Introduction
According to Wes Morriston, proponents of divine command theory (DCT) are vulnerable to the very sort of subjectivism their view purports to avoid. Instead of providing moral principles that are ‘above the shifting sands of human preference and convention’, the DCT-ist's inability to constrain what God might command leads to the possibility that ‘divine commands and preferences may be arbitrary or even cruel’ (Morriston (Reference Morriston2009), 249). So, for example,
What if … God were to command the annual sacrifice of randomly selected ten-year-olds in a particularly gruesome ritual that involves excruciating and prolonged suffering for its victims? According to the simplest and most straightforward version of divine-command meta-ethics, it would be morally obligatory to sacrifice many children in the prescribed way. (ibid.)
Call a random annual sacrifice of this sort a ‘Sacrificial Scenario’ (SS, for short).
As Morriston sees things, DCT implies that we could be obligated to participate in an SS – at least if God could command one, which Morriston thinks a distinct possibility. But this is simply outrageous: ‘only a terrible deity – one who does not deserve our obedience – would command such a terrible thing’ (ibid., 250). It would be a terrible world indeed, much worse than the one in which we find ourselves, if there were such a deity. The conclusion to be drawn is that DCT is a rank falsehood.
As Morriston rightly observes, however, the friends of DCT are not without response here, ‘[t]he first and most obvious [being] to deny that God could issue such a cruel command, on the ground that it is incompatible with the perfection of His nature’ (ibid.). In this article, we defend this response against Morriston's core complaint – that even if God cannot do evil, nevertheless the following ‘Divine Command Counterfactual’ (DCC) is true: if God were to command an SS, it would be obligatory. We argue first that a sensible DCT-ist will think of DCC as a counterpossible (a counterfactual with an impossible antecedent). Unfortunately, Morriston's semantic method for assigning truth to DCC is neither incumbent upon the savvy DCT-ist, nor does it permit Morriston to reject DCC as false whether or not God is essentially good. We then show that it is plausible to think the sensible DCT-ist will incline to theistic activism, according to which if there were no God, nothing would exist. This metaphysical platform provides a far more discriminating basis for parsing counterpossibles involving God's non-existence, one neatly avoiding the scandal of DCC in a principled way.
‘Sacrificial’ obligations?
‘So, then’, Morriston asks, what if God commanded an SS? ‘Would we have a moral obligation to obey?’ (ibid.). One line of response is to maintain that since God has Moral Perfection and Essential Goodness (MPEG), we needn't worry about this possibility. For in fact it is an impossibility:
Let us consider, first, the suggestion that the perfection of God's nature makes it impossible for Him to command [an SS]. If God is a perfect being, then He is essentially good, and there is no possible world in which He commands anything like [an SS]. (ibid.)
In other words, there is no answer to the question ‘What would our obligations be, if God were to command an SS?’ The question is improper – a bit like asking what God would be like if he didn't exist. We couldn't possibly have an SS-obligation because God couldn't possibly issue an SS-command. There's little more to say. For his part, Morriston is fairly unimpressed with this line of response. He writes:
At first glance, this may seem to be quite a weak response to the critics' charge. Even if God couldn't command [an SS], doesn't the DCT still have the counterintuitive implication that if He did command [an SS], [an SS] would be morally obligatory? (ibid.)
In one sense, this is a question that scarcely needs asking. For as Morriston stipulates from the outset, ‘[a]ccording to the simplest and most straightforward version of divine-command meta-ethics’, if God were to command an SS, ‘it would be morally obligatory to sacrifice many children in the prescribed way’ (ibid., 249–250). By hypothesis, therefore, a proponent of DCT is committed to the ‘Divine Command Conditional’:
DCC: If God were to command that an SS be performed, then an SS would be morally obligatory.
But an SS is a ‘terrible thing’, Morriston tells us, and only a ‘terrible deity’ would command it. Such a God ‘does not deserve our obedience’, in which case we can hardly be expected (much less be obligated) to take part in divinely sanctioned ‘hunger games’. Sadly then, we must say that DCC is false. For under the assumption of its antecedent, our intuitions tell us plainly and emphatically that its consequent is false. What they support, rather, is
~DCC: If God were to command that an SS be performed, then an SS would not be morally obligatory.
Accordingly, if DCT implies DCC, but DCC is false, we have a decent reason for throwing divine command theory overboard.
