Here's my almost-twenty-years-old memory of Rachels' argument (I read it in an introduction-to-ethics class at a community college):

  1. If God existed, there would be a being more important, morally, than our moral autonomy.
  2. Nothing can be morally more important than our moral autonomy.
  3. Therefore, God doesn't exist.

This echoes one horn of the trilemma of the problem of evil: "If God is perfectly good, It is not (given the actual world) maximally powerful." Can that horn be subverted(?)? So to say: "If God is not perfectly good, then God is not maximally powerful." Then:

  1. If God were perfectly good, God would be maximally powerful (this is a different premise than the subversion: the presupposition is that having the power to be perfectly good would mean being maximally powerful ipso facto (c.f. Kant's attribution of final power to the divine nature on the basis of the requirement that God would be able to balance the entire negative and positive causal orders in favor of the good, which is an act "no greater than which can be conceived"...)).
  2. If God were maximally powerful, God would prevent x amount of evil (that we don't see being prevented) (premise).
  3. Therefore, God isn't maximally powerful.
  4. Therefore, God isn't perfectly good.
  5. Or, then: there is no perfectly good and maximally powerful being at all.

Is ultimate goodness a prerequisite for eternal power? Now if divine knowledge were a prerequisite of divine goodness such that any divine knower would be maximally good in turn, then (5) can be expanded to and there isn't a divine knower, either.

Can the problem of evil be transmuted into an argument against dystheism as much as, or even more than, an argument against normal theism? It would seem as if Rachels' reasoning testifies against a god of good and a god of evil, then: "Nothing is absolutely worse than everything else; so God is not absolutely worse than everything else; so there's no absolutely and necessarily worst being; so a god of evil doesn't exist." If it would take maxipotence to be the worst of all possible gods,E the subversively neo-scholastic argument goes through all over again, too, then (it seems).

EAlternatively stated, "If there is a worst of all possible beings, this being is yet not maxipotent (neither maxiscient), wherefore..." It would be peculiar if there were (possibly) a worst possible being but not a best possible one, however.

  • Well known Meme
    – Rushi
    Aug 22 at 4:34
  • 1
    "Nothing can be morally more important than our moral autonomy" sounds pretty presumptuous, and very plausibly false. The premise before the obversion is not what is written in 1. (a typo?). And 1. is not one of the horns of Epicurus's trilemma either, it seems. That would be: "if God were perfectly good then he'd be willing...". How does "willing" turn into "able" (all-powerful)?
    – Conifold
    Aug 22 at 4:42
  • @Conifold I assume Rachels was invoking the spirit of Kantian autonomy, not just a generic sense of "being a law unto oneself." The "obverse" conditional, well, I don't know if that's the right word for it (I said "reverse" and "inverse" first, in my head, but those didn't sound right either). Now the first premise on the second list is not supposed to be from the classical trilemma but is assumed vs. Kant's grounding of "all-powerful" in "able to bring about the sum of good," or the old idea that God can't do evil because God is all-powerful: if God did evil, It would "lose" Its power. Aug 22 at 5:16
  • I should add that I don't have much of a sense of plausible-vs.-implausible, at least not anymore. Once I got to the point where I was willing to try doubting the universality of principles like noncontradiction or arithmetic or whatever, my willingness to appeal to either my own, or other people's, judgments of philosophical plausibility mostly dissolved except insofar as I'm now willing to found many arguments that I use in discussion, on the basis of premises that I pre-theoretically would have found implausible. "Whatever works" is closer to my "motto" now, I guess... Aug 22 at 5:33
  • But the "premise above the obversion" says:"If God is perfectly good, It is not (given the actual world) maximally powerful". And 1. says:"If God were perfectly good, God would be maximally powerful". The consequents are negations of each other (assuming "It" should be "he"). And yes, 1. is not the obverse of what's right above it, but rather its inverse.
    – Conifold
    Aug 22 at 5:34

2 Answers 2


There are multiple ways to approach this question:

Response 1: Critique Rachels' argument. Neither premise 1 nor 2 of Rachels' argument is obviously true. Nor can premise 3 be derived from the prior ones, unless one assumes moral necessitarianism, which is also problematic.

For premise 1, why would God be more important than us? Both rights ethics, and utilitarian ethics tend to weight all moral agents and patients at equal weighting. I personally think that some sort of scaled moral weighting is necessary for any moral system to be useful, but even with that, moral weighting need not have an infinity in it for God.

For premise 2, it is implausible that freedom of action is the absolute of moral values. Freedom is certainly valuable, but how could it be more valuable than, say, existence, or say the discernment of the love/hatred dichotomy, or say the experience of irredeemable despair? That all the others either do or do not occasion are pre-requisites for useful freedom.

For 3, the instated premise of moral necessitarianism is unsupported. Necessity is EXCEEDINGLY difficult to demonstrate even based on abstracts such as logic, the extension of it to moral arguments, which are difficult to justify even objectively exist, appears to be an exceedingly over-optimistic inference.

Response 2: I am aware of a current theologian who accepts premise 2, and uses it to characterize what kind of God could exist. This is Thomas J Oord, and the work of his that I think most directly addresses this issue is "God Can't". Oord does not agree with Rachels conclusion, however.

Oord's version of God is capable of omnipotence (and presumably omniscience), but because Oord and his God accept premise 2, he argues that God self-limits because of Its moral nature, and therefore does not act in this world as that would take away freedom from us, and the other agents within it. Oord's conception of God does not allow God to be a creator of the universe, as an act of creation would constrain the freedom the particles of such a universe. Oord's God is therefore not a creator God.

