No CrossRef data available.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 22 April 2025
Brian Leftow’s observation that the logical problem of evil is largely resolved, mainly due to the success of free-will defenses, is not universally accepted.1 Graham Oppy, for instance, acknowledges the current shortcomings in formulating a logical argument from evil but contends that deeming such efforts futile is premature.2 Likewise, James Sterba’s recent work shows continued attempts to develop this argument.3 These views prompt a critical question: Is constructing a successful logical argument from evil futile, or will persistence yield success as it has in other philosophical inquiries?
This article introduces the ‘Layers of Reality Response’ to the logical problem of evil, aiming to conclusively resolve the debate for classical theists, independent of free-will defenses. This response establishes two points:
1. Common linking principles in logical arguments from evil are not evidently true and fail to apply across different layers of reality.
2. There is no non-question-begging linking principle for proponents of the logical argument from evil, making a successful formulation inherently unachievable.
If correct, this renders the logical argument from evil not just a philosophical failure but an argument that cannot, in principle, be successfully reformulated – it is not only dead but ‘unresurrectable’.
1 Brian Leftow, God and Necessity (Oxford: Oxford University Press, 2012).
2 Graham Oppy, ‘Logical Arguments from Evil and Free-Will Defences’, in The Cambridge Companion to the Problem of Evil, ed. by Chad Meister and Paul K. Moser (Cambridge: Cambridge University Press, 2017), pp. 45–46.
3 Our argument is aimed at two groups: first, those actively advancing a logical problem of evil, such as Sterba, and second, those, such as Oppy, who, while unconvinced by existing formulations, remain open to the possibility of a successful one emerging. In other words, we seek to demonstrate not only why current formulations fail but also why no future formulation can ultimately succeed.
4 Graham Oppy, ‘Logical Arguments from Evil and Free-Will Defences’, in The Cambridge Companion to the Problem of Evil, ed. by Chad Meister and Paul K. Moser (Cambridge: Cambridge University Press, 2017), pp. 45–46.
5 Specifically, our work seeks to advance an idea introduced by James Ross in Philosophical Theology (Indianapolis: Bobbs-Merrill, 1969) that has not received sufficient attention – namely, how the distinction between God’s metaphysical causality and creaturely causality undermines traditional arguments from evil. Ross initially suggested that inferences about divine responsibility for evil should be restricted when metaphysical dependence is at play. We build upon this by clarifying his principle of limited inferences, strengthening his argument by answering unanswered questions or filling obvious gaps, and addressing objections he left unexplored. Additionally, we refine Ross’s approach by dropping certain examples he used that, in our view, weaken his case and are ultimately unnecessary. Another paper that might seem related to ours is Samuel Lebens’ God and His Imaginary Friends. However, several key differences set our approach apart. We do not argue that creation is God’s dream or fictional story, nor do we rely on analogy. Rather, we show inductively that moral principles required for a logical problem of evil fail when applied across ontological levels.
Moreover, Lebens’ first theodicy suggests God lacks full control over creation, like a dreamer over their dreams – an idea we reject as incompatible with Classical Theism. His second, treating history as God’s fictional story, is also distinct from our view. We do not consider creation fictional but rather as existing at an ontologically different level of reality, yet (we would suggest) as related to God in a sui generis way.
While analogies such as dreamers and dreams can illustrate how moral principles break down across ontological layers, they do not define the God–world relation. Instead, we maintain that once creation’s distinct ontological status is fully appreciated, no moral linking principle can sustain a logical problem of evil. Such principles either fail inductively or beg the question, as we aim to show. Samuel Lebens, ‘God and His Imaginary Friends: A Hassidic Metaphysics’. Religious Studies, 51(2) (2015), 183–204. https://doi.org/10.1017/S0034412514000298.
6 For a recent defense of this thesis, see Gaven Kerr, ‘Goodness and Being, Transcendentals, Participation’, in The New Cambridge Companion to Aquinas, ed. by Eleonore Stump and Thomas Joseph White (Cambridge: Cambridge University Press, 2022), pp. 85–106.
7 Brian Davies, The Reality of God and the Problem of Evil (New York: Continuum, 2006).
8 Although in common parlance ‘fictional’ is treated as synonymous with ‘unreal’, we contend that a more accurate framework is to consider ‘fictional’ as less fundamental, i.e., existing in a mode that is dependent on limited minds. It is wrong to say, for instance, that Sherlock Holmes doesn’t exist, or that there isn’t really a Sherlock Holmes. Rather, Sherlock Holmes really is the fictional character of Arthur Conan Doyle’s novels.
9 We mean there is no story as being imagined by Tolkien without Tolkien actively, simultaneously imagining it. This ignores the issue of Tolkien setting his story down on some medium, such as paper, which preserves in potency a certain range of meaning which can later be actualized by other readers who understand enough of the content and language.
