, in the last section of his (1963), Pike construes Philo’s second position as an evidential argument from evil, thus prefiguring the direction of the subsequent literature on the topic.) This, of course, is not to suggest any impropriety on the part of Plantinga (or anyone else), but merely to provide some examples in support of a claim from the introduction—namely that Pike’s contributions to the philosophy of religion are under appreciated.

Demonstrating that (17) is a necessary truth is perhaps a tall order, but we are pushed in that direction when we consider the different kinds of reasons that are typically counted as morally sufficient for allowing an instance of suffering. Pike offers five general circumstances that fit this description:

• (A) Lack of physical ability;

• (B) Lack of awareness (of the suffering);

• (C) Lack of awareness of physical ability;

• (D) Belief that the suffering will effect some future outweighing good;

• (E) Belief that the preventing the suffering will prevent some prior outweighing good.

These last two are worth elaborating on. The reason why a mother allows her son to go through the painful process of being vaccinated is her belief that the vaccination will prevent her son from contracting a variety of serious illnesses; and that counts, in virtue of falling under category (D), as a morally sufficient reason for allowing suffering. To illustrate category (E), Pike provides an example of the father who allows his daughter to eat a piece of birthday cake, knowing that her eating the cake will result in her feeling slightly ill later in the day. In this case, the father judges that the feeling of slight illness is outweighed by the good of enjoying birthday cake.

Now we are in a position to see just how formidable this sharpened challenge is. To the extent that we are reluctant to ascribe reasons of type (A)–(E) to God, we should be reluctant to affirm both (12) and (16). And of course (A)–(C) must be ruled out immediately; to affirm (12) is to deny that God can have any such reason for allowing suffering. What about reasons of type (D) and (E)? Well, as Hume’s Philo argues, reasons of type (D)—often associated with “soul-making” theodicies—don’t seem to be available to God, because presumably an omnipotent being could produce the desired good in some other way that wouldn’t involve suffering (Pike (1963, 187).

What about reasons of type (E)? Free will is one specific good that is supposed to count as a prior outweighing good, and thus worth having even if its presence entails the existence of suffering. But, as Mackie (1955) and others have argued, an omnipotent God might be able to create a world in which free creatures never go wrong. (Call free creatures who never do wrong “do-gooders.”) If there exists a possible world in which only do-gooders exist, then free will is not an outweighing good that would justify a reason of type (E) for allowing suffering.

Consider two responses to this appeal to possible worlds which contain only do-gooders. The first can be found in Smart (1961)—which Pike collects in his (1964)—who challenges the claim that the possibility of do-gooder worlds disqualifies free will as an outweighing good. He challenges that claim by disputing the notion that a world containing only do-gooders would be better than a world containing agents who can go wrong. His own claim is that it’s not clear what sort of moral value we can assign to creaturely actions in a do-gooder world. For agents never to go wrong, they would either have to be radically different from us (e.g., they would be immune to temptation), or they would have to be placed in circumstances that are radically different from ours (e.g., they would never face temptation). Is a world that fits one of these two descriptions better than a world that contains suffering? Perhaps not—for perhaps a do-gooder world (despite its label) has no positive moral value, and is thus no better than a world that contains some suffering but on balance more good than evil.

The second response can be found in Plantinga (1965, 217; quoted in Pike 1966a, 94), and it consists of the following argument, which utilizes the notion of “transworld depravity”:

• (12) There exists an almighty and omniscient being who is a perfectly good person and who is God.

• (18) God creates some free persons.

• (19) Every possible free person performs at least one morally wrong action.

• (20) Every actual free person performs at least one morally wrong action. (12, 18, 19)

• (21) God creates free persons who perform morally evil actions. (18, 20)

Plantinga’s claim, then, is that (12), (18), and (19) are compossible and entail (21) (via (20)). If this is right, then it follows that (12) and (21) are consistent. And if (12) is consistent with (21), then it follows that (12) and

• (16) The world contains instances of suffering.

cannot be contradictory, and as a result (17) is not the necessary truth that it was alleged to be (and that it needs to be to generate the problem of evil).

