Meanwhile, I was delighted to see that John Schellenberg is due to publish A New Logical Problem of Evil. He offers a different way of prosecuting the logical problem, but one which shares with my paper the general approach of using God’s purity and maximal goodness to forestall attempts at theodicy. Well worth reading.

.. but this is just horrific. Feel free to share it if you have the same sense of disgust and distress in realizing that some fellow humans condone this barbarity.

“The reader in search of knock-down arguments in favor of my theories will go away disappointed. Whether or not it would be nice to knock disagreeing philosophers down by sheer force of argument, it cannot be done. Philosophical theories are never refuted conclusively. (Or hardly ever. Gödel and Gettier may have done it.) The theory survives its refutation—at a price. Argle has said what we accomplish in philosophical argument: we measure the price. Perhaps that is something we can settle more or less conclusively. But when all is said and done, and all the tricky arguments and distinctions and counterexamples have been discovered, presumably we will still face the question which prices are worth paying, which theories are on balance credible, which are the unacceptably counterintuitive consequences and which are the acceptably counterintuitive ones. On this question we may still differ. And if all is indeed said and done, there will be no hope of discovering still further arguments to settle our differences.

It might be otherwise if, as some philosophers seem to think, we had a sharp line between “linguistic intuition,” which must be taken as unchallengeable evidence, and philosophical theory, which must at all costs fit this evidence. If that were so, conclusive refutations would be dismayingly abundant. But, whatever may be said for foundationalism in other subjects, this foundationalist theory of philosophical knowledge seems ill-founded in the extreme. Our “intuitions” are simply opinions; our philosophical theories are the same. Some are commonsensical, some are sophisticated; some are particular, some general; some are more firmly held, some less. But they are all opinions, and a reasonable goal for a philosopher is to bring them into equilibrium. Our common task is to find out what equilibria there are that can withstand examination, but it remains for each of us to come to rest at one or another of them. If we lose our moorings in everyday common sense, our fault is not that we ignore part of our evidence. Rather, the trouble is that we settle for a very inadequate equilibrium. If our official theories disagree with what we cannot help thinking outside the philosophy room, then no real equilibrium has been reached. Unless we are doubleplusgood doublethinkers, it will not last. And it should not last, for it is safe to say that in such a case we will believe a great deal that is false.

Once the menu of well-worked-out theories is before us, philosophy is a matter of opinion. Is that to say that there is no truth to be had? Or that the truth is of our own making, and different ones of us can make it differently? Not at all! If you say flatly that there is no god, and I say that there are countless gods but none of them are our worldmates, then it may be that neither of us is making any mistake of method. We may each be bringing our opinions to equilibrium in the most careful possible way, taking account of all the arguments, distinctions, and counterexamples. But one of us, at least, is making a mistake of fact. Which one is wrong depends on what there is.”

~ David Lewis, Philosophical Papers Vol. 1

(1) Every contingent fact has an explanation. (The Principle of Sufficient Reason, or PSR)
(2) There is a contingent fact that includes all other contingent facts.
(3) Therefore, there is an explanation of this fact.
(4) This explanation must involve a necessary being.
(5) This necessary being is God.

Before I get to my criticism of this kind of argument, I must mention that ‘explanation’ here is always explanation of a certain sort. Put vaguely, an explanation of some fact serves to make that fact comprehensible, to make it less mysterious and surprising. But two ways to accomplish this. The first way is via descriptive explanation, which tells us in detail what the explanandum¹ is, and in doing so makes comprehensible the truth of what is explained – to believe that some fact is true requires a clear conception of the nature of that fact, and descriptive explanations facilitate this semantic prerequisite of belief. The second way is via causal explanation, which tells us why the explanandum is. Unlike descriptive explanations, causal explanations involve the postulation of entities, those to which the explanans and explanandum refer. And unlike descriptive explanations, in which the explanans is identical to the explanandum, a causal explanans is always external to the explanandum. It cites a cause, or set of causes, which make the truth of the explanandum comprehensible by showing us how it is the necessary or likely product of some state of affairs which does not include it².
Here is an example of what I mean: suppose I request an explanation of Republicanism in the United States from you. Your answer might involve a description of the political system, in terms of its three tiers of representation, the separation of its legislative, judicial and executive branches, the Constitution, and so on. Such an explanation is a descriptive explanation, for it tells me in what Republicanism in the United States consists. But you might instead give me an account of early American history, including the social influences that lead to the American Revolutionary War, and onwards to the drafting of the Constitution. This second explanation would be a causal explanation, for you presume me to know what Republicanism here is, but interpret me as asking after the conditions from which Republicanism could be seen to follow. Both answers count as explanations, as both serve to make comprehensible Republicanism in the United States, but they are obviously explanations of different kinds³.
With the distinction in hand, we see immediately that the sense of explanation relevant to the Argument from Contingency is that of causal explanation. What the PSR is intended to assert is not that every contingent fact has a description which would make it comprehensible, since one could not infer the existence of anything from this, but that for every contingent fact there is something external to it from which the truth of the fact can be seen to follow. So, in what follows I’ll be discussing causal explanations, and I’ll add the qualification to quotations. We proceed to a criticism of the PSR.

Van Inwagen’s Modal Fatalism Argument

Peter Van Inwagen, a theist himself, offers a reductio of the PSR:

(11) No necessary proposition [causally] explains a contingent proposition. (Premise.)
(12) No contingent proposition [causally] explains itself. (Premise.)
(13) If a proposition [causally] explains a conjunction, it [causally] explains every conjunct. (Premise.)
(14) A proposition q only [causally] explains a proposition p if q is true. (Premise.)
(15) There is a Big Conjunctive Contingent Fact (BCCF) which is the conjunction of all true contingent propositions, perhaps with logical redundancies removed, and the BCCF is contingent. (Premise.)
(16) Suppose the PSR holds. (For reductio.)
(17) Then, the BCCF has an [causal] explanation, q. (By (15) and (16).)
(18) The proposition q is not necessary. (By (11) and (15) and as the conjunction of true contingent propositions is contingent.)
(19) Therefore, q is a contingent true proposition. (By (14) and (18).)
(20) Thus, q is a conjunct in the BCCF. (By (15) and (19).)
(21) Thus, q [causally] explains itself. (By (13), (15), (17) and (19).)
(22) But q does not [causally] explain itself. (By (12) and (19).)
(23) Thus, q does and does not [causally] explain itself, which is absurd. Hence, the PSR is false.

Both this argument and the basic argument we met earlier have been lifted from Alexander Pruss’s extensive chapter on Leibnizian Cosmological arguments, from the Blackwell Companion to Natural Theology. And as we would expect, Pruss has something to say about Van Inwagen’s reductio. He thinks that the theist should not accept (11). He tells us:

The main reason to accept (11) is the idea that if a necessary proposition q [causally] explained a contingent proposition p, then there would be worlds where q is true but p is false, and so q cannot give the reason why p is true. This sketch of the argument can be formalized as follows:

(24) If it is possible for q to be true with p false, then q does not [causally] explain p. (Premise)
(25) If q is necessary and p is contingent, then it is possible for q to be true with p false. (A theorem in any plausible modal logic)
(26) Therefore, if q is necessary and p is contingent, then q does not [causally] explain p.

Instead of attacking (11) directly, I shall focus my attack on (24).. [which] seems to capture just about all the intuition behind (11). By contraposition (24) is equivalent to:

(27) If q [causally] explains p, then q entails p.

But Pruss thinks (27) is false. Briefly, it is false because a causal explanation need not entail that which it explains. In fact, a causal explanation need not even make probable what it explains, as in the explanation of syphilis by paresis: whereas the cause of paresis is known to be untreated syphilis, it is only in a small percentage of cases of untreated syphilis that paresis does result. But untreated syphilis is a causal explanation of paresis, even if not a sufficient (entailing) one, and so it follows that causal explanations need neither entail nor make probable their explanandum.
Pruss is right, of course, to say that (27) is false, and so that (24) is false. But showing that (24)-(26) is an unsound argument does not show that (11) is false; at most, it shows that one way of attempting to justifying (11) is a failure. So let’s try another way.

(A) Suppose that some necessary proposition q causally explains a contingent proposition p. (Premise for reductio)
(4) A proposition q only causally explains a proposition p if q is true.(Premise)
(B) So, q is a necessary truth. (From A and 4)
(C) The probability of a necessary truth is 1. (Premise)
(D) So, P(q) is equal to 1. (From B and C)
(E) Then, P(p|q) is equal to P(p). (From D⁴)
(F) But, if q causally explains p, then P(p|q) > P(p). (Premise)
(G) So, q does not causally explain p. (From E and F)
(H) Then, q both does and does not causally explain p, which is absurd. (From A and G)
(11) (Therefore) No necessary proposition causally explains a contingent proposition.

