Philosopher James Sterba just came out with a book of that title, arguing for the conclusion that a good God is logically impossible, given present observations. At the same time, Michael Shermer and Brian Huffling published in Skeptic Magazine a closing exchange in reaction to their debate on the same question.

Both are very interesting reads; Sterba’s even quite useful, if you want to hone your ability to explain why the Argument from Evil cannot be dismissed with any of the tactics delusional believers attempt in order to avoid admitting to themselves the monstrosity of their imagined god. His every chapter is thoroughly argued and has timely and detailed bibliographies, making it an excellent reference and guide.

I just debated this subject quite extensively myself with Dr. Wallace Marshall, an acolyte of William Lane Craig’s Reasonable Faith community (see that debate’s index, under my “Argument from Indifference” against God; though also somewhat pertinent was our exchange on the “Moral Argument” for God). That debate also contains models and ideas you can benefit from in making the point clear. But here I will add what didn’t come up there.

Defining the Problem

The Problem of Evil has two forms, the Logical and the Evidential (a.k.a. “Inductive,” “Empirical,” or “Probabilistic”). The Internet Encyclopedia of Philosophy and the Stanford Encyclopedia of Philosophy online both have detailed discussions of the Logical form. Until Sterba’s book, the best and latest defenses of the Logical Argument from Evil have been collected in The Impossibility of God (still a worthy read). The form most usually argued is, however, the Evidential. Because it is the stronger argument and much harder to evade with any excuse.

The distinction is important to know, because the most common rhetorical tactic Christian apologists use in the face of the Argument from Evil is to pretend you made a Logical argument when in fact you made an Evidential one, and then “refute” the Logical argument, and claim to have refuted you. For example, every time they argue some excuse is merely possible, they are only rebutting the Logical Argument from Evil. To rebut the Evidential Argument requires proving the excuse is probable. Otherwise, there is no avoiding the conclusion that God is at least as improbable as that excuse. Often the excuses proposed are quite improbable indeed; certainly never is any evidence presented that would render any excuse likely.

In fact, quite usually, no excuse is even proposed—because literally none can be imagined! Theists will simply insist “there could be an excuse, even though we can’t think of it” (and “we just can’t think of it because we’re limited mortals” or something). Which actually amounts to admitting the probability of there being such an excuse is extremely low; for were it at all probable, we’d have thought of it by now. Humans are, after all, the same species who discovered Game Theory, Set Theory, Relativity, Quantum Mechanics, and the Standard Model. Humanity is no dunce.

So this is how the delusion game gets played. And it’s sometimes played dishonestly, as I showed with Timothy Keller; or illogically, as I showed with Wallace Marshall. But this really just reinforces the conclusion that there is no Logical Argument from Evil yet revealed, any more than there is a Logical Argument against the Supernatural—however much we might suspect there is one, and even have good reasons to suspect that, we still have no deductive proof of it. “Proof” here meaning an actual formal syllogism, none of whose premises can be false and whose conclusion validly follows—rather than “proof” in the more colloquial sense of “enough evidence to warrant a conclusion.” The latter is an empirical, not a logical, proof.

So as clever and useful I find Sterba’s argument to be, I don’t see it providing a logical proof of the impossibility of a good God. And I mean that even apart from the arbitrariness with which one can define “good.” That problem is surmountable by requiring any consistent belief structure: if a theist would accept something as good or bad coming from a human, he must accept it as good or bad coming from a god. As good is good, and bad is bad. If you wish to redefine god as “good” in a different sense from how you define everyone and everything else, then you are just playing a semantic game to conceal the real truth, a truth entailed even by your own belief system: that your god is a monster—and you just want to avoid admitting that by changing what words mean. But, alas, you can’t change reality by changing what you call it.

So, for example, for all of Sterba’s elegant reasoning, one can always say of a key premise in his arguments, such as that God cannot have an overriding excuse for allowing (or indeed even making possible) some particular example of evil, that that is never known to 100% certainty. However small the probability, there is still some nonzero probability we’re mistaken, that we’ve overlooked an excuse, some set of circumstances that would indeed warrant the behavior observed even from a good person, and at the same time remain compatible with god’s other necessary attributes, such as omnipotence and omniscience.

