C.S. Lewis Is Great But You Should Be Reading Alvin Plantinga

C.S. Lewis Is Great But You Should Be Reading Alvin Plantinga

Analytic Philosophy, The Problem of Evil & Christian Belief

By Professor Michael Almeida (University of Texas, San Antonio)

April 5, 2015         Picture: Alvin Plantinga (University of Notre Dame)

Many theists find resources to defend their beliefs in the important literary and apologetic work of C.S. Lewis. There is much to be said for Lewis’s popular style, rhetoric, and incisiveness. It is persuasive and pleasurable reading, and perfectly effective when we are in what we might call the popular room. But philosophers and laymen sometimes—perhaps often, these days—find the standards on cogent argumentation raised, and indeed want them raised.

Consider, for instance, the powerful critical assessment, by academic philosophers, of the currently popular arguments against theistic belief. Popular atheological arguments wither under such scrutiny.[1] Of course, serious atheists may be unimpressed. The standards for atheological argumentation also go much higher than anything imagined in Dawkins, Hitchens, Dennett, and a host of other popular writers.

The work of atheistic thinkers such as John Mackie, Jordan Howard Sobel and William Rowe, for example, is much more powerful, if much less popular. So, how well does theistic belief do when the standards and expectations on good reasoning are at their highest? How well does religious belief do when, as David Lewis described it, we are in the philosophy room?

In the philosophy room, the otherwise impressive work of C.S. Lewis is much less helpful. His work will not serve to defend theistic belief in the face of the sheer critical power and breadth of, say, John Mackie’s The Miracle of Theism (Oxford: Oxford University Press, 1982) or Jordan Sobel’s Logic and Theism (Cambridge: Cambridge University Press, 2003).

For illustration, it will be useful to consider the interplay between John Mackie’s most forceful—and often underestimated—challenge to theistic belief, and Alvin Plantinga’s ingenious—and often misunderstood—response to that challenge. The debate concerns a single issue, but the standards of reasoning invoked in addressing the issue are worthy of emulation.

In the well-known essay, Evil and Omnipotence‘, John Mackie reformulated a familiar challenge to theistic belief that has come to be known as the logical problem of evil.[2] It’s not uncommon in philosophy that a familiar argument receives a powerful and compelling reformulation.[3] Mackie argued that (1) – (3) form an inconsistent triad:


(1) God is omnipotent, (2) God is wholly good and (3) Evil exists

Any two of the propositions entails the falsity of the third. How does the inconsistency arise?According to Mackie, the inconsistency is explicit once we add the quasi-logical rules that (4) necessarily, a good thing always eliminates evil as far as it can and (5) necessarily, there are no limits to what an omnipotent thing can do.

There have been counterexamples offered to the propositions in (4) and (5)—these are purported exceptions that show the principles are not perfectly general. Perhaps a good thing might not prevent a person from drowning when he can do so. Perhaps he might fail to do so in order to save another person. And perhaps an omnipotent being is limited by the metaphysically possible or the logically possible—he cannot make 2 + 2 = 5 and he cannot make Plantinga identical to Mackie.[4]

Of course, Mackie really intends (4) and (5) to be applied to an absolutely perfect being. But even granting the suggested restrictions on perfect goodness and omnipotence, it follows from (5) that an absolutely perfect being can do anything that is metaphysically (or, broadly logically) possible. But since it is surely metaphysically possible to eliminate all evil, it seems to follow from (4) and (5) that an absolutely perfect being would bring about a world that includes no evil at all.

But Mackie concludes that, since it is evident that there is evil in the world, it must be false that (1) God is omnipotent or that (2) God is wholly good. Mackie does not suggest that there is no solution to this problem. He does suggest that an adequate solution to the logical problem of evil requires abandoning the traditional view of God as both omnipotent and wholly good.