The thought behind (24)-(27) was that causal explanations entailed what they explained. The thought behind the argument I propose is different, and is captured by (F): though causal explanations may neither entail nor make probable what they explain, it is at least true of causal explanations that their being true increases the probability of that which they explain⁵. Consider again the syphilis-paresis case: though the presence of untreated syphilis does not make paresis probable, as it only rarely leads to paresis, still someone’s having syphilis makes paresis more likely than it otherwise would be, and plausibly this is required in order for it to be explanatory. By contrast, were it not the case that syphilis raised the probability of paresis, then it would be difficult to see in what sense syphilis would make the existence of paresis more comprehensible, less mysterious or surprising. It would be difficult to see what causal relevance syphilis had to paresis at all. We can bring this thought out with the following argument:

(I) To causally explain some fact X, one must cite some fact Y other than X which is yet relevant to X’s obtaining.
(II) To causally explain X is to simultaneously causally explain X’s being true.
(III) So, to causally explain X is to cite some Y other than X’s being true, which is relevant to X’s being true.
(IV) Some Y is relevant to X’s being true only if it either conduces to the truth of X, making it more probable, or conduces to the falsity of X, making it less probable.
(V) Whatever conduces to the falsity of X does not causally explain X.
(VI) Therefore, to causally explain X is to cite some Y other than X’s being true, which conduces to the the truth of X, making X more probable.

So, that a causal explanation increases the probability of its explanandum simply falls out of its being an explanation of some fact’s being true. And now we can see why a necessary fact fails to explain a contingent fact. If a necessary fact did explain a contingent fact, then it would explain why that contingent fact was true. So a necessary fact would have to offer something relevant to the truth of the contingent fact. And whatever is relevant to the truth of the contingent fact, which we could describe as causally explanatory, would have to conduce to the truth of that contingent fact. But a necessary fact is not conducive to the truth of any contingent fact for, what is equivalent, the truth of a necessary fact does not increase the probability of that contingent fact’s obtaining - a necessary fact will obtain whether or not the contingent fact also obtains, and so it tells us nothing about whether the contingent fact also obtains. Hence, Van Inwagen’s modal fatalism argument is sound: the Principle of Sufficient Reason is false, and all arguments which assume its truth are fatally flawed.

Conclusion

If the above argument is sound, then the PSR is false, and so one more argument for the existence of God is unsound. But though that is of interest, there is more we can conclude from the defense of Van Inwagen’s argument: it follows from the fact that no necessary proposition causally explains a contingent proposition that, if God is necessary, then God does not causally explain the universe. And given that part of what we understand by the term ‘God’ is ‘the cause of the universe’, it follows that if God is necessary, God does not exist. In that case I recommend that theists give up the notion that God is necessary, as well as the principle of sufficient reason⁶.

Addenum

There are, it is true, potential counterexamples to the principle that “if q causally explains p, then P(p|q)>P(p)“. I’ll address two, taken from the SEP’s article on Probabilistic Causation. The first:

(i) Probability-lowering Causes. Consider the following example, due to Deborah Rosen (reported in Suppes (1970)). A golfer badly slices a golf ball, which heads toward the rough, but then bounces off a tree and into the cup for a hole-in-one. The golfer’s slice lowered the probability that the ball would wind up in the cup, yet nonetheless caused this result.

In response, I deny that it is true that the golfer’s slice lowers the probability of a hole-in-one. That the golfer’s slicing the ball as opposed to hitting the ball cleanly lowers the probability of a hole-in-one is undoubtedly true, and in that case it is true that the golfer’s slicing the ball as opposed to hitting it cleanly does not causally explain the hole-in-one. On the other hand, slicing the ball simpliciter does raise the probability of a hole-in-one, since a hole-in-one is more likely on a slice than on no information regarding the initial conditions at all. Hence slicing the ball simpliciter is causally explanatory of the hole-in-one. The second case:

(ii) Preemption. A different sort of counterexample involves causal preemption. Suppose that an assassin puts a weak poison in the king’s drink, resulting in a 30% chance of death. The king drinks the poison and dies. If the assassin had not poisoned the drink, her associate would have spiked the drink with an even deadlier elixir (70% chance of death). In the example, the assassin caused the king to die by poisoning his drink, even though she lowered his chance of death (from 70% to 30%). Here the cause lowered the probability of death, because it preempted an even stronger cause.

My response here is of the same sort. It is not true that the assassin’s poisoning the king’s drink lowers the probability of the king’s death. What is true is that the assassin’s poisoning the drink as opposed to leaving the poisoning to their associate lowers the probability of the king’s death, and so, it is true that the assassin’s poisoning the drink as opposed to leaving the poisoning to an associate is not causally explanatory. However, as the assassin’s poisoning the king’s drink simpliciter does raise the probability of the king’s death, we can say that the assassin’s poisoning the king’s drink causally explains the king’s death.
As the SEP article notes, both of these are cases of singular causation (they make reference to particular individuals, places, and times). I suggest that this is not coincidental: To judge whether or not some X counts as a cause, we need to come to a conclusion about what causal powers X has, and this can only be done if we abstract away from the particular situation and advert to some comparison class of similar situations including X. But to ask after a cause in a case of singular causation leaves open just which details we are to hold constant in determining our comparison class, and so makes possible the kind of inconsistent treatment above – the probabilistic evaluation of a purported cause erroneously based on the irrelevant features of a singular case.

Notes

¹ For convenience, I’ll be using the Latin terms ‘explanandum’ and ‘explanans’. An explanandum is what is explained, whereas an explanans is what does the explaining.

² One response to the Argument from Contingency, and indeed all cosmological arguments, is to observe that they would appear to require a cause of the Big Bang. But, so the objection goes, it is part of the definition of a cause that it precedes its effect in time, and as nothing can precede the Big Bang in time, nothing can be its cause. For this post I’ll assume that the notion of cause is flexible enough to accommodate a timeless cause, which does not precede its effect.

³ There are two other classes of explanation sometimes added to my two. The first is intentional explanation – explanations which cite the beliefs and desires of agents as producing some effect. For my part, I see no good reason to assume that intentional explanations are not a species of causal explanation: though some philosophers may believe intentional explanations do not fit the scientific criteria of a cause, I take the lesson here to be not that they are non-causal explanations, but that there is a broader notion of cause upon which the scientific model is a restriction.
The second potential class is that of justification – explanation in terms of reasons for belief in what is explained. This too, I believe to be a species of causal explanation, for in giving reasons for belief one is often citing such reasons as causally efficacious in producing one’s belief. Cases where this does not happen, say, when reasons for belief are given merely to recommend a belief to others, or where they are cited in support of one’s belief without being productive of that belief, are not cases of genuine explanation – the former because explanation is factive, whereas one may recommend a belief to others without their being some belief which would be explained; the latter because it is an instance of rationalization, but rationalization merely masquerades as explanation.
In any case, it does not appear that insisting on either class in addition to our two would favor the Argument from Contingency: if the PSR is to be interpreted as demanding an intentional explanation for every contingent fact, then the argument from contingency would seem to beg the question; alternatively, if the PSR is interpreted as demanding a reason why one should believe any given contingent fact, then a logical demonstration would suffice to explain the contingent fact which includes all other contingent facts.

⁴ This follows from Bayes theorem. According to (D), P(q) is 1. Then, P(q|p) is also 1, and we can deduce as follows:
P(q|p) = P(p|q) x P(q) / P(p) [Bayes Theorem]
1 = P(p|q) x 1 / P(p) [Substitution]
1 = P(p|q) / P(p) [Elimination]
1 x P(p) = P(p|q) / P(p) x P(p) [Multiplication of both sides by common factor]
P(p) = P(p|q) [Result]

⁵ I stress that this is but necessary condition of a causal explanation, and is not intended as an analysis of what causal explanations are.

⁶ I feel I should add that this concession may not be theologically significant: often, the terms ‘necessary’ and ‘contingent’ are taken to be synonyms of ontological independence and dependence respectively. But as Ex-Apologist has argued, these concepts are not synonymous, and since it seems to me that it is ontological independence which is part of the core conception of God as opposed to necessity, I believe that this is an adjustment the orthodox theist can accept.

Some seriously clever stuff. I was impressed enough to order the book, of which you can find a summary of some main ideas here.

EDIT: Best version and last version here.

____________________________________________________

Groundwork for the Argument

I begin by picking up a thread of inquiry which Alvin Plantinga follows in God, Freedom, and Evil¹. There we find Plantinga examining the logical problem of evil, as given by J.L. Mackie in his 1955 paper, Evil and Omnipotence. According to Plantinga, Mackie takes the following propositions to form an inconsistent set..

(1) God is omnipotent
(2) God is wholly good
(3) Evil exists.
(19) A good thing always eliminates evil as far as it can
(20) There are no limits to what an omnipotent being can do.