Pike is not convinced by this argument, and the bulk of his critique involves an analysis of the notion of a “possible person.” A discussion of the details would take us too far afield, but the upshot according to Pike is that Plantinga’s analysis of “possible persons” is going to allow for both morally perfect and morally imperfect possible persons. And if there are both perfect and imperfect possible persons, then God could actualize a world containing only the former.

Pike offers further criticism of the notion of transworld depravity in his (1979). His first main point is that a world full of free creatures who are transworld depraved is a risky thing to create. Plantinga appeals to “middle knowledge”—knowledge of what every possible creature would freely do in every possible circumstance—to mitigate the risk, but Pike identifies two problems with this appeal. The first is that on the Molinist picture (i.e., on the picture that appeals to middle knowledge), it seems that transworld depravity “is not just an affliction, it is an infliction” (Pike 1979, 465). What grounds this claim is the idea that for every possible creaturely essence (a notion which I will take to be roughly equivalent the notion of a possible person), it is possible that that essence be transworld depraved but also possible that that essence not be transworld depraved. In other words, there are possible worlds that contain only do-gooders; these are worlds containing felicitous instantiations of creaturely essences that are non-transworld-depraved. The question then becomes why God didn’t actualize one of those worlds rather than the one we find ourselves in. The most plausible answer seems to be simply to deny the claim that there are possible worlds that contain only do-gooders. But this is a bold claim.

It’s here that Pike incorporates some insights from Robert Adams (1977). Against the Molinist picture, Adams disputes the existence of the sorts of person-properties that would give God knowledge of what a creature would freely do in certain circumstances. God can know what a creature would probably freely do in every circumstance, but that’s the best he (or anyone) can do. This claim cuts both ways with respect to Plantinga’s appeal to transworld depravity. On the one hand, it would explain why it’s not possible for God to actualize a world containing only do-gooders. On the other hand, it would reintroduce the risk that God faces when creating the world: If all he has access to are “would probably” truths, then there’s always a possibility that things could go badly in that world, with the result being that the overall balance of value favors the bad. And given that possibility, it’s not at all clear that creating a world with free will is worth the risk.

Let us return now to “Hume on Evil.” Although Pike doesn’t find Plantinga’s argument convincing, he does find some promise in a suggestion arising out of Demea’s and Aquinas’ theodicies. The suggestion can be illustrated by considering a set of blocks, which exist in different shapes and can be put together to form various composite shapes. Suppose that the T-shaped block has the most aesthetic value, and that the L-shaped block has the least aesthetic value. Suppose further that a square is the most aesthetically valuable combination of blocks (including combinations that contain only one block). And suppose, finally, that there is only one arrangement of blocks that will produce a square—an arrangement which includes an L-shaped block. This, then, is a situation in which the least (aesthetically) valuable ingredient is a necessary component of the most valuable state of affairs, and obviously a model from which a (Demean) theodicy can be extrapolated. (Pike 1963, 189–90)

Pike points out an interesting shortcoming of this theodicy, namely that it presupposes God’s existence. For when we ask what reason we have for believing that the best of all possible worlds must necessarily contain suffering, it’s hard to see any reason other than the assumption that God exists. Despite this obvious limitation, this Demean theodicy does seem sufficient to unseat even the sharpened version of the argument from evil. For if it’s possible that the best possible world contains suffering as a necessary component (and this does seem possible, even if difficult to establish apart from the assumption that God exists), then

• (17) An omnipotent and omniscient being would have no morally sufficient reason for allowing instances of suffering.

is not a necessary truth and the truth of (16) does not imply the falsity of (12). Pike concludes that it is “far from clear that God and evil could not exist together in the same universe” (1963, 192).