John Mackie’s version of the logical problem of evil presents a spectacular challenge to the philosophically serious theist. How do we address the challenge? On the one hand, we don’t want to abandon the traditional view of God as omnipotent and wholly good. What would it mean to abandon such a view anyway? Wouldn’t that just be abandoning the position that God exists in favor of the position that some pretty good and pretty powerful being exists? On the other hand, there appears to be no seam in Mackie’s argument to slip through: it’s as seamless an argument as any in the literature. So there’s no obvious way to avoid his conclusion without giving up (1) or (2).

“There appears to be no seam in Mackie’s argument to slip through: it’s as seamless an argument as any in the literature”

Alvin Plantinga’s response to John Mackie’s logical problem of evil does not obviously involve abandoning either (1) or (2).[5] Plantinga instead gives us some interesting reasons to abandon (even the restricted versions of) (4) and (5).

Plantinga aims to show that, contrary to (5), possibly, an omnipotent being cannot do everything that is metaphysically possible. Strictly, he aims to show that, possibly, an omnipotent being cannot do everything that is both metaphysically possible and consistent with being wholly good, all knowing, and necessarily existing.

So, Plantinga is not merely aiming to show that God cannot do things like fail to keep a promise or fail to recall some event or cease to exist. Rather God cannot do some things that are metaphysically possible and perfectly consistent with his nature.

Does abandoning (5) entail abandoning (1)? Mackie would surely insist that it does.[6] He’d no doubt urge that abandoning (5) is just a way of saying that God is not really omnipotent. And are there any good reasons to abandon (5)?

Again, Mackie would surely insist that there are not; the reasons Plantinga advances assume that God cannot bring about certain metaphysically possible states of affairs, e.g., the free actions of rational agents. But that is just to assume—not to argue—that (5) is false. But more on this in what follows.

In addition, Plantinga aims to show that, contrary to (4), possibly, a wholly good being does not always eliminate evil as far as it can. There are evils that serve no larger purpose—evils that can be eliminated without moral cost—that a wholly good being can blamelessly fail to eliminate. So, even the restricted version of (4) is false, if Plantinga is right.

For all the ingenuity in Plantinga’s argument, the idea behind his response to Mackie—the famous Free Will Defense—is rather straightforward. The central assumption is that, possibly, there are significantly free agents or creatures. A significantly free agent is one that is free to perform a morally impermissible action A or, instead, to perform a morally permissible action ~A.[7]

But, according to Plantinga, in order for the action involved to be genuinely free and morally valuable, it must be undetermined or uncaused by anything outside the agent—it must be uncaused by nature, for instance, and uncaused by God. If you perform some action because God caused you to do it, or nature caused you to do it, you did not perform the action freely.

Now Plantinga asks you to imagine whether, before creation, God could have found himself in the following situation. God is wondering what sorts of things to create, and he decides to create all of the wonders and beauties in the vast universe. And he then wonders whether he should create any beings like you and me. He wonders whether he should create beings that are both rational and free.

“If you perform some action because God caused you to do it, or nature caused you to do it, you did not perform the action freely”

The decision might be a difficult one; he already has a great universe. And God might know that if he were to create me, and place me in this universe, that I would freely do something wrong. He might know that I would do something (perhaps many things) morally evil.

If God creates me with genuine freedom, then, of course, he could not cause me to go morally right. But God might also know that rational and free agents are good things to create, even if they sometimes go wrong. And God might know that, if he creates me, it will involve some moral good and some moral evil, but the good will outweigh the evil. Perhaps he might know all of that.

But even those considerations will not be enough for God to decide to create me. God might wonder whether he could create some other rational and free being that would never go wrong. God might then have his cake and eat it too. He’d have all the value of the created rational and free beings, and none of the disvalue. In that case, God would surely not create me.

Now Plantinga asks you to imagine whether, before creation, God could have found himself in a much worse situation. God might know that, no matter which rational and free being he creates, and no matter what circumstances he creates that being in, that being will freely do something wrong. And Plantinga asks whether, in those circumstances, God could create some rational and free beings. The answer does seem to be yes, despite the fact that creating some free and rational agents—such as ourselves—will ensure that there is some moral evil.