.. all of which seem plausible on theism. Yet if the set is inconsistent at least one proposition must be rejected on pain of contradiction. Since (3) is obvious, while (19) and (20) appear to be sound definitions, it appears that either (1) or (2) must go, or both, if the presupposition of God’s existence is false. But Plantinga contests (19) on several grounds, guiding a cascade of revisions, the most important amongst these being that if some evils are logically required for the existence of a greater good, then we would expect that a good being would not eliminate the evil, as this would also eliminate the good. Eventually, he settles on an inconsistent set of propositions from which a valid argument from evil may be constructed:

(1) God is omnipotent.
(2) God is wholly good.
(2′) God is omniscient.
(3) Evil exists.
(19c) An omnipotent and omniscient good being eliminates every evil that it can properly eliminate.
(20) There are no nonlogical limits to what an omnipotent being can do.
(21) If God is omniscient and omnipotent, then he can properly eliminate every evil state of affairs.

So, having constructed this set, why does Plantinga doubt that a logical argument succeeds? Because, he tells us, (21) is possibly false, whereas it needs to be a necessary truth in order to ensure that the existence of God and the existence of evil are incompatible. I aim to show that, pace Plantinga, something like (21) is necessarily true, though I shall first reformulate Plantinga’s set into an argument, making two improvements along the way.

The first change I make to the set above is to replace talk of ‘proper elimination’ with that of ‘denial of existence’. According to Plantinga, “a being properly eliminates an evil state of affairs if it eliminates that evil without either eliminating an outweighing good or bringing about a greater evil”². But this notion appears to be inadequate, for it seems to be consistent with unjustifiable evil existing in the world, so long as God eventually gets around to dispensing with it. A fortiori, Plantinga’s formulation places no restriction on what God creates, good or evil. By contrast, the requirement that God should deny existence to every evil does restrict the scope of what God would create and allow³.

One further adjustment is needed. Plantinga indicates that a premise like (21) can easily be falsified, since there might be some state of affairs S, such that S is good overall, but where a constituent of S is an arbitrary evil – for example, S might represent the conjunction of “Alvin is deliriously happy” with “Paul has a minor abrasion”. A conjunction entails its conjuncts, and so (21) is false if any good outweighs some evil, and if these are compossible. The obvious way in which (21) goes wrong (or in which it is ambiguous) is that the “goods” and “evils” cover complex states of affairs, whereas it is not specifically these that we think of as goods and evils in the context of the argument from evil. What usually comes to mind are goods or evils conceived as basic, the kind of goods and evils which are at once atomistic and general. Being atomistic, they do not divide into conjuncts responsible for their value as a whole; being general, they describe types of good or evil which have particular tokens. In short, I think we require good-making and evil-making properties, which I define thus:

Good-making Property: A property which (i) is objectively good, and (ii) cannot be decomposed into constituent properties in virtue of which it is good.

Evil-making Property: A property which (i) is objectively evil, and (ii) cannot be decomposed into constituent properties in virtue of which it is evil.

For brevity’s sake, I’ll call these GMPs and EMPs in what follows, and ‘properties’ will refer to these kinds of properties. With these adjustments, and some minor tweaks to presentation, I can now advance this essay’s main argument.

The Argument from Evil

(1) If God exists, then God is an omnipotent, omniscient and wholly good being.
(2) An omnipotent, omniscient and wholly good being denies existence to every evil-making property whose existence is not entailed by some greater good-making property.
(3) Every evil-making property is such that its existence is not entailed by some greater good-making property.
(4) Therefore, if God exists, every evil-making property is such that God denies it existence.
(5) Evil-making properties exist (since evil exists).
(6) If God exists, then some evil-making property exists and it is the case that God denies it existence (which would be contradictory).
(7) Therefore, God does not exist.

(NB: References to numbered premises will henceforth be to this argument). The argument is valid. But is it sound? Well, (1) is a uncontroversial statement of traditional theism, and (2) seems unobjectionable as well. (5) has as good a claim to truth as any other moral assertion. (4), (6) and (7) are logical consequences. That leaves (3), which makes an obviously outrageous claim. And yet I think standard philosophical theism is committed to this claim:

A Sub-Argument for (3)

(A) God exists. (Assumption)
(B) If God exists, it is possible that he should exist, without any other thing existing.
(C) If God exists, then he instantiates all good-making properties, and no evil-making properties.
(D) Thus, it is possible that God alone should exist, instantiating all good-making properties and no evil-making properties.
(E) Then, there is no good-making property whose instantiation entails instantiation of an evil-making property.
(F) Therefore, every evil-making property is such that its existence is not entailed by some good-making property.
(3) Every evil-making property is such that its existence is not entailed by some greater good-making property.

Let’s run through the premises. (A) assumes God’s existence, which every theist must do also. (B) asserts the ontological independence of God, which various notions of God support. For example, the notion of God as ‘the ground of all being’ stipulates the being of God as prior to, and supportive of, the existence of anything else: this is God’s sovereignty and aseity. The conception of God as a necessary being, particularly as used in the Cosmological Argument, is taken to imply his ontological independence. And lastly, it would seem that ontological independence is a perfection, which a perfect-being such as God would exemplify⁴. Less abstractly, imagine the state of affairs prior to God’s creation ex nihilo: supposedly God alone exists, and afterward performs the acte gratuit of creating the universe. If actual, this scenario is certainly possible, and so (B) would be true.

(C) explicates a plausible interpretation of God’s being wholly good: of any GMPs that he has them, and of any EMPs that he lacks them⁵. The strongest way in which this premise may be justified is if one takes God to be goodness itself, so that ‘God’ and ‘good’ are interdefinable. Since goodness itself cannot be evil, God would have no EMPs, and since goodness itself cannot fail to include some GMP (else that property would not be good), neither would God lack any GMPs. A supporter of this view is William Lane Craig, here responding to the Euthyphro dilemma:

“God’s character is definitive of moral goodness; it serves as the paradigm of moral goodness. Thus, the morally good/bad is determined by reference to God’s nature; the morally right/wrong is determined by reference to his will… If the non-theist should demand, “Why pick God’s nature as definitive of the Good?” the answer is that God, by definition, is the greatest conceivable being, and a being which is the paradigm of goodness is greater than one which merely exemplifies goodness.”⁶

Notice that Craig appeals to perfect-being theology as a reason to identify God with the good, an orthodoxy with distinguished luminaries (Anselm, Aquinas, Leibniz.. many more⁷). But even leaving aside this argument, (C) is compelling: we understand God as being without flaw (hence without EMPs), and as being good in every way that something can be good. Of course, this does not mean that each of God’s properties are good, for some may be neither good-making nor evil-making, but value-neutral.

(D) gathers (A), (B) and (C) together. (E) draws the consequence of (D). To see this, suppose the contrary, that some GMP does entail an EMP: then, if the GMP exists, so must the evil-making one. But (D) states that this is not so, that every GMP can exist (as God) in the absence of every EMP. So if (E) is false, then (D) is false, thus (D) entails (E). Similarly, (F) is a deductive consequence of (E). Suppose the contrary, that there is some evil-making property whose existence is entailed by a GMP. Then there would be a GMP which entailed it, whereas (E) states that no GMPs entail EMPs. If (F) is false, (E) is false, so (E) entails (F). (3) is a special case of (F) – if every EMP is not entailed by a GMP, then a fortiori, they are not entailed by greater GMPs. So there we have it, support for (3). And with it, I believe, a successful argument from evil.

Objections

(I) I disagree with (C): God does not instantiate every good-making property, but on the contrary, only most good-making properties – those which are compossible.

Reply: No argument from evil covers every conception of God, and this argument is unexceptional in that regard⁸. Nevertheless, I believe the arguments here cover standard philosophical theism, which endorses the thesis that God is a perfect-being. This is because God’s perfection seems to entail the having of every GMP:

(I) God is a perfect-being.
(II) A perfect-being instantiates every perfection.
(III) A perfection is an optimal fulfillment of some evaluative dimension.
(IV) So, God optimally fulfills every evaluative dimension.
(V) To instantiate a good-making property is to optimally fulfill an evaluative dimension – namely, the binary dimension which concerns the having or not of the good-making property.
(VI) So, God instantiates every good-making property.

(I) expresses perfect-being theology. (II) is a definitional expansion of (I) – a Cartesian claim which, so far as I’m aware, is accepted by all perfect-being theologians, this being the premise which allows them to infer that God has any individual perfection (e.g. necessary existence, omniscience, perfect goodness)⁹. (III) expresses perfection. (V) includes its explanation – whatever else a GMP is, it is evaluative, and being evaluative, it implies some dimension (scale) on which it is evaluative – at least the GMP/~GMP dimension. (IV) and (VI) are logical consequences, the latter confirming the suspect fragment of (C). The second conjunct in (C), that God instantiates no evil-making property is, I presume, uncontroversial. I therefore take it that perfect-being theology is committed to (C).

(II) Theodicy/Defense X solves the logical problem of evil. Since your argument does not deal with theodicy/defense X, you have not made your case.

Reply: Supposing the Argument and the Sub-Argument are sound, I have indeed made my case. For a sound deductive argument necessitates its conclusion, and so, whatever else there may be to say about why a given theodicy fails, the arguments here make it given that it does so. But since I realize that many readers will want more of an explanation before they feel convinced, I will offer some general remarks.