10 This point is highly relevant to our forthcoming consideration of moral principles, since any linking premise deployed in a logical argument must select a moral principle that the classical theist is himself committed to. Why so? Because, to be frank, what other option is there? Ethical principles are, as everyone knows, matters of considerable controversial, and (pace certain Kantians) none of them are analytic, and there are many moral anti-realists to boot. So, absent a moral principle that is analytic, the linking premise must involve some objective moral principle already internal to the classical theistic worldview, which is to say, some principle the classical theist already accepts or should accept as a consequence of their theory (see note 1). Urging some other principle that the classical theist is definitely not committed to – say, some consequentialist principle (‘the greatest good for the greatest number’, say) – is not going to serve the project, since the theist can simply reject that principle outright, no matter how committed the objector may personally be to it. Put simply, we are concerned with the problem of evil as an internal world-view critique.
11 For a helpful overview, see Patrick Flynn and Enric F. Gel, ‘Is Grounding Essentially Ordered Causation?’ The Review of Metaphysics, 77(2) (2023), 247–273. https://muse.jhu.edu/article/915461.
12 James P. Sterba, Is a Good God Logically Possible? (New York: Palgrave Macmillan, 2019). See also Sterba, ed., ‘Is the God of Traditional Theism Logically Compatible with All the Evil in the World?’ Religions, 11(5) (2020), Special Issue. <https://www.mdpi.com/journal/religions/special_issues/god_compatible_evil> [Accessed 10 April 2024].
13 J. L. Mackie, ‘Evil and Omnipotence’, Mind, 64 (1955), 200–201. While Mackie abandoned the logical problem of evil in his later work, he leaned toward the evidential version in The Miracle of Theism: Arguments for and Against the Existence of God (Oxford: Oxford University Press, 1982).
14 Graham Oppy, for example, is one such atheist who challenges Mackie’s principle. Graham Oppy, ‘Logical Arguments from Evil and Free-Will Defences’, 45–46.
15 We acknowledge that Mackie appears to have abandoned his own version of the logical problem of evil in his later work. Nevertheless, we find his formulation useful in illustrating our central point. Sterba, by contrast, has not only maintained his commitment to the logical problem of evil but has continually refined and defended it against criticisms. His work has even led to a dedicated special issue on the topic. See James P. Sterba, ed., Do We Now Have a Logical Argument from Evil? Special issue, Religions, 12(4) (2021), https://mdpi.com/books/reprint/8645-do-we-now-have-a-logical-argument-from-evil.
16 Note that the principle necessarily entails, for instance, that the moral character of a creator is dependent on his creation. Now, we will consider a way in which Tolkien’s moral character is plausibly dependent on his creation, but not qua creator but qua voyeur of his creation. As the God of classical theism cannot attend to creation voyeuristically, we argue that this aspect of moral contingency simply does not apply to God.
17 We should mention that this result is so far unsurprising from within the classical theistic paradigm. After all, within classical theism – certainly the classical theism held by thinkers like Thomas Aquinas – the moral law is something attendant upon human nature. From human nature and its inherent ends do our moral obligations flow. God does not have a human nature and has no end beyond himself required for his perfection. Thus, God does not have moral obligations, nor can it be sensibly said that God has virtues (or vices). More on this in a moment. Also, see Davies, The Reality of God and the Problem of Evil.
18 One might wonder whether the key moral difference is phenomenal consciousness – that it is precisely the possession of phenomenal consciousness (which fictional beings presumably lack) that makes the problem of evil a genuine issue. However, we maintain that this is, at best, a partial restatement of the objection we are addressing – namely, that the entities we have considered so far are fictional (or imaginary) rather than real. While the presence or absence of phenomenal consciousness is one factor distinguishing fictional from real beings, it is not the only one.
To be more precise: By swapping real beings for beings with phenomenal consciousness, the skeptic has not introduced a fundamentally new case – they have merely narrowed their principle to a subset of the same reality. But this does nothing to avoid the original problem because the principle still only applies to God and creation.
Since the principle (‘If some being who is the metaphysical cause of some other being with phenomenal consciousness is in a position to prevent the suffering or sin of that other phenomenally conscious being without incurring some cost clearly worse, they should do so. Otherwise, they are morally at fault’), when reformulated to focus on beings with phenomenal consciousness rather than real beings, still has no other substitution instances aside from God and the world, the skeptic has not provided independent support for it. Instead, they have merely restated the conclusion they need to prove in slightly different terms. Thus, the reformulated argument, like its predecessor, remains question-begging.
19 We can now look back at Sterba’s formulation to see a similar issue. In one of his Requirements, Sterba specifies rational beings, but if his formulation specifies, as it must to avoid breakdown across layers of reality, real as opposed to fictional or imaginary rational beings that are produced by a deep metaphysical cause, it too becomes question begging. Without such specifications it is not obviously true as absolute but in fact rejectable on inductive grounds.
20 This point is made by James Ross in Philosophical Theology, 271.
21 Our claim is not that there cannot be any moral obligations from a higher plane of reality to a causally dependent created layer. Rather, our claim is that it is the atheologian’s burden to establish a linking premise and there simply cannot be any good warrant to suppose that there are such obligations. Any proposals are either drawn from horizontal causal relations and applied by bad analogies or they beg the question, as we have suggested above.
22 See James E. Dolezal, ‘Defending Divine Impassibility’, in Classical Theism: New Essays on the Metaphysics of God, ed. by Jonathan Fuqua and Robert C. Koons (London: Routledge, 2023),
23 Here we are just following Peter Geach, ‘Good and Evil’, Analysis 17(2) (1956), 33–42.