Has James Sterba Established a Logical Argument from Evil or Just a Very Good Evidential One?

Abstract

James Sterba’s new treatise advancing a logical argument from evil against the existence of God fails in one respect and succeeds in another. As with all claimants to having found such a thing before him, Sterba fails in properly achieving a logical argument from evil. But he succeeds in producing one of the most undefeatable evidential arguments from evil yet published. Elegantly dispatching all the common defenses, Sterba shows that there is no way to avoid the force of his argument against the existence of God without adopting extraordinarily improbable hypotheses that theists can’t even intelligibly articulate.

In Is a Good God Logically Possible? (Palgrave Macmillan 2019), James Sterba strives to establish a logical argument from evil against the existence of a (good) god and determinedly rebuts the likes of Marilyn Adams, Alvin Plantinga, Michael Bergmann, and Brian Davies, providing throughout a very useful bibliography.1 In the process, Sterba maintains that his is not merely an air-tight “evidential” argument from evil, but a bona fide “logical” argument from evil. I do not believe this is the case. I do believe, however, that his book establishes beyond a reasonable doubt that a good god, in any meaningful sense, cannot exist on our current observations and information. However, that conclusion depends on “our current observations and information”, any of which has a nonzero probability of being false (there may be facts that have escaped our observation, or elements of our information might be incorrect), which makes his argument an evidential one, not a logical one. Yes, if our information is complete and correct, then a good god cannot logically exist. That much I believe Sterba demonstrates. But this logical impossibility is contingent on premises whose truth is not known to a logical certainty. Those premises depend on evidence, and that renders the whole argument evidential, which form of the argument is also sometimes referred to as “Inductive”, “Empirical”, or “Probabilistic”, to distinguish it from the strictly “Deductive.”2 Regardless, Sterba’s book has touched a nerve: in just the four years since its publication there have been over twenty-five attempts to rebut it.3

There are, of course, a number of semantic labyrinths that theists can try to use as an escape here (and they can be found repeated across many of the responses to Sterba that have been published so far), such as redefining the word “good” so that even the most monstrous of sociopaths would qualify. Sterba mostly avoids such trickery by relying on the theist’s own internal logic: rather than try to “settle” a definition of “good” on one thing or another, Sterba correctly maintains that theists must mean by “good” when applied to God the same thing they mean of any person, or else they are affirming a contradiction. Since contradictory states of affairs cannot exist, their God then becomes implicated in a contradiction, and thereby becomes logically impossible, which is a victory no theist can allow. To extract God from this tragic fate requires resolving the contradiction, and there are only two ways to do this: admit to the principle (God is only good if he is good in the same sense anyone can ever be good, and thus is good according to the believer’s own standards of goodness as they apply to all persons of whatever their knowledge and power, such that even any situationally different ethics deriving from greater knowledge and power must still apply equally to any person as to any god), or admit that evil people are good (and thus abandon the theist’s entire ability to determine right from wrong).

Sterba’s point is that if God can commit mass murder and genocide, afflict germ warfare against masses of innocent children, endorse slavery, be a material accomplice to rape, order terrorist acts and contract hits on innocent people, build environments he knows will collapse and kill their residents, fill the world with pollution he knows will afflict widespread serious harm, and the like, and still be called “good”, then these behaviors, these choices, are thereby being endorsed as good, and anyone who conducts them is then “good” by the theist’s own construction. This is, of course, too terrifying and censurable a course to take for the theist, as it would immediately establish them as no mere villains, but as the enemies of all human society. They would be more honest, then, to turn to worshiping the gods of H.P. Lovecraft. In the same fashion, redefining “God” so as to deprive that entity of greater knowledge and power than even humans collectively possess would simply negate any legitimate purpose to employing the word “God”; and anyway, this is not a popular recourse for theists either.

Accordingly, usually different defenses for God are deployed than these, which actually appeal to the same defenses anyone could appeal to in the same circumstances, such as “necessity” or “unavoidability” or some form of “greater good” logic. This amounts to accepting Sterba’s principle that if it is a defense for God, it must be a defense for any person in relevantly similar circumstances.4 But then Sterba dispatches all these more usual defenses of God. After his Introduction (Chapter 1), he conclusively demonstrates that there is no logically coherent “free will” defense (Chapter 2), that the prospects of soul-building or an afterlife do not actually resolve anything as to current evidence of any extant God’s moral character (Chapters 3 and 4), that the theist’s insistence that different moral standards apply to God (again redefining evil as good) is only a covert way of conceding Sterba’s entire argument (Chapters 6 and 7), and that appealing to “natural evil” does not get God off of any hook either (Chapter 8). Sterba then wraps with his Conclusion (Chapter 9).