Now as Morriston notes, one way of subverting this line of argument – he considers it ‘a weak response to the critics’ charge’ (ibid., 250) – is to make the ‘Standard Move’. Following Edward Wierenga (Reference Wierenga1983, Reference Wierenga1989), we note that (on classical theism) God is essentially morally good; hence he cannot command an SS, in which case the antecedent of DCC is a logical impossibility. But that means that DCC is no ordinary conditional; it is what David Lewis calls a counterpossible: a counterfactual conditional with an impossible antecedent. According to the standard semantics for these items, all counterpossibles (without exception) are trivially true.Footnote 1 (For ease of reference, let's call this the ‘Trivial Truth Thesis’.) On this view, not only is DCC true, so is its negation, ~DCC. And thus the proponent of DCT can actually bite the bullet here, affirming Morriston's (Reference Morriston2012, 21) ‘alarming-sounding’ counterpossible, DCC, but with the caveat that it expresses no substantive metaphysical truth – either about the commands of God or about our moral obligations. As Wierenga (Reference Wierenga1983, 395) puts it, an axiological counterpossible like DCC ‘is harmless’. Its truth is purely stipulative and semantic.
Morriston thinks we can reject or at least bypass the Standard Move. There are two things to note. First, he says, invoking the Lewis–Stalnaker semantics, championed by Edward Wierenga (Reference Wierenga1983; Reference Wierenga1989; Reference Wierenga1998), isn't ‘a particularly helpful suggestion’. This is because ‘there are lots of nonvacuously true [or false] “if per impossibile” counterfactuals’ (Morriston (Reference Morriston2009), 250). And then, second, there is ‘a more perspicuous way’ to press the objection against DCT – ‘one that does not involve counterfactuals with impossible antecedents’ (ibid.). Here Morriston claims that the following ‘simple argument’ (ibid., 251) does just that:
(P1) The DCT entails that whatever God commands is morally obligatory.
(P2) God could command [an SS].
(P3) So if the DCT is true, [an SS] could be morally obligatory.
(P4) But [an SS] could not be morally obligatory.
(P5) Therefore, the DCT is false.Footnote 2
Unfortunately, while it may have the advantage of being ‘more perspicuous’, this argument fails to avoid the counterpossible question. For consider (P1). It says that the proposition Whatever God commands is morally obligatory is an entailment of DCT. Now if Morriston is right, this proposition is non-contingent; something's being commanded by God is sufficient for its being a moral obligation. Thus (P1) is more accurately rendered as
(P1*) The DCT entails that (necessarily) for all actions A, if God commands that A be performed, then A is morally obligatory.
A quick glance at the argument (P1)–(P5), however, reveals that Morriston will gladly allow this substitution instance of (P1*):
(P1**) The DCT entails that (necessarily) if God commands that an SS be performed, then an SS is morally obligatory
which in turn implies
(P1***) The DCT entails that (God commands that an SS be performed > an SS is morally obligatory)
where ‘>’ expresses the counterfactual connective. So the first premise of Morriston's ‘simple argument’ involves the claim that DCT strictly implies DCC – the Divine Command Conditional. But in the presence of MPEG, DCC has an impossible antecedent, in which case it is a counterpossible – a ‘harmless’ one if Wierenga is right.Footnote 3
The upshot is that Morriston cannot casually dismiss the Standard Move without entering the fray. Contra Lewis and Stalnaker, he must show that there are non-trivial counterpossibles. And if DCC is a counterpossible (or if the DCT-ist simply thinks it is), Morriston will have to identify the method to be used for assigning its truth value. Whatever that method is, it will be essential that it meet the following conditions: (i) it assigns truth values in a principled, non-arbitrary way; (ii) it guarantees the truth of DCC; and (iii) its use is compulsory for the DCT-ist; it is the method she is obliged to use in evaluating DCC.
In what follows, we show that Morriston's method meets conditions (i) and (ii) in such a way that not only may the DCT-ist disregard it, but the method actually sabotages Morriston's own case against DCT.
Wierenga's way out
One of the gaps in Morriston's coverage of divine command theory is his failure to engage fully the details and technicalities of the Standard Move. As Wierenga (Reference Wierenga1998) notes, the most promising argument on offer for the Trivial Truth Thesis is one devised by Linda Zagzebski. Although she eventually rejects it, Wierenga claims that not only is Zagzebski's argument ‘pretty good’, it is ‘convincing’ and ‘successful’ – her protests to the contrary notwithstanding (ibid., 91–92). It proceeds as follows:Footnote 4
(1) For all propositions p and q, if p is impossible then p entails q.
(2) For all propositions p and q, if p entails q then p counterfactually implies q.
Hence,
(3) For all propositions p and q, if p is impossible, then p counterfactually implies q. [Trivial Truth Thesis]
Now in Zagzebski's opinion, this argument is actually inconclusive. (1), she is prepared to grant, is ‘obviously true’ – indeed trivial – if either ‘entails’ means ‘strictly implies’ (in C. I. Lewis's sense) or if every impossible proposition has an explicitly contradictory structure. In the former case, to say that p entails q is just to say that p strictly implies q, that is, ~ ◊ (p & ~ q) where ‘◊’ is the possibility operator. This seems intuitively correct, as Alfred Freddoso, another theistic defender of (1), explains:
To say that p strictly implies q is to say that it is metaphysically impossible that p be true and q false – a condition trivially satisfied if p is itself necessarily false. So it would be foolhardy indeed to doubt that a necessarily false proposition strictly implies any proposition you please. (Freddoso (Reference Freddoso1986), 44)
Hence, the conjunction of a necessarily false proposition with any proposition is broadly logically impossible. And this is reflected in the fact that ~ ◊ p → ~ ◊ (p & q) is a theorem in the standard modal systems (for example, in Feys's system T and the Lewis systems S4 and S5).