But it will do us good to pause for a moment and ask whether we’ve been blinded by all the details. Let’s back up and ask: why doesn’t God just actualize a morally perfect world? Why doesn’t he just bring about a world that includes free moral agents who always do the right thing? God’s omnipotent, isn’t he, why couldn’t he just do that? This is in fact a question that Mackie continued to press in his debate with Plantinga.

“If God has made men such that in their free choices they sometimes prefer what is good and sometimes what is evil, why could he not have made men such that they always freely choose the good? If there is no logical impossibility in his freely choosing the good on one or several occasions, there cannot be a logical impossibility in his freely choosing the good on every occasion. God was not, then, faced with a choice between making innocent automata and making beings who, in acting freely, would sometimes go wrong; there was open to him the obviously better possibility of making beings who would act freely but always go right. Clearly his failure to avail himself of this possibility is inconsistent with his being omnipotent and wholly good”.[8]

Plantinga agrees that it is not (broadly) logically impossible that God should create rational and free agents that always go right, and he agrees that God is omnipotent, but insists that God might be unable to do what it is logically possible for him to do.

There are morally perfect worlds—that is, worlds where every free agent always goes morally right—and it is (broadly) logically possible that God actualizes a morally perfect world. But God might still be unable to actualize a morally perfect world.

Let’s revisit how this is possible. Plantinga asks us to imagine the possibility that every rational and free being that God could create would freely do something wrong, no matter what circumstances they were created in. In the world we are imagining, all of these beings suffer from what Plantinga describes as transworld depravity.

God knows, for instance, that if he creates Almeida and places him in difficult and trying circumstances, he’d freely do something wrong. And he knows that, if he creates Almeida, and places him in nice and comfortable circumstances, he’d still freely do something wrong. No matter where he creates this guy, he is going to freely do something morally wrong. And, unfortunately, in the world we are imagining, the same goes for anyone else he might create.

What God cannot do is create Almeida, or anyone else, and cause him to go freely right. He cannot do that since, if God causes someone to go right, then they do not freely go right.

So here is God’s dilemma: If I create rational and free beings, the world might thereby be improved, but I’m guaranteed that they do something wrong. There is nothing I can do about that; I cannot cause these free beings to go right. My options seem to be (1) Choose not to create any of them or (2) Choose to create the ones that, at least, won’t go horribly wrong. It is not among my options, contrary to Mackie, to choose to bring about a morally perfect world. The rational and free agents I create could bring about such a world—they could all choose always to go morally right—but I’m not able to do it.

“What God cannot do is create Almeida, or anyone else, and cause him to go freely right. He cannot do that since, if God causes someone to go right, then they do not freely go right”

Given God’s dilemma, what should he do? Plantinga argues that God is permitted to create rational and free beings that he knows will sometimes go wrong if it is true that some world with such beings—even granting the moral evil that they will bring about—is better than any world with no rational and free beings.[9]

If God might face a choice in which some world with such fallible beings is better than any world without them, then, Plantinga concludes, it is possible that God coexists with evil. That, in brief, is Alvin Plantinga’s Free Will Defense.

Is the Free Will Defense successful? There are at least two conditions that must be met for it to succeed. First, the Free Will Defense must be consistent with God’s omnipotence. Recall that God is omnipotent only if, necessarily, God can do anything that is (broadly) logically possible.[10] It might not be unreasonable to consider, instead, the weaker notion that God is omnipotent only if, necessarily, God can do anything that is (broadly) logically possible and consistent with his nature as essentially omniscient, perfectly good, and necessarily existing.

But, second, the Free Will Defense succeeds only if it is consistent with God’s perfect goodness. There are many reasonable conditions on God’s perfect goodness. Here’s one, rather weak, condition: God is perfectly good only if, necessarily, God actualizes a morally perfect world.

Now Mackie might complain that the Free Will Defense violates both of these conditions. According to the Free Will Defense, it is possible that God is unable to bring about a (broadly) logically possible state of affairs that is consistent with his divine nature. It is possible, according to Plantinga, that God is unable to actualize a morally perfect world. But there are morally perfect worlds! And actualizing such a world would not entail that God lacks any of the traditional attributes!

