B. Sin and Knowledge

The most important cognitive consequence of sin, therefore, is failure to know God. And this failure can have further cognitive consequences. At present and especially in academia, there is widespread doubt and agnosticism with respect to the very existence of God. But if we don’t know that there is such a person as God, we don’t know the first thing (the most important thing) about ourselves, each other, and our world. That is because (from the point of view of the model) the most important truths about us and them is that we have been created by the Lord and utterly depend upon him for our continued existence.²⁶⁰²⁶⁰   In this connection, consider the despised creationists, who believe that the world is only ten thousand years old: they are ignorant, pitifully ignorant about when God created the world. From the point of view of the model, this ignorance pales into utter insignificance compared with that of many of their cultured detractors, who foolishly believe that there is no God and thus (naturally enough) are ignorant of the vastly more important fact that the world was, indeed, created by God. We don’t know what our happiness consists in, and we don’t know how to achieve it. We don’t know that we have been created in the image of God, and we don’t grasp the significance of such characteristically human phenomena as love, humor, adventure, science, art, music, philosophy, history, and so on.

Can we take things a step further yet? According to John Calvin, “As soon as ever we depart from Christ, there is nothing, be it ever so gross or insignificant in itself, respecting which we are not necessarily deceived.”²⁶¹²⁶¹   Commentaries on the First Book of Moses, Called Genesis, tr. John King (Edinburgh: Calvin Translation Society, 1847); reprinted by Baker Book House (Grand Rapids, 1979). Perhaps Calvin means only what we have already noted: one who doesn’t know God fails to know the most important truth about anything else. He may mean to go even further, however: perhaps he means to say that those who don’t know God suffer much wider ranging cognitive deprivation and, in fact, don’t really have any knowledge at all. (This view is at any rate attributed (rightly or wrongly) to some of his followers, for example, Cornelius van Til.) That seems a shade harsh, particularly because many who don’t believe in God seem to know a great deal more about some topics than most believers do. (Could I sensibly claim, for example, that I know more logic than, say, Willard van Orman Quine, even if I can’t do any but the simplest logic exercises, on the grounds that at any rate I know something about logic and he, being an unbeliever, knows nothing at all about that subject or indeed anything else?) As it stands, this suggestion is desperately wide of the mark; surely many nontheists 218do know some things, for example, their age to the nearest year or so, to whom if anyone they are married, and which university it is that employs them. (If this weren’t so, contemporary academia would display even more confusion than it does.)

1. Sin and Skepticism

A couple of less sweeping views however have a great deal to be said for them. One who is agnostic about the existence of God may also be agnostic about his origin and his place in the universe. In this section, I shall argue that one who displays a certain kind of agnosticism with respect to his origin and place in the universe, and also grasps a certain cogent argument, will not, in fact, know anything at all; nothing he believes will have warrant sufficient for knowledge. To explore this suggestion, we may begin by considering the Scottish philosopher David Hume. Thomas Reid, Hume’s great contemporary and antagonist, took Hume to be a skeptic with respect to external objects, an enduring self, other minds, causality, the past, and so on.²⁶²²⁶²   Although Reid’s view has been the majority opinion with respect to Humean exegesis, there has always been a minority opinion according to which Hume really wasn’t a skeptic at all. This striking divergence is testimony to the fact that Hume is a black enigma: a certain surface clarity masks a deep underlying murkiness that makes confident interpretation impossible. As Reid sees him, Hume thinks that there is something wrong in believing the things we ordinarily do: it isn’t as if Hume simply announces that as a matter of fact we don’t really know all we think we know about external objects, causal relations, our own selves. Perhaps that would be bad enough, but there is something much deeper.

We can see what by considering the Hume of the conclusion of Book I of the Treatise.²⁶³²⁶³   Treatise of Human Nature, ed. L. A. Selby-Bigge (Oxford: Clarendon Press, 1951; first published in 1739), pp. 263ff. Subsequent page references to the Treatise are to this edition. Here he isn’t coolly announcing, as a mildly interesting fact about us, that fewer of our beliefs constitute knowledge than we ordinarily think. Instead, he finds himself in a sort of existential crisis; he simply doesn’t know what to believe. When he follows out what seem to be the promptings and leading of reason, he winds up time after time in a black coal pit, not knowing which way to turn:

Where am I, or what? From what causes do I derive my existence, and to what condition shall I return? Whose favour shall I court, and whose anger must I dread? What beings surround me? and on whom have I any influence, or who have any influence on me? I am confounded with all these questions, and begin to fancy myself in 219the most deplorable condition imaginable, inviron’d with the deepest darkness, and utterly depriv’d of the use of every member and faculty. (p. 269)

Of course this is Hume in his study, sometime before he emerges for that famous game of backgammon. Nature herself, fortunately, dispels these clouds of despair: she “cures me of this philosophical melancholy and delirium, either by relaxing this bent of mind, or by some avocation, and lively impression of my senses, which obliterate all these chimeras. I dine, I play a game of backgammon, I converse, and am merry with my friends” (p. 269).

Still, the enlightened person, Hume thinks, holds the consolations of Nature at arm’s length. She knows she can’t help acquiescing in the common illusion, but she maintains her skepticism of “the general maxims of the world” and adopts a certain ironic distance, a wary double-mindedness: “I may, nay I must yield to the current of nature, in submitting to my senses and understanding; and in this blind submission I shew most perfectly my sceptical disposition and principles” (p. 269). This is the irony of the human condition: those who are enlightened can see that what nature inevitably leads us to believe is false, or arbitrary, or at best extremely dubious; they also see, however, that even the best of us simply don’t have it in them to successfully resist her blandishments. We can’t help believing those “general maxims,” or if we can, it is only for brief periods of time and in artificial situations. No one can think Humean thoughts about, say, induction, when under attack by a shark or when clinging precariously to a rock face high above the valley floor. (You won’t find yourself saying, “Well, I do of course believe that if this handhold breaks out, I’ll hurtle down to the ground and get killed, still [fleeting sardonic, self-deprecatory smile] I also know that this thought is just a deliverance of my nature and is therefore not really to be taken seriously.”) Still, in other circumstances, one can take a sort of condescending and dismissive stance with respect to these promptings of nature; in reflective moments in my study, for example, I see through them. As a rational creature, I can rise above them, recognizing that they have little or nothing to be said for them. Indeed, I see more: this skepticism is itself a reflexive skepticism; it arises even with respect to this very thought; this very doubt, this feeling of superiority, this seeing through what our natures impose on us, is itself a deliverance of my nature and is thus as suspect as any other. The true skeptic, says Hume, “will be diffident of his philosophical doubts, as well as of his philosophical conviction” (p. 273).²⁶⁴²⁶⁴   And this leads to the scandal of skepticism: if I argue to skepticism, then of course I rely on the very cognitive faculties whose unreliability is the conclusion of my skeptical argument.