The real problem with the argument, says Zagzebski, lies with its second premise – the Principle of Logical and Counterfactual Entailment (PLACE), as we might call it. This principle hasn't been adequately defended. Here she notes that John Pollock treats PLACE as an undefended axiom of counterfactual logic. He remarks only that it ‘seems evident’ (Pollock (Reference Pollock1976), 39). David Lewis, on the other hand, comments that PLACE ‘ought always to be true’ (Lewis (Reference Lewis1973), 24), while noting that where the antecedent of a counterpossible is explicitly contradictory it logically implies any consequent.
No doubt these considerations are less than wholly coercive. But this fact (taken alone) doesn't give us a sufficient reason to reject PLACE. For from the fact that it hasn't yet been properly defended, it doesn't follow that it is indefensible. And in fact, Wierenga (Reference Wierenga1998, 92) has constructed a very clever argument for this principle. His premises are:
(4) For all propositions p, q, and r, if p counterfactually implies q, and q entails r, then p counterfactually implies r
and
(5) For every proposition p, p counterfactually implies p.
These are taken to be obvious truths. If we let P and Q be particular propositions, then from (4) it follows that
(6) If P counterfactually implies P and P entails Q, then P counterfactually implies Q;
and from (5) we have
(7) P counterfactually implies P.
But (6) and (7) jointly entail.
(2) For all propositions p and q, if p entails q then p counterfactually implies q.Footnote 5 [PLACE]
By all accounts, this constitutes a serious defence of PLACE, one that (unlike the terse remarks of either Pollock or Lewis above) cannot be easily brushed aside. On the face of things, then, it seems that we have sound support for the Trivial Truth Thesis, which nicely positions the proponent of DCT to dismiss Morriston's alarming counterpossible as harmless.
Morriston's manoeuvre
So what is Morriston supposed to say at this point? Although he offers no discussion of Wierenga's argument, we can hazard a guess at his response. On Morriston's reckoning, for example, there are multiple counterexamples to the Trivial Truth Thesis. Just suppose, he says,
that it is indeed a metaphysically necessary truth that God is good. Then it is impossible for God to be evil. But surely it is true that if (per impossibile) God were evil, He would be not be good, and false that if (per impossibile) God were evil, He would be good. (Morriston (Reference Morriston2009), 266)
In other words, if we assume MPEG, we have grounds for affirming the non-trivial truthFootnote 6 of
(E1) If God were evil, he would not be good.
and the falsity of
(E2) If God were evil, he would be good.
Intuitively and at first glance, that seems right. The question arises, however: on what basis do we assign differing truth values to (E1) and (E2). In Morriston (Reference Morriston2009), there is no answer. More recently, however, Morriston tells us that ‘the assignment of truth values in such cases need not be arbitrary’ (Morriston (Reference Morriston2012), 20). Fair enough. But then how is the principled cut to be made? Approximately as follows:
To take just one example, it seems to me that if – per impossibile – a completely truthful and omniscient being said that two-plus-two is five then two-plus-two would be five. What's driving my intuition in this case is that the antecedent does not entail the consequent merely in virtue of being impossible. It also does so because of the way in which its content is logically related to the consequent. Whatever a completely truthful and omniscient being says must be true, and what we have here is a straightforward substitution instance of that pattern. (ibid., 20–21)
So we can add another counterpossible to our stock of counterexamples to the Trivial Truth Thesis, namely,
(O1) If a completely truthful and omniscient being said that two-plus-two equalled five, then two-plus-two would equal five.
No doubt Morriston also means to ascribe falsity to (O1)'s companion
(O2) If a completely truthful and omniscient being said that two-plus-two equalled five, then two-plus-two would not equal five.