The conspicuous premise of the Argument is (3), and it is conspicuous precisely because it is designed to cover all theodicies: every viable theodicy purports to cite a greater good for whose sake evil may be justified. So my general reply to those who wonder why their pet theodicy fails is given in the support for (3), the Sub-Argument. That argument works by exploiting a double standard inherent in theodicies: on the one hand, theists claim the existence of a perfect-being, who has every perfection and is without flaw; on the other hand, a theodicy claims the regrettable but necessary imperfectibility of the created world. But whatever is actual is possible, therefore, if perfection is actual in the form of God, perfection is possible. Since an omnipotent God can bring about any states of affairs which are logically possible, God can bring about a perfect creation. And so theodicies are destined to fail.

The free-will defense perhaps deserves special mention here. It might be thought that, contrary to the usual form of a theodicy, where the greater GMP straightforwardly entails an EMP, free-will merely makes evil possible – a denial of (2) rather than (3). However, if possibility is all that follows from the good of free-will, then the connection between the GMP of free-will and any EMP is not necessary, and so it seems open to an omnipotent being to bring about free-will without allowing evil. Hence, contemporary apologists attempt to show that (for all we know) free-will does entail evil via certain ‘counterfactuals of freedom’ – conditionals which state what free persons would do if placed in particular circumstances – and in doing so dispute (3). Therefore our treatment will be the same as other theodicies: find and exploit the double standard at the heart of the theodicy. On the one hand, the theist wants to say that human free-will opens up the possibility of evil; but on the other hand, the theist wishes to maintain that this possibility is not open to God, at least not in a way that he would avail himself of it. But if God does not have free-will, then free-will can’t be a GMP, else God would have it. Alternatively, if God has free-will, but of a different sort to our own, then God’s free-will is a GMP, and human free-will is only good insofar as it approximates God’s sort of free-will. Either way, the free-will defense fails because of its differential treatment of God and man. However…

(III) Your argument assumes that creation could be God-like, in that God could avoid creating anything which, in departure from his own nature, would introduce evil into the world. But you have overlooked the one property which God does not instantiate, but which must be instantiated by his creation: that of being created by God. This property, perhaps together with other good-making properties, does entail evil.

Reply: Quentin Smith offers an argument temperamentally similar to my own, in A Sound Logical Argument from Evil. He begins by distinguishing three different varieties of freedom..

A person is externally free with respect to an action A if and only if nothing other than (external to) herself determines either that she perform A or refrain from performing A.
A person is internally free with respect to an action A if and only if it is false that his past physical and psychological states, in conjunction with causal laws, determine either that he perform A or refrain from performing A.
A person is logically free with respect to an action A if and only if there is some possible world in which he performs A and there is another possible world in which he does not perform A. A person is logically free with respect to a wholly good life (a life in which every morally relevant action performed by the person is a good action) if and only if there is some possible world in which he lives this life and another possible world in which he does not.¹⁰

.. and notes that, whilst God has internal and external freedom, his omnibenevolence precludes logical freedom. From this infers he that logical freedom is not metaphysically valuable (else God would have it), though internal and external freedom may be. He then constructs an argument from evil asserting that God could’ve created necessarily good (thus logically unfree) but internally and externally free beings, like himself, and therefore need not and would not have created a world in which moral evil exists¹¹. Alexander Pruss, in The Essential Divine-Perfection Objection to the Free-Will Defence, takes up our objection and extends it:

The initial form of my argument is very simple. If Patricia is a creature who lacks logical freedom with respect to a wholly good life, then by Smith’s definition either it is a necessary truth that if Patricia exists, Patricia leads a wholly good life, or it is a necessary truth that if Patricia exists, Patricia does not lead a wholly good life. For concreteness, take the first case: that Patricia exists entails that Patricia leads a wholly good life [...] Then, that God creates Patricia entails that Patricia exists. Therefore, that God creates Patricia entails that Patricia leads a wholly good life. But surely that means that Patricia is determined to lead a wholly good life by something external to her, namely by God’s creating her. Hence, she is not externally free with respect to leading a wholly good life.¹²

At first pass this response looks decisive. It is undeniable that God has the property of being uncreated, whereas any of his creations must have the property of being created. It is undeniable that, as God is part of a causal chain leading to any particular of Patricia’s actions, God qualifies as a cause of them. And it further appears that some notion of libertarian free-will will support the idea that causal origination in God would be freedom-canceling. However, Pruss’s argument fails because it applies indiscriminately to all creatures, whether logically free or not.

To begin with, notice that God’s entailment of Patricia’s virtuous behavior proceeds by way of her counterfactuals of freedom. Pruss tells us that Patricia’s existence entails her leading a wholly good life, and this is only true if Patricia’s existence entails those counterfactuals that ensure every morally relevant action she performs is good. So we can gloss Pruss’s argument in a way which makes this salient: God creates Patricia in circumstances C, she has certain counterfactuals of freedom specifying a good action A in C, and so God determines her doing of A in C. But since the same is true of every action Patricia performs, and since every action Patricia performs is good, it is (surely!) the case that God determines that she lead a wholly good life. Hence, she is not externally free with respect to leading a wholly good life.

Yet if this is how the entailment is derived, then it is not just necessarily good creatures who are determined by God, but creatures of every stripe. For consider Manuel, a person-essence whose counterfactuals of freedom support a mixture of right and wrong actions. That God creates Manuel in circumstances C, having counterfactuals of freedom specifying his doing A in C, means that God determines his doing of A in C. The same is true of every action Manuel performs. To complete the symmetry, we can let ‘M-life’ denote the kind of life that Manuel would live in C given his counterfactuals of freedom – then, that God creates Manuel in C (surely!) determines that he lead an M-life. Hence, Manuel is not externally free with respect to leading an M-life. Thus, Pruss’s counter to Smith would seem to lay waste to the external freedom of God’s creation generally.

It isn’t difficult to see where your typical free-will defender will think this goes awry. The free-will defense typically distinguishes between strong and weak actualization, where to strongly actualize a state of affairs is to be the cause of that state, and to weakly actualize a state of affairs is to strongly actualize (cause) a subset of some state of affairs containing free beings, who complete that state with their free acts. The point of such a distinction, as I understand it, is to allow us to separate the issue of entailment (which weak actualization implies) from that of causation and hence responsibility (which strong actualization implies). But if the distinction is to do its job it must be the case that, though God’s bringing about C would counterfactually entail some person-essence’s action A, this entailment is not sufficient for God himself to cause and therefore be responsible for A. Plausibly this is so: despite entailment, we think that God only controls an initial segment of the causal chain specified in C, and that consequently he does not determine the outcome of the process in the relevant sense – his creatures do. But then, by creating necessarily good beings, God need not control all parts of the causal chain either – he can leave the counterfactuals of freedom to be determined by his necessarily good beings^{† 13}.

So it would seem that creaturehood cannot be used to support the free-will defense after all – it either cancels external freedom for all creatures, or mistakenly overlooks the distinction between strong and weak actualization. But could another defense be mounted on the basis of it? I have no argument that it cannot. All I can say here is that the free-will defense appears to be the only apologetic which might put the difference to use. This being so, I take my argument for the non-existence of God to be complete.

Conclusion

I began this essay by looking at Alvin Plantinga’s God, Freedom, and Evil, wherein we find the general form of a successful argument from evil. I made two adjustments to this form: first, by eschewing talk of ‘proper elimination’ in favor of ‘denial of existence’ talk; and second, by bringing in the notions of good-making and evil-making properties. With these changes, I proposed a valid argument from evil. I then noted that, as the other premises were unobjectionable, the weight of the argument fell on premise (3), the proposition that “Every evil-making property is such that its existence is not entailed by some greater good-making property“. I offered a sub-argument for this premise, using the notion of God’s ontological independence (i.e. that “it is possible that [God] should exist, without any other thing existing.”) and his perfection (i.e. that God “instantiates all good-making properties, and no evil-making properties”) to show that, if theism is true, then (3) is also true. But if (3) is true, of course, then the argument from evil proposed succeeds – so theism, if true, is false, therefore it is false.

Finally, I considered three objections. The first of these objected to the premise stating God would instantiate every good-making property, and no evil-making property. I conceded to this objection the point that my argument would not disprove every conception of God, but bolstered my argument that it would apply to God conceived as a perfect-being, and thereby to the God of orthodox philosophical theism. I then considered the objection that my argument ignores the various attempts to reconcile evil and the existence of God, making it inadequate. My reply here was to give a general answer, grounded in the sub-argument, such that by its lights any theodicy would have to be guilty of a double-standard. Lastly, I considered an exception to the argument, in the form of a property which all God’s creation must have, but which God himself would not have – that of being created by God. Because God does not have this property, we cannot use the sub-argument to show that it does not entail evil, nor that its combination with other good-making properties would not entail evil. In this vein I consider the response of Pruss to Smith’s logical argument from evil, and assess Pruss’s counter as failing to support the free-will defense. Given the failure of these objections, and being unaware of any others, I conclude that the theist must accept the soundness of the Argument From Evil. Thus, standard philosophical theism is untenable.

