The evidential argument as championed by William Rowe asserts that it is not possible in principle both for God to exist and for gratuitous evil to occur; and yet, it seems that there is gratuitous evil, so there probably is no God. The first premise, what might be called the Standard Claim – that it is not possible in principle that God exist and gratuitous evil occur – while treated as self-evident by Rowe, has been challenged by several philosophers, including a clever and original objection presented by Peter van Inwagen. Van Inwagen's objection has at its heart the proposition that:

Proposition (N) is called the No-Minimum Claim.Footnote ¹ In a previous article I argued that (N), understood in a way relevant to van Inwagen's argument, was false. Concisely put, I argued that there is a difference between diminishing something and diminishing something in a morally significant way – to diminish pain and suffering in a morally significant way requires a detectable diminishing, but pain and suffering, even if diminishable ad infinitum, cannot be diminished in a detectable way ad infinitum.Footnote ² In a later article, I contended that the inference of (J) from (N) was unsound:

The argument was that suffering is a moral concern, so for any two amounts of suffering, such that the second involves more suffering than the first, one has a strong reason to opt for the first over the second.Footnote ³

The overall dialectic can be summarized by employing these abbreviations:

G:

God exists

P:

pointless or gratuitous evil

H:

human pain

Taking the Standard Claim to be that no possible world contains G & P, van Inwagen's argument contra the Standard Claim can be represented as:

1. 1: Every possible world with H has P.

2. 2: There is a possible world in which God may allow H. So,

3. 3: There is a possible world with G & P.

The primary support for premise (1) is proposition (N), the No-Minimum Claim. I argued that given the human sensory apparatus, there is a practical lower limit to the sensitivity of felt pain – heat at certain levels generates felt pain, at others it does not. There being a practical lower limit to felt pain implies that (N) is false.

One might seek to salvage the argument by replacing (N) with:

But, no matter ‘where’ God sets the practical lower limit (makes it more or less sensitive), it is still a lower limit, and thereby falsifies (N). And, of course, replacing (N) with (R) is not relevant as regards the Standard Claim, since the Standard Claim has to do with gratuitous evil and not with the contingent and variable levels of pain sensitivity.Footnote ⁵

Chris Dragos has recently offered two interesting objections to my arguments.Footnote ⁶ Dragos notes that my arguments understand evil as pain and suffering, while van Inwagen's No-Minimum argument is framed as dealing with evil-in-general. And as evil-in-general is a wider concept than pain and suffering, Dragos objects that ‘showing that pain and suffering are not infinitely diminishable is insufficient for Jordan's purposes; Jordan must show that evil-in-general is not infinitely diminishable, since, crucially, van Inwagen is interested in evil-in-general’.Footnote ⁷ Dragos's second objection is that my argument contra the inference of (J) from (N) does not ‘adequately engage’ van Inwagen's argument if it is understood as involving vague concepts. Besides the two objections, Dragos also suggests that the best way to engage with van Inwagen is via an investigation of the merits of divine satisficing.

Consider, first, Dragos's objection that evil understood as pain and suffering is not as broad as evil-in-general and that showing the former is not infinitely diminishable does not imply that the latter is not infinitely diminishable. This objection can be dealt with quickly as it is irrelevant to the issues. To see this, let's distinguish between those evils that consist of pain and suffering (call these ‘class 1 evils'), and those that do not (call these ‘class 2 evils’). Instances of the latter class might include acts of malice, or unrecognized misfortunes, or a loss of liberty. Class 1 evils and class 2 evils jointly cover every evil. Employing this distinction, my objection was that van Inwagen's ‘No-Minimum’ argument did not succeed with class 1 evils, even if it did with class 2 evils. Typically, the evidential argument proceeds via a claim that there is more evil in the world than one would expect if this were a God-created world. Van Inwagen's ‘no-minimum’ objection, if it is intended as a general defeater of the claim that there is more evil than one would expect if this were a God-created world, must deal with both class 1 evils and class 2 evils. It does not. Dragos's objection oddly overlooks this point even while acknowledging, in effect, the distinction between class 1 evils and class 2 evils, and acknowledging that the ‘No-Minimum’ argument is intended as a general reason against the evidential argument.

One might contend that as long as van Inwagen's No-Minimum argument works for class 2 evils, then the Standard Claim is false, as it denies there is any possible world in which G and P occur. So, even if the ‘No-Minimum’ argument does not provide a general reason against the evidential argument, it does show that the Standard Claim is false.

But this contention presents us with a route taking us nowhere significant – for one thing, recalling the distinction between diminishing evil and diminishing evil in a detectable way, it is far from clear that class 2 evils are diminishable in a detectable way, even if diminishable. Without good reason to think that class 2 evils are diminishable in a detectable way we are not entitled to hold that the No-Minimum argument works for class 2 evils. More importantly, let's distinguish between:

And:

Even if the No-Minimum argument works contra (SC2), it does not work contra (SC1), and it is the latter which motivates many important versions of the evidential argument from evil, in particular Rowe's evidential argument.