In these passages, therefore, Hume isn’t shamefacedly confessing an epistemic weakness or flaw, rather as a victim of neurosis or mental disease might. (“Doctor, I often find that I simply can’t bring myself to believe that induction will continue to work, or that I myself have existed for a good long time, or that there really are other people or external objects.”) No; this multiply skeptical position, he thinks, is somehow the right one, the one that the man of sense (at least the man of philosophic sense) will adopt. The rest of us who unthinkingly acquiesce in the promptings of nature, who without a thought believe in causal connection, induction, persistent selves, external objects—the rest of us are from this perspective naive or foolish, unwitting dupes of our own nature. Hume is a sort of Presbyterian of the intellect; we are all, sage and ingenue alike, enmeshed in the toils of an original sin of the mind (and here perhaps we can see a lingering influence of the Calvinism of his youth). Of course Hume might claim that at least he has the advantage of recognizing that (ordinarily) he is a dupe. In this regard, he may seem like the publican in Jesus’ parable, who at any rate had the grace to confess that he was indeed a sinner. But the fact is Hume is really more like the Pharisee. He isn’t confessing a frailty or shortcoming, hoping for a cure; he is arguing, as he sees it, from a position of strength or at least insight; the rest of us who unthinkingly accede to the promptings of nature are the ones who suffer from intellectual shortcoming. More than that, we are irrational, in the Humean view, in that reason, carefully preserved from the corrupting influence of everyday attitudes, enjoins this skepticism upon us. To fail to accept it is to fail to follow reason, to go against its teachings, and in that sense to fall into irrationality.

Now Thomas Reid takes issue with Hume (at any rate Hume as he sees him) at just this point. He sees Hume as standing with Descartes in thinking that the deliverances of perception, memory, induction, sympathy, testimony, and any other faculty we might have must be validated before the bar of reason and consciousness. That is, none of these faculties can reasonably be trusted until it has been shown to be reliable by an argument that meets two conditions. First, the argument in question must start from premises that are either self-evident (like elementary truths of arithmetic), or else deliverances of consciousness: such propositions about my own mind as that I seem to see a horse, or am appeared to redly, or believe that the Orkney Islands are north of Aberdeen. Second, the argument must be such that each of its steps is self-evidently valid.

Now Descartes thought that in fact the other sources of belief could be legitimated by reason and consciousness. He thought first to establish the reliability of reason itself by giving a reasoned (rational) proof that we have been created by a benevolent God who is nondeceptive 221(and here we fall into that distressing Cartesian circle), but God would be a deceiver if the world weren’t very much like our perceptual faculties reveal it to be. As Reid sees it, Descartes is mistaken at several points; the point of present interest, however, is Descartes’s confidence that the reliability of those other sources can be established by reason. It took the work of modern philosophy from Descartes to Hume, so Reid thinks, to show that this is in fact a chimera, a will-o’-the-wisp; it simply can’t be done. (The inevitable failure of this Cartesian project was therefore wholly evident to Reid some two hundred years or so before Rorty and Quine took this failure as a reason for proclaiming the death of epistemology [Rorty²⁶⁵²⁶⁵   See his Philosophy and the Mirror of Nature (Princeton: Princeton University Press, 1979).] or its transmogrification into empirical psychology [Quine²⁶⁶²⁶⁶   As in “Epistemology Naturalized,” in Ontological Relativity and Other Essays (New York: Columbia University Press, 1969).].)

Now one reaction would be to see this condition as interesting and perhaps even mildly regrettable, but of no real importance: these other sources of belief are perfectly acceptable, whether or not we can find arguments of the above sort for their reliability. Reid’s Hume, however, takes quite a different tack; he takes it to be a sign of foolishness or error or dupery (in any event, part of the deplorable human condition) to accept the testimony of any source whose veracity hasn’t been (or, worse, can’t be) established by way of consciousness and reason. He therefore concludes that the rational course is to reject these beliefs (given that we can’t show in the way in question that their sources are reliable), even if because of nature’s imperious edicts we can’t actually follow that austere prescription.

This strikes Reid as a piece of consummate arbitrariness:

The sceptic asks me, Why do you believe the existence of the external object which you perceive? This belief, sir, is none of my manufacture; it came from the mint of Nature; it bears her image and superscription; and, if it is not right, the fault is not mine: I ever took it upon trust, and without suspicion. Reason, says the sceptic, is the only judge of truth, and you ought to throw off every opinion and every belief that is not grounded on reason. Why, sir, should I believe the faculty of reason more than that of perception?—they came both out of the same shop, and were made by the same artist; and if he puts one piece of false ware into my hands, what should hinder him from putting another?²⁶⁷²⁶⁷   Thomas Reid’s Inquiry and Essays, ed. Keith Lehrer and Ronald E. Beanblossom (Indianapolis: Bobbs-Merrill, 1975), pp. 84–85.

I believe that Reid is substantially right here; the Humean skeptic is arbitrary.²⁶⁸²⁶⁸   But perhaps not entirely arbitrary; see Warrant: The Current Debate, pp. 100ff. But this is not the place for a discussion of this point: what I want to argue instead is that Hume has a different reason for his skepticism, a reason shared by anyone who concurs with him in agnosticism about our origin and the origin of our cognitive faculties. Suppose, for one reason or another, you give up this idea that we have been created by a benevolent deity. Perhaps with Hume you adopt instead a thoroughgoing agnosticism: there is simply no way to know whether there is any being at all like God, no way to know whether there is a divine being who created the world, no way, indeed, to know anything about the ultimate origin of the world or of the ultimate origin of ourselves and our cognitive faculties. “Our experience,” he says, “so imperfect in itself and so limited both in extent and duration, can afford us no probable conjecture concerning the whole of things.”²⁶⁹²⁶⁹   Dialogues concerning Natural Religion, ed. Richard Popkin (Indianapolis: Hackett Publishing, 1980), p. 45. Perhaps the world owes its existence to intelligent design: just as likely, though (at least as far as we can tell), it owes it to animal or even vegetative generation (perhaps comets are seeds and our world has arisen from one); and there are a thousand other possibilities, some of them canvassed with grace and style in the Dialogues concerning Natural Religion. Hume’s²⁷⁰²⁷⁰   Or at any rate Philo’s; I make no pretense to settle the question of who speaks for Hume in the Dialogues, something Hume artfully conceals. conclusion there, it seems, is that

In such questions as the present [cosmogony, the origin of the universe], a hundred contradictory views may preserve a kind of imperfect analogy, and invention has here full scope to exert itself. Without any great effort of thought, I believe that I could, in an instant, propose other systems of cosmogony which would have some faint appearance of truth: though it is a thousand, a million to one if either yours or any one of mine be the true system. (Dialogues, p. 49)

He adds a bit later that on this topic, “A total suspense of judgment is here our only reasonable resource” (p. 53). Hume so understood has no idea at all how the world got here, how rational creatures such as we ourselves have arisen, and what the origin and provenance of our rational or belief-producing faculties might be.

Now turn to the question whether our cognitive faculties are reliable and do, in fact, produce for the most part true belief. Given Hume’s complete agnosticism about the origins of his cognitive faculties, something like his deeply agnostic attitude to that question is no more than sensible. For suppose Hume asks himself how likely it is that our cognitive faculties are reliable, given his views (or rather lack of views) about the origin and provenance of ourselves and those 223faculties. What is the probability that our faculties produce the considerable preponderance of true belief over false required by reliability, given his views of their origin and purpose (if any)? I should think he would have to say that this probability is either low or inscrutable—impossible to determine. From his point of view, there are innumerable scenarios, innumerable ways in which we and our cognitive faculties could have come into being: perhaps we have been created by God, but perhaps we and the world are the result of some kind of vegetative principle, or a result of copulation on the part of animals we have no knowledge of, or the result of Russell’s accidental collocation of atoms, or of . . . . On many of these scenarios, our cognitive faculties wouldn’t be reliable (although they might contribute to fitness or survival); perhaps on others they would be reliable; on balance, one just wouldn’t know what to think about this probability.