____________________________________________________

Notes

¹ Plantinga, p. 12-29.
² Plantinga, p. 20.
³ Another reason is that Plantinga’s formulation leads one to think that the problem of evil is a problem of God’s non-intervention, given that the world is the way it is and contains evil. But the problem of evil is much broader than that of non-intervention in a ready-made world: it is a problem of the world’s having evil in the first place.
⁴ There’s an objection from Platonism I ignore here: that abstract objects exist necessarily, and so God could not exist alone. The objection is easily enough handled, for I need only modify (B) so that God is ontologically independent of concreta, and add as a premise: “Only concrete objects can instantiate EMPs”. This adjustment aside, there might be good reason to take abstract objects as part of God, which would support (B).
⁵ This may not be obvious if one is thinking of moral goodness alone. However, it is said that God’s goodness does not merely consist in how he acts, but in what he is: goodness has a metaphysical sense. “God is said to be good in a wider or narrower sense; wider, when this indicates the fullness and completeness of his being, his self-sufficiency and freedom from want or deficiency of any kind. In this sense of “ perfect goodness ” it has the same reference as “ perfect being, ”though a different sense. Divine perfection provides the conceptual link between being and goodness in God’ s case; God alone is, and can be, good. In the narrower sense God’ s goodness is an aspect of his moral character, and he communicates this goodness to his creatures in acts of creation and redemption.” (Paul Helm, ‘Goodness‘, in The Blackwell Companion to the Philosophy of Religion, p. 263).
⁶ Craig, p. 182.
⁷ Anselm: “Now, one thing is necessary, viz., that one necessary Being in which there is every good – or better, who is every good, one good, complete good, and the only good” (Proslogion, Ch. 23). Aquinas: “All the perfections of all things are in God” (1964, pt. I, q. 4, art. 2). Leibniz: “41. Whence it follows that God is absolutely perfect; for perfection is nothing but amount of positive reality, in the strict sense, leaving out of account the limits or bounds in things which are limited. And where there are no bounds, that is to say in God, perfection is absolutely infinite. (Theod. 22, Pref. [E. 469 a; G. vi. 27].)”.
⁸ As Jordan Howard-Sobel tells us, “The logical problem of evil is a problem for perfect-being theologies only” (p. 479).
⁹ More support: “According to traditional Western theism, God is the greatest being possible in virtue of possessing a complete set of great-making qualities or perfections”. Hoffman & Rosenkranz, p.15.
¹⁰ Smith, “A Sound Logical Argument from Evil” in Ethical and Religious Thought in Analytic Philosophy of Language (New Haven CT: Yale University Press, 1997), p. 149. (Emphasis added).
¹¹ On Smith’s view, such necessarily good beings are not humans who have been restricted to the good, nor are they humans whose counterfactuals of freedom just happen to ensure their perfect goodness in the actual world. Instead, they are to be understood as a class of beings over and above humans, internally-externally free but logically determined, but who might qualify as ‘human’ in a loose sense as being similarly rational persons. Though Smith doesn’t say so directly, his reference to a class of necessarily good beings suggests that their goodness is essential to them, as it is to God. Pruss (see below) does not always observe this difference between humans and these ‘humans’ in arguing for his position, crucial though it is: “[Patricia] has a certain nature, and God has created Patricia as having that nature. But surely then God has determined her to act rightly.” (Pruss, p. 441, emphasis mine).
¹² Pruss, p. 435-6.
¹³ Pruss further argues that creaturehood and a lack of logical freedom entails a lack of internal freedom, chiefly on the grounds that logical determination would have to be grounded in internal dispositional states. I do not treat this here because (i) Pruss does not show that internal freedom can be had in the absence of logical determination, which would be necessary to justify God’s creating logically free beings as securing the supposed greater good of internal freedom, and (ii) Pruss needs to show that his arguments against internal freedom in logically determined creatures do not also apply to God, but does not do much more than gesture towards a Thomistic metaphysics.
^† Spelt out: Let “P c Q & Q c R” symbolize “P causes Q, and Q causes R”. Then, that God causes Patricia, who in turn causes action R fits this form. Pruss derives from God’s causing Patricia, and Patricia’s causing R, that therefore God causes R, i.e. from “P c Q & Q c R” that “Q c R”, which is perfectly valid – causation is transitive, and so the predecessor in a causal chain will always be a (not necessarily the) cause of a later event in the chain. However, this is not the causation required for cancelling freedom. What is needed instead is that God causes Patricia to cause action R, for him to cause her counterfactual of freedom. So, what Pruss really needs to deduce from “P c Q & Q c R” is “P c (Q c R)”. But this derivation is fallacious. For example, though true that “smoking causes cancer, and cancer causes death”, it is not true that smoking causes the causal relation between cancer and death itself, for the causal relation between cancer and death obtains regardless of whether there is anybody who smokes. Likewise, it does not follow from God’s causing Patricia, and Patricia’s causing action R, that God causes Patricia to cause that action.

Bibliography
Aquinas, Of God And His Creatures: An Annotated Translation Of The Summa Contra Gentiles Of Saint Thomas Aquinas. New York: Kessinger Publishing, Llc, 2010.

Anselm, Proslogion, with the Replies of Gaunilo and Anselm. Indianapolis, IN: Hackett Publishing Company, 2001.

Helm, P., ‘Goodness’, in The Blackwell Companion to the Philosophy of Religion (Blackwell Publishing Ltd, 2010). pp. 263-269.

Hoffman, J., and Rosenkrantz. G.S.. The Divine Attributes (Exploring the Philosophy of Religion). Malden, MA: Wiley-Blackwell, 2002.

Craig, W.L., Reasonable Faith: Christian Truth and Apologetics. Leicester, England: Crossway Books, 2008.

Leibniz, G.W., The Monadology and Other Philosophical Writings. New York City: Cornell University Library, 2009.

Mackie, J.L., “Evil and Omnipotence”, Mind 64 (1955); reprinted in Louis P. Pojman (ed), Philosophy of Religion: An Anthology, 3rd edition (Wadsworth, 1998), pp. 186–193.

Plantinga, A.: God, Freedom, and Evil (Grand Rapids: Eerdmans, 1974).

Pruss, A., “The Essential Divine-Perfection Objection to the Free-Will Defence”, Religious Studies 44, (Cambridge University Press, 2008). pp. 433–444.

Smith, Q., “A Sound Logical Argument from Evil” in Ethical and Religious Thought in Analytic Philosophy of Language (New Haven CT: Yale University Press, 1997).

Sobel, J.H., Logic and Theism: Arguments For and Against Beliefs in God. 1 ed. (New York: Cambridge University Press, 2009).

(C) If God exists, then he instantiates all good-making properties, and no evil-making properties.

.. which I had taken to be an unproblematic consequence of perfect-being theology. Here I’d like to provide some motivation for the connection. For ease of exposition, I assume that God exists, and as in the previous posts, ‘good-making property’ refers to types rather than tokens. 

The phrase “There is no such thing as a perfect X” may amount to two different kinds of denial. First, the denial may be purely existential: it just so happens, as a matter of contingent fact, that no perfect X exists. Second, the denial may be also conceptual: there is no such thing as a perfect X because there can be no such thing as a perfect X.

This latter option further breaks down. Perhaps there can be no such thing as the perfect X because there can be no such thing as an X. Or perhaps there can be no such thing as a perfect X because there are no evaluative standards for an X which a perfect X would meet. It might be because there are multiple evaluative standards, and so whilst there is such a thing as being a perfect W-X, there is no such thing as being a perfect X simpliciter. Finally, it may be because, although some single standard is operative over an X, maximal fulfillment of every evaluative dimension of that standard is impossible to meet.
Given that these options cover the various ways in which the denial of a perfect X can be true, we can parlay these observations into a set of claims that the oppposite position, perfect-being theology, makes:

(i) A perfect-being exists (i.e. God).
(ii) The concept of a perfect-being is logically possible.
(iii) There is an evaluative standard by which a being is to be judged.
(iv) There is only one such standard.
(v) The maximal fulfillment of every evaluative dimension of that standard by some individual is possible.

At this point one wonders just where the evaluative standards come from. Obviously, they come from us humans in some way or another (it’s our language after all), but by the same token, we can’t cheat and use any old standard – that would render a claim of perfection meaningless. So I think a reasonable suggestion here is that the evaluative standards are those found in the X specifying the kind of perfection in question: a perfect knife is to be evaluated by the standards of knives, and so on.

Again, there are two ways this might go, according to how X determines the evaluative standard. X might provide content that is otherwise absent: nothing is really perfect simpliciter¹, but perfect in this or that way, and so, e.g. the ‘knife’ in “a perfect knife” tells us in what way it is perfect, according to the evaluative standards for knives. But this approach doesn’t bode well for perfect-being theology: the concept ‘being’ is inspecific, providing little or no content, and so it is hard to see how an evaluative standard can be extracted from it. The second way of thinking about this is to suppose that there is such a thing as perfection simpliciter, and that the X determines content by providing a restriction on this perfection. So, ‘knife’ restricts the concept of perfection in “a perfect knife” to certain dimensions of evaluation particular to knives.