Furthermore, we are given a glimpse at the principle driving his alethic intuitions. The antecedents of these four counterpossibles have different entailments. (E1) and (O1) are true because their antecedents ‘entail’ their respective consequents. Not so for (E2) and (O2). In saying this, of course, Morriston is thinking of entailment as something other than strict implication, for the antecedents of these counterpossibles do strictly imply their consequents (since they strictly imply every proposition). Rather, the notion he has in mind seems to be that of Content Inclusion (C).Footnote 7 If the content of the antecedent includes that of the consequent, we've got a true counterpossible on our hands; otherwise, not. A bit more precisely, where p is an impossible proposition, q is any proposition, and ‘©→’ expresses C-entailment, the Morriston truth test for counterpossibles is this: if p ©→ q, then p > q. And if ~ (p ©→ q), then ~ (p > q). Pretty straightforward.Footnote 8
Now how does this bear on Wierenga's argument for PLACE, the key premise supporting the idea that counterpossibles (DCC included) are nothing but harmless trivialities? As follows. If Morriston would deny this principle, he must give up one of the premises leading to it: either (6) or (7). To be as clear as possible, suppose we construct specific instances of these premises, using the propositions figuring in Morriston's (E2). Where ‘⇒’ is the symbol for strict implication, that gives us:
(6i) If ( (God is evil > God is evil) & (God is evil ⇒ God is good) ), then God is evil > God is good
and
(7i) God is evil > God is evil.
How shall we go about assessing these? Well, since the proposition God is evil ©-entails itself, Morriston will no doubt affirm the truth of (7i). As he also seems to endorse MPEG, he will think God is evil is an impossible proposition – one strictly implying God is good in C. I. Lewis's sense. Accordingly, both the left and right conjuncts in (6i)'s antecedent come out true. So far then, so good.
What about the consequent? Surely, Morriston will deny it; for the counterpossible ‘God is evil > God is good’ is false; it is unthinkable that its antecedent ©-entails its consequent. And so there is every reason for us to reject this counterpossible – that is, (6i)'s consequent – as false. Contra Wierenga, then, we have a sufficient reason for rejecting PLACE, thereby giving us an undercutting (if not a rebutting) defeater for the Trivial Truth Thesis. Consequently, it is not at all obvious that the Standard Move is what the DCT-ist needs to ward off the hounds.
The question in all of this (one to which we shall return) is why a proponent of DCT can't follow Morriston in rejecting PLACE and the Trivial Truth Thesis. More exactly, why can't she join the chorus of protests against the long-standing Lewis–Stalnaker semantics, but then go on to argue that while the admittedly alarming DCC might be an implication of naive divine command theory, a sensible DCT-ist need take no responsibility for it?
Signs of trouble
Methodological misadventure
So let's return to that originally troubling counterpossible:
DCC: If God were to command that an SS be performed, then an SS would be morally obligatory.
There are two serious problems with Morriston's handling of this conditional. The first has to do with the mechanics of Morriston's semantic method. He wants to say that DCC is false. However, there is a snag; his truth value test for counterpossibles won't allow it. We can see this as follows. Begin by recalling Morriston's justification for assigning non-trivial truth to
(O1) If a completely truthful and omniscient being said that two-plus-two equalled five, then two-plus-two would equal five.
(O1) is true, he says, because ‘Whatever a completely truthful and omniscient being says must be true, and what we have here is a straightforward substitution instance of that pattern’ (Morriston (Reference Morriston2012), 20–21). Well, that's nearly right. To be exact, (O1) is an entailment of a substitution instance of that general truth. That is, beginning with the obvious analytic truth
(8) Necessarily, for all x, if a completely truthful and omniscient being says x, then x is true,
a substitution instance of which is
(9) Necessarily, if a completely truthful and omniscient being says two-plus-two equals five, then ‘two-plus-two equals five’ is true
we then infer (O1), the relevant counterfactual (by the principle p ⇒ q; so p > q).
The problem here is that we can construct a parallel argument for the truth of DCC, using this same method. For surely, whatever God, a perfectly morally good being, commands must be morally obligatory. And what we have with DCC is a (nearly) straightforward substitution instance of that pattern. We can then sharpen things up as follows:
(10) Necessarily, for all x, if God, a perfectly morally good being, commands x, then x is morally obligatory.
Hence
(11) Necessarily, if God, a perfectly morally good being, commands that an SS be performed, then an SS is morally obligatory
And thus
(12) DCC is true.
In other words, using Morriston's method, we can show that an SS's being morally obligatory is included in the concept of God's having commanded an SS. So what we have here is actually a Humean relation of ideas. But this spells disaster for Morriston's case against DCT. For it commits him to the truth of DCC, the very conditional whose scandal was supposed to be our reason for bidding adieu to DCT. ‘Even if God couldn't command [SS]’, Morriston says, ‘doesn't the DCT still have the counterintuitive implication that if He did command [an SS], [an SS] would be morally obligatory?’ (Morriston (Reference Morriston2009), 250). It's a rhetorical question; a positive answer is expected. And the thing to see is that Morriston's Humean method will assign truth to DCC whether or not its antecedent is impossible. It matters not if, as he thinks, ‘God could command’ an SS (ibid., 251). DCC will still come out true, since this is guaranteed by virtue of the fact that its antecedent's content includes that of its consequent.
