August 28, 2011
An evidential POE argument from Peter Hurford of Greatplay.net:
1. Needless suffering, by definition, is any suffering that doesn’t exist because of a higher good.
2. Needless suffering, by definition, could be eliminated with no consequences.
3. Any all-good entity desires to eliminate all needless suffering.
4. Any all-knowing entity would know of all needless suffering, if any needless suffering exists.
5. Any all-powerful entity would be capable of eliminating all needless suffering.
6. Our world contains needless suffering.
7. Therefore from 1, 2, 3, 4, and 5, and 6, an all-good, all-powerful, and all-knowing entity cannot exist.
8. God, as described by the major religions is all-good, all-powerful, and all-knowing.
9. Therefore from 7 and 8, God as described by the major religions does not exist.
I recently said that all the POE arguments I’ve heard reduce to arguments from incredulity, and this argument is no different. Inability to conceive of a higher good is the only thing grounding the claim that any given instance of suffering is needless. 6 is a naked assertion sustained only by incredulity. That alone invalidates the argument in my opinion, but I can make a stronger case.
One thing I find troubling about POE arguments in general is that both sides rarely tailor them to a specific theology. This promotes generic arguments and rebuttals, so let’s identify some specific theologies and evaluate each against Peter’s argument. I will paint somewhat broadly to avoid getting bogged down in theological disputes.
On universalism, all sufferers eventually inherit eternal paradise, and no sufferers suffer eternally. On exclusivism, only a minority of sufferers inherit eternal paradise, and the majority suffer either finitely [annihilation], or infinitely [eternal torment]. The argument map varies according to the ontology under consideration, and I’d like to avoid appealing to a subjective value judgment [e.g., Peter might not think the suffering is worth it, but God might]. Though certainly a valid reply—and perhaps ultimately inevitable—I think we should attempt an “as objective as we can get” method of evaluation.
There are at least two ways to try this: we can evaluate tokens of suffering vs. tokens of joy, or we can evaluate the number of agents experiencing suffering vs. the number of agents experiencing joy. For simplicity’s sake, I might use U for universalism, E/A for exclusivism with annihilation, and E/ET for exclusivism with eternal torment.
Peter attempts to buttress 6:
If suffering is not needless, then eliminating that suffering also eliminates the higher good, and we are worse off overall by eliminating that suffering. This means that any suffering we are better off without is needless suffering. And it turns out there is a lot of suffering we are better off without; any notion that suffering exists for some greater good is a notion that crippling polio and smallpox was necessary, and that the world is now worse because we eradicated these diseases. Clearly, “God allows suffering for greater good” and “God wants us to work to ameliorate suffering” are entirely incompatible statements.
From there, he formulates the following supporting argument:
10. If an instance of suffering that is necessary (because of a higher good) were prevented, then that higher good would also be prevented.
11. Therefore from 10, preventing necessary suffering makes us worse off.
12. There are some instances of suffering that were prevented where we did not become worse off.
13. Therefore from 11 and 12, needless suffering exists (and 6 is true).
I’m fairly comfortable granting 10 and 11, but how can anyone know that we are better off without polio and small pox? 12 appears to be a naked assertion with no evidence or argumentation to support it. It might seem intuitively true, but many things that seem intuitively true are apparently false [cf. geocentrism]. Conversely, many things that seem intuitively false are apparently true [cf. quantum mechanics]. Like 6 in the main argument, 12 is a naked assertion sustained only by incredulity: “I can’t imagine how we would be better off with polio and small pox, ergo we were better off preventing them.” There is more that could be said here—for example, we could talk about whether the higher good has to follow every instance of suffering and obtain in this life, as I explore here—but per Peter’s definition of “needless suffering,” the truth or falsity of 6 hinges on one criterion only: whether or not suffering entails a higher good. For the sake of argument, I will grant the fallacious premises.
Peter clarifies 1:
If someone gets more benefit from the suffering than they suffer, then the suffering is not needless — instead it exists because of a higher good.
On that logic, if tokens are the primary criterion and tokens of joy outnumber tokens of suffering, the scale tips to the “higher good” side and the argument seems defeated. Conversely, if tokens of suffering outnumber tokens of joy, the scale tips to the “needless suffering” side and the argument seems intact. This is an evaluation of net suffering vs. net joy.
If agents are the primary criterion, as long as agents experiencing joy outnumber agents experiencing suffering, the scale tips to the “higher good” side and the argument seems defeated. Conversely, as long as agents experiencing suffering outnumber agents experiencing joy, the scale tips to the “needless suffering” side and the argument seems intact. This is an evaluation of net sufferers vs. net jubilants [with jubilant being used atypically as a noun denoting those who experience joy; I’m open to a better word if you can think of one].
It seems safe to say that by any conceivable unit of measure, a sufferer who inherits eternal paradise “gets more benefit” in comparison to their temporary suffering. Therefore, no sufferer who inherits eternal paradise experiences needless suffering [according to Peter’s definition]. Consequently, if universalism is true, the argument is irrevocably defeated whether we evaluate by tokens or agents. If all agents eventually inherit infinite tokens of joy, then finite tokens of suffering entail a higher good and the discussion ends there.
What if exclusivism is true? What if only a minority of sufferers inherit eternal paradise? Is a minority in eternal paradise “worth” the majority who aren’t? The answers vary according to whether we evaluate E/A or E/ET. They also vary according to whether we use tokens or agents as the primary criterion.
If E/A is true, and tokens are the primary criterion, the argument seems defeated again, even if only one person inherits eternal paradise. After all, no matter how many tokens are experienced before God annihilates sufferers, that number is necessarily finite and therefore less than the infinite number of tokens experienced by the one who inherited eternal paradise. The discussion ends there.
However, if E/A is true and agents are the primary criterion, things start to get a little fuzzy. I defined exclusivism as the ontology in which a minority inherit eternal paradise, and the majority suffer either infinitely or finitely [finitely in this case]. Is 10 agents experiencing infinite joy “better than” 100 agents experiencing finite suffering before annihilation? I think so, but unfortunately, that seems to reduce to a subjective value judgment. Can anyone think of a more objective way to judge this case?
Moving along, if E/ET is true, tokens of suffering and tokens of joy are both infinite. This seems to force an evaluation of agents, and unfortunately, this progression also seems to lead inevitably towards a subjective value judgment. Is 10 agents experiencing infinite joy “better than” 100 agents experiencing infinite suffering? I don’t think so, but at the same time, I can’t deny the possibility. God might think so, but I’m skeptical, and I lean towards calling that a loss. IOW, the argument would seem provisionally intact [presuming we grant the fallacious premises].
So, to recap: this argument contains fallacious premises, and even if we grant them, U undeniably falsifies 6 and defeats the argument as currently stated. Evaluated by tokens, E/A also undeniably falsifies 6 and defeats the argument as currently stated. Evaluated by agents, E/A seems strong enough to provisionally defeat the argument, but I would not say it undeniably does so because one might object to the subjective value judgment [that 10 agents experiencing infinite joy is “better than” 100 agents experiencing finite suffering before annihilation]. We cannot reliably evaluate E/ET by tokens, but an evaluation by agents seems to leave the argument provisionally intact [presuming we grant the fallacious premises and accept the subjective value judgment]. At best, this argument contains fallacious premises and only makes a provisional case against E/A and E/ET when evaluated by agents. At worst, it can’t even get off the ground [because it contains fallacious premises].
Where might we go from here?