Of the two, this latter seems most promising for perfect-being theology: since ‘being’ is the widest category which can be drawn, that ‘perfect-being’ applies to some individual allows us to draw the sort of substantial conclusions which the theologians are wont to draw. Further, it appears to make ‘perfect-being’ synonymous with ‘perfection’ itself, as no restriction is made to differentiate the terms. As this second option is salutary to the perfect-being theologian, and there appear no workable alternatives, I shall assume it as correct. Therefore, I take perfect-being theology to endorse the further claim:

(vi) A perfect-being maximally fulfils every evaluative dimension.

(No need to relativize this to a standard: on the present account the standard is as wide as it can be). From here, it’s a short argument to (C), which I targeted at the beginning of this post:

(vi) A perfect-being maximally fulfils every evaluative dimension.
(*) God is a perfect-being.
(#) To fulfil every evaluative dimension is to instantiate every good-making property.
So, (C’) God instantiates all good-making properties.

Perhaps (#) might need an extra word. Suppose the contrary: that an individual which fulfils every evaluative dimension does not instantiate some good-making property. Since, being a good-making property, this property is evaluative, there is some evaluative dimension which requires it for fulfilment, namely the binary dimension which concerns this property solely. So an individual which lacked this property would not fulfil every evaluative dimension after all, and therefore (#) is true.

(C’) is the comparatively controversial fragment of (C) – I take it there are no objections to the conjunct which states that God instantiates no evil-making properties. And so, I think that (C) is a commitment of perfect-being theology.

¹ This is not to say that we never use the concept of perfection unqualified, just that wherever we do, the qualification would be implicit.

In the last post, Joyce’s conception of moral inescapability was introduced, and moral imperatives were identified as strongly categorical. In this post, Joyce uses the insights gained to sketch an argument for a moral error theory. A theory of practical rationality is given, and used to refine the argument.

The Argument for Moral-Error Theory: First Pass

This brings us to our first sketch of Joyce’s argument:

1. If x morally ought to φ, then x ought to φ regardless of whether he cares to, regardless of whether φing satisfies any of his desires or furthers his interests.
2. If x morally ought to φ, then x has a reason for φing.
3. Therefore, if x morally ought to φ, then x has a reason for φing regardless of whether φing serves his desires or furthers his interests.
4. But there is no sense to be made of such reasons.
5. Therefore, x is never under a moral obligation.¹

Premises 1,2 and 4 do the work here. 1 expresses the conclusion we arrived at earlier, that moral imperatives are categorical, rather than merely hypothetical. Joyce adds to the justification for 1 in passing, noting that once we know that a certain action has been performed, we take ourselves to have all the information we need to pass moral judgment on the perpetrator. Indeed, it is not uncommon that we purposely ignore the desires of the transgressor as both irrelevant to and distracting from our moral assessments. The point can be put more forcifully: there are no unusual facts about the desires and interests of Hitler that would falsify the claim that he ought not to have enacted the Final Solution. Premise 2 is a special case of Mackie’s Platitude (MP), mentioned in the last post. However, Joyce rather confusingly wants ‘reason’ to mean something stronger than the institutional reasons we met there, for obviously ascription of these reasons do make sense, which 4 would appear to deny. He wants ‘reason’ to mean a real reason, one which cannot legitmately be ignored, the sort of inescapable reason promised by the interpretation of moral imperatives as strongly categorical.
So we seem to have good grounds for 1 and 2, whereas 4 tells us that the kind of reasons indicated are incoherent. But are they? To decide what sorts of reasons are legitimate grounds for action, we need to settle on a theory of reasons, of practical rationality.

Practical Rationality

We begin with a commonsense distinction between two kinds of reasons:

Objective Reasons: S has an objective reason to φ if and only if φing will further S’s ends.

Subjective Reasons: S has a subjective reason to φ if and only if S is justified in believing that she has an objective reason to φ.²

So, suppose young Thomas desires to ride the rollercoaster – he knows that other people find rollercoasters fun, and wants to join in. What Thomas assuredly has is a subjective reason to ride the rollercoaster, since he is justified in believing that he will have fun, which is one of his ends. In usual run of things this will also amount to an objective reason – Thomas really will have fun and suffer no ill effects from the ride. But now suppose that Thomas has gorged himself on one too many hotdogs, and his riding will result in violent illness, spoiling his day.Here subjective and objective reasons come apart, much like justification and truth do, and so although Thomas has a subjective reason to ride, he no longer has an objective reason. Indeed, he has an objective reason against riding, since not riding would actually further his ends. This distinction suggests a proposal:

Instrumental Rationality: S is practically rational to the extent that she is guided by her subjective reasons.³

Supposing Thomas chooses to ride knowing that the ride is fun, he is practically rational, even if it should happen that his objective reasons would recommend a different course of action to this. On the other hand, someone is practically irrational when they act for no reason at all, or when the reasons for which they act are not justified reasons (reasons which have the epistemic support of one’s other beliefs). Practical irrationality is then a form of epistemic negligence, a failing of taking into account one’s evidence in deciding how to act. Various phenomena may be responsible for this failure – “rage, passion, depression, distraction, grief, physical or mental illness”⁴ – which can prevent agents from being motivated and/or acting as they should.
Prima facie, this Humean Instrumentalism causes trouble for the moral realist. If objective reasons are defined by their connection to an agent’s ends, then there appears to be no latitude for the idea that we have reasons to perform actions regardless of our desires or interests. The only hope here appears to identify a universal desire of some kind, perhaps self-interest, which, whilst not strictly qualifying as categorical in the Kantian sense, would be the next best thing. However, it is highly doubtful that moral prescriptions always coincide with self-interest: the very ideas of ‘what one wants to do’ and ‘what one ought to do’ are often presented as contrastive. As Joyce notes, the search for universal desire “has all the appearance of a lost program”⁵.

An Alternative Theory

However, the above view of rationality is subject to objection. The objection runs as follows: if all that is required for practical rationality is that the agent should be guided by their subjective reasons, then any action will count as rational so long as the agent has a relevant desire. But, it would seem, at least some of our desires are themselves irrational, and thereby indict the actions produced for their sake. In particular, the problem comes to the forefront in cases where the agent is naturally described as suffering under akrasia or ‘weakness of the will’ – a disposition to act contrary to one’s own considered judgment about what it is best to do. Since this typically occurs when one of the agent’s desires comes to briefly overpower the others, this incontinence will usually meet the condition of according with the agent’s subjective reasons. But weakness of the will is generally classified as a species of irrationality.
The phenomenon is endlessly puzzling if we continue to think of the situation as one in which the agent has two equally worthy desires, one of which happens to win out. So a helpful distinction we could bring in is between what an agent values, and what he merely desires, where valuing X may be defined in terms of a second-order desire in favour of desiring X, or alternatively the having of a belief that X is desirable. Where a value conflicts with a mere desire, and the agent is guided by the desire rather than the value, the desire is irrational, and so the agent acts irrationally.
There’s another problem with our simple view of practical rationality: it is plausible to think that what it is rational for an agent to do depends upon what he or she would do after the exercise of their faculty of reason, i.e. after adequate deliberation. But such deliberation may supply one with new values, desires, and beliefs, or eliminate some that one has. One may come to believe that a particular belief is false, for instance. Or one may come to see that some new desire or value organizes our previous desires into a coherent whole, such that it would be preferable to adopt it.
Such problems lead Joyce to concede the earlier view in favor of what he calls ‘Non-Humean Instrumentalism’. It adds to the above distinction between subjective and objective reasons a rational/irrational dimension, and seems to break down like so:

Objective Rational Reasons: S has an objective rational reason to φ if and only if φing will further the ends of S’s idealized rational counterpart S+, where S+ is a flawlessly deliberating version of S. (All other objective reasons are irrational).

Subjective Rational Reasons: S has an subjective rational reason to φ if and only if S is justified in believing that she has an objective rational reason to φ. (All other subjective reasons are irrational).