And there is always room to move there as well: maybe God is not perfectly omniscient or omnipotent, but just powerful and cognizant enough to still rate as a god, and his limitations are part of his excuse—although we still cannot honestly think of such an excuse even on that assumption. And indeed we can’t, for as Sterba notes when addressing this very point, a god must always be by definition more powerful and cognizant than humans, just to be a god, and yet humans exhibit better and wiser behaviors than god must be evincing, and we face no excuses impeding us—so how can we, the weaker species, be free of impediments holding back even a god? Such would imply humans are more powerful or wiser than god, which negates any claim to his being a god. Thus, we’re back to logical impossibility.

So there is a logical conundrum here. But it only establishes the extreme improbability of a solution. Something we’ve overlooked that would qualify as a valid excuse remains very improbable; even absurdly improbable. Indeed unfathomable—for how can having more power make a good person less able to act? What could be stopping them? And if something’s stopping them, how can they then claim to be the supreme being—or even a god at all? Wouldn’t that giant juggernaut completely hand-tying them then be the actual god, and what we are calling God its mere quivering bitch? I confess this seems logically impossible to me.

But we haven’t proved that. Not formally. So it’s still “possible”—in an absurdly, unfathomably remote sense. It’s a vanishingly small probability to be sure. But that’s why I’ve yet to see any Logical Argument from Evil that wasn’t really just another Inductive Argument from Evil. But Sterba does prove it’s pretty damned close. One can bite the bullet and admit God is not good, but you still have no evidence even such a god as that exists. And since the evidence for a good god is catastrophically weaker than for an indifferent one, you are thoroughly empirically screwed if you want to honestly maintain any rational belief in such a god.

The Sterba Case

One elegant feature of Sterba’s cumulative argument is that it does wall off a number of common rhetorical escapes. For instance, you can’t argue that maybe we are mistaken in how much evil we observe, because a good god by definition would not allow us to be; just as a good god by definition would inform us of the moral reason for all the evil we do observe. Thus Sterba makes sure the escapes are whittled down to just one and the same thing every time: the need of an excuse for god to act apparently immorally, when no similarly situated good person would ever do so.

But there of course is where I see the flaw in Sterba’s argument, the flaw that every time converts his purportedly logical argument into an inductive one. Just as I observed that Kant’s every attempt to justify a non-hypothetical imperative just always ends up covertly converting it into just another hypothetical imperative, Sterba’s logical argument ends up just being an inductive one. Albeit an incredibly strong one.

Indeed, as such, Sterba topples the most common attempts to propose a justifying excuse for God, such as Plantinga’s “free will defense.” And that makes his book quite useful, even just for advancing a thoroughly robust inductive argument—as Sterba definitely shows Plantinga’s “excuse” is extremely improbable, even self-contradictory. We have, in fact, very little reason to believe it would ever justify god; as the same argument does not justify anyone else. Quite simply, God cannot act differently than any other good people would in exactly the same circumstances and still qualify as being good himself. And that’s where attempts to defend God all fall apart.

There Is No Free Will Defense

The important contribution Sterba makes to the “Free-Will Defense” debate is to articulate clearly how self-refuting it really is. Allowing freedom actually consequentially results in greatly reducing it.

For example, allowing tyrants freedom to take power and oppress a population results in a vast net reduction in freedom. If physically restraining me from doing my will violates my free will, then allowing someone to physically restrain me violates my free will. God therefore cannot use “free will” as an excuse not to intervene. To the contrary, if “free will” is really so supremely important, that fact would compel him to intervene a great deal more than he is observed to do, specifically in order to maximize the abundance of free will.

After all, a free society contains more unrestrained free will than an oppressed one, obviously. That one must violate some free will to maintain a free society is a logically necessary fact (e.g. tyrants, killers, slavers must be thwarted in their designs); but if the goal is to reduce violations of free will, a good ruler must do this. Not doing so has the opposite result—and thus cannot be justified by any principle on which maximizing free will is the greater good.