Someone might object that there is a vast difference between the omniscience and moral goodness cases. The necessary connection in (8) between being true and what a completely truthful and omniscient being says is patent. Not so for (10). It is not at all obvious that being commanded by a perfectly morally good being and being morally obligatory are necessarily connected. Only the DCT-ist believes that. The rest of us believe that moral duties are determined by more mundane considerations.
But here there is a slight problem. The objector has confused causes and conditions. (10) doesn't actually say that God's commands cause, bring about, or in any way determine our moral obligations. Perhaps they do; but then again perhaps not.Footnote 9 For all that (10) says, it could be that God's commands simply highlight or reinforce obligations we already have on independent grounds. Something similar goes for (8). A completely truthful and omniscient being's saying something is a sufficient condition for the truth of what is said. However, it doesn't automatically follow that God's having said that p is the reason p is true. Sometimes this is the case, of course, as when we are told, ‘And God said, “Let there be light”, and there was light’ (Genesis 1:3, emphasis added). But it certainly isn't true across the board. No doubt God also says (of himself) that he exists; he doesn't thereby cause his own existence.
Signs along the way
There is a second crippling feature of Morriston's semantic method, stemming from his undue emphasis on concept inclusion as the sole factor for sifting true from false counterpossibles. This is surely inadequate. For here we're not merely interested in what conceptual content is (or isn't) included in some impossible antecedent. Rather, we are asking what the extra-conceptual world would have been like had that antecedent been true. Thus Linda Zagzebski asks, ‘Even if it is assumed that some state of affairs could never have obtained, can anything interesting be said about what would have been the case if it had? (Zagzebski (Reference Zagzebski and Beaty1990), 165). For present purposes, we want to know what the world would have been like, if God had commanded one of Morriston's Sacrificial Scenarios. From a classical theistic perspective, it is radically incomplete (and incorrect) simply to retort, ‘Well in that case, I suppose a Sacrificial Scenario would have been obligatory’. This is the reply of the naive DCT-ist, as well as anyone wielding Morriston's semantic apparatus. Both approaches are desperately misguided. For if being perfectly morally good is a property God could not have failed to possess without failing to exist, then any counterpossible whose antecedent involves God's not being perfectly good is a state of affairs involving God's non-existence (a SIGN, for short).
Now what difference does that make? Well, it depends on what you think would be lost if God didn't exist. Apparently, Morriston thinks not all that much. Strangely, he praises Linda Zagzebski for her ‘interesting suggestions about how [a non-standard] semantics for counterfactuals might accommodate counterpossibles’ (Morriston (Reference Morriston2012), 33). But he neglects to inform his reader that Zagzebski's reason for thinking impossible states of affairs have different implications
is connected with the fact that … it is a mistake to think of one impossible situation as the same as any other. In particular, it is reasonable to think that some necessary states of affairs can enter into relations, including causal relations, with other states of affairs. (Zagzebski (Reference Zagzebski and Beaty1990), 173)
Zagzebski goes on to say that Christians ‘are probably committed to this view’, citing Thomas Morris's theistic activism, which advances the null world hypothesis: the thesis that if God did not exist, nothing whatsoever would exist.Footnote 10 Thus Morris:
from the perspective of any thoroughgoing theism – any theism according to which God is necessarily the creator of anything that might exist distinct from himself … if God were, per impossibile, to fail to exist, nothing else would exist either. (Morris (Reference Morris1987), 170)
Brian Leftow makes substantially the same claim: ‘To activism, if God does not exist, nothing else necessary or contingent exists either … This renders God's non-existence unlike any other impossible state of affairs' (Leftow (Reference Leftow1990), 196).
What these remarks suggest is a more discriminating theistic basis (and what DCT-ist isn't a theist?) for tracing out the counterfactual implications of impossible propositions involving God's non-existence;Footnote 11 in which case it becomes apparent (on theistic activism) that DCC is false. For given MPEG together with the obvious fact that commanding that an SS be performed is a mark of gross moral imperfection, it follows that
(13) Necessarily, God commands that an SS be performed if and only if God does not exist.Footnote 12
However, (13) and DCC jointly entail
(14) If God did not exist, then an SS would be morally obligatory
(by the principle (p ⇔ q) & (p > r); hence (q > r) ). Against the backdrop of the null world hypothesis, it is easy to see that (14) is non-trivially false. Its antecedent is a SIGN; however, contrary to what would be the case if God didn't exist (namely, that nothing would exist), its consequent implies the existence of at least one thing: a sacrificial scenario with the property of being obligatory.
Turning next to
~DCC: If God were to command that an SS be performed, then an SS would not be morally obligatory
the thoroughgoing theist will initially insist that ~DCC is ambiguous. On the one hand, it can be read as the de re claim that
~DCC1: If God were to command that an SS be performed, then there would be an SS having the property being non-morally-obligatory.