We can see this more fully as follows. Let R be the proposition that our cognitive faculties are reliable: now what is the likelihood of R? As Reid points out, we all instinctively believe or assume that our cognitive faculties are indeed reliable; but what is the probability of that assumption, given the relevant facts? Well, what are the relevant facts? First, they would be facts about those faculties: the probability of R given (relative to) the population of China would not be relevant. And presumably the relevant facts would be facts about how these faculties originated; whether they were designed; if so, by whom and with what end in view; what constraints governed their development; and what their purpose and function is, if, indeed, they have a purpose and function. Were they, as Reid thought, created in us by a being who intends that they function reliably to give us knowledge about our environment, ourselves, and God himself—all the knowledge needed for us to attain shalom, to be the sort of beings God intended us to be? On that scenario, the purpose of our cognitive faculties would be (in part, at least) to supply us with true beliefs on those topics, and (given that they are functioning properly) there would be a high probability of their doing just that.

Did they, by contrast, arise by way of some chance mechanism, something like the mindless swerve of atoms in the Democritian void? What is the likelihood, on that possibility, that our cognitive faculties are reliable? Well, you might think it pretty low. More likely, you may think that you simply can’t say what that probability is: perhaps it is high (though presumably not very high), perhaps it is low; you simply can’t tell.²⁷¹²⁷¹   We aren’t thinking here of Bayesian personal probability but of some kind of objective probability, the sort of probability Hume has in mind when he says that “it is a thousand, a million to one if either yours or any one of mine be the true system.” There will be many more such scenarios, says Hume, some involving vegetative origin, some copulative origin, 224some still other kinds of origin; with respect to them, too, the probability that our cognitive faculties are reliable is simply inscrutable. So first, Hume thinks his grasp of the whole set of relevant scenarios is at best infirm; second, with respect to many of these scenarios, those possible origins, the probability of R is inscrutable; and finally, the probability with respect to any of these scenarios that it is in fact the truth of the matter is also, as far as Hume is concerned, quite inscrutable.

But that means that the probability of R, given Hume’s agnosticism, is also inscrutable for Hume. Let F be the relevant facts about their origin, purpose, and provenance: my claim is that, for Hume, P(R/F) (the probability of R on F) is inscrutable. He simply doesn’t know what it is and has no opinion about its value, although presumably it wouldn’t be very high. Another way to put it: the probability of R, given Hume’s agnosticism, is inscrutable.

And that gives Hume a reason to be agnostic with respect to R as well; it gives him a reason to doubt that R is, in fact, true. For our cognitive faculties, our belief-producing mechanisms, are a bit like measuring instruments (more exactly, measuring instruments under an interpretation). Our faculties produce beliefs; for each belief, there is the content of that belief, the proposition believed, a proposition that is true if and only if the belief is true. Now a state of a measuring instrument (relative to a scheme of interpretation) can also be said (in an analogically extended sense) to have content. For definiteness, consider a thermometer and suppose its pointer is resting on the number 70. Given the natural scheme of interpretation, this state can be said to have the content that the ambient temperature is 70˚ F. And of course a thermometer is reliable only if the propositions it delivers in this way are for the most part true, or nearly true.

Imagine, then, that you embark on a voyage of space exploration and land on a planet revolving about a distant sun. This planet has a favorable atmosphere, but you know little more about it. You crack the hatch, step out, and immediately find something that looks a lot like a radio; it periodically emits strings of sounds that, oddly enough, form sentences in English. The sentences emitted by this instrument express propositions only about topics of which you have no knowledge: what the weather is like in Beijing at the moment, whether Caesar had eggs on toast on the morning he crossed the Rubicon, whether the first human being to cross the Bering Strait and set foot on North America was left-handed, and the like. A bit unduly impressed with your find, you initially form the opinion that this quasi radio speaks the truth: that is, the propositions expressed (in English) by those sentences are true. But then you recall that you have no idea at all as to what the purpose of this apparent instrument is, whether it has a purpose, or how it came to be. You see that the probability of its being reliable, given what you know about it, is for you inscrutable. Then (in the absence of investigation) you have a defeater 225for your initial belief that the thing does, in fact, speak the truth, a reason to reject that belief, a reason to give it up, to be agnostic with respect to it. Relative to your beliefs about the origin, purpose, and provenance of this apparent instrument, the probability that it is a reliable source of information is low or (more likely) inscrutable. And that gives you a defeater for your original and hasty belief that the thing really does speak the truth. If you don’t have or get further information about its reliability, the reasonable course is agnosticism about that proposition.

The same goes, I think, in the case of Humean views (or non-views) about our origins and the origin and purpose, if any, of our cognitive faculties. Suppose I join Hume in that agnosticism. Then P(R/F) is for me inscrutable (as for Hume); I have no idea what the probability of my faculties being reliable is, given the relevant facts about their origin and purpose. But then I have a defeater for my original belief or assumption that my faculties are in fact reliable. If I have or can get no further information about their reliability, the reasonable course for me is agnosticism with respect to R, giving it up, failing to believe it. It isn’t that rationality requires that I believe its denial, but it does require that I not believe it.

Suppose, therefore, that I am agnostic with respect to R: I believe neither it nor its denial. And now consider any belief B I have: that belief, of course, will be a deliverance of my cognitive faculties. However, I don’t believe that my cognitive faculties are reliable—not because I’ve never thought about the question, but because I have thought about it and seen that P(R/F) is inscrutable for me. Well, what does rationality require with respect to this belief B? The clear answer seems to be that I have a defeater for this belief too, a reason to withhold it, to be agnostic with respect to it. Perhaps it isn’t possible, given my nature, that I be agnostic with respect to it, at least much of the time; as Hume says, nature may not permit this. Still, this agnosticism is what reason requires, just as Hume suggests (though for different reasons). And we can take one further step with Hume. Because B is just any belief I hold—because I have a defeater for just any belief I hold—I also have a defeater for my belief that I have a defeater for B. This universal, all-purpose defeater provided by my agnosticism is also a defeater for itself, a self-defeating defeater.²⁷²²⁷²   Of course this raises problems: if I have a defeater-defeater (a defeater for my defeater for R), then don’t I thereby lose my defeater for R? Am I back where I was before I acquired the defeater for R? No; for my defeater-defeater is also a defeater for R. For explanation and detail, see part IV, section E, “The Dreaded Loop,” from my “Naturalism Defeated,” presently unpublished. And hence this complex, confusing, multilayered, reflexive skepticism Hume describes, a skepticism in which I am skeptical of my beliefs and also of my doubts, and of the beliefs that lead to those 226 doubts, and of my doubts with respect to those doubts, and the beliefs leading to them. Thus the true skeptic will be skeptical all the way down; he “will be diffident of his philosophical doubts, as well as his philosophical conviction.”