In the middle of all this (Chapter 5), Sterba addresses a completely different defense: the argument from Skeptical Theism that we cannot claim to know if God is good or evil because God possesses knowledge that we do not (he “sees the big picture”). This is the most important chapter of the book, for here lies the problem with claiming to have established a logical argument from evil rather than an evidential one. Sterba’s other chapters adequately refute all defenses of God; none of them hold up even in respect to logic. So, as far as the rest of the book is concerned, I believe Sterba can claim to have proved a logical contradiction between presently observed states of affairs and any meaningful idea of a “good” God. The only problem is at this very juncture: whether we have all pertinent information, and whether all our information is correct. Sterba deploys good rebuttals here, but they are not as comprehensive as in his other chapters. His main foil is Michael Bergmann, and it can fairly be said that Sterba adequately dispatches all of Bergmann’s arguments. However, there remains a weak version of Bergmann’s case that remains immune to Sterba’s rebuttals: if we take the position that, as Bergmann argues, scenarios we cannot imagine at present are possible, in which we will be mistaken as to any extant God’s moral character but then abandon the inalienable requirement of Bergmann’s argument that any of these scenarios be even remotely probable, then we end up rejecting the conclusions of both Bergmann and Sterba.

In other words, because it remains logically possibly the case that some Bergmann scenarios still exist, Sterba cannot establish a logical incompatibility between current observations and a good God; however, because none of those scenarios are even remotely probable, we are in no way justified in believing any obtain. As a result, this does not rescue God as Bergmann hoped, but rather establishes the extraordinary improbability of his existence, which warrants our abandoning belief in God. Hence, while taking this “out” rescues us from the horn of Sterba’s logical argument from evil, it throws us on the horn of an evidential argument so powerful that we still ought to abandon belief in any such God.

Skeptical Theism rests on the fallacy of “possibiliter ergo probabiliter”, meaning “possibly, therefore probably”. But “possibly” does not get you to “probably”. Therefore, it cannot get you to “probably there is a good God”. To the contrary, Sterba’s demonstration of the extreme improbability of any Bergmann scenario being true entails the opposite conclusion: “that there is a good God is extremely improbable”. This is no mere agnosticism but an accomplished proof that such a God’s existence actually is extremely improbable based on present evidence. Sterba has composed a very good evidential argument from evil—arguably a decisive one—but it still technically is not a logical argument from evil. It is not logically necessarily the case that a good God does not exist; it is just extraordinarily improbable that one does.

This is still an important achievement. Take, for example, the fact that the Bible (Old Testament and New) universally endorses slavery (indeed, in the Old Testament, even outright sex slavery).5 No one can produce any reason why a moral person would allow that to happen when they have every means to ensure it does not. God can simply tell every living soul “that’s not my book”, or literally change every such reference back to condemning slavery every time any scribe attempted to alter God’s Book. Either way, a God can ensure their will is accurately represented without violating anyone’s liberties (because no one is at liberty to commit fraud, nor can any moral society exist that allows it without even remark, much less redress). What is the probability that a benevolent God would have a valid excuse not to say one correct heavenly word to his devoted believers about this? And what evidence do we have supporting that probability? From our background knowledge of benevolent beings (billions of humans) with the power to speak unharmed (millions of humans), a valid excuse not to speak up is so rare that we never see one single instance of it. And “rare” is just a synonym of “very infrequent”, and “very infrequent” is just a synonym of “very improbable”. It is at least, in fact, millions to one against. There is no way to turn this probability around. Theists simply have no evidence that such excuses as would here be needed are any more likely than that.

So, theists like Bergmann will insist, “there could be an excuse for being totally silent about this, even though we can’t think of it”, and even propose, “none of us can think of it because we are all limited mortals”, or something to these effects. But this does not respond to the point. To the contrary, it amounts to admitting that the probability of there being such an excuse is extremely low, for were it at all probable, we would have thought of it by now. Humans are, after all, the same species who discovered Game Theory, Set Theory, Relativity, Evolution, and Quantum Mechanics. Humanity is no dunce. And we openly denounce slavery all the time without moral impediment, so how can humans have more power and wisdom in this matter than God? Yes, however small the probability, there is still some nonzero probability we are mistaken here, that we have overlooked an excuse, some set of circumstances that would indeed warrant a good person of godlike knowledge and means allowing slavery to be universally endorsed in their name for thousands of years and never condemning it in any communication from them whatever. But it is plain to see that that is extraordinarily unlikely, particularly for a God. So, if no one can come up with a reason—even in concept, much less adduce any evidence that that reason is even true for God—it cannot be claimed that “probably there is a reason”. To the contrary, this failure is stalwart evidence that probably there is not.