One could extend this even to natural evils—which the Free-Will Defense has never plausibly accounted for. Indeed Sterba has a closing chapter on natural evil that addresses the difficulties it adds to theism. Merely among which are the fact that such evils produce violations of free will on a massive scale: killing people without their consent certainly violates their free will; as does inflicting them with disease; crippling someone without their consent—especially mentally but more obviously also when physically—certainly impedes their freedoms, no differently than binding them in chains or beating them with sticks; taking away one’s hard-earned property by disaster leaves them less free than people who were able to avoid that same fate by having access to protective resources; and so on. If a good God really would value “free will” above all else, then we would not see the world around us that we do—directly contrary to Plantinga’s thesis. His excuse is thus self-refuting. It leads not to a justification of god, but to an even more glaring logical contradiction between his god as-defined and what we observe.

Sterba demonstrates this result holds even for moral evil. Thus his chapter on this is aptly titled, “There Is No Free-Will Defense.”

Sterba further points out that thwarting the effects of immoral acts does not even qualify as violating anyone’s free will. Making me invulnerable to disease with a vaccine does not violate my free will. Just as making me invulnerable to a murderer’s bullets would not. Which means God’s just making us all immune to disease wouldn’t violate anyone’s free will, and thus his failure to do so cannot be justified. But it would not even violate a murderer’s will to make me as immune to bullets as to disease—any more than putting me in a bullet proof vest would. No one would dare argue bullet proof vests are immoral and should be banned because they violate the free will of gunmen who wish to shoot at people. They can shoot at them all they want. Their will is free. Their evil is also thwarted. The Free-Will Defense is thus defunct. It can no more justify God, than it would justify you preventing anyone wearing a bullet proof vest. Indeed your doing so would make you evil, not good. Plantinga’s argument thus does likewise to God.

The perverse flaw in Platinga’s Free-Will Defense is that it entails a nightmarish anarcho-libertarian worldview wherein if I have the physical means to kill or abuse you, I therefore in effect have the moral right to do so, because no one ought take any measure to stop me—in fact it would be their doing so that’s evil, not the act of a good person. Because maintaining my freedom is a greater good than thwarting my victim’s fate. But this flips around the entire meaning of good and evil, turning every act we deem good, evil; and every act we deem evil, good.

Plantinga’s Free-Will Defense is thus another semantic trick: he actually is defining his god as a horrific monster, and simply “chooses” to call being a horrific monster “good.” All just to avoid admitting to himself that he admires a horrific monster. Placing freedom above justice quite simply is evil. For evil is by definition the absence of justice. Just as placing freedom above respecting people’s well being is the absence of goodness, precisely because respecting people’s well being is what goodness means. Freedom is only good when thwarting it would do more harm than exercising it, thus we thwart murder but not speech, rape but not consent, theft but not commerce, and so on; recognizing this distinction is literally what goodness and justice are.

Sterba thus shows that Plantinga’s god as-defined cannot be good. It is a logical impossibility. Therefore one must abandon that definition and retreat to some other conception of a good god, for there to be any logical possibility of one. (See also Laura Ekstrom’s God, Suffering, and the Value of Free Will.)

And Other Escapes Barred

Sterba also goes on to show the same for “soul-building” theodicies (finding that any conceivable model of it is logically incompatible with observations); and also for “no intervening agency” theodicies (finding that a good god could and would have then solved all problems of evil at the instant of creation—e.g. by not even making death or disease a thing); and also for traditional “redemptive history” theodicies (finding them all to be internally incoherent for a good god on present observations); and also for “doing-evil-for-good” theodicies.

The latter is a doctrine Sterba cleverly notes Romans 3:8 condemns; and his analysis shows that exceptions to the rule “never do evil that good result” only stem from lack of power, an excuse god by definition never sufficiently has. Sterba thereby demonstrates the logical impossibility of this excuse too. Sterba likewise soundly refutes arguments, like Wallace Marshall’s, that a god who intervened at all would intervene “too much.” Because a good god by definition would not do that; but would still by definition do more than we observe. So that excuse is illogical.