Here we can imagine the following response: “Hey, hang on a minute! You said Hume and any similarly situated agnostic has a defeater for R, a belief to which he is inclined by nature—and you added that the rational course for them therefore is to give up belief in R—provided they have no other information about the reliability of their faculties. But what about that strong natural inclination to believe that our faculties are in fact reliable? Doesn’t that count as ‘other information’?” According to Reid (who might object to being pressed into service in defense of Hume), this belief in the reliability of our faculties is a first principle:

Another first principle is—That the natural faculties, by which we distinguish truth from error, are not fallacious. (275)

He goes on:

If any truth can be said to be prior to all others in the order of nature, this seems to have the best claim; because, in every instance of assent, whether upon intuitive, demonstrative, or probable evidence, the truth of our faculties is taken for granted. . . . (277)

Surely there is truth here: this conviction is one normal human beings ordinarily have, and, as Reid gleefully points outs, even skeptics also seem to assume, in the course of ordinary daily living, to be sure, but most poignantly when proposing their skeptical arguments, that their faculties are functioning reliably. Very few skeptics, in offering their skeptical arguments, preface the argument by saying something like, “Well, here is an argument for general skepticism with respect to our cognitive faculties; of course I realize that the premises of this argument are themselves produced by cognitive faculties whose reliability the conclusion impugns, and of whose truth I am therefore extremely doubtful.”

But our question is whether this belief can sensibly be pressed into service as information that can defeat the defeater provided for R by Hume’s agnosticism about the origin and provenance of ourselves and our faculties. As Reid clearly sees, it cannot. If the general reliability of our cognitive faculties is under question, we can’t hope to answer the question whether they are reliable by pointing out that these faculties themselves deliver the belief that they are, in fact, reliable. “If a man’s honesty were called into question,” says Reid, “it would be ridiculous to refer it to the man’s own word, whether he be honest or not” (276). Concede that it is part of our nature to assume R; concede further that it is part of our nature to take R in the basic way, so that this conviction is not given or achieved by argument and evidence but comes with our mother’s milk; concede still further, if 227you like, that this belief is produced by our cognitive faculties functioning properly. None of this, clearly enough, can serve to defeat the defeater for R provided by Hume’s agnosticism. That is because any doubt about our cognitive faculties generally is a doubt about the specific faculty that produces this conviction; therefore we can’t allay such a doubt by appealing to the deliverances of that faculty.²⁷³²⁷³   The same goes, naturally enough, for the suggestion that we try to determine by scientific means whether our cognitive faculties are reliable; any such attempt could proceed only by reliance on the very faculties whose reliability is at issue.

2. Naturalism and Lack of Knowledge

Agnosticism with respect to our origins is one way to reject the theistic belief that we human beings have been created in the image of God: as we have seen, agnosticism with respect to origins destroys knowledge. There is another way to reject the belief in question: by accepting a belief incompatible with it, for example, philosophical or metaphysical naturalism. As Bas van Fraassen notes, it isn’t easy to say precisely what naturalism is;²⁷⁴²⁷⁴   See his “Science, Materialism, and False Consciousness,” in Warrant in Contemporary Epistemology: Essays in Honor of Plantinga’s Theory of Knowledge, ed. Jonathan Kvanvig (New York: Rowman and Littlefield, 1996). for present purposes, suppose we take it to be the view that there is no such person as God, nor anyone or anything at all like him (it isn’t that you believe, for example, that there are one or more finite gods). Paradigm cases of naturalism would be the views of Daniel Dennett in Darwin’s Dangerous Idea²⁷⁵²⁷⁵   New York: Simon and Schuster, 1995. or Bertrand Russell in “A Free Man’s Worship”: you think that “man is the product of causes which had no prevision of the end they were achieving, that his origin, his growth, his hopes and fears, his loves and his beliefs, are but the outcome of accidental collocations of atoms.”²⁷⁶²⁷⁶   In Why I Am Not a Christian (New York: Simon and Schuster, 1957), p. 107. (Perhaps you even go so far as to add, with Richard Dawkins, that the very idea that there is such a person as God is really a kind of cognitive virus, an epistemic sickness or disease, distorting the cognitive stance of what would otherwise be reasonable and rational human beings.²⁷⁷²⁷⁷   “Viruses of the Mind,” in Dennett and His Critics: Demystifying Mind, ed. Bo Dahlbom (Oxford: Blackwell, 1993), pp. 13ff. As evidence for the virulence and tenacity of this virus, Dawkins cites the fact that it took Sir Anthony Kenny (as learned and sapient a person as we can easily find), a very long time to fight his way clear of it. Others may wonder whether the virus is all Dawkins says it is, given that Dawkins himself apparently escaped it long ago.) Unlike Hume, therefore, you are not agnostic as to whether there is such a person as God or any being at all like him; you think there is not.

There is likely to be a further difference between you and Hume. Having rejected theism, Hume had no comparable story to put in its place: he was left with no idea as to how humanity arose, under what conditions our cognitive faculties came to be, and so on. The contemporary naturalist, however, is in a different condition; for naturalism now sports a shared myth or story about ourselves and our origins, a set of shared beliefs about who we are, where we come from, and how we got here. The story is familiar; I shall be brief. We human beings have arrived on the scene after millions, indeed, billions of years of organic evolution. In the beginning, there was just inorganic matter; somehow, and by way of processes of which we currently have no grasp, life, despite its enormous and daunting complexity at even the simplest level, arose from nonliving matter, and arose just by way of the regularities studied in physics and chemistry. Once life arose, random genetic mutation and natural selection, those great twin engines of evolution, swung into action.²⁷⁸²⁷⁸   Various other mechanisms (e.g., genetic drift and neutral evolution) have been proposed, but these two remain the favorites. These genetic mutations are multiply random: they weren’t intended by anyone, of course, but also were not directed by any sort of natural teleology and do not arise at the behest of the design plan of the organism. They are “not in a response to the needs of the organism” (Ernst Mayr); they just unaccountably appear. Occasionally, some of them yield an adaptive advantage; their possessors come to predominate in the population, and they are passed on to the next and subsequent generations. In this way, all the enormous variety of flora and fauna we behold came into being.

Including ourselves and our cognitive systems. These systems and the underlying mechanisms have also been selected for, directly or indirectly, in the course of evolution. Consider, for example, the mammalian brain in all its enormous complexity. It could have been directly selected for in the following sense: at each stage in its development, the new stage (by virtue of the structures and behaviors it helped bestow) contributed to fitness and conferred an evolutionary advantage, giving its possessors a better chance of surviving and reproducing. Alternatively, at certain stages new structures (or new modifications of old structures) arose, not because they were themselves selected for, but because they were genetically associated with something else that was selected for (pliotropy). Either way these structures were not selected for their penchant for producing true beliefs in us; instead, they conferred an adaptive advantage or were genetically associated with something that conferred such an advantage. And the ultimate purpose or function, if any, of these belief-producing mechanisms will not be the production of true beliefs, but survival—of the gene, genotype, individual, species, whatever.