If God is omnipotent, Mackie might urge, then, necessarily, he can actualize such a world. Mackie can agree that God’s omnipotence is compatible with, possibly, God’s being unable to cause significantly free agents to always go right. There are, we agree, no possible worlds where that happens. But it is not compatible with God being unable to actualize a morally perfect world. After all, there is a world where that happens! Or, so Mackie might insist.

Mackie might also complain that an essentially morally perfect being necessarily actualizes a morally perfect world. If that is true, then there simply are no possible worlds that are not morally perfect. The existence of such worlds is not compatible with the existence of God! But the Free Will Defense simply assumes that there are morally imperfect worlds and aims to show that, possibly, God is unable to actualize a morally perfect world.

Of course, Mackie’s complaints can be resisted in a variety of ways. I will not pursue those lines of response here. My aim has been to show that the level of discussion among serious theists and serious atheists is extraordinarily high. Clearly, arguments for and against theism (and for and against atheism) can be carried on at a level of discourse that leaves popular discussion of these views far behind, beside the point, and largely irrelevant.

Atheists would do well to emulate the work of, among others, John Mackie, J. Howard Sobel, William Rowe, and Michael Tooley [See Oppy on the work of academic atheists]. Theists would be equally well served to emulate the work of, among others, Alvin Plantinga, Peter van Inwagen, Brian Leftow, and Richard Swinburne. These are among the best representatives of their respective positions. Of course, these thinkers are also much less accessible and much more demanding and, alas, much less popular.

Footnotes & References

[1] For a particularly nice example, see Alvin Plantinga’s review of Richard Dawkins here http://www.booksandculture.com/articles/2007/marapr/1.21.html.

[2] See John Mackie, ‘Evil and Omnipotence’ in Michael Rea and Louis Pojman (eds.)Philosophy of Religion: An Anthology, Sixth Edition (Boston; Wadsworth, 2012) 299-306.

[3] Compare, for instance, Peter van Inwagen’s consequence argument for incompatibilism in his An Essay on Free Will (Oxford: Oxford University Press, 1983), and Nelson Pike’s argument for theological fatalism in his ‘Divine Omniscience and Voluntary Action’, Philosophical Review 74 (1) (1965) 27-46.

[4] These and many other exceptions are discussed in Alvin Plantinga, God, Freedom, and Evil (Grand Rapid: Eerdmans, 1977).

[5] Plantinga presents the argument in several places. See his The Nature of Necessity(Oxford: Oxford University Press, 1974) p. 165 ff. and his God, Freedom, and Evil (Grand Rapids: Wm. B. Eerdman’s Publishing, 1974) p. 29 ff. The version with the fewest controversial assumptions is in James Tomberlin and Peter van Inwagen (eds.) Profiles: Alvin Plantinga (Dordrecht: D. Reidel Publishing, 1985) p. 36 ff.

[6] ‘Evil and Omnipotence’, op. cit., p. 300 ff. on Adequate and Fallacious Solutions.

[7] This assumption alone, Mackie might rightly complain, entails that the logical problem of evil is invalid. It follows from the assumption that, possibly, there are significantly free moral agents that there is possible world in which there is evil and God exists.

[8] Ibid., p. 304

[9] Actually, Plantinga takes the position that God might actualize a world that includes rational and free agents that he knows will do something wrong if he knows that it is not possible to actualize a world with moral good without actualizing one with moral evil. See The Nature of Necessity op. cit. p. 167. The position is too weak, since, it might be the case that every world with moral good is on balance bad. So, it must be possible that there is some world with moral good and moral evil that is at least as good (or better) than any world with neither moral good nor moral evil. See Michael J. Almeida,Freedom, God and Worlds (Oxford: Oxford University Press, 2012) p. 50 ff.

[10] Note that ‘ God is omnipotent only if, necessarily, God can do anything that is (broadly) logically possible’ is equivalent to ‘necessarily, God is omnipotent only if God can do anything that is (broadly) logically possible’ under the assumption that God is essentially omnipotent, omniscient, morally perfect and necessarily existing.

Michael Almeida
Michael Almeida
Dr. Michael Almeida is Professor of Philosophy at the Department of Philosophy and Classics at The University of Texas at San Antonio. He is the author of The Metaphysics of Perfect Beings (Routledge, 2008) and Freedom, God, and Worlds (Oxford University Press, 2012). He is also published in, among other places, The Australasian Journal of Philosophy, Analysis, Philosophical Studies, Philosophia, The Monist, Erkenntnis, Midwest Studies in Philosophy, Faith and Philosophy, Oxford Handbook of Free Will, Journal of Ethics and Social Philosophy, Oxford Studies in Philosophy of Religion, Theoria, International Journal for Philosophy of Religion, and Religious Studies.
Recent Posts
  • Alan

    Michael–very, very nice post.

    A comment:

    “But the Free Will Defense simply assumes that there are
    morally imperfect worlds and aims to show that, possibly, God is unable to
    actualize a morally perfect world.”

    As with Hick’s “soul-making argument” I’d argue that
    this assumes that this actual world is the best
    of all possible “imperfect worlds” for free beings to exist as a
    result of God’s morally perfect creation, even given the modal constraints you’ve
    imposed on that creation pace
    Mackie. What this means to me is that
    there is no other possible-world that (i) contains the kind of free beings
    posited in this actual imperfect world and (ii) contains at least one less kind
    of evil attributable to such free beings in virtue of at least one factor of
    agency that is absent in that possible world that is present in our actual
    world. What I mean by (ii) is for
    example one less part of the self in that possible world that is present in our
    actual world that results in actual-world evil.
    It may be as minor as a form of psychological fetish–a fixation on toes
    that allows one to choose (by freedom hypothesis) to pursue prurient behavior
    and (say) stomp on others’ toes (there are real cases here)–that may be
    avoided by creating free-choosing creatures without toes. On the one hand this raises the question of
    what a minimally-endowed creature
    must be to be physically and psychologically free in the relevant way necessary
    for being a morally free creature. But
    more relevantly for this post it raises the question why we must believe that this world maximally instantiates the kind of creature thus free–for only if
    God created such a maximal world of choice with the minimum amount of possible
    evil that might result from such choices may God be said to have created the
    best of all imperfect possible worlds of choice. And given God’s posited omnipotence and
    omniscience he should have created such as the actual world.

    But at least to me, this is not obviously the best such imperfect
    possible world. And I can think of no
    way to prove that it combines the maximum of moral choices leveraged against
    the minimum of possible evil consequences of such choices.

    • Mike Almeida

      Hi Alan,

      Thanks for this comment! You say,

      “. . .there is no other possible-world that (i) contains the kind of free beings
      posited in this actual imperfect world and (ii) contains at least one less kind
      of evil attributable to such free beings in virtue of at least one factor of
      agency that is absent in that possible world that is present in our actual

      But both Plantinga and Mackie agree that there are such worlds. There are worlds that include all of the actual world agents and that do not include any moral evil. There are mildly morally evil worlds that include all of the actual world agents and include much less moral evil. If I’m reading you right, you are denying that there are such worlds. That’s an important claim that neither Plantinga nor Mackie make. On the supposition that agents are libertarian free, then you can get close to proving that there are morally perfect and mildly morally imperfect worlds.

  • josh

    This really isn’t any higher argument than you will find in popular atheist or apologetic books.

    • Mike Almeida

      Right. At the level of conflation at which you’re reading, I’m sure it all sounds pretty much the same.

      • josh

        If you’re going to reply this late you’d best come up with something more substantial than this.

        • Mike Almeida

          You can’t tell the difference between the op and popular discussions of the problem. That’s pretty substantial. I don’t think there’s much left to say without wasting my time. Indeed, having now imprudently replied twice to trollish comments, I’ll have the thread closed.

          • josh

            Go on, enlighten me. What do you think is substantially new and different in Mackie’s version of the problem of evil that popular atheists are unaware of? What in Plantinga’s free will defense is different from the long-standing free will defense one regularly finds in popular apologetics?

          • Mike Almeida

            Josh, that task involves some very heavy lifting. Setting aside all of the stage-setting and clarifications, I can say that it is because of the precision of Mackie’s formulation of the problem and the precision of Plantinga’s reply that we can see clearly that the assumption that, possibly, there are libertarian free agents begs the question against Mackie. We’d never be able to see that had the problem not been so clearly formulated. Mackie’s argument refined down to its essence is that God cannot exist in a pluriverse that includes libertarian-free beings. Who’d have guessed that God’s existence has metaphysical implications for the nature of free will? That’s hard to see without all the work Mackie and Plantinga did. I’d never have seen it.

          • josh

            Mike, thanks for replying rather than closing me out. I’m a little unclear on what you are trying to argue here. Are you saying that Plantinga’s defense fails because it begs the question? Mackie’s formulation of the problem says nothing directly about free will, although he considers it as a defense (since of course it was common before Plantinga). Mackie would say that libertarian free will is incoherent. He would also say that God’s omnipotence/omniscience directly conflicts with free will, although that is not the essence of his argument. He could also argue that free will, if it were a thing, in no way implies the existence of evil, a fact which seems to hold directly in Plantinga’s own theology.

            But without going through the whole song and dance, none of this was particularly new to me when I first heard Mackie’s and Plantinga’s names attached to it, and none of it is foreign to the popular names of atheism. Dennett and Harris, for example, have written extensively on free will. Perhaps you like Plantinga’s and Mackie’s presentation particularly, but content-wise you haven’t pointed out anything beyond the ambit of the more popular writers.

            It might be worth pointing out that Mackie was a fan of Dawkin’s The Selfish Gene. While Plantinga’s “Evolutionary Argument Against Naturalism” is mostly warmed-over Lewis, probably by way of Chesterton.

          • staircaseghost

            As far as I am aware, Plantinga was the first apologist to introduce the rubric of the then-new theory of possible-worlds semantics to the LPoE, arguing that there might be possible worlds in which we have libertarian free will which God’s goodness will nonetheless not let him actualize.

            While you are of course right to see this as basically the street-level “unknown purposes defense” jumped up by way of Rube Goldberg, it is precisely at the level of these technical details that he has invented new and exciting ways to commit fallacies of equivocation, category error, and outright self-contradiction. Have a look at the first sentence of this comment, and you can see all three of them in order: there “might be” such worlds, for all we know (equivocating between epistemic modality and the logical modality he forgot he was supposed to be talking about; you might as well say “for all we know, we are making some subtle error when we add One and One, therefore Two being the sum of One and One is not logically necessary”); he speaks of having libertarian free will in individual worlds, an obvious category error (even accepting for the purposes of argument that LFW is right and compatiblism is wrong, which is wrong) because you cannot talk about the existence of worlds where an agent does otherwise in that same world — either all versions of that agent have LFW or none of them do, by virtue of the existence or nonexistence of even a single logically possible world where the agent does otherwise; and his confused nonsense about someone or something “actualizing” worlds leads him to the flagrant contradiction of “possible worlds it is impossible to actualize”.

            He also argued that individual beliefs could be subject to natural selection, since individual genes code for individual beliefs. Really let that one roll around in your mouth for a moment.

            In other words, he managed to commit no fewer than three (3) trivially fatal errors, in his master argument in modal logic, his alleged area of expertise, and to commit a blunder in biology so elementary it is second among philosophers only to Aristotle’s thoughts on women’s tooth count, all while keeping legions of fans (believers and nonbelievers alike) hailing him as the paradigm example of the Sophisticated Theist The Dumb Gnuatheists Won’t/Can’t Engage With.

            I know your professional background is scientific and not in the humanities, but that is no excuse for smugly dismissing these astonishing accomplishments.

          • josh

            I knew Plantinga phrased things in possible worlds language but I only find that to be new in a “spoiler on a Pinto” kind of way. He could have dressed it up in Kripke semantics if he really wanted the bells and whistles.

Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt

Start typing and press Enter to search