This proposition will almost certainly be repudiated by the DCT-ist who cleaves to theistic activism; for its consequent is existence-entailing whereas its antecedent involves a SIGN. But there is another reading in the neighbourhood; we can also take ~DCC to be the de dicto claim
~DCC2: If God were to command that an SS be performed, then it is not the case that there would be a morally obligatory SS.
And that the sensible DCT-ist can happily endorse, as its consequent merely denies the existence of a particular state of affairs – a semantic outcome perfectly in line with the null world hypothesis.
The conclusion to be drawn, therefore, is that Morriston's ‘relation of ideas' method is radically insufficient for his purposes. The problem, fundamentally, is that it isn't properly aimed at truth. It doesn't ask the salient alethic question: ‘What if the impossible had been actual?’Footnote 13 It fails to inquire into what the world would be like if there were no God. This is no small shortcoming.
Two objections
Objection 1
‘Your defence of a sensible DCT requires non-trivial truth values for counterpossibles. In particular, it turns on the falsity of DCC. But your reason for thinking that DCC is a counterpossible traces to your endorsement of MPEG – the idea that (necessarily) God is essentially morally good. That's why you think God's commanding an annual ritual sacrifice is impossible. You therefore flagrantly beg the question against Morriston, since (P2) in his ‘simple argument’ (Morriston (Reference Morriston2009), 251) says that God could command a sacrificial scenario’.
Reply: the question is whether it's a cheat to appeal to God's necessary moral perfection to block the inference to his having the power to command an SS. Here it is important to remember where we're at in the dialectic. The ‘simple argument’ is designed to show the DCT-ist something about her beliefs – that they're allegedly inconsistent. So Morriston must shoulder a burden of proof; it is up to him to show the DCT-ist that despite her fondness for MPEG, the concept of God's power she holds (or ought to hold) logically commits her to (P2).
How does Morriston propose to do that? In short, with an a priori argument:
I believe that quite a strong case can be made for saying that omnipotence entails the ability to command [an SS], in which case premise [(P2)] must be true. Admittedly, there is an unproblematic sense in which even an essentially good God has the ability to command [an SS]. If He chose to command [an SS], He would succeed in doing so. What an essentially good God could not do is choose to exercise this power. Given His essential goodness, such a choice is impossible for Him. (ibid.)
Morriston goes on to argue that, intuitively, a being that could choose to command an SS would be ‘more powerful overall’ than one that couldn't, so that if the DCT-ist thinks God is omnipotent, she should also think he can choose to command a sacrificial scenario.
This is a fascinating little argument. Not surprisingly, perhaps, it turns on how we assess a certain ‘lynchpin’ counterfactual – in this case
(15) If God chose to command an SS, an SS would be commanded by God.
The idea is that since God is omnipotent, he can actualize (15)'s antecedent and thus also its consequent. This is said to deliver ‘quite a strong case’ for the conclusion that ‘omnipotence entails the ability to command [SS]’ (ibid.).
Of course, all of this presupposes that (15) is true. But why should we believe that? More to the point, why think the DCT-ist is wedded to (15)? Morriston tells us: ‘This [i.e., (15)] is another example of a non-vacuously true “if per impossibile” counterfactual’ (ibid., 266, fn. 6). Now if that's right, Morriston is treating (15) as a non-vacuously true counterpossible (again contrary to his claim that (P1)–(P5) are counterpossible-free). Like the DCT-ist, then, he believes the antecedent of (15) is impossible. And that will be the case on his view, presumably, because (15) is an instance of the general truth that (necessarily) whatever an omnipotent God chooses to command, he does command – or what comes to the same thing: the content being commanded by God (an omnipotent being) is included in being chosen as a command by God.
Unfortunately, Morriston provides no argument to show that this procedure for parsing counterpossibles is incumbent upon the DCT-ist. What we have, at most, is the mere assumption that it is. In the present context, however, this is little more than question-begging. For as we noted earlier, the DCT-ist, if she is sensible, will affirm MPEG and thus see (15)'s antecedent as a SIGN. She will therefore reject (15) as false, since its consequent is existence-entailing.
Second, in his attempt to demonstrate ‘that there are lots of nonvacuously true “if per impossibile” counterfactuals’ (ibid., 250), Morriston leads with the pair
(E1) If God were evil, he would not be good
and
(E2) If God were evil, he would be good.
But his stated reason for thinking these are counterpossibles requires that we ‘[s]uppose that it is indeed a metaphysically necessary truth that God is good’ (ibid., 266). That is to say, if we suppose that MPEG is true, then (E1) and (E2) count as counterpossibles. That being settled, we can then call upon Morriston's handy semantic method to assign them differing truth values.