Non-Humean Instrumentalism: S is rational to the extent that S acts upon his or her subjective rational reasons.⁶

So, putting these distinctions to use:

“Suppose Molly has a rational desire to refrain from eating a piece of cake, but an irrational desire tugging her to eat it. I see no grounds for insisting that only one desire provides her with a reason. It seems preferable (or at least permissible) to say that she has reasons both to eat the cake and not eat the cake, but the latter is privileged – it “trumps,” if you will – in virtue of deriving from a rational desire. (Alternatively, she may not have a desire to refrain at all, but still have a trumping reason to refrain if such a desire would be created in the process of deliberation.) This distinction between rational and irrational reasons is consistent with the earlier one between subjective and objective reasons. There are those actions Molly might perform which really would satisfy her rational desires, and those that she is justified in believing would. So, for example, she may well be justified in believing that refraining from eating the cake is the action that will satisfy her rational desires (where her desiderative set includes a desire to avoid getting chubby – one which would survive deliberation – and she notes, reasonably, that her cake-eating habits have been pushing her in that undesirable direction). In that case she has a subjective rational reason to refrain from eating the cake. However, suppose that she is oblivious of the fact that tomorrow she will be marooned on a desert island, and the extra calories from the cake would put her in good stead for an extra day or two, perhaps even contributing importantly to the highly desirable end of saving her life. In such a case she has an objective rational reason to eat the cake. In sum: she has (i) an objective rational reason to eat, (ii) a subjective rational reason to refrain from eating, and (iii) a quite irrational desire (providing her with an “irrational reason,” we might say) to eat.”⁷

The Argument for Moral-Error Theory: Second Pass

With a theory of practical rationality in hand, we can now embark on an investigation of whether it can endorse anything like a moral reason. In order for that to happen, our Non-Humean Instrumentalism must support (i) (strong) categorical imperatives, and (ii) those categorical imperatives including paradigmatic moral prescriptions – “proscribing stealing, enjoining promise-keeping, and so forth”⁸. But before we get to this, Joyce has some adjustments to make to his argument:

1. If x morally ought to φ, then x ought to φ regardless of what his desires and interests are.
2. If x morally ought to φ, then x has a reason for φing.
3. Therefore, if x morally ought to φ, then x can have a reason for φing regardless of what his desires and interests are.
4. But there is no sense to be made of such reasons.
5. Therefore, x is never under a moral obligation.⁹

Whereas premise 1 had previously read “If x morally ought to φ, then x ought to φ regardless of whether he cares to, regardless of whether φing satisfies any of his desires or furthers his interests”, the ‘regardless’ clause now has wider scope. This reflects the expansion of our theory of rationality to include idealized rational counterparts; a necessary change since unless the entire sphere of practical rationality is covered by the ‘regardless’ clause, there is latitude for moral imperatives to be trumped by other rational considerations, which could contravene their categorical nature. Since an idealized rational counterpart is constructed from the desires and interests of the agent whom it is counterpart of, the simple “.. regardless of what his desires and interests are” substitution does the trick. Premise 3 is adjusted to match 1, and premise 4, containing a demonstrative, covertly changes with it.

That’s it for now. The next post concerns relativity of reasons, and with it, the prospects for strong categorical imperatives.

¹ Joyce, p.42.
² Joyce, p.53.
³ Joyce, p.54.
⁴ Joyce, p.58.
⁵ Joyce, p.61.
⁶ These are my own explications of what Joyce says on p.71-5.
⁷ Joyce, p.72-3.
⁸ Joyce, p.75.
⁹ Joyce, p.77.

In the last post, we left off with a method of determining  whether an error-theoretical stance should be taken toward moral discourse: first, find one or more non-negotiable propositions implied by the discourse; second, attempt to ascertain whether these non-negotiable propositions are true. In this post, Joyce begins to build his case for an error theory.

Moral Inescapability & Prudential Oughts

Joyce’s argument for error theory centers on what he calls ‘moral inescapability’.  He views moral discourse in the tradition of J.L. Mackie, as being objectively prescriptive:

“..it is the idea that there are actions which we “have to do, regardless” that underlies the claims of objective prescriptivity. The problem of ordinary moral discourse is not a matter of what motivations accompany our moral judgments – it is, rather, that we think that people are “bound” even if they make no moral judgments at all. Even the person who has rejected that whole realm we still think of as being under the jurisdiction of morality.” ~ Joyce, p. 31.

These ideas about being “bound”, about having to do something “regardless” seem hopelessly vague. However, we can begin to see just what they mean by comparison with non-binding and circumstantial ‘prudential reasons’:

“Moral inescapability is an elusive notion. Imagine the child asking why he mustn’t pinch his play mate. The parent replies “Because it’s wrong.” The child continues “But why mustn’t I do what’s wrong?” The parent might give an exasperated “Because you mustn’t!” It is the attempt to clarify the inadequate parental response that is the task of this chapter. Of course, there are all sorts of kinds of reasons that the parent might give: “If you pinch Violet Elizabeth, then she might pinch you back” or “. . . then she won’t play with you” or “. . . then you’ll be sent to your room.” All good (and possibly effective) prudential reasons. But prudence, I shall argue, is not what underwrites moral prescriptivity. Regarding any type of prescription which can be justified on prudential grounds, we can always imagine an unusually situated or unusually constituted agent who “escapes” the prescription. Perhaps the child doesn’t want Violet Elizabeth as his friend, perhaps he doesn’t mind being pinched or being sent to his room. Or perhaps these costly consequences are things he has the power to avoid. In such a case what becomes of the injunction against pinching? On prudential grounds, it must evaporate. But moral proscriptions do not evaporate, regardless of how we imagine the agent situated. If it is not pinching, but torturing that is at stake, then there is no escaping. But the thought that torturing is always proscribed on prudential grounds is just silly. That it always is as a matter of fact is a case that might be made – but that it must be, even in situations where the philosopher gets to stipulate the costs and benefits (let’s say without breaking any laws of nature) is nothing but groundless optimism.“ ~ Joyce, p. 31.

So, being “bound” by moral prescriptions can be understood in terms of inescapability, where no unusual facts about an agent can suffice to exempt them from such prescriptions. Their rational force does not evaporate upon discovery of an agent’s desires or circumstances, unlike prudential prescriptions.

Of course, this remarkable difference between prudence and morality hasn’t prevented philosophers from arguing that moral prescriptions are to be fulfilled due to their practical consequences. After all, it is difficult to see just how else moral prescriptions are to be supported. But aside from the above quoted concern, that prudential grounds cannot always be available to support a moral prescription, this longstanding practice appears to give the wrong answer to the question of why one ought not to perform immoral acts: e.g. if the reason why one ought not to murder is because murdering is morally wrong, and a prudential reason is offered as the reason why one ought not to murder, then this seems to identify moral reasons with those prudential grounds. Yet although it may be true that eternal damnation, guilt, imprisonment, etc., are reasons not to murder, their threat is not what makes a prescription moral. So, useful though it may be to catalogue the prudential reasons in favor of a moral cause,  it appears that this project does cannot tell us what reasons stand behind moral oughts qua moral oughts. A better model for what they are is afforded by ‘institutional oughts’.

Institutional Oughts

Another way to gloss the prudential prescriptions considered above is to take them as supporting hypothetical imperatives, where these prescribe action to an agent on a hypothesis of that agent’s supposed desires or ends¹. As alluded to earlier, such prescriptions are dependent for their application upon facts about an agent, and so may not always apply to “unusually situated or unusually constituted” agents. By contrast, categorical imperatives declare ““an action as objectively necessary in itself apart from its relation to a further end.”“². So, lacking a dependency on contingent facts about the agent to whom it is applied, a categorical imperative cannot be escaped by those for whom the contingent facts do not hold, and thus its prescriptive force is non-evaporable too. As these two features are shared by moral imperatives, it therefore seems that they must be categorical in nature.

Interestingly, moral imperatives aren’t the only sort of categorical imperatives. Various systems of rules, or ‘institutions’ – those of ettiquette, of a social club, of gladitorial combat, even – give rise to imperatives which are likewise non-evaporable. For instance, suppose it is the rule of one’s social club that each member should bring a plate to their regular meetings: in such a case this rule supports a corresponding imperative “You ought to bring a plate” and accounts for the correctness of its use. That this is so even if the member has good prudential reasons for not bring a plate, that the imperative cannot be discharged by producing those reasons or any other facts, marks the imperative as categorical. But suppose the prudential reasons were compelling, that perhaps a member has failed to bring a plate because her house has burnt down and this makes bringing a plate an inconvenience as well as a financial burden? Well then, we feel the need to make a distinction: in one sense, the member ought to have brought a plate, in another sense, she needn’t have done so. Of course, we think that the prudential considerations here trump the institutional considerations, but by the same token, the fact that the rules of the social club remain in place regardless require us to continue to recognize the correctness of the original prescription. (Rules are rules, after all). And so it is with morality – so far, so good.

However, this analogy obviously has its limits. Perhaps the social club member doesn’t particularly value their membership: if so, then although it would still be true to say that the member ought to bring a plate, we could no longer think of these rules as binding her, as being her rules, as really saying what she should do. It would be legitimate to ignore the rules in favor of her own prudential considerations. But moral imperatives cannot be legitmately ignored, and cannot be escaped in the event that an agent doesn’t particularly care for morality.

Categorical Imperatives, Weak & Strong

So what accounts for the difference between the weak categorical imperatives generated by systems of rules and the strong categorical imperatives of morality? One Kantian answer is that strong categorical imperatives bring with them reasons for action, and such reasons may not be legitimately ignored. But don’t weak categorical imperatives too imply reasons for action? With J.L. Mackie we might think that part of what we mean by ‘ought’ is ‘has a reason’:

MP: It is necessary and a priori that, for any agent x, if x ought to φ, then x has a reason to φ³.