If you are a naturalist and also believe these things, then you are what I shall call an ordinary naturalist.²⁷⁹²⁷⁹   Daniel Dennett’s book Darwin’s Dangerous Idea is a paradigm of ordinary naturalism as well as naturalism simpliciter; the same goes for Richard Dawkins’s The Blind Watchmaker (London: W. W. Norton, 1986). For animadversions on Darwin’s Dangerous Idea (and on Darwin’s dangerous idea), see my “Dennett’s Dangerous Idea,” Books and Culture (May-June 1996); for a powerful animadversion on the first but not the second, see Jerry Fodor’s “Deconstructing Dennett’s Darwin,” in Mind and Language 11, no. 3 (September 1996), pp. 246–62. In chapter 12 of Warrant and Proper Function (WPF), I argued that an ordinary naturalist is like Hume in that she has a defeater for any belief she holds—including, ironically enough, ordinary naturalism itself, so that ordinary naturalism is self-defeating.²⁸⁰²⁸⁰   See James Beilby, ed. Naturalism Defeated? Essays on Plantinga’s Evolutionary Argument against Naturalism (forthcoming) for fascinating objections to and critical comments on this argument, along with my reply. I shall not repeat that argument; instead, I will take this opportunity to make some corrections, simplifications, and additions.

First, a correction. In chapter 12 of WPF, there are really two arguments: a preliminary argument and a main argument. The main argument is for the conclusion that naturalism is self-defeating (and hence not rationally acceptable); the preliminary argument is not for that conclusion, but is, instead, a straightforward (probabilistic) argument for the falsehood of naturalism. The preliminary argument is also straightforwardly incorrect.²⁸¹²⁸¹   Here I was helped by Branden Fitelson and Elliott Sober; see their paper “Plantinga’s Probability Arguments against Evolutionary Naturalism,” Pacific Philosophical Quarterly 79 (1998), pp. 115–29. We can see this as follows. It began with an argument for the conclusion that P(R/N&E&C) is fairly low. Here R is the proposition that our cognitive faculties are reliable, N is metaphysical naturalism, E is the proposition that our cognitive faculties have developed by way of the mechanisms to which contemporary evolutionary theory directs our attention, and C was an unspecified proposition describing our noetic systems. In fact, C is dispensable, so in what follows I shall suppress it.

After arguing that P(R/N&E) is low, I went on:

Suppose you do estimate these probabilities in roughly this way: suppose you concur in Darwin’s Doubt, taking P(R/N&E) to be fairly low. But suppose you also think, as most of us do, that, in fact, our cognitive faculties are reliable (with the qualifications and nuances introduced above). Then you have a straightforward probabilistic argument against naturalism—and for traditional theism, if you think these two the significant alternatives. According to Bayes’s Theorem,

P(N&E/R) = P(N&E) × P(R/N&E) / P(R)

where P(N&E) is your estimate of the probability for N&E independent of the consideration of R. You believe R, so you assign it a probability of 1 (or nearly 1); and you take P(R/N&E) to be no more than 1/2. Then P(N&E/R) will be no greater than 1/2 times P(N&E), and will thus be fairly low. No doubt you will also assign a very high probability to the conditional if naturalism is true, then our faculties have arisen by way of evolution; if so, then you will judge that P(N/R) is also low. But you do think R is true; you therefore have evidence against N. So your belief that our cognitive faculties are reliable gives you a reason for rejecting naturalism and accepting its denial (p. 228).

A very pretty little argument: too bad it contains a serious flaw. Here is the problem: in this argument, I was confusing the absolute (logical or anyway objective) probability of R with its probability conditional on our background information B; that is, I was confusing P(R/B) with P(R) simpliciter. (For simplicity, I shall henceforth also suppress E, annexing it to N, so that henceforth N will stand for ordinary naturalism, the conjunction of naturalism simpliciter with E.) We can see this by considering the argument both ways: first, relativizing the probabilities to our background knowledge B, and second, not so relativizing them.

First interpretation: if we relativize the probabilities in question to B, then the relevant application of Bayes’s Theorem will be

P(N/R&B) = P(N/B) × P(R/N&B) / P(R/B)

Here I can set P(R/B) very high, just as I say on p. 228. But I can’t sensibly claim that P(R/N&B) is low. That P(R/N) is low is what I argued: I didn’t argue that the probability of R is low on N plus background knowledge. In the argument that P(R/N) is low, I was abstracting from what we ordinarily think we know (for example, R itself). So I can’t, without further argument, anyway, claim that the probability of R on N together with our background knowledge, is low.

Second interpretation: if we don’t relativize the argument to B (or anything else), the relevant application of Bayes’s will be

P(N/R) = P(N) × P(R/N) / P(R)

as I said on p. 228. But if we are thinking of the absolute probability of R (conditioned only on necessary truths), then I can’t claim (as I did) that P(R) is high: how would I know what proportion of the space of possible worlds is occupied by worlds in which R is true? In particular, the fact that R is true in fact is no reason for assigning it a high absolute (logical) probability. So either way the argument fails.

Fortunately, there is a repair. We are comparing theism (T) and N. So the relevant applications of Bayes’s will be

P(N/R) = P(N) × P(R/N) / P(R)

and

P(T/R) = P(T) × P(R/T) / P(R)

where we are thinking of absolute or logical probabilities. P(R) will have the same value in each expression; so the question is, how do

(a) P(N) × P(R/N)

and

(b) P(T) × P(R/T)

compare in value? Well, P(R/N) is low, as I had argued. However, P(R/T) is not; R is just what we’d expect, given T. (At any rate we’ve got no reason for thinking P(R/T) low.) So (given that we don’t assign N a considerably higher absolute probability than T) we should take the probability of T on R to be greater than that of N on R. But we do, in fact, believe R. So we have a reason to prefer T to N. Not perhaps a very strong reason (this doesn’t tell us a whole lot about the probabilities of T and N on our total evidence) but a reason nonetheless. (It’s the same sort of reason the atheologian has for preferring atheism to theism, given that he thinks it unlikely that a world created by God would display all the evil the world does, in fact, display.)

In essence, the main argument is for the conclusions that P(R/N&E&C) (which I’ll abbreviate as P(R/N); see small print above) is either low or inscrutable; in either case, so I argued, one who accepts N (and also grasps the argument for a low or inscrutable value of P(R/N)) has a defeater for R. This induces a defeater, for him, for any belief produced by his cognitive faculties, including N itself; hence, ordinary naturalism is self-defeating. Now I argued that P(R/N) is low or inscrutable by noting first that natural selection isn’t interested in true belief but in adaptive behavior (taken broadly), so that everything turns on the relation between belief and behavior. I then presented five mutually exclusive and jointly exhaustive possibilities for the relation between belief and behavior, arguing with respect to each possibility P_i that P(R/N&P_i) is low or inscrutable, yielding the result that P(R/N) is low or inscrutable.