And when it comes to Skeptical Theism (see Stanford and IEP), Sterba soundly dispatches it by showing it violates a basic moral principle of “informed consent,” and thereby, again, requires god to be not a good person. A good god would seek our informed consent before treating us so callously as he in fact must be given observed facts.

Allowing every evil, and explaining none, is simply not logically compatible with a good person, certainly not one who has no practical limits to their knowledge and power (and thus, a god). Nothing prevents a god from informing us; nor prevents them asking our consent to the process they find necessary before subjecting us to it. And no good god would create beings incapable of the requisite understanding for this either; that would itself be evil, akin to deliberately lobotomizing your children so they never grow up and can never have competence to consent to anything you do to them or allow to happen to them. Skeptical Theism therefore cannot restore a logical compatibility between a good god and observed evil.

There still remains, to my mind, a logical possibility of some other inconceivable excuse for God. One that must prevent God from informing us, seeking first our consent, or even making us capable of informed consent; as well as justify god’s allowing every horrific evil, natural and moral, to proceed in all its terrible effects; and somehow still strangle God in all these ways despite God also being all powerful (!). I agree that after Sterba’s survey we can conclude no such excuse has ever been produced—all the contenders are in fact logically incompatible with either a good god or observed facts, or both. And it certainly seems impossible any such excuse could ever be contrived. But that no one has put forward an excuse that survives Sterba’s analysis is not logically equivalent to there being no such excuse. It does render the probability of there being one absurdly low, however. And thus so also is the probability of a good god absurdly low. No one is therefore rationally justified in believing there is one.

The Shermer-Huffling Exchange

In the exchange in Skeptic Magazine, Shermer’s entry hardly requires additional comment (Skeptic 24.2: 42-48). It’s a decent boiler-plate defense of the Argument from Evil. What we can learn the most from is how Huffling attempts to defend his god in response to it (Skeptic 24.2: 49-54). But both lack the same feature: a sound use of the logic of probability to illustrate their points. Such would expose the errors of Huffling and shore up the points of Shermer. Both could thus benefit from knowing the relevance and power of one simple truth: prior odds times relative likelihood equals final odds. In other words, Bayes’ Theorem.

The prior odds are of course also a problem for theism, empirically and analytically: background evidence, having only ever found physical explanations for everything, does not make things like gods at all a probable explanation of anything; and the probability-space occupied by God, indeed especially a hyper-specific one like the Christian God, must necessarily be extremely small—as many far simpler things, as well as a vast range of alternatives, share most of the rest of that space in the absence of specific evidence otherwise.

But we are here talking about a set of evidence, and thus the likelihood of that evidence, on the competing alternatives of “there is a good God” and “there is not a good God.” Whatever the prior odds are, the relative likelihoods are a separate question. And here, the question is, what is the likelihood (the probability) of the evils we observe—both natural and moral; and both historically contingent and innate to the universe’s (and our) very design. And in particular, what is that likelihood on “there is a good God” and what is that likelihood on “there is not a good God.” The ratio between those two probabilities is the relative likelihood of the evidence.

So, for example, if the evils we observe are twice as likely when God does not exist than when he does, then the prior odds against God are thereby doubled. If those evils are a billion times as likely without a God than with one, the prior odds against are likewise magnified by a billion. When looking at the evils we see, and have record of, and look at the means a god must by definition have, and what a good person would not fail to do with those same resources qua being good, we cannot but conclude the odds against a good God existing are countless trillions to one against. Even by the most generous conceivable estimate.

Because we’ve seen trillions of acts of good people by now; and never do they have an excuse to say and do nothing, when the means to do something is available to them; and the means available to a god are vastly greater than could even be marshaled by all the human race; vastly more so, a single individual. The obstacles to good people acting are far greater than to God; yet good people routinely speak and act for goodness and justice. If even they are not impeded, God certainly cannot be. Ergo, for something to exist that completely and thoroughly cripples God but somehow manages to be overcome and outdone by millions of human beings, trillions of times, would require something so extraordinarily rare, and thus so extraordinarily improbable, that we have not witnessed it even once in countless trillions of trials.