But isn't there a problem here? The claim to be defended is categorical: there are non-trivial counterpossibles. Yet the supporting example is conditional: if MPEG is true, then (E1) and (E2) are counterpossibles. To be sure, if you do embrace MPEG, you're in business; you can transform this conditional support into the desired categorical conclusion. As it turns out, Morriston is quite reserved about MPEG. Thus he writes:
[T]he appeal to God's essential goodness is the divine-command theorist's best bet. Whether it can produce a satisfying result depends mainly on our ability to give an account of God's goodness that does not trivialize it and that does not make God subject to an independent standard of goodness. I do not see how to do that. (ibid., 266)
But in that case, Morriston is not in a position to use (E1) and (E2) as counterexamples to either PLACE or the Trivial Truth Thesis. Since she happily and unreservedly affirms MPEG, the sensible DCT-ist faces no such injunction.
Finally, it's difficult to know just how seriously to take Morriston's defence of (P2). He initially proposes ‘that omnipotence entails the ability to command [SS], in which case premise [(P2)] must be true’ (ibid., 251). However, it quickly becomes apparent that what he has in mind is a fairly strong notion of omnipotence. For in the next breath, Morriston notes that the DCT-ist could ‘replace’ the claim that God has power of this sort with ‘something weaker – perhaps with the claim that God has as much power as is compatible with His essential goodness’ (ibid.). Right. That is what one would expect the sensible DCT-ist to say. But wherein lies the problem? This is not clear.Footnote 14 Indeed, it's hard to imagine the DCT-ist saying anything but ‘yea and amen’ to Morriston's earlier words on the matter:
Perhaps the greatest possible being is (a) necessarily morally perfect; and (b) as powerful as is logically consistent with (a). That would make it very powerful indeed – powerful enough to create the world, powerful enough to perform all sorts of (good) miracles … it would have enough power not to detract in any way from God's greatness or make Him unworthy of unconditional worship and devotion. Even with the restrictions on power that flow from necessary moral perfection, the greatest possible being would have enough power to satisfy the Psalmist. ‘Great is your power’, he wrote, but then added, ‘your wisdom is immeasurable’. (Morriston (Reference Morriston2001), 158)
Just so. But then, arguably, (P2) is false.
Objection 2
‘You appeal to MPEG in defending your brand of DCT, but you fail to provide “an account of God's goodness that does not trivialize it and that does not make God subject to an independent standard of goodness” (Morriston (Reference Morriston2009), 266). Until you do that, the jury must remain out on the adequacy of your sensible divine command theory so-called’.
Reply: here there is confusion. The DCT-ist, at least as we have represented her, appeals to MPEG merely to allay the concern that nothing would (or could) prevent God's commanding a sacrificial scenario, thereby making it obligatory. So long as God has the property being perfectly morally good, and has that property essentially – that is, in every world in which he exists – there won't be any worlds in which he commands an SS. But ‘why is God good?’, Morriston asks. What is your account of God's goodness? Is he good because he commands that he is? Well, obviously not. Here Morriston is right: if God could ensure his own goodness simply by commanding it, ‘it would [then] be only too easy for a “good” God to command [an SS]’ (ibid., 252). The DCT-ist concurs. God's commands don't define what is good; rather, they clarify, determine, or reinforce our moral obligations – a different matter entirely. It is simply God's nature to be good.
Here Morriston will reply that this only raises further concerns. For if we identify ‘the Good’ with God's nature, then there must be ‘a cluster of moral good-making properties that God perfectly and essentially instantiates. God is necessarily (morally) good insofar as He necessarily possesses such properties as loving-kindness, mercy, compassion, justice, and so on’ (ibid., 253), in which case we must face the following Euthyphro-like dilemma: ‘Is God good because He has these good-making properties? Or are they good-making because God has them?’ (ibid).
The answer is: neither. There are a couple of things to note here. First, the assumption behind the question is mistaken. It assumes (without argument) that being morally good isn't a basic property of God: a property he has, but not in virtue of possessing some other property. No doubt many of God's properties are non-basic. He has the property having created Adam by virtue of possessing having the power to create. He has knowing that Eve freely takes the apple in virtue of having being knowledgeable. And so on. The question is whether all of God's properties are like this. Why should we think so? What prevents us from saying that God's being good is more like an electron's having a charge? There is no ‘how’ or ‘why’ about it. Thus Plantinga:
An electron has basic properties such as spin and a negative charge. But then the question ‘How does an electron manage to have a charge?’ is an improper question. There's no how to it … Its having a negative charge is rather a basic and immediate property of the thing. (Plantinga & Tooley (Reference Plantinga and Tooley2008), 59)
The same thing goes, we submit, for God's being good. A sensible DCT-ist won't explain this in terms of other properties God has. Her claim will be that God has this property in a basic way – the way in which that electron has its spin and charge.