Here we can simplify by dropping two of the five possibilities, leaving just epiphenomenalism, semantic epiphenomenalism (perhaps ‘content epiphenomenalism’ would be a more felicitous name), and the common sense (‘folk psychological’) view of the causal relation between belief and behavior. The first possibility (call it ‘P₁’) is epiphenomenalism, the proposition that belief (conscious belief) isn’t involved in the causal chain leading to behavior at all. This view was 232named and suggested by T. H. Huxley (“Darwin’s bulldog”).²⁸²²⁸²   “It may be assumed . . . that molecular changes in the brain are the causes of all the states of consciousness. . . . [But is] there any evidence that these stages of consciousness may, conversely, cause . . . molecular changes [in the brain] which give rise to muscular motion? I see no such evidence. . . . [Consciousness appears] to be . . . completely without any power of modifying [the] working of the body, just as the steam whistle . . . of a locomotive engine is without influence upon its machinery” (T. H. Huxley, “On the Hypothesis That Animals Are Automata and Its History” [1874], chapter 5 of his Method and Results [London: Macmillan, 1893], pp. 239–40.) Later in the essay: “To the best of my judgment, the argumentation which applies to brutes holds equally good of men; and therefore . . . all states of consciousness in us, as in them, are immediately caused by molecular changes of the brain-substance. It seems to me that in men, as in brutes, there is no proof that any state of consciousness is the cause of change in the motion of the matter of the organism. . . . We are conscious automata” (243–44). (Note the occurrence here of that widely popular form of argument, I know of no proof that not-p; therefore, there is no proof that not-p; therefore, p.) In contrast to Huxley, I am here using the term ‘epiphenomenalism’ to denote any view according to which belief isn’t involved in the causal chain leading to behavior, whether or not that view involves the dualism apparently part of Huxley’s version. Although epiphenomenalism runs counter to our commonsense ways of thinking, it is nonetheless widely popular among those enthusiastic about the “scientific” study of human beings. According to Time, a few years ago the eminent biologist J. M. Smith “wrote that he had never understood why organisms have feelings. After all, orthodox biologists believe that behavior, however complex, is governed entirely by biochemistry and that the attendant sensations—fear, pain, wonder, love—are just shadows cast by that biochemistry, not themselves vital to the organism’s behavior.”²⁸³²⁸³   December 28, 1992, p. 41.

And the same can be said for conscious belief: if “behavior, however complex, is governed entirely by biochemistry,” there seems to be no room for conscious belief to become involved in the causal story, no way in which conscious belief can get its hand in; it will be causally inert. Furthermore, if this possibility were, in fact, actual, then evolution would not have been able to mold and shape our beliefs, or belief-producing structures, weeding out falsehood and encouraging truth; for then our beliefs would be, so to speak, invisible to evolution. Which beliefs (if any) an organism had, under this scenario, would be merely accidental as far as evolution is concerned. It wouldn’t make any difference to behavior or fitness what beliefs our cognitive mechanisms had produced, because (under this scenario) those beliefs play no role in the production or explanation of behavior. What then is the probability of R on this scenario? That is, what is P(R/N&P₁)? What reliability requires, of course, is that a large preponderance of our beliefs be true. Now most large sets of propositions 233do not meet that condition; but one large set of beliefs—at any rate, of beliefs we human beings are capable of having—would seem to be about as likely as any other on this scenario. Hence we couldn’t claim with a straight face that there is a high probability, on this scenario, that most of our beliefs are true. Perhaps the verdict is that this probability is relatively low; just for definiteness, let’s say it’s in the neighborhood of .3 or so. Alternatively, we might think that the right attitude here is that we simply can’t make a sensible estimate of this probability, so that P(R/N&P₁) is inscrutable.

The second possibility as to the relation between belief and behavior (call it P₂) is semantic epiphenomenalism. From a naturalistic point of view, the natural thing to think is that human beings are material objects.²⁸⁴²⁸⁴   Though it isn’t easy to say just what a material object is (as Bas van Fraassen emphasizes in “Science, Materialism, and False Consciousness”; see above, footnote 274). For present purposes we need not try to address that project; we can simply narrow our focus to the claim that beliefs are neural events or processes of some sort. Well, suppose that’s what they are: then what sort of thing will a belief—perhaps the belief that Cartesian dualism is false—be? Presumably it will be a long-standing neural or neuronal event of some kind. This neural event will have electrochemical properties: the number of neurons involved; the way in which the neurons involved are connected with each other, with other neuronal events, with muscles, with sense organs, and so on; the average rate and intensity of neuronal firing in various parts of this event and the ways in which this changes over time and with respect to input from other areas. (Call these the ‘syntax’ of the belief.) Of course it is easy to see how these properties of this neuronal event should have causal influence on behavior. A given belief is neurally connected both with other beliefs and with muscles; we can see how electrical impulses coming from the belief can negotiate the usual neuronal channels and ultimately cause muscular contraction.

Now if this belief is really a belief, then it will also have other properties, properties in addition to its syntax or neurophysiological properties. In particular, it will have content; it will be the belief that p, for some proposition p—in this case, the proposition Cartesian dualism is false. But how does the content of this neuronal event—that proposition—get involved in the causal chain leading to behavior?²⁸⁵²⁸⁵   A question just as pressing, of course, is ‘How does this neuronal event have a content at all?’ What is it that assigns to this neuronal event the proposition that Cartesian dualism is false, as opposed, for instance, to the proposition that it is true, or interesting, or obsolete, or vaguely obscene? Under this scenario, it will be difficult or impossible to see how a belief can have causal influence on our behavior or action by virtue of its content. 234Suppose the belief had had the same electrochemical properties but some entirely different content, perhaps the proposition Cartesian dualism is true; would that have made any difference to its role in the causation of behavior? It is certainly hard to see how: there would have been the same electrical impulses traveling down the same channels, issuing in the same muscular contractions. The neurophysiological properties seem to have swept the field when it comes to the causation of behavior; there seems to be no way in which content can get its foot in the door. Of course, it is the content of my beliefs, not their electrochemical properties, that is the subject of truth and falsehood: a belief is true just if the proposition that constitutes its content is true. As in the epiphenomenalist scenario, therefore, the content of belief would be invisible to evolution. Accordingly, the fact that we have survived and evolved, that our cognitive equipment was good enough to enable our ancestors to survive and reproduce—that fact would tell us nothing at all about the truth of our beliefs or the reliability of our cognitive faculties. It would tell something about the neurophysiological properties of our beliefs; it would tell us that, by virtue of these properties, those beliefs have played a role in the production of adaptive behavior. But it would tell us nothing about the contents of these beliefs, and hence nothing about their truth or falsehood. On this scenario as on the last, therefore, we couldn’t sensibly claim a high probability for R. As with the last scenario, the best we could say, I think, is that this probability is either low or inscrutable; P(R/N&P₂) is low or inscrutable, just as is P(R/N&P₁).

Finally, what is the probability of R, given N&P₃, the commonsense (folk psychological) view as to the causal relation between behavior and belief? According to folk psychology, belief serves as a (partial) cause and thus explanation of behavior—and this explicitly holds for the content of belief. I want a beer and believe there is one in the fridge; that belief, we ordinarily think, partly explains those movements of that large lumpy object that is my body as it heaves itself out of the armchair, moves over to the fridge, opens it, and extracts the beer.