The probability of a good God then simply becomes “the probability that a good person with godlike means to tell us they don’t endorse slavery, nevertheless wouldn’t, and would even instead let the world claim they were all for it, for thousands of years”. A mere human in this predicament can claim ignorance, a lack of resources, or “I was dead at the time”. God has no available excuses because all known excuses are born of the limitations that, by definition, do not exist for a god. This is just one example of evident moral failure. Add to this all the other endorsements to crime and terrorism attributed to God in the Bible and all the death, pain, disease, corruption, and unthwarted crime allowed in the world, indeed even caused by its very design, and the probability that an excuse exists for every single one of these seemingly immoral decisions—an excuse not merely to do nothing about any of these things, but to remain utterly silent on the matter in every single case—becomes astronomically compounded well toward zero. This cannot be gainsaid by raising mere possibilities. It is the probability that is at issue.

As Sterba notes when addressing every attempt to invent excuses for God in the face of overwhelming evidence of his crimes (should he even exist to be responsible for them), a god must always be by definition more powerful and cognizant than humans, indeed all of collective humanity, just to be a god; and yet, humans exhibit better and wiser behaviors than God must be evincing in every one of these cases. We who oppose slavery say so and openly oppose anyone who would defraud the public with false claims about our position on it; we who oppose terror and murder and disease and corruption and crime say so and even act to suppress or fix it. Plenty of us face no excuses impeding us, and even those who do face only impediments that a God would not (like being liable to being injured or killed for our troubles, not having enough money, or not knowing what to do). So, how can we, the far weaker and less informed species, be free of impediments holding back even a god? Such would imply that humans are more powerful or wiser than God, which negates any claim to his being a god. This takes us back to logical impossibility. The only escape for the theist is to admit the only possible way a good God can exist is on the supposition of extraordinarily improbable conditions that are nowhere in evidence. This logically entails God’s existence is extraordinarily improbable.

Sterba does wish to deny this. He thinks his dispatch of all of Bergmann’s arguments suffices to render all Bergmann scenarios logically impossible. But nowhere is that step reached in Sterba’s argumentation. He makes a sound case against Bergmann’s insistence that such scenarios are at all probable, but that is not the same thing as establishing them to be impossible. Consider, for example, what Sterba reports to be Bergmann’s premises. For example, on p. 72, Sterba argues that Bergmann’s “ST1” holds that “We have no good reason for thinking that the possible goods we know of are representative of the possible goods there are”. ST2, ST3, and ST4 advance similar assertions from different angles. Sterba refutes the “no good reason” element of all of these premises. We actually do have good reason to believe our understanding is representative. So, all of Bergmann’s arguments fail, but they fail only to rescue God as a probable entity. They remain intact if they are weakened to only admit the possibility rather than the probability of what Bergmann proposes in each case; so, “It is possible that there are goods we don’t know of”, and (which must be added on top of that), “It is possible one or more of those unknown goods justifies God’s total silence and inaction in every case of naturally engineered and unchecked evil across the board”. One can construct the same of unknown evils (T2), unknown entailment relations between goods and evils (T3), and even an unknown “total moral value” of complex sets of affairs (T4). But taking this step renders the existence of God improbable on Sterba’s rebuttal, so Bergmann fails, but so does Sterba—unless you reset Sterba’s conditions of success from “God is logically impossible” to “God is extraordinarily improbable”. Then he succeeds.

Sterba cannot cancel this consequence by proposing reasons why we should conclude that the existence of such unknowns is unlikely to excuse everything about God’s inaction, even his complete silence, even though Sterba is right about the premise. As he notes, at no point does God seek informed consent from anyone for all the horrible things he causes and allows to happen to them, including all the horrible things he allows the authors of the Bible to claim he approves of and opposes, nor does God ever give anyone a morally acceptable answer as to why he is doing all this, nor does he ever help anyone in need, despite that being in consequence of all these things (most of which he is directly responsible for, such as viruses, bacteria, parasites, cancers, genetic disorders, tsunamis, earthquakes, floods, mudslides, volcanoes, freezes and heatwaves, vicious animals, and even susceptibility to harm and mortality itself). That is all true. But Sterba is wrong about the conclusion. Not being able to think of a reason that would ever justify this complete silence is not even functionally equivalent to proving such a reason logically impossible. It is at best equivalent to proving it extraordinarily improbable, and that is simply an evidential, not a logical argument from evil.