In support of MP (Mackie’s Platitude) it might be pointed out that linguistic constructions of the form “You really ought to φ, though you have no reason to φ” sound incorrect. Moreover, it seems perfectly in order to respond to an imperative’s invocation by asking for a reason in its favor, and refusing the prescription if no reason is supplied. Supposing MP is true, then, the difference between weak and strong categorical imperatives must be found elsewhere.

Still, why should it be initially plausible that weak categorical imperatives fail to imply reasons? Perhaps because the reasons which are offered in favor of, say, etiquette, often amount to an impotent invocation of that framework: suppose I ask you not speak with your mouth full, and you ask for a reason – here it would seem pointless of me to point out that you would violate etiquette in doing so, as your question makes it unlikely that you would accept the dictates of etiquette, were they explicated. So it is not that weak categorical imperatives do not imply reasons, but rather that the reasons they provide are in some sense internal to a framework, and so such reasons may be escaped by a person who places themselves outside the framework. In contrast, one cannot step outside of the moral framework in the same way. One can question the framework (‘why should I be moral?’) or signal a lack of commitment to it (by using relativistic locutions like “According to our moral framework, X is wrong”), but even so, we take this framework to bind reasons to persons and make demands of them which cannot be legitimately ignored. Morality cannot be escaped “by simply flying the skull and cross-bones and renouncing altogether the aim of belonging to the moral community”⁴.

It is this inescapability, then, that distinguishes strong categorical imperatives from their weak counterparts. It is this inescapability which is peculiarly moral, and the target of Joyce’s substantive criticism.

¹ Hypothetical imperatives take the form of a conditional: “If φ, then ψ”, where φ might be something like ‘..you want to dunk a basketball..’, and ψ ‘..perform calf raises, squats, and plyometrics exercises daily.”. Since they only apply to an agent if φ is true of the agent, assertion of a bare hypothetical imperative, or its consequent (φ) typically presuppose that φ is true of the agent. The presupposition can be cancelled by an explicit denial of φ.

² Joyce, p. 35.

³ Joyce, p. 38. Strictly MP needs sharpening: sometimes ‘ought’ is used in a predictive sense, as when we say “John ought to arrive any minute now”. MP only applies to non-predictive uses of ‘ought’.

⁴ Joyce, p. 42, quoting David Wiggins.

Anyhow, Emil of ‘Clear Language, Clear Mind‘ thinks settling the question is worthwhile, and to that end explores two commonly adopted nomenclature. The first of these he calls the ‘Old Nomenclature’:

• Atheism means “the denial of the existence of God or gods.“
• Theism means belief/faith in the existence of God or gods.
• Agnosticism means either “the belief that there can be no proof either that God exists or that God does not exist” or is the lack of belief “either way”.

In contrast, the ‘New Nomenclature’:

• “Atheism” means the lack of belief in God or gods.
• “Weak/negative atheism” means the lack of belief ‘either way’.
• “Strong/positive atheism” means the denial of the existence of God or all gods.
• “Theism” means belief in the existence of God or gods.
• “Agnosticism” means either the belief that there is no knowledge about God or gods, or the belief that knowledge of God or gods is impossible.

But how do we decide between them? One argument Emil marshals is that we would prefer the terms of our nomenclature to be exclusive of each other – though he doesn’t say why, I expect that exclusivity is a virtue because if our terms are exclusive, then this an aid to both clarity and and efficient communication. So, given that the Old Nomenclature allows ‘agnostic’ to be combined with both ‘theist’ and ‘atheist’ categories², the New Nomenclature is preferable – right? Well, no. In fact, the New Nomenclature is worse than the old in this regard – not only can ‘agnostic’ be combined with ‘atheist’, but the addition of ‘weak’ or  ’negative’ makes for a triple-barreled label of overlapping terms.

A better argument he gives is simply that the Old Nomenclature is unclear. In particular, the root problem seems to be that it ascribes two meanings to ‘agnostic’ which are either compatible with the ‘atheist’ label (i.e. the “no proof” meaning”) or apparently incompatible (i.e. the lack of belief “either way”). This means that in any given case where someone self-identifies as agnostic, we have no idea what the relevant contrast positions are.

A third argument, which Emil considers is based in etymology: since the term ‘atheism’ comes from the Greek “theos”, meaning “god”, and since the prefix “a-” means “not” or “without”, we should construe ‘atheism’ as meaning “without God” or more awkwardly, “not God”. That is, we should construe ‘atheism’ as the negation of ‘theism’. Concerned to avoid the etymological fallacy, Emil rejects the argument; however, I find it reasonable – as Emil notes, etymology is a guide to meaning, and for my part I think it stands in favor of the New Nomenclature against the old that it gives a definition consonant with the contribution “a-” makes to other words (‘apolitical’, ‘agnostic’, ‘atemporal’, etc.). Certainly we have a fallacy if a straight line is drawn from ancient language to modern definitions, but where the original meaning is cited to show how the prefix works elsewhere, I don’t think there’s a problem.

Leaving behind Emil’s post, I think there are a couple more reasons to discard the old terminology. First, the overlapping of categories is much worse than it first seems. Imagine a person living in one of the Scandinavian bastions of irreligion, who publicly denies the existence of God for reasons of conformity, but privately believes that God exists, though he cannot be known to exist. Such a person would be, according to the Old Nomenclature, a ‘theistic agnostic atheist’. But this is absurd, for surely when one identifies as ‘a theist’, this excludes being ‘an atheist’, and vice versa! Yet this strange mixture is allowed by the identification of ‘atheism’ with performing a particular speech-act, that of denial, rather than contraposing it with belief in God. Unfortunately, the New Nomenclature generates equal absurdity – our Scandinavian would be a ‘theistic agnostic positive atheist’. Worse, since the denial of the existence of God does not here couple with a lack of belief in God, the Scandinavian would not count as an ‘atheist’, despite being a ‘positive atheist’!

Another problem is that the old terminology seems fundamentally unfair on the atheist. Supposing a theist to be someone who believes that God exists, a person who identifies as a theist does not commit themselves to being able to provide any reasons for that belief – they are merely self-reporting. On the other hand, if an atheist denies the existence of God, then identifying as an atheist involves making a truth-claim (It is not true that “God exists”), and like all truth-claims, there is an expectation that the claimant has good reason to back that claim up³. I take this asymmetry to be undesirable – the point of our terminology is identify distinct attitudes to God’s existence, not to impose an evidential burden on one party or other. Insofar as it does impose a burden, the asymmetry tends to distort our picture of what people’s attitudes are, as some will be more willing to take on this burden than others.

At any rate, it seems that both the New and Old Nomenclature are problematic. So how should we define our terms? I humbly suggest the following definitions:

These definitions are hardly earth-shattering. However, that’s the point. For the sake of clarity, the formal definitions of the three terms ought be pretty close to what we’ve been working with – else we’ll just be adding more confusion to their use. Therefore ‘theism’ stays exactly the same. ‘Atheism’ is now symmetrical to ‘theism’, and together these two terms are mutually exclusive and exhaustive: one is either a theist, or an atheist, that’s it.

The big change, of course, concerns ‘agnosticism’. As with the other terms, it’d be best if it matched previous definitions, but failing that, a revisionary definition would do well to accord with common use, which is what my definition does: I take it that far and away the most common use of ‘agnostic’ is to designate the proverbial fence-sitter, rather than the refined philosophical skeptic. As such, ‘agnostic’ should denote a middle ground between ‘theism’ and ‘atheism’, something which seems at odds with the mutual exclusivity and exhaustiveness of those labels. Defining an  ’agnostic’ as a person who lacks knowledge of their own God-beliefs lets us have our cake and eat it too – though it must be true that the person either has a belief in God or lacks it, and thus is either a ‘theist’ or an ‘atheist’, we wouldn’t expect the self-identified agnostic  to be able to commit to one side, given that they lack the knowledge to say which. Further, the definition has the added benefit of being (somewhat) faithful to philosophical use of the term – ‘agnostic’ here does indicate a lack of knowledge, though the particular lack it indicates differs from the traditional one. Outstanding.

Well, that’s my recommendation about how we should define these terms – I think it overcomes all the muddle-headed difficulties, Old and New, and all at the small cost of revising the one term out of the three which is most abused. Think it’ll catch on?

¹ Yes, once there were forums. It was awhile back.

² I’ll freely use ‘theist’, ‘agnostic’ and ‘atheist’ throughout, rather than the mass terms, since it is easier to explain the pros and cons by citing cases. I take these terms to apply to persons, by the way, thus avoiding the problem that  terms ‘agnostic’ and ‘atheist’ on some definitions might queerly be thought to apply to inanimate objects.

³ An easy argument for this: utterances of the form “X, but I don’t know X” sound strange to most ears, indicating that some linguistic expectation has been violated. Since it doesn’t seem contradictory, it’s likely to be some pragmatic expectation that’s been cancelled, resulting in the odd pitch – in this case, the expectation that one has adequate evidence for one’s belief is cancelled by denying one’s knowledge of X.