Worse, not only have we yet to encounter any such excuse, we cannot even conceive of what it could be—which makes what it would have to be even more bizarre, even rarer, and thus even more improbable. For the number of things humans can imagine getting in their way is vastly greater than the number of things they’ve actually overcome. And we are mere humans—it therefore must necessarily be vastly less likely any obstacle could impede a god. Stack the odds. They plummet to near zero. The relative likelihood of a good god given observed evils is therefore not just small, it’s literally inconceivably small.

This becomes more obvious when you look at the converse evidence condition: a world in which a god exists and actually does act like a good person. We’d then see routine divine messaging and explanation when sought, and routine acts of help and aid and good will from on high. We’d see better, smarter design of our minds and bodies and world. The evidence a god existed and was good would then be overwhelming.

Think of how much such evidence would increase the probability of such a god being real. Surely it would make a good god’s existence trillions of times more likely. Now realize the mathematical consequence of that: as that evidence is removed in our world, the probability of such a god existing in our universe must be reduced by as much as it is in that other world increased. So if it is trillions of times more likely in that world, it is logically necessarily trillions of times less likely in our world.

There is no rational escape from this conclusion.

Science and Evidence

Huffling starts with contradictory claims about science. At one point admitting science can study God though his effects, but then claiming instead that science can’t study God through his effects, asserting “we still couldn’t measure God as the cause” of anything in the world.

Huffling thus proposes the following:

If science is indeed a study of the natural world by means of the senses, then by definition it is locked into studying the physical universe but not anything that is not part of the physical universe.

That’s entirely false. Science can study anything that affects or has ever affected the physical universe. Including anything and everything supernatural there might happen to be (as I demonstrated long ago in Defining the Supernatural). As with atoms, photons, gravity, magnetic fields, the Big Bang, science routinely studies things we cannot see by studying their effects. And it is logically impossible that God has no effects on the universe. To claim that would be to claim that he did not cause or design it, and that he has never done anything in it—like never inspired a prophet, never performed a miracle, never answered a prayer, or even worse for Huffling, never become incarnate.

Worse, it would be logically contradictory to claim science cannot study the complete inaction of God. That God does not act is a scientific observation. It has consequences. It tells us something about God—should there be one. Indeed it can even tell us something about whether there is one. Inaction is an observation. An observation of the senses. That’s evidence. Conclusions follow. What is the probability of a good God being a total do-nothing no show—not even having created or arranged anything about the design or contents of the universe, nor ever once observably acting in aid of justice or compassion? Absurdly low. Therefore, so is the probability of there being a good God absurdly low. Q.E.D.

If a thing has no effects whatever on the universe, and has never had any effects on the universe, its existence and properties are literally unknowable. Belief in such can never be warranted. Least of all the more elaborately absurd a thing you then define it to be—as the more you propose is true of it, the less probable it becomes, by inevitable consequence of the logical division of the available possibility-space. That there is an unknowable superman, for example, is always vastly less probable from the moment you utter the proposal than that there is an unknowable rock or atom or vacuum. Adding the properties that this superman is ethereal, yet has all possible powers, and all possible knowledge, and designed and caused this universe, and is a good person, and yet is so thoroughly crippled by some other unknown and unfathomable cosmic force as to be completely incapable of acting to facilitate justice and goodness in the very universe he designed and governs, only reduces his probability more and more and more and more and more and more and more and more.

See my discussion of this fact in respect to Cartesian Demons, for example; and of gods in particular; and the supernatural generally. It’s even worse if you want to define this superman as the Christian God, as that entails adding even more ad hoc claims, to the point that, as John Loftus rightly says in a whole chapter of the very title, “Christianity Is Wildly Improbable,” in The End of Christianity.

So if you want to warrant belief in this extraordinarily remarkable thing, you need evidence. In fact a lot of it. So that there is none is rather decisive empirical proof that this particular superman’s existence is not merely unknowable, but absurdly improbable. And that’s as sure a conclusion as anything science can yet tell us about the cosmos. It’s most sure of all with regard to that one vexing attribute: being a good person. You’d have far better luck defending a horrifically indifferent superman. Though still as a remarkably-propertied superman void of evidence, even that’s still effectively impossible to rationally believe.