Compare, along with Pike (1979, 472–3), this theodicy to the approach that Plantinga takes. Plantinga claims, on the basis of the possibility of transworld depravity, that instances of suffering are an unfortunate consequence of creaturely freedom. Augustine, on the other hand (who was another proponent of this type of theodicy), simply claims that wrong actions somehow contribute to the ultimate good. This claim follows directly from the possibility that evil is an essential component of the best possible world, which itself follows from the assumption that (12) is true—i.e., that the world is governed by an omnipotent and omniscient being who is perfectly good. Augustine’s theodicy, then, while perhaps not likely to convince his opponent, is at least notable for its elegance, simplicity, and lack of logical machinery.

There is another issue that divides theodicies: the question of whether human sin is imputable to God. The Thomistic theodicy, which Pike considers in his (1983), answers in the negative, but this answer is difficult to justify. If God is omnipotent (and omniscient), then it would seem that he is liable for the suffering caused by sin because he is fully capable of determining whether or not it occurs. The Augustinian theodicy, as noted above, ignores the imputability question and instead focuses on evaluating permitted evils in light of supposed greater goods. Its success does not require that we attempt to absolve God of responsibility for sin; he could even be solely responsible for sin. Thus we see that the Augustinian theodicy stands alone: it shares neither the questionable commitments nor the drawbacks of the Thomistic (and Plantingian) theodicies.

Omnipotence and God’s Ability to Sin

Pike’s (1969) begins with an argument for the incompatibility of omnipotence and perfect goodness, which runs as follows (1969, 209): Consider any consistently describable state of affairs that involves intense, undeserved suffering which serves no greater good. It seems clear that an individual who knowingly brings about this state of affairs would be morally reprehensible. An omnipotent being can bring about such a state of affairs (because it’s consistently describable), but a perfectly good being cannot bring it about (because bringing it about would render an individual morally reprehensible). And yet God is supposed to be both omnipotent and perfectly good, which would imply a contradiction. So it seems impossible that a being be both omnipotent and morally good.

Pike considers four proposals for dissolving this tension. The first proposal, which comes from Aquinas and is also developed by Anselm, is that omnipotence implies an inability to sin, because an ability to sin is a failing, or falling short. An omnipotent being cannot fall short, and thus cannot sin. But Pike responds that the failing here is a moral failure rather than a failure of power.

The second proposal—also from Aquinas—suggests that

• (22) God can sin if he wants to

is true, even though both the antecedent and consequent are impossible. Pike responds by pointing out that propositions like (22) don’t behave the way that conditional statements do. He offers two additional examples (1969, 211) of what are now, in honor of J. L. Austin, often called “biscuit conditionals”:

• (23) Jones can wiggle his ear if he wants to;

• (24) Jones has an ace in his hand if he wants to play it.

What makes these propositions unlike conditional statements is that if the consequent is false, then the whole statement is false whether or not the antecedent is true. If Pike is right about this, then (22) is, contra Aquinas, false in virtue of the consequent’s being necessarily false.

The third (Thomistic) proposal is that there are some situations that seem evil to us, but would not be evil if God were to bring them about. There are two problems with this proposal. First, if we are claiming that a “perfectly good” being can perform actions that by our lights seem evil, then we are using a notion of “perfectly good” that applies only to God. And whatever may be the case with respect to this special notion of perfect goodness, God would not be perfectly good in the ordinary sense. Second, we cannot make this move on behalf of St. Thomas because his method of explicating the divine attributes involves taking the ordinary senses of those attributes and then removing imperfections. According to this method, if an individual fails to satisfy the criteria for application of the ordinary term, then it follows that he will fail to satisfy the criteria for the perfected term. (Pike draws a geometrical analogy: if a shape doesn’t count as a triangle according to a relaxed conception of “triangle,” then it won’t count as a triangle according to a strict conception of “triangle.”)

Pike then considers a fourth suggestion, which is that a state of affairs in which God acts in a morally reprehensible way is not a consistently describable state of affairs. Unfortunately, this definitional gambit won’t work, as it would allow a being to be omnipotent even if he is constrained by some arbitrary and trivial limitation. If, for example, “Gid” is a being who makes only leather sandals (cf. Pike 1969, 214), then a state of affairs in which Gid makes a leather belt is not consistently describable and thus doesn’t rule him out as a candidate for omnipotence. But this can’t be right—because if Gid lacks the ability to make a leather belt, then he is clearly a limited being. And the same, it seems, should be said about a God who does not have the ability to act in a morally reprehensible way.