Consider the following scenario. It so happens, unbeknownst to us, that it is logically impossible for God to create a paradisiacal world without a concomitant purchase through a particular array of suffering. Accordingly, the reason God cannot undo this feature of existence is that it cannot be undone; no power can logically exist that would overcome it. It so happens that if God alleviates any of that suffering by intervening or even speaking to the persons who, collectively, must pay this price, the effect is at once undone, like touching an electrical current to ground. This, too, unbeknownst to us, is logically necessarily the case, and thus no power of any god can undo it. But God is a good God, so he creates a number of people, as competent adults, and tells them all of this until they fully comprehend it (as, being God, he knows when they have), and he gives them a choice: you can have your memory erased and be born and raised into a world where a certain random amount will be suffered by each, purchasing the balance of karmic energy God needs to secure you in an eternal paradise after, or God can set you in another universe where you will remember all this but endure for as long as you choose a lesser degree of mixed suffering and pleasure, neither horrendous nor wonderful, with no paradisiacal outcome (think “The Medium Place” in the afterlife satire series The Good Place). Again, the reason that this is the best God can otherwise do is that it is, unbeknownst to us, logically impossible for God to create or manage any better outcome in a balanced karmic product, other than through the more brutal but brief scheme to secure a place in paradise. The only catch is that if you say yes to the paradise scheme, your memory of this choice, and in fact of everything whatever, must be erased, and God can never tell you any of this while you grow up from a helpless baby through perhaps an adulthood of random length and go through a brief mortal life of random suffering—because otherwise the suffering cannot earn the purchase. And you are told all of this before choosing.

In this scenario, no one exists who did not give their full informed consent to their fate. Everyone in this world, unbeknownst to them, already fully consented to be; everyone who did not consent to this universe is in another (somewhat less interesting) universe we will never meet. The limitations on God are not of his choosing (any alternative effort he makes is thwarted by logical necessity), yet he remains godlike in power (he can create people and universes and set up this entire scheme). This explains every observation and maintains a morally good God at the helm of it all. Granted, this is a God more consistent with the one of Douglas Adams’ Hitchhiker novels, who left a sign on a distant planet saying, “Sorry for the Inconvenience”, than with the God of any popular theism today; however, in this scenario, none of those religions really come from God anyway but are just a part of the random karmic misery we must endure on the road to paradise.

To be clear, the scenario I just described is ridiculous and bears no appreciable probability of being true, and there is no evidence whatever that it is true, or even that a single one of its premises is true. But it has one meagre epistemic merit: it is logically possible. At least it is so far as we know. And indeed, I have no reason to believe this is the only scenario that could answer here. It really is not reasonable to think that I, or even all humanity, has thought of every possible thing that could be. Though it is reasonable to think that I, and certainly humanity as a whole, have thought of every probable thing that could be, which is what undercuts all of Bergmann’s premises. The impossibility of our having thought of every possible thing that could be undercuts Sterba’s claim to have established the logical impossibility of anything that might answer and could yet be true. Yes, it would have to be some truly bizarre thing like I just proposed, but that only gets us to improbability, not impossibility. It only gets us to an evidential argument from evil—albeit a very strong one.

This conclusion still follows even if you can come up with some genuine proof of the logical impossibility of the scenario I just described because you still have infinitely more unknown scenarios to similarly disprove before you can prove them all impossible. Perhaps one day someone will come up with a sweeping formal proof that establishes all such unknown scenarios impossible; perhaps that will complete at last the logical disproof of a good God’s existence. But that day has not yet come. There is no such proof in Sterba. All he argues against is the conceivable. He has nothing really to say about the inconceivable; nor likely could he, as being inconceivable, he cannot have conceived of such things so as to formulate any objection to them. This is precisely what forestalls the conclusion that all Bergmann scenarios are impossible. At present, at least, we cannot possibly know that.

Despite this single technical objection, James Sterba’s treatise is required reading for anyone aiming to advance or defeat the argument from evil against the existence of God. He corners and casts down every usual defense against it and leaves only one escape route: an admission of an extremely improbable state of affairs contrary to present observation, which renders the existence of a good God extremely improbable. But this does not quite achieve the esoterically specific goal of finally establishing a logical argument from evil. It does, however, more firmly establish an evidential argument from evil than any treatise heretofore. As such, Is a Good God Logically Possible? will still have to be reckoned with by any theist still bent on rescuing their God from this fate, and it will benefit any atheist or philosopher with the converse goal to take lessons from, cite, and draw upon this work in aid thereof.