Can we mount an argument from the evolutionary origins of the processes, whatever they are, that produce these beliefs to the reliability of those processes? Could we argue, for example, that these beliefs of ours are connected with behavior in such a way that false belief would produce maladaptive behavior, behavior which would tend to reduce the probability of the believers’ surviving and reproducing?²⁸⁶²⁸⁶   Thus Quine: “There is some encouragement in Darwin. If people’s innate spacing of qualities is a gene-linked trait, then the spacing that has made for the most successful inductions will have tended to predominate through natural selection. Creatures inveterately wrong in their inductions have a pathetic but praiseworthy tendency to die before reproducing their kind” (“Natural Kinds,” in Ontological Relativity and Other Essays [New York: Columbia University Press, 1969], p. 126). No. False belief doesn’t by any means guarantee maladaptive action. Perhaps a primitive tribe thinks that everything is really 235alive, or is a witch or a demon of some sort; and perhaps all or nearly all of their beliefs are of the form this witch is F or that demon is G: this witch is good to eat, or that demon is likely to eat me if I give it a chance. If they ascribe the right properties to the right witches, their beliefs could be adaptive while nonetheless (assuming that in fact there aren’t any witches) false.²⁸⁷²⁸⁷   Objection: in any event, these tribespeople would be ascribing the right properties to the right things, so that their beliefs are, in some loose sense, accurate, even if strictly speaking false. Reply: by further gerrymandering, we can easily find schemes under which their beliefs would lead to adaptive behavior (thus being functionally equivalent with respect to behavior to the true scheme) but are not accurate even in this loose sense. There are schemes of this sort, in fact, in which the properties ascribed are logically incapable of exemplification. They think everything is a witch; perhaps, then, their analogue of property ascriptions involves ascribing certain sorts of witches (rather than properties). (One of these witches, for example, is such that, as we would put it, if a thing has it, then that thing is red.) Then their beliefs will not be accurate in the above sense and will indeed be necessarily false. Also, of course, there is the fact that behavior, if it is partly produced by belief, is also partly produced by desire: it is belief and desire, along with other things, that together produce behavior. But then clearly there could be many different systems of belief and desire that yield the same bit of adaptive behavior, and in many of those systems the belief components are largely false; there are many possible belief-desire systems that yield the whole course of my behavior, where in each system most of the beliefs are false. The fact that my behavior (or that of my ancestors) has been adaptive, therefore, is at best a third-rate reason for thinking my beliefs mostly true and my cognitive faculties reliable—and that is true even given the commonsense view of the relation of belief to behavior. So we can’t sensibly argue from the fact that our behavior (or that of our ancestors) has been adaptive, to the conclusion that our beliefs are mostly true and our cognitive faculties reliable. It isn’t easy to estimate P(R/N&P₃); if it isn’t inscrutable, perhaps it is moderately high. To concede as much as possible to the opposition, let’s say that this probability is either inscrutable or in the neighborhood of .9.

Note that epiphenomenalism simpliciter and semantic epiphenomenalism unite in declaring or implying that the content of belief lacks causal efficacy with respect to behavior; the content of belief does not get involved in the causal chain leading to behavior. So perhaps we can reduce these two possibilities to one: the possibility that the content of belief has no causal efficacy. Call this possibility -C. What we have so far seen is that the probability of R on N&-C is low or inscrutable, 236and that the probability of R on N&C is also inscrutable or at best moderate. Now what we are looking for is P(R/N). Because C and -C are jointly exhaustive and mutually exclusive, the calculus of probabilities tells us that

P(R/N) = P(R/N&C) × P(C/N) + P(R/N&-C) × P(-C/N),

that is, the probability of R on N is the weighted average of the probabilities of R on N&C and N&-C—weighted by the probabilities of C and -C on N.

We have already noted that the left-hand term of the first of the two products on the right side of the equality is either moderately high or inscrutable; the second is either low or inscrutable. What remains is to evaluate the weights, the right-hand terms of the two products. So what is the probability of -C, given ordinary naturalism: what is the probability that one or the other of the two epiphenomenalistic scenarios is true? Note that according to Robert Cummins, semantic epiphenomenalism is in fact the received view as to the relation between belief and behavior.²⁸⁸²⁸⁸   Meaning and Mental Representation (Cambridge: MIT Press, 1989), p. 130. That is because it is extremely hard to envisage a way, given materialism, in which the content of a belief could get causally involved in behavior. If a belief just is a neural structure of some kind—a structure that somehow possesses content—then it is exceedingly hard to see how content can get involved in the causal chain leading to behavior: had a given such structure had a different content, its causal contribution to behavior, one thinks, would be the same. By contrast, if a belief is not a material structure at all but a nonphysical bit of consciousness, it is hard to see that there is any room for it in the causal chain leading to behavior; what causes the muscular contractions involved in behavior will be states of the nervous system, with no point at which this nonphysical bit of consciousness makes a causal contribution. So it is exceedingly hard to see, given N, how the content of a belief can have causal efficacy.

It is exceedingly hard to see, that is, how epiphenomenalism—semantic or simpliciter—can be avoided, given N. (There have been some valiant efforts, but things don’t look hopeful.) So it looks as if P(-C/N) will have to be estimated as relatively high; let’s say (for definiteness) .7, in which case P(C/N) will be .3. Of course we could easily be wrong—we don’t really have a solid way of telling—so perhaps the conservative position here is that this probability, too, is inscrutable: one simply can’t tell what it is. Given current knowledge, therefore, P(-C/N) is either high or inscrutable. And if P(-C/N) is inscrutable, 237then the same goes, naturally enough, for P(C/N). What does that mean for the sum of these two products, i.e., P(R/N)?

Well, we really have several possibilities. Suppose we think first about the matter from the point of view of someone who doesn’t find any of the probabilities involved inscrutable. Then P(C/N) will be in the neighborhood of .3, P(-C/N) in the neighborhood of .7, and P(R/N&-C) perhaps in the neighborhood of .2. This leaves P(R/N&C), the probability that R is true, given ordinary naturalism together with the commonsense or folk-theoretical view as to the relation between belief and behavior. Given that this probability is not inscrutable, let’s say that it is in the neighborhood of .9. And given these estimates, P(R/N) will be in the neighborhood of .41.²⁸⁹²⁸⁹   Of course these figures are the merest approximations; others might make the estimates somewhat differently; but they can be significantly altered without significantly altering the final result. For example, perhaps you think P(R/N&C) is higher, perhaps even 1; then (retaining the other assignments) P(R/N) will be in the neighborhood of .44. Or perhaps you reject the thought that P(-C/N) is more probable than P(C/N), thinking them about equal. Then (again, retaining the other assignments) P(R/N) will be in the neighborhood of .55. Suppose, however, we think the probabilities involved are inscrutable: then we will have to say the same for P(R/N). Therefore, P(R/N) is either relatively low—less than .5, at any rate—or inscrutable.