Dragos's second objection is that my argument contra the inference of (J) from (N) ‘simply denies (J) instead of engaging with van Inwagen's argument for it’.Footnote ⁸ But my 2011 objection went beyond simply denying (J) – I argued that the alleged inference of (J) from (N) was invalid. But set this aside, because the move away from infinite diminishability in favour of vague concepts is not as promising as Dragos thinks, as employing vagueness very probably does not liberate one from invoking infinite diminishability.Footnote ⁹ Consider van Inwagen's analogy of the judge who issues a ten-year sentence for a criminal, when in fact a nine-year and 364-day sentence is equally sufficient for the purposes of rehabilitation.Footnote ¹⁰ If rehabilitation is the sole goal of the sentencing, and if the presiding judge knows that the shorter sentence is just as effective as the longer one, yet opts for the longer one, then the judge has not acted justly.Footnote ¹¹ If α is worse than β, while both are equally effective in achieving a permissible end, one morally should opt for β rather than α, if one opts for either. And failing to do so, while knowing all the relevant facts, would be a moral failing. The only way to evade this is by claiming that for any and every effective alternative there is another which is just as effective, but slightly better. But, when pain and suffering are at issue, this evasion route is unavailable, as pain and suffering are not diminishable in a detectable way ad infinitum. As with his first objection, Dragos's second objection fails.

Finally, let's take a look at satisficing. In book V of the Odyssey the nymph Calypso seeks to persuade Odysseus to remain as her consort, rather than resuming his quest to return to Ithaca:

Faced with a prospect of gaining immortality and consorting thereafter with an ever beautiful woman, and even acknowledging the overall superiority of this prospect, Odysseus prefers an ordinary life.Footnote ¹³ His reason is that Calypso never ‘persuaded the heart that was in me’.Footnote ¹⁴ As Homer has it, Odysseus is no optimizer. We'll say that one is an optimizer if she holds that a person is irrational whenever that person knowingly chooses an alternative all-things-considered inferior to another within his power:

According to (O), for any rational individual who knows that she has at least two alternatives in her power, α and β, such that she can morally choose either, and she judges that α is better than β, then she will choose α rather than β, if she chooses either. Of course, even with (O), there can be rational reasons to choose which do not make an outcome better. One may rationally choose between a domestic and an import when thirsty, since neither brings about a better outcome than the other. Choice among alternatives with equal outcomes is compatible with (O). If (O) were true, then Odysseus is in fact irrational.Footnote ¹⁵

In contrast to (O) Homer might be interpreted as having resourceful Odysseus conform to:

Let's call proponents of (S) ‘satisficers’ as, according to (S), a rational individual who knows that he has at least two alternatives in his power, α and β, such that he can morally choose either, and he judges that α is better than β, can rationally choose β rather than α.Footnote ¹⁷ A rational agent can, in short, knowingly select a suboptimal alternative. While (O) gives rise to an imperative – select the best available – principle (S) generates a corresponding permission:

Proposition (SP) is indexed to certain kinds of circumstances or cases. No satisficer holds that one should forego optimizing in every case, holding instead that there are cases in which satisficing is not irrational. The dispute between optimizers and satisficers arises since satisficing holds that one may judge that α is better than β, and yet knowingly choose β – that a person's choice need not align with his judgement. Rational judgement and rational preference diverge in satisficing cases.

There are strong reasons to hold that one can rationally satisfice, since nearly every human is in fact a satisficer with regard to what he loves most, and to act otherwise – to engage in optimizing – would preclude total or wholehearted commitment. The virtues of love and loyalty, that is, would be lost.Footnote ¹⁸ Optimizing, in other words, is descriptively and normatively inadequate. But what about ‘moral satisficing’ as mentioned by Dragos?Footnote ¹⁹ The case here is not so clear. Certainly, a principle like (U) presents no obstacle to moral satisficing:

The idea of (U) is that one's supererogatory action producing a lesser good rather than another producing a greater good, while not morally wrong, contributes less to one's goodness than would an action producing the greater good. In brief, one's moral standing increases as the good produced by one's supererogation increases. With something very much like (U) in hand, Rowe argues that God must select the best if God is essentially morally perfect.Footnote ²⁰ But (U) may be dismissed out of hand as false: suppose Jones is dining with his grandmother, who insists on buying dinner. Even though he would greatly enjoy the expensive steak much more than the inexpensive burger, Jones orders the burger in deference to his grandmother's budget. Although Jones knows that the aggregate of good produced by ordering the steak (Jones's extreme enjoyment of the steak added to his grandmother's near mystical elation at Jones's obvious gustatory ecstasy) exceeds that of ordering burger, even after the respective costs are deducted, Jones has not, contrary to (U), decreased his moral standing by ordering a burger.

On the other hand, a principle like (D) cannot be overcome so easily, primarily because it is true:

The idea of (D) is that knowingly violating a moral requirement lowers one's moral standing. So, if moral satisficing implies that performing a lesser supererogatory action rather than one greater has no bearing on the agent's moral status, then moral satisficing makes sense. And in so far as Rowe's argument of ‘no-best world’ depends on a premise relevantly similar to (U), then divine moral satisficing may provide the theist a way of escape.Footnote ²¹ On the other hand, in so far as Rowe's argument can be framed on a premise relevantly similar to (D), then divine moral satisficing provides no escape.

Is the concept of divine satisficing relevant to proposition (N) and the No-Minimum argument as Dragos suggests? It does not seem to be relevant in any straightforward way, as satisficing has to do with rational decision-making, while van Inwagen argues, in effect, that permitting gratuitous evil is unavoidable for God, given that divine goals require permitting significant evil.Footnote ²² Divine satisficing, if defensible, would be relevant as an objection to Rowe's ‘no-best world’ argument, since given divine satisficing, God could bring about a world less good than another he could have brought about instead. But suppose that van Inwagen's No-Minimum argument is sound: then it may be, for all we know, that God must bring about the best world that he can, and the best world that God can bring about is a world with gratuitous evil. One can endorse (N) without endorsing divine satisficing, and one can deny (N) without denying that God can satisfice.Footnote ²³