Notice, however, that there is some logical wiggle room here. For even if a state of affairs in which God acts in a morally reprehensible way is not consistently describable, it doesn’t follow that God lacks the ability to act in a morally reprehensible way. It merely follows that there is no possible world in which he does act in such a way. And this insight can be applied to the original problem. What we need is a way of interpreting

• (25) God cannot sin

without making it a logical truth (which would make the concept of omnipotence too permissive) and without limiting God’s abilities (which would make the concept of omnipotence too restrictive). Pike offers us this third interpretation by construing (25) as expressing “material assurance” that God, while having the ability to sin, nevertheless will not sin (1969, 215). On this understanding of (25), God is strongly disposed not to sin: “he cannot bring himself to act in a morally reprehensible way” (1969, 216). God has the ability to sin, but we can have complete assurance that he won’t.

Mysticism

Pike’s work on mysticism represents something of a change of gears from his treatment of the divine attributes, but it exemplifies the same relentlessly lucid approach that characterizes the rest of his writing. In what follows, I will focus on his examination of the epistemic status of mystical visions—i.e., whether and in what sense they can be sources of knowledge.

Pike’s first published contribution to the literature on mysticism is his “On Mystic Visions as Sources of Knowledge” (1978). In this piece he provides a helpful taxonomy of mystic visions, or apprehensions (I will use the terms interchangeably), and lays down some principles for evaluating their reliability. He also considers a critique (MacIntyre 1964) of the attempt to rely upon mystic visions as evidence for religious belief.

He begins by distinguishing between two kinds of revelation: public and private. The public revelation consists primarily of the Scriptures, whereas private revelations consist primarily of God’s communication with individuals through mystic visions. These private revelations can be divided according to their content: sometimes their content overlaps with existing Christian doctrine, and sometimes it does not. Of these private revelations whose content does not overlap with Christian doctrine, we can further distinguish between those that contain theological content and those that contain non-theological content. The epistemic value of a mystic vision will partly depend on which category the vision falls into, but there are two categories that are of particular interest. The first category includes those visions that produce new doctrine: these would be private revelations whose content is theological in nature but does not overlap with Christian doctrine. The second category of interest includes visions whose content re-confirms (and thus overlaps with) existing doctrine: such revelations will of course contain theological content as well.

Pike then introduces a guiding principle that plays an important role in traditional mystical theology (1978, 216):

• (26) If a mystic vision containing a revelation is produced by God, then the vision in question is a reliable source of information: the revelation therein contained is true.

How is one to determine whether a particular mystic vision is produced by God? Pike introduces two tests. The first is a “spiritual effects” test: a vision is produced by God only if it generates positive affective states and dispositions toward virtue (1978, 219). The second test is a doctrinal test: a vision is produced by God only if it is consistent with received doctrine. (One implication of this second test is that a private revelation is incapable of forcing a revision or correction of a doctrine that has been formulated on the basis of public revelation.)

Implicit in the doctrinal test is an important point that comes out even more clearly in Pike’s (1992)—namely that the “articles of standard dogmatics are taken as axiomatic” for the mystical theologian (1992, 35). In other words, it is a fundamental assumption in Christian mystical theology that the teachings of Church doctrine are true: they are the epistemological foundation of the mystical framework.

MacIntyre (1964) criticizes this framework, arguing that it leads to circular reasoning about the evidential value of a mystic vision. MacIntyre claims, moreover, that mystic visions cannot provide any degree of support for the truth of a particular religious belief. The target of this complaint is a mystical pattern of argument in which a vision is verified with reference to the congruence between the content of that vision (the “revelation therein”) and the content of Christian doctrine. MacIntyre takes this “congruence” to involve entailment (cf. Pike 1978, 223–4): a vision is genuine (i.e., trustworthy) if the revelation therein is entailed by some point of doctrine. But it is not hard to see that this pattern of argument is circular:

