Sceptical theism and its dialectical role

Sceptical theism is best understood, in the first instance, as a family of responses to evidential arguments from evil that challenge our entitlement to infer, from the evils we observe, what God would or would not be likely to permit. As the Stanford Encyclopedia of Philosophy explains, sceptical theists maintain that, even if theism were true, we should be sceptical of our ability to predict God’s plans for organizing the world, including plans concerning the amount and nature of evil (Perrine Reference Perrine, Zalta and Nodelman2023). In that familiar role, sceptical theism functions not as a theodicy, but as an undercutting reply: it does not aim to identify God’s reasons for permitting suffering so much as to deny that our failure to discern such reasons licenses the conclusion that there are none. This is also how the broader literature has typically treated sceptical theism: not merely as piecemeal scepticism about a few isolated inferences, but as an undercutting response to evidential arguments from evil under conditions of significant cognitive limitation (Bergmann Reference Bergmann2001; Perrine Reference Perrine, Zalta and Nodelman2023).

For present purposes, however, not every form of sceptical theism is equally relevant. A comparatively weak version might say only that human beings are badly placed to make crude moral judgements about what a perfect being would do in particular cases. A more robust version places pressure on a broader range of expectation-guided inferences. It suggests that our epistemic situation is too limited to warrant confident judgements about what sorts of states of affairs would be unsurprising, improbable, or disconfirming if theism were true. A merely minimal sceptical theism would do little more than caution against simplistic theodicies; it would not explain why sceptical theism has been thought to undercut serious evidential arguments from evil in the first place.

The version of sceptical theism relevant here need not be maximally strong. My argument does not require the extreme thesis that virtually all conditional probabilities involving divine intentions are inscrutable. It requires only a form of sceptical theism robust enough to deny that, in a significant range of cases involving evil, suffering, and divine permission, our evaluative position licenses the relevant expectation-guided probability judgements. The sceptical-theist idea at issue is not merely that we should be cautious. It is that our cognitive limitations are deep enough to defeat or undercut substantially many ordinary attempts to infer from observed suffering to what would be likely, unlikely, expected, or surprising if theism were true. This is the sense in which sceptical theism bears not only on arguments from evil but also on any evidential argument that depends on claims about what God would be likely to produce, permit, or prefer.

Once matters are put in this way, the connection to the fine-tuning debate becomes immediate. If sceptical theism constrains our ability to form justified expectations about what would be unsurprising under theism, then its significance extends beyond negative arguments from evil and reaches positive evidential arguments for theism as well. This is why the recent debate between Donahue and Boyce matters. Donahue argues that the considerations supporting sceptical theism undercut the fine-tuning argument, at least in its standard form, while Boyce attempts to show that sceptical theists can nevertheless endorse a suitably revised version (Donahue Reference Donahue2026, 164–172; Boyce Reference Boyce2026). The dispute is therefore best seen as a test case for the broader dialectical reach of sceptical theism.

Fine-tuning and the need for theistic expectations

The fine-tuning argument, in the form relevant here, is best approached not as a dispute over cosmological detail, but as a comparative evidential claim. Its central contention is that some feature of the universe – most familiarly, that life-permitting conditions obtain only under highly stringent constraints – confirms theism over a rival hypothesis such as naturalism. For present purposes, the important point is not the physics as such, but the logic of confirmation that the argument presupposes. Both Boyce and Donahue frame the issue in broadly evidential terms, as a question about whether the relevant datum supports theism once its probabilistic structure is made explicit (Boyce Reference Boyce2026, 1–3; Donahue Reference Donahue2026, 165–168).

Once the argument is cast in that evidential form, one premise becomes indispensable. It is not enough to say that life-permitting conditions would be very improbable under some rival hypothesis. To generate support for theism, the argument also requires the claim that such conditions are not very improbable given theism. This is precisely the premise on which recent discussion has focused. Boyce explicitly characterizes standard presentations of the fine-tuning argument as relying on a theism-friendly conditional probability of this sort, and he takes the sceptical-theist challenge to arise because support for that probability is usually drawn from claims about what God would have reason to do (Boyce Reference Boyce2026, 2–4).