Second, let's suppose that God does have the basic property of being good. Morriston still wants to know about God's other good-making properties – his mercy, forgiveness, compassion, and the like. Does their making the things that have them good derive solely from the fact that the God who has them is good? Morriston thinks not. ‘Why should it make any difference to the good-makingness of compassion’, he says, ‘if there is (or isn't) a supremely compassionate God?’ (Morriston (Reference Morriston2009), 253). Apparently, Morriston's idea is that it makes no difference. We take it, then, that he would affirm
(16) If God didn't exist, compassion would still be essentially good-making.
The claim to the contrary – that compassion would no longer be good-making in the absence of God – ‘seems incredible to me’, says Morriston. ‘It implies that if there were no God who perfectly exemplified [these properties], [they] would count for nothing’ (Morriston (Reference Morriston2012), 22). Furthermore, no ‘discerning atheist [would] have to think a thing like that’ (ibid.).
Morriston is right or partly right: atheists can certainly be expected to believe (16) and thus to reject its denial. Indeed, a theist might hold that God himself has created compassion (or instances of it) in such a way that anyone (theist or not) can grasp its essence, and see that it includes goodness. So clearly, you don't have to believe in God for compassion to ‘count’ for something.
The real point, however, lies in a different direction. It concerns the metaphysical question: is (16) true? Would compassion still exist if God didn't? No doubt that depends on what sort of thing you think compassion is. Morriston doesn't say just how he thinks of it, and there isn't the space here to explore the matter properly.Footnote 15 In a way, that doesn't matter. The more important question is why we should think of (16) as having any purchase on the DCT-ist (armed with theistic activism). After all, on her view, (16) isn't true; it's a false counterpossible; for it implies that something would exist even if God did not.
Suppose, like Morriston, you hold that neither God's character nor his commands determines our moral obligations. Perhaps you think obligations arise from the social contracts we enter into; or maybe they're just brute givens about how decent human beings should treat one another. In any event, you see them as in some way supervening on human affairs. But now suppose that you're also a theist. Presumably, you won't then think that this material world in which we ‘live and move and have our being’ could have existed in absolute independence of God's causal activity. (What theist ever held that?) Instead, you'll think that if per impossibile there were no God, there wouldn't be a physical world: no planets, moon, or stars, no evolution of homo sapiens, no possibility for human interests, rights, or good-making traits like compassion to emerge. But in that case how could there be any moral obligations? Obligations to whom? There simply wouldn't be anyone to have them to: human or divine.
To be clear: this is not an argument against atheistic moral realism. An atheist can consistently affirm that the things we've said wouldn't exist if there were no God would in fact exist in his absence.Footnote 16 But can a theist say that? You can set aside divine command theory and its alleged ills. Morriston's (16) is false on theism alone.Footnote 17 Accordingly, DCT-ists who are theists (very few aren't these days) will be wholly unmoved by his attempt to use the truth of (16) to make trouble for the idea that being perfectly morally good is a basic and essential property of God. If she has her wits about her, the DCT-ist will see this as little more than a flagrant case of begging the question against theism itself.
Conclusion
It may be true, as Morriston says, that ‘the simplest and most straightforward version of divine-command meta-ethics' has it that if God were to command an SS, ‘it would be morally obligatory to sacrifice many children in the prescribed way’ (Morriston (Reference Morriston2009), 249). And he's right that only a ‘terrible deity’ would command such ‘a terrible thing’ (ibid., 250), so that there could scarcely be any obligation to obey on our part. But how does it follow (as Morriston says that it does): ‘that the divine command theory is false’ (ibid.)? Surely that is a gross non sequitur. At best what must be sacrificed here is that ‘simplest’ version of DCT, not DCT simpliciter. What we have called sensible DCT remains wholly unscathed.
To his credit, Morriston doesn't ‘pretend to have settled all the relevant issues’ (ibid.). One of these – a rather tricky one – is what to do with counterpossibles, having once dismissed Wierenga's Standard Move. The fact of the matter is that Morriston's ‘relation of ideas' method for carving up the counterpossible terrain is the silent partner in virtually every one of the key transactions in his case against DCT. It is therefore somewhat ironic that it ends up committing him to the very thing he is intent on disavowing – that deplorable DCC counterfactual.
By contrast, the DCT-ist subscribing to MPEG might (for any number of reasons) find herself attracted to theistic activism, in which case she will have at her disposal principled reasons for rejecting DCC – reasons emerging quite naturally from a robust conception of DCT itself. Thus, there is no ‘terrible’ deity, no ‘terrible’ commands, and no reason to think that if the DCT-God did exist, the world would be far worse than it already is. To make trouble for this line of reasoning, Morriston needs to persuade the DCT-ist to take his flawed semantics on board, or at least convince her she's not entitled to her own. That's a fairly tall order – one that despite all his useful work on this topic, Morriston has yet to fulfil.Footnote 18