The trouble here, I think MacIntyre is saying, is that the argument for reliability makes use of the assumption that [the] doctrine is true and thus that the revelation contained in the relevant apprehension—which is entailed by the very doctrine for which the apprehension is supposed to be supplying support—is true. (Pike 1978, 224)

While this objection is a good one as far as it goes, Pike points out that it misses the mark if designed to undercut actual mystical practice. For, as we saw above, the important question to answer about a particular revelation is not whether it’s true, but whether the corresponding vision was produced by God. And whether the vision was produced by God is determined by the spiritual effects test and the doctrinal test. More importantly, though, it appears that MacIntyre has misconstrued the doctrinal test. The doctrinal test is one of consistency. This means that if the propositional content of a particular private revelation contradicts a publicly revealed doctrine, then that private revelation fails the test. All other propositions—even those not entailed by Christian doctrine—pass the test.

As long as a private revelation produces the right sort of affective states and dispositions, and the propositional content of the revelation is not equivalent to the negation of a proposition entailed by church doctrine, then that revelation (argues the mystic) is reliable. This argument form applies whether or not the content of the revelation overlaps with doctrine—and even when there is overlap, the premise used to establish reliability says nothing about the truth of the propositional content. Thus it’s clear that the argument for the reliability of the relevant vision need not be circular.

Two additional remarks are worth emphasizing. First, as noted above, these tests for the reliability of a mystic vision clearly presuppose the truth of a body of doctrine. Thus an argument for the reliability of a particular vision will only be effective if its audience includes someone who shares those antecedent doctrinal commitments. Second, we might ask what support can be offered for the principle in (26). In other words, why should we assume that a revelation produced by God always contains true information? Aren’t there some situations in which God—even a morally perfect God—might have good reason to produce a revelation whose propositional content is false? This, according to Pike (1978, 232) is the element of the theory that needs the most attention.

In a later piece on mystic visions (1986), Pike moves from a defense of the mystical procedure for establishing the reliability of visions to a development of a positive view about the epistemic significance of such visions. Pike’s positive proposal, which is an extrapolation from the writings of John of the Cross, is that the epistemic significance of a mystic vision lies not in the information that it communicates, but rather in the depth of understanding that it generates.

John’s view on mystic visions is, on its face, surprising: in essence he recommends to the mystic that she ignore and perhaps even reject the information communicated in a vision. This is because the meaning of whatever message God intends to convey is a spiritual meaning that cannot be captured using the ordinary meanings of terms in our language. (The same holds for the ordinary meanings that we might derive from visual imagery.) Mystic visions are thus, according to Pike, “opaque as regards their meaning” (1986, 22). One of the examples that John cites is that of Abraham—who was told by God, and thus came to believe, that he would rule the land of Canaan. And yet he never ruled in Canaan. One might be tempted to conclude that the revelation was false, but instead John diagnoses a miscommunication. The message that God intended to convey was that Abraham’s descendants would rule Canaan, not that Abraham himself would rule Canaan. But, as Pike points out, this cannot be the correct diagnosis of the case (at least if John is right about the spiritual meaning of messages from God). If the meaning of God’s messages cannot be captured in ordinary language, then any attempt to clarify that message in ordinary language (e.g., by pointing out that the promise was intended to apply to Abraham’s descendants rather than to Abraham himself) is going to fail. Pike (1986, 23) puts the point eloquently: “It is one of those special cases in which the import of the theory undermines the argument used in its support.”

Of course, the theory itself might be adequate even though the argument cannot stand. But Pike marshals a medieval argument against John’s theory (1986, 23–4), the upshot of which is that an omnipotent and omniscient God could not fail to communicate a message that he wanted to communicate. The problem is that John is making two claims (about the epistemic value of mystic visions) that are incompatible. He is claiming on the one hand that God’s purpose in producing a mystic vision is to convey information, but also claiming on the other hand that mystic visions cannot accurately convey the spiritual meaning of God’s message. But he can’t have it both ways; he needs to “either go the distance or return to the fold” (Pike 1986, 29). Pike offers a proposal, on John’s behalf, that allows him to “go the distance” while still preserving epistemic importance for mystical visions.