Oord focuses on omnipotence, and I fear does not think thru the power of Omniscience, as an omniscient deity could convince any of us to do whatever we ought to, as omniscience gifts the knowledge and wisdom to provide absolutely convincing arguments. If one is informed of the good, and chooses to do good, this is complete freedom per his formula, and given the agency he extends to all of the pieces of the universe, all a deity would need is omniscience rather than omnipotence to lead to the universe's moral maximization thru the power of God's persuasion.

Oord's God would therefore have to be also self-muted, not just self-crippled, for our morally flawed universe to be explained by it. As he is a Christian, who thinks God walked the earth and communicated Christianity, his God is at best only partially self-muted, as well as only partially self-crippled. I don't think this theology works, but Oord provides an interesting example of how one can run with premise 2, and arrive at a different place than Rachels.

Response 3: the scholastic approach to the Problem of Evil, which is to try to work thru the concepts of absolutes and necessity, is not how the problem was originally formulated. Epicurus's statement of the POE is an empirical test case. This is consistent with Kant's, and post-Kantian focus in philosophy, on the limits of rationalism.

What one can do empirically with God hypotheses, is derive predicted consequences from them, then examine our world to see if it is consistent with that hypothesis or not. One can postulate an Omni-God, and the consequence of that postulate is that our universe would be the best of all possible worlds. The moral failings of our world are the POE test case for this, and one of many failed test cases. Our universe is also not optimally beautiful, optimally variable, optimally law-like, optimally well communicated, optimally configured for life, optimally understandable, optimally mysterious, or any other possible predicted optima based on inferred Divine objectives.

Note that for each of the above, one need not even have omnipotence nor omniscience for the failed test cases to be problematic. IF there is a single dominant entity working in our universe, even if that entity is not the universe's designer, then that entity would be able to apply willing to the structure of the universe, and cause it to progressively approach a desired optimal state. Even a God of limited power or knowledge should have caused our universe to approach any desired optima. It is only thru absolute self-restraint of some kind such as Oord proposes, or alternatively due to a form of insanity (an inability to maintain thoughts over time on the part of a damaged God would also lead to ineffectiveness) that even a limited God would also not lead to a universe closely approaching optimum.

Response 4: Yes, contra to the scholastics, and instead consistent with treating all God hypotheses as on the table, a God could be a Dystheist God. However, our universe is also not maximally evil, so the same empirical test cases that Omni-Gods fail for omnibenevolence, other God hypotheses fail for omni-malevolence. This is also true for any mixed moral character as our universe is not just not optimized morally, it also is not optimized in any other way. Whatever character a monotheist God has, it would still optimize, or near optimize, based on the objectives its character calls for.

Response 5: The only God hypotheses that pass these empirical optimization tests are Deist Gods who have no objectives, insane Gods who cannot maintain a coherent thought direction, an extremely multitudinous polytheism with weak Gods such that our universe has too many competing small actions on it that it cannot develop a character, or di-theism with two almost identically powerful and hostilely competing agents in it.

  • Reason for accepting the answer: I had been arguing that the myriad first-order divine properties could be extracted from one divine property, but not a second-order one as I asked about in a previous post, but one of the first-order ones (viz. primarily divine goodness in this case). But you mentioned the notion of "insane gods" whose insanity forestalls their effective use of their vast powers (alone or in concert), and who do not think to use (or do not even have!) an additional power-to-stabilize-their-minds, maybe. This is an interesting narrative option (and not unknown in literature). Aug 22 at 18:24
  • Also, the overview of Oord touches upon a line of argument, in favor of the goodness of God, that I have long pursued: an ought-implies-can argument like, "Creating the world from nothing would be good; God created the world from nothing; therefore, God did something good." And moreover, "Since something's being possible-because-it-would-be-good is the only way to make a possibility out of nothing else, it's the only way to make an actuality out of nothing else, too." But I suppose that Oord is to some extent more invested in free-will theodicy than I am... Aug 22 at 18:42

My own small contribution to this party follows:

  1. If God existed, there would be a being more important, morally, than our moral autonomy.
  2. Nothing can be morally more important than our moral autonomy.
  3. God doesn't exist.

Insofar as I'm concerned, God (exists) = Nothing can be morally more important than our moral autonomy.

If you agree then ...

premise 1 becomes self-contradictory gibberish:

If nothing can be morally more important than our moral autonomy then there would be something morally more important than our moral autonomy

This squares with the free will defense against the problem of evil: God turns a blind eye and a deaf ear toward evil because moral autonomy is paramount.

  • 1
    I do wonder about the stability of abstract uses of "exist," since here we seem to have a case of Nothingness being a thing, which is then more important than our moral autonomy (or whatever), and if God doesn't exist, then if God is this Nothingness, then God is more important than morality after all? (The typical analytic philosopher will say: "Ah, but this is an example of why existence is not a property...") Aug 22 at 18:32
  • Nice play on words there, a linguistic trap I often fall into. As to the issue of moral autonomy & God, the two are synonymous to my reckoning and so Rachel's argument disintegrates. Aug 23 at 1:41
  • 1
    It's possible to stipulatively define "moral autonomy" in a certain way that equates to God (and this is not altogether infelicitous even on Kant's account, for he himself stipulates that we are to take the commands of our pure will for divine commands), but I think Rachels meant to refer to something distinguishable from God, at least if the word "God" is defined as more than "that which we are to obey above all else" (if that were all there were to the definition, the terms would be synonymous, though). Aug 23 at 1:48
  • 1
    @KristianBerry, some words I gather, like the word "perfect" have no specific intensional definition. They're just meant to direct attention to some kind of endpoint for a given trajectory a particular set of qualities is on. To me "God" is such a word and if at any point one believes the highest moral good is autonomy then God is that autonomy. Aug 23 at 2:17

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .