From John Mackie's “Evil and Omnipotence” in 1955 through Alvin Plantinga's God, Freedom and Evil in 1974 and “Supralapsarianism, or 'O Felix Culpa'” in 2004 the problem of evil has had a pretty good run, and while a lot creative thought has gone into this philosophical hickory nut for those fifty years, the lines have pretty much hardened into a fairly dull (if occasionally vitriolic) stalemate. To review, the structure of the problem of evil is roughly as follows
Almeida attempts to open new ground by calling into question three assumptions that are almost universally accepted by theists and atheists alike. The three related calms that are the subject of his refutation are:
D1. Necessarily, God can actualize the best possible world only if God does actualize the best possible world.
D2. Necessarily, God can actualize a morally perfect world only if God does actualize a morally perfect world
D3. Necessarily, God can actualize a good enough world only if God does actualize a good enough world.1
In the above the words “actualize” and “world” have a technical and specific meaning. Worlds are made of states of affairs—states of affairs being the way things are or might be. A world is a maximal state of affairs, the way everything is or might be or more technically,
x is a possible world if, and only if, x is a possible state of affairs & for any possible state of affairs y, x either includes or precludes y2
The world that we inhabit is the actual world. So to actualize is to move a state of affairs (or world) from the possible to the actual. If the structure of D1-D3 seem a bit obtuse, perhaps the following would be of some help,
D1'. Necessarily, if God could actualize the best possible world, then God would actualize the best possible world.
D2'. Necessarily, if God could actualize a morally perfect world, then God would actualize a morally perfect world
D3'. Necessarily, if God could actualize a good enough world, then God would actualize a good enough world.
Casting D1-3 in the subjective mood does muddy things a bit, but seeing them in the more conventional conditional clause structure is helpful nonetheless. What is important to note is that while Plantinga would accept D1-3, he argues against the antecedent. The free will defense as found in God, Freedom, and Evil and The Nature of Necessity argues that God cannot actualize a best possible world (because, he argued, there is no such thing), and God cannot actualize a morally perfect world (the crux of the freewill defense). D3 is left open, presumably this is a good enough world (that is a world with a better balance of good than evil), though it does leave open the question about whether worlds that are not good enough are even possible. In doing so, Plantinga is denying (3). Almeida challenges D1-3 altogether and moves the response to the problem of evil from (3) to (4).
To lay the groundwork for this challenge, Almeida proposes to make a distinction between metaphysical necessity and epistemic necessity. Where he draws the like between the two is that of Lewis' system S5 (which he uses to analyze metaphysical necessity) and S4 (for epistemic necessity). The difference comes to that while S4 accepts the following axioms.
*T. □(A→B) → (□A→□B)
*4. □A→□ □A.
While S5 has the following,
*T. □(A→B) → (□A→□B)
If one looks at metaphysical modalities, one sees a fairly straigh-forward application of possible worlds as the way things are or might be, broadly speaking. The system S5 seems to fit well. Normally, epistemic modalities are thought out in terms of what is known. What one knows would could be represented by '□' and what might be known, but whose truth is in question would then be represented by '◊.' For instance, as of this writing, Goldbach Conjecture is either true or false, and whichever it is, it is necessarily the case. However, from an epistemic point of view it might be true or false because it has not been proven.
Almeida has a slightly different take on what he is calling epistemic possibility. His is in terms of conceivability. For instance, 500 pound pasley ants are conceivable (they may violate some laws of biology, perhaps physic, and certainly good taste, but are nonetheless conceivable). Such concepts are pretty innocent. Now take a look a possible world that includes only rabbits who suffer horribly. There is nothing a priori inconceivable about such worlds, but such a world would seem to exclude the existence of God.
Here Almeida lays down a second plank in his foundation. One can find from Anslem through Kant and beyond, that the important attributes of (especially those seen in (1)) not only are metaphysically necessary but a priori (that is epistemically necessary).4 Almeida argues that knowledge of the divine attributes, (including God's necessary existence) are known a posteriori. This position he labels Moderate Anselmianism.