At this point, the need for what I shall call theistic expectations comes into view. By this, I mean epistemically licensed judgements about what evidence would be relatively unsurprising on the assumption that theism is true. To say that a life-permitting universe is not very improbable given theism is, at minimum, to say that such a universe would not be especially surprising if God exists. That judgement need not amount to a fully developed account of divine psychology, but it does require some bridge between theism and expected evidence. Without such a bridge, the theistic conditional probability becomes evidentially inert. Compatibility is not enough; confirmation requires some reason to think that the evidence is less surprising under theism than under the relevant rival hypothesis (Donahue Reference Donahue2025, 794–801).

This point should not be overstated. The present claim is not that a fine-tuning argument must presuppose a richly detailed moral portrait of God. A defender of fine-tuning may insist that only a modest assumption is needed – perhaps that a perfectly good God would have some reason to value life, consciousness, or a world hospitable to rational agency. But even this leaves the underlying structure unchanged. Whether strong or weak, direct or indirect, the argument still depends on some expectation-guided connection between theism and the observed universe. Standard formulations cannot dispense with such a connection altogether.

Once this expectation-guided element is made explicit, the relevance of sceptical theism becomes immediate. If sceptical theism places pressure on our entitlement to form judgements about what would be unsurprising under theism, then standard fine-tuning arguments are structurally exposed to sceptical-theist worries. Donahue’s recent argument makes exactly this point: the same considerations that support sceptical theism undercut standard versions of the fine-tuning argument (Donahue Reference Donahue2026, 164–172). Even if Donahue’s full conclusion is rejected, the pressure he identifies remains. The standard fine-tuning argument is not insulated from sceptical-theist concerns by virtue of being a positive argument for theism. On the contrary, it appears especially vulnerable to them because its probative force depends on a comparison that cannot proceed without some account of what would be expectable under theism.

This does not yet show that no fine-tuning argument is available to the sceptical theist. It shows only that any such argument will require a more careful account of how the relevant theistic expectations are to be licensed. If the standard route relies on a priori judgements about divine intentions that sceptical theists are reluctant to endorse, then some alternative epistemic route must be found. This is exactly the opening Boyce exploits.

Boyce’s reformulation

Boyce’s proposal is best understood not as a rejection of the underlying problem, but as an attempt to solve it by changing the evidential and epistemic structure of the fine-tuning argument. He explicitly grants that standard formulations of the argument are difficult for sceptical theists to endorse. The reason, as he presents it, is that those formulations typically rely on a priori moral judgements about what God would have reason to do, for instance, by assuming that a perfectly good God would be likely to create embodied conscious agents or a life-permitting cosmos. If sceptical theism places pressure on precisely such judgements, then standard fine-tuning arguments appear to rely on a kind of insight that sceptical theists cannot straightforwardly claim to possess (Boyce Reference Boyce2026, 2–3, 7–9).

Boyce’s proposal is revisionary at two distinct levels. Firstly, it modifies the evidential datum. Instead of treating the relevant evidence simply as the fact that the universe is life-permitting, he shifts attention to the stringency of the conditions required for a life-permitting universe. Secondly, and more importantly, he alters the epistemic route by which the theistic side of the evidential comparison is supported. Rather than grounding the relevant conditional probability in unaided a priori reflection on what God would likely do, he appeals to a body of historical information, D, which includes alleged cases of divine disclosure and is supposed to make certain claims about divine intentions epistemically respectable. The attraction of the move is that it purports to distinguish between illicit speculative moral projection and evidentially mediated theological judgement (Boyce Reference Boyce2026, 13–15, 18).