Huffling wants to have it both ways, of course, claiming we do have evidence, while claiming science can never verify any of that evidence or what causes he proposes for it. This is irrational. If it cannot be verified by any reliable process, the conclusion you are claiming cannot be thus verified either. There is no path to knowledge here. Not of God. Not of anything. You may as well claim everything you believe on no evidence is a real miracle was caused by sorcery, faeries, aliens, or naturally human “psychic powers.”

The rest of us, who only base beliefs on evidence, conclude more rationally that there are no real miracles. Because we observe every claim to such that could be investigated has turned up false. Exaggerated and fake tales, misreported observations, completely natural and calculatedly expected coincidences, are all vastly more probable causes of “miracle reports” than “God did it.” If the latter were ever true, we’d have confirmed some cases by now. Particularly if God is “good.” Precisely because a good person of unlimited power and knowledge could not fail to have acted far more than we observe. The evidence has thus instead gone the other way. Rational people follow the evidence.

Sterba’s point carries here. Huffling is presenting us with a logically contradictory thesis: that a God exists who does intervene in the world (performing miracles, communicating with people, designing the natural world), that this God is a good person, and that this God does nothing a similarly empowered good person would do. Those three properties cannot logically co-exist. Unless there is some other unknown god-crushing force out there, some property of existence even God cannot overpower or remove, that completely and thoroughly cripples God’s power to speak and act, even to the point that he speaks and acts less than a kind but mentally and physically enfeebled human being does—and thus is not a god by any definition. It’s extremely hard to even come up with a logically possible God in all this; and the effort would entail a God so improbable no rational person should ever believe in it.

Trying to Define “Good” Out of Existence

Huffling tries to define “good” out of any empirical existence so as to avoid the consequences of his imagined God never doing any. To this end he declares “when we call something good we are referring to something beyond its physical properties,” and therefore we cannot discern empirically where it does or does not exist. This is false. “Good” can only ever be defined in terms of effects—which are observable. If I am a good person, that entails empirically observable facts about me—about how I think and act. This is testable and measurable. Even by science. Ditto God.

Indeed, even what we should call “good” is empirically discoverable, as it is entirely the empirically measurable consequence of what people want for themselves and their lives. What no one wants, and would never want no matter how rational and informed they are, is never meaningfully definable as “good.” Rather, the only possible, useful definition of “good” is precisely only one thing: what all rational and informed persons would want, what they would prefer over the contrary. Which is an empirical object of scientific study, in psychology, sociology, and anthropology. (See my thorough defense of this point in my debate with Wallace Marshall and articles linked therein.)

Thus, for example, harming someone vs. aiding them is empirically observable; thus so is failing to come to someone’s aid. Science certainly can test the existence and non-existence of entities aiding or harming people or neglecting to aid them. The absence of action is an observable, empirical, documentable fact. As is what people want most when at their most rational and informed, and thus what the most rational and informed people recognize as good and thus wish and need there to be more of.

Instead, Huffling tries to pervert the very definition of “good” by essentially identifying even evil as good, declaring “existence itself is good.” False. Evil’s existing does not make it good; existing is not a good-making property of evil. To the contraty: evil existing is precisely what is not good about it. Cancer is good because it exists? Bullshit. It is not good because it exists. It lacks the property of goodness by existing. Only its not existing would be a good-making property. “That only exists in fiction” would be good. “It actually exists” will never be so.

Such attempts to redefine evil as good typify delusional theists. Rational people are not fooled. Changing what you call it, doesn’t change what it is. Evil is evil. And it is evil because it exists. It acquires no goodness by existing. Only that which is good is good. So you can’t escape the fact that you’re actually defining your god as evil by verbal trickery. You’re just irrationally excusing your admiration for evil as somehow “good.” It’s not.