The general flow of Almeida's presentation can be broken down as follows,
Plantinga's Free Will Defense is very tight and very subtle. Don't expect any justice to be done to it here. Even in his explication, Almeida expects his readers to be acquainted with the argument. Here is a very brief (and very inadequate) summary. Plantinga starts by distinguishing between what God can strongly actualize and what God can weakly actualize. The former God does directly, the latter indirectly through agents (the agents are, however, directly actualized). For instance, God could strongly actualize a state of affairs where Fred mows his lawn at 3:00 p.m. on 21 June, but God can only weakly actualize a state of affairs where Fred freely mows his lawn at 3:00 p.m. on June 21. Only Fred can strongly actualize that state of affairs. One would also presume that God knows whether Fred would or would not actualize the state of affairs that includes him mowing the lawn. Now instead a state of affairs that includes mowing the lawn (or declining to mow the lawn), Fred is in a world where he actualizes a state of affairs that includes his launching a massive e-mail scam that brings him illicit gain. Fred freely acted and it is not within God's power to change that. What if there is no state of affairs that God could strongly actualize where everyone (or even most people) does what was right.5
Note that Plantinga is not suggesting that this is the case, but for all we know it could be the case. Put another way, it is conceivable that God cannot create a universe with no evil, so, without the necessary entailment, the logical problem of evil falls short. Plantinga is working to establish a consistency defense, not a proof. Clearly, Plantinga's Free Will Defense accepts some variation of D1-3 and attacks the logical problem of evil at premise (3).
It is the modesty of conceivability that is both the strength and weakness of this defense. The strength lies in the fact that very little needs to be proven. The weakness in the defense is that because it is based on an a priori framework anything else conceivable within that framework can eclipse Plantinga's key point.
Given God's omniscience, Almeida ask why it is that one of the things God cannot strongly actualize is God prediction of what Fred (or anyone else) will do. If so, why cannot God simply predict that Fred (and everyone else) will do what is right. Objections that this would somehow infringe on created freewill fall short on the same grounds that objections of God's foreknowledge fall short. The same arguments that support Middle Knowledge6 also support this sort of Divine Prediction. One cannot argue on the basis of God's Middle Knowledge of essences because that is a metaphysical argument. One would have to argue not that notions such as transworld depravity are epistemically possible or conceivable but that they describe the way things are.
It is at this juncture that Almeidia deploys his own proposal fo the impossibility argument. Assume for the moment that there is a best of all possible worlds, that this best of all possible worlds contains the maximum of moral good and no moral evil. Now assume that not only can God actualize such a world but from,
God is obligated to actualize such a world. Since the concept of God includes him existing in all possible worlds, that combination renders this world (among an infinity of others) not only impossible but inconceivable. However, such worlds do appear conceivable. Almeidia proceeds with D1-D3 in a remarkably similar fashion. The Impossibility Argument—like Argument from Transworld Depravity—is a consistency argument. The central idea is that, unlike Leibniz and those that followed, God is under no moral obligation to actualize a world that is the best possible, a morally perfect world, or even a world with a favorable balance of good and evil.
Lurking underneath this analysis, and never even mentioned, is a modal method of examining moral statements. One can take 'O' for “is obligated” and 'P'7 for “is permitted and use them in a fashion similar to '□' and '◊.' By extension, such an analysis works better within the constraints of S4 than S5.8 The parallel to Almeida's epistemic analysis is striking. If God is obligated to perform some action A, then God is compelled by perfect goodness to perform A.
Here it could be objected that while God's goodness extends beyond what is morally permitted or forbidden. Although God may be under no obligation to actualize a world such that none greater can be wrought, surely God would want to instantiate such a world. Although Almeida deals with such an objection only indirectly when discussing divine freedom, one should first note that this falls outside the logical constraints of logical problem of evil. We can also ask why God should prefer any possible world to any other possible world. At this stage, what God wants, what God's character dictates would not be a matter of a priori deduction but a posteriori discovery.
One might then ask whether God has any obligations to this or any world? To a large extent, Almeida addresses this in his reply to Plantinga's O Felix Cupla defense. Before looking at this directly, however, some background is required. Almeida draws a distinction between gratuitous evils and irredeemable evils—terms that are have been used interchangeably by Plantinga or Marilyn McCord Adams. A gratuitous evil is one that serves no higher purpose. There are many instances of suffering in the world that lend themselves to something more noble, eg., compassion, patience, and so on. Other evils are such that it hardly seems possible that there could be any good that would justify their existence. Some theistic philosophers appear accept the existence of gratuitous evil. Terrence Tilley has argued that any attempt justify every evil, trivializes such evils, leads to a quiescence concerning them, and can even be used as a justification for our employing them. Peter van Inwagen has suggested that at least some gratuitous evil may be unavoidable. Almeida simply falls back on his consistency argument that since gratuitous evil is conceivable, it is possible and God is under no obligation not to instantiate a world that includes it.
This bring Almeida to the idea of redeemable evil and with it redeemed worlds. A redeemed world is one which may or may not include gratuitous evil, but one for which there is no evil that is not atoned for.9 One might hear echo's of Julian of Norwich “All shall be well and all shall be well and all matter of things shall be well.” Given this distinction, we can proceed to Almeida's critique of Plantinga's Felix Culpa defense.
Very briefly, the Felix Culpa argument maintains that God permitted the fall in order to demonstrate God's mercy and love, which would be a greater good than if no evil had occurred at all. One can see with Plantinga's argument there is no real gratuitous evil because every evil is for a greater good. Almeida, however, argues that in doing so Plantinga has God violating a basic principle in ethics, that of double effect. That is (following Kant) that it is immoral to inflict upon a moral agent, without that agent's consent, a wrong—even if that wrong would lead to a greater good. So far, Almeida has not advanced a critique of Plantinga past Adam's accusation of Plantinga casting God as someone afflicted with Munchausen by Proxy Syndrome. The key difference is that where Plantinga has God actualizing a state of affairs with the express purpose of redeeming it, Almeida has God redeeming a world simply because God is good.
Given that Almeida presented only consistency argument, it would be too much to expect it to address our visceral grief, anger, loss, and confusion in the face of evil. It would have been interesting had Almeida explored his distinction between metaphysical and epistemic possibility further. Though he trained this distinction on what the infinite being can do, does not the same distinction apply to finite being? One could conceive of a best possible world or a world without moral evil if worlds are infinitely malleable, but no possible world is infinite. Given a full conception of possible worlds, what is metaphysically possible may well be more restrictive that we think. Perhaps, after all, all the feasible worlds of any significance that there are, are ones that fall.
1 pp. 1-2
2 Plantinga, Alvin; The Nature of Necessarily. Oxford: Clarendon Press, 1974. pp. 44-46.
3 '□' is necessity operator while '◊' designates possibility. The axioms may be read as T: If necessarily A then B, the if necessarily A then necessarily B; M: If necessarily A, then A; 4. If necessarily A, then necessarily necessary A; 5. If possibly A, then necessarily possibly A. Note that S4 can be derived from S5., so that 4 is included in S5
4 I might suggest, however, that there have been important exceptions to this, especially as found in Thomas Aquinas.
5 Plantinga actually has something much stronger in mind, namely the notion of transworld depravity such that any state of affairs that God can strongly actualize and which includes moral goodness is one where everyone goes wrong. Almeida argues that this is much stronger than anything Plantinga actually needs to face his critics.
6 Again all to briefly, as formulated by Luis de Molina, Middle Knowledge (knowledge of what free creatures would do) stands between God's Natural Knowledge (all possibilities, including what free creatures can do) and God's Free Knowledge (what in fact does happen, given God's free decision to create the world he did). Since Plantinga is dealing with possible worlds and counterfactuals, the dependence on the notion of Middle Knowledge is evident.
7 Some would include a third operator 'F' for “Forbidden,” but this can be handled either by ~P (not permitted) or O~ (obligated not). By extension, it is fairly easy to translate ~P~ to O and ~O~ to P.
8 For instance, normally one is obligated not to slice open a fellow human being, but under certain circumstances a surgeon is permitted to do exactly that.
9 Depending on what one means by it, atonement seems a bit weak. It could simply require to make things right. Adams is actually more forceful on this point. She would insist that the power of evil must in the end be broken. That is to say, not only make things right, but place God's creatures in a position where evil is never again a live possibility. This may underlie the strongest difference between Adams and Almeida on this point. Almeida is working from the point of view of generic theist. Adams would argue that such attempts are too anemic, since they fail to use the resources of one's particular religious heritage. As such, she adopts a specifically christological approach. One may see here why Adams does not draw the distinction between gratuitous and irredeemable evil. A broken evil would not ultimately be gratuitous.