The proposal, in short, is that mystic visions are not intended to communicate information but rather to produce spiritual growth; they do for the soul what vitamins do for the body (Pike 1986, 31). One way they do this is by adding an additional dimension or depth to one’s understanding of an already-known truth. Pike develops an evocative example in the service of this proposal, which I will summarize as follows. Imagine a dream experience, after which I find myself thinking, “My father loved me very deeply.” This proposition is not expressed in the dream, and none of the imagery of the dream suggests the proposition; it is just that when I wake up I am entertaining that proposition in a salient way. This is a proposition that I already know, but the dream strengthens and deepens my grasp of that proposition. And this, according to Pike’s proposal, is the way mystic visions work as well. They are not essentially communicative, and any communication that happens is causal rather than interpretive. The auditory or visual content of the vision is unimportant; what’s important instead are the cognitive and conative effects of the vision—the cognitive wallop that it delivers (Pike 1986, 36). This proposal is plausible in its own right, but it also harmonizes with two tenets of mystical theology. The first tenet (as noted above) is the foundational assumption of the truth of Church doctrine. Given this assumption, it is not surprising that the primary function of mystical visions would be to reinforce or supplement these axiomatic truths. Second, this proposal accords with the fact that mystic visions or experiences are largely passive: “they are imposed or, as is said, ‘infused’ by God and are not to be thought of as achievements on the part of experiencing mystic” (Pike 1992, 77). Removing the interpretive element from mystic visions thus respects the passive nature of mystical experience in general.

Nelson Pike’s treatment of the epistemic value of mystic visions recapitulates his general approach—and exemplifies his contribution—to the philosophy of religion: it consists of a charitable and incisive reading of an important yet incomplete argument, culminating in a cautious positive proposal that is admirable for its clarity and elegance.

Pike Bibliography

Below is a list of papers and books that were written by Nelson Pike and which are cited above.

1958. “God and Evil: A Reconsideration.” Ethics 68: 116–24.

1963. “Hume on Evil.” The Philosophical Review 72: 180–97.

1964 (ed.). God and Evil: Readings on the Theological Problem of Evil. Englewood Cliffs, NJ: Prentice-Hall.

1965. “Divine Omniscience and Voluntary Action.” The Philosophical Review 74: 27–46.

1966a. “Plantinga on the Free Will Defense: A Reply.” The Journal of Philosophy 63: 93–104.

1966b. “Of God and Freedom: A Rejoinder.” The Philosophical Review 75: 369–79.

1969. “Omnipotence and God’s Ability to Sin.” American Philosophical Quarterly 6: 208–16.

1970. God and Timelessness. New York: Shocken Books.

1977. “Divine Foreknowledge, Human Freedom, and Possible Worlds.” The Philosophical Review 86: 209–16.

1978. “On Mystic Visions as Sources of Knowledge.” In S. T. Katz (ed.), Mysticism and Philosophical Analysis, 214–34. New York: Oxford University Press.

1979. “Plantinga on Free Will and Evil.” Religious Studies 15: 449–73.

1983. “Over-Power and God’s Responsibility for Sin.” In A. J. Freddoso (ed.), The Existence & Nature of God, 11–36. Notre Dame, IN: University of Notre Dame Press.

1984. “Fischer on Freedom and Foreknowledge.” The Philosophical Review 93: 599–614.

1986. “John of the Cross on the Epistemic Value of Mystic Visions.” In R. Audi and W. J. Wainwright (eds.), Rationality, Religious Belief, and Moral Commitment: New Essays in the Philosophy of Religion, 15–37. Ithaca, NY: Cornell University Press.

1990. “Alston on Plantinga and Soft Theological Determinism.” International Journal for Philosophy of Religion 27: 17–39.

1992. Mystic Union: An Essay in the Phenomenology of Mysticism. Ithaca, NY: Cornell University Press.

1993. “A Latter-Day Look at the Foreknowledge Problem.” International Journal for Philosophy of Religion 33: 129–64.