Huffling attempts the same trickery again when he tries to argue God is not a “moral being” but is still “good” (p. 52). He is simply switching around what we call what, pledging belief in a completely amoral monster and, like Plantinga, simply “choosing” to call it “good.” When all the while, what he is describing in actual fact is indisputably evil. This is all a game. A game that can only fool an irrational man. Reasonable people aren’t taken in by semantic trickery like this. They know a thing cannot be changed in what it really is by the perverse device of merely changing what it’s called.

And We Come Full Circle

Huffling’s entire tack—apart from his delusional irrationality about science, evidence, and identifying goodness in the world—indeed collapses under a standard fallacy of possibiliter ergo probabiliter: God’s goodness remains possible despite all evidence to the contrary, “therefore” we should conclude it’s probable. I identify how that line of argument is both common and illogical in Proving History (index, “possibility fallacy”). The fact is, the possibility Huffling is desperately clinging to is absurdly improbable. And that’s why his belief in it is irrational.

Eventually Huffling says Shermer was presenting the Logical Argument from Evil and so all Huffling has to do is argue for a possibility of an excuse. But that’s not true. Possibility doesn’t get you to probability; and so it doesn’t get you to any warranted belief. Shermer wins if all you have is an absurdly improbable “possibility” of something being overlooked. Belief in a good God simply cannot be rational in that case. No matter how much you want the absurdly improbable to be true, it remains absurdly improbable. And so therefore does your wished-for God. It is irrational to conclude otherwise.

Shermer’s argument also wasn’t really the Logical Argument. It was thoroughly empirical; that’s why he kept mentioning science and facts and probability. He did not argue we can be certain God has no excuse; he argued there is no reason to believe any such excuse exists. That’s an inductive argument. As on p. 46: “From my perspective, if you don’t know” why God allows evil then “you don’t have a case,” meaning you don’t have a case for there being a God who is good. Huffling confused the two arguments for the same reason maybe Shermer did: because they don’t realize all Logical arguments from evil are just disguised Empirical arguments from evil. As I explained earlier.

So Huffling errs repeatedly here, ever-leaning on the same fallacy, irrationally thinking all he needs to warrant believing a good God exists is the mere remotest possibility that one does. And then, satisfied by that complete failure of logic, he just hand-waves away the Inductive Argument (on p. 51), calling it “weaker,” when in fact it’s far stronger, precisely because it does not require the elusive formal proof the Logical form does. It rather rests on vast reams of irrefutable evidence. Indeed it’s hardly possible to find a stronger argument for any conclusion anywhere.

Conclusion

Science can test any model that will have observable effects on the world. And any system can be described with a model. God plus a universe constitutes a single system. We can therefore model that system—of a universe with a god making, sustaining, and interacting with it. We can then play out the model, and thus predict what will happen in that model, and thus what will be observed in any system described by that model.

It is therefore not, as Huffling falsely asserts, a question of what a good God “should” do. It’s a simple question of what a good God would do, by any consistent definition of “good.” We can model that. And that model entails predictions. If God is not a person, he is not a god. And if he is not a good person, he is not good. So we have our model. Give God the mind of any good person, and the means and resources of a god. The outcome is reasonably predictable.

Lo, observation simply doesn’t fit what Huffling’s model predicts. And when he confessed to Shermer that he “doesn’t know” why God allows evil, he is admitting that neither he nor anyone on earth, in the whole of human history, can even conceive of any possible thing we could put in that model that would make it fit observation—other than removing God’s goodness or omnipotence, which means: either removing a god from the model, or admitting the only model that fits observation is one with a god devoid of any measurable goodness.

And that’s why the Logical Argument from Evil is compelling. Not because it comes with a deductive proof that nothing can fix the model; but because nothing we can think of yet fixes the model; and worse, as a matter of well-established empirical and analytical fact, the only thing that could fix it would have to be so bizarre, so inconceivable, so unimaginably strange, that it would have to be one of the most improbable things ever thought of or encountered by humankind. Therefore so is Huffling’s God. And it’s simply not rational to believe in anything so improbable as that.

-:-

My formal critique of Sterba’s case is now available (among several others in the same and adjoining issues): Richard Carrier, “Has James Sterba Established a Logical Argument from Evil or Just a Very Good Evidential One?” in Religions 14.3.307 (2023).