In either case, however, doesn’t the ordinary naturalist—at any rate, one who sees that P(R/N) is low or inscrutable—have a defeater for R, and for the proposition that his own cognitive faculties are reliable? I say he does. To see how, we must note some analogies with clear cases. First, there are the analogies I mentioned in WPF (229–31); here are a couple more. Return (pp. 224) to that voyage of space exploration and the radio-like device that emitted sounds that constitute English sentences, sentences that express propositions of whose truth value you are ignorant. At first, you were inclined to believe these propositions, if only because of shock and astonishment. After a bit of cool reflection, however, you realize that you know nothing at all about the purpose, if any, of this instrument, or who or what constructed it. The probability that this device is reliable, given what you know about it, is low or inscrutable; and this gives you a defeater for your initial belief that the instrument indeed speaks the truth. Consider another analogy. You start thinking seriously about the possibility that you are a brain in a vat, being subjected to experiment by Alpha Centaurian cognitive scientists in such a way that your cognitive faculties are not, in fact, reliable. For one reason or another, you come to think this probability is greater than .5; then you have a defeater for your belief that your cognitive faculties are reliable. Suppose instead that you think this is a genuine possibility, 238but you can’t make any estimate at all of its likelihood, so that you can’t make any estimate at all of the probability that your faculties are reliable: as far as you can tell, the probability could be anywhere between and 1. Then too you have a defeater for your natural belief that your cognitive faculties are reliable.

The same goes for the naturalist who realizes that P(R/N) is low or inscrutable. With respect to those factors crucially important for coming to a sensible view of the reliability of his belief-producing mechanisms—how they were formed and what their purpose is, if any—he must concede that the probability that those faculties are reliable is at best inscrutable. Unless he has some other information,²⁹⁰²⁹⁰   And how could he have or get other information? Any such information would consist in beliefs that were a product of his cognitive faculties, but he has a defeater for the reliability of those faculties and hence for any belief produced by them. the right attitude would be to withhold R. But then something like Hume’s attitude toward my beliefs would be the appropriate one. I recognize that I can’t help forming most of the beliefs I do form; for example, it isn’t within my power, just now, to withhold the belief that there are trees and grass outside my window. However, because I now do not believe that my cognitive faculties are reliable (I withhold that proposition), I also realize that these beliefs produced by my cognitive faculties are no more likely to be true than false: I therefore assume a certain skeptical distance with respect to them. And, because my doubts about my beliefs themselves depend on my beliefs, I also assume a certain skeptical distance with respect to these doubts, and with respect to the beliefs prompting those doubts, and with respect to the beliefs prompting the doubts about those doubts. . . . The ordinary naturalist, therefore, should join Hume in this same skeptical, ironic attitude toward his beliefs. This holds, of course, for N itself; for this reason, we might say that N is self-defeating, in that if it is accepted in the ordinary way, it provides a defeater for itself, a defeater that can’t be defeated.²⁹¹²⁹¹   See chapter 12 of WPF, and “Naturalism Defeated.” The defeater can’t be defeated because any defeater would arise from the very faculties or belief-producing processes in question. For example, the defeater might take the form of an argument, perhaps for the conclusion that those belief-producing processes are reliable after all. But then I would have the same defeater for each of the premises of this argument, as well as for my belief that if the premises are true, then so is the conclusion.

We can briefly extend this result to the case where I am agnostic about ordinary naturalism. I don’t really believe it; either it seems to me to be about as likely as its denial, or its probability is inscrutable for me. In either case, once more, I have a defeater for R, just as in the case of the ordinary naturalist. To see this, consider once again an analogy, and just to preserve continuity, make it another instrumental analogy. You 239are confronted with a measuring instrument of some kind—a barometer, say. You believe that this barometer is in one or the other of two conditions, C1 and C2; the probability that it is in either is for you either inscrutable or about .5. The probability of its being reliable, given that it is in C1, is high, certainly high enough so that if you believed that it was in C1, you would unhesitatingly accept its deliverances. However, the likelihood that it is reliable, given that it is in condition C2, is inscrutable so far as you are concerned: it could be high, but it could also be low; you just don’t know what to think about that probability. Would it be reasonable to accept the deliverances of this instrument? I should think not. You know that if it is in C1, it is reliable; but the probability that it is in C1 is (for you) either about .5 or inscrutable. Either way, the rational attitude is to withhold the belief that it is reliable, accepting neither it nor its denial. And then (given that you have no other source of information) the same goes for the output of the barometer: for any proposition in its output, the rational course for you would be agnosticism with respect to that proposition. The pointer points to thirty inches; still (if you have no other information), you will not on that account believe that the ambient atmospheric pressure is thirty inches. Of course you won’t form a belief inconsistent with that one either: you will withhold the proposition.

It is easy enough to make the application to agnosticism as between theism and ordinary naturalism. If I am such an agnostic, the probability of ordinary naturalism is either in the neighborhood of .5 or inscrutable for me. Suppose the former: what attitude should I take toward R? Well, there is a fifty-fifty chance that my cognitive faculties were produced in a way with respect to which the probability of R is low or inscrutable; if so, however, I have a defeater for R, good reason to withhold. Suppose the latter: then I can’t rule out any probability for ordinary naturalism. Because the probability of R on ordinary naturalism is also inscrutable, I can’t rule out any probability for R; in particular, I can’t rule out a low probability for R. But again, that gives me a defeater for my ordinary and instinctive belief that R. In either case, therefore, I acquire a defeater for R; unless I have or can come up with a defeater-defeater for this defeater,²⁹²²⁹²   And again (see footnote 291), how could I? Any such defeater-defeater would be subject to the very defeater defeating R in the first place. I should be agnostic with respect to R. And if I am agnostic with respect to R, then just as Hume sees, the rational attitude is to be agnostic with respect to any of the deliverances of my cognitive faculties. I may not, in fact, be able to be agnostic with respect to them, but agnosticism is what rationality requires. Of course I also recognize that the beliefs involved in my coming to this agnosticism—such as the belief that the relevant probabilities are inscrutable—are themselves products of my cognitive faculties, and no better off than any other such products. Hence that multilayered reflexive Humean skepticism.

By way of conclusion: the noetic effects of sin don’t necessarily include failure to know anything; Calvin (if that is what, in fact, he thought) goes too far. Still, something in the same general neighborhood is true. If I reject theism in favor of ordinary naturalism, and also see that P(R/N) is low or inscrutable, then I will have a defeater for any belief I hold. If so, I will not, if forming beliefs rationally, hold any belief firmly enough to constitute knowledge. The same goes if I am merely agnostic as between theism and ordinary naturalism. And the same goes if I am agnostic about my origin and the origin of my cognitive faculties. So rejection of theistic belief doesn’t automatically produce skepticism: many who don’t believe in God know much. But that is only because they don’t accurately think through the consequences of this rejection. Once they do, they will lose their knowledge; here, therefore, is another of those cases where, by learning more, one comes to know less.

In this chapter, we have begun to explore the extended model by exploring the nature of sin and some of its cognitive consequences. These consequences extend further than one would ordinarily think; indeed, insofar as sin interferes with the sensus divinitatis and thus with our knowledge of God, it can easily lead to a noetic condition where what rationality demands is that complex, many-layered Humean skepticism. But here a nasty problem looms. According to the A/C model of chapter 6, knowledge requires proper function, and knowledge of God requires proper function of the sensus divinitatis. According to the extended model, however, this belief-producing process has been damaged because of sin, so that it no longer functions properly: how then (on this model) can we have knowledge of the existence and character of God? In the next chapter, we turn to the question how specifically Christian belief, not just generically theistic belief, can have warrant; in answering that question we will also see how the sensus divinitatis is repaired.