Seen in this light, Boyce’s proposal has real philosophical interest. It recognizes that the sceptical-theist challenge to standard fine-tuning arguments is not trivial, and it avoids the unpromising response of simply insisting that sceptical theists may continue to rely on the same sort of judgements they elsewhere treat with suspicion. Instead, Boyce attempts to distinguish between two epistemic sources: on the one hand, speculative a priori moral reflection about what God would likely do; on the other hand, historically mediated information internal to a theistic tradition, including putative disclosures of divine purposes.

To appreciate the strength of this move, it is useful to distinguish two possible readings of Boyce’s claim. On a weaker reading, he shows only that sceptical theists are not logically prevented from endorsing some form of the fine-tuning argument. On a stronger reading, he shows that sceptical theists can endorse a revised fine-tuning argument while preserving the central dialectical benefits of sceptical theism in responses to the problem of evil. The weaker reading is comparatively easy to defend. The stronger reading is the philosophically interesting one. For if Boyce secures compatibility only by substantially narrowing the role sceptical theism is allowed to play, then the original tension has not been resolved in the fullest sense. It has only been managed within a more restricted evidential framework (Boyce Reference Boyce2026). The remainder of this paper targets the stronger reading.

Residual theistic expectations

Boyce’s reformulation is philosophically significant because it does not merely restate the standard fine-tuning argument in more cautious language. It attempts to alter the epistemic route by which the relevant theistic probability judgement is supported. Instead of relying on a priori moral reflection about what God would have reason to do, Boyce appeals to D in order to make certain claims about divine intentions epistemically respectable. If this manoeuvre succeeds, then the sceptical theist would seem to have a way of endorsing a fine-tuning argument without relying on the very kind of direct moral insight that sceptical theism places under pressure. Any adequate assessment must begin by taking that attraction seriously.

The force of the present objection, however, depends on distinguishing several different kinds of theistic expectation. At one extreme are merely formal expectations: the minimal comparative assumptions without which no evidential argument can get started at all. These are too thin, by themselves, to generate any distinctive difficulty for sceptical theism. More substantial are what may be called axiological expectations: judgements to the effect that certain broad features of reality, such as life, value, rational agency, or intelligibility, would not be especially surprising if theism were true. Stronger still are intention-involving expectations: judgements not merely that some valuable outcome is compatible with theism, but that God would have reason, aim, or intention to bring about such an outcome. The present paper does not claim that Boyce simply reintroduces the standard a priori moral premise of familiar fine-tuning arguments unchanged. The claim is more precise: Boyce still requires expectations of the third sort, or ones close enough to them to do genuine evidential work. His innovation lies in changing their epistemic source, not in eliminating their evidential role. Only the third kind is dialectically central here, since sceptical theism is not seriously threatened by merely formal comparative structure, but it is threatened once evidential force depends on sufficiently scrutable claims about what God intends, aims at, or has reason to bring about (Donahue Reference Donahue2025, 795–803; Boyce Reference Boyce2026, 5–8).

Once this distinction is in view, Boyce’s strategy can be stated more exactly. His proposal does not rely merely on formal comparative structure, nor only on very weak axiological assumptions. Rather, it aims to make certain intention-involving judgements epistemically respectable by appeal to D. Boyce’s revised argument, which does not merely require that life or value be broadly consistent with theism. It requires that D make sufficiently scrutable a more specific claim about divine intention – namely, that God intentionally produced, or had reason to produce, a life-permitting universe. In Boyce’s own framework, what matters is not simply that a life-permitting universe is compatible with theism, but that the evidence can support a theism-friendly probability judgement because D renders certain claims about divine purposes sufficiently scrutable. That is why Boyce introduces not only theism but also a more specific, intention-laden proposition concerning God’s intentionally producing a life-permitting universe. The issue, therefore, is not whether Boyce preserves the standard premise of fine-tuning in its original form. It is whether the revised intention-involving expectations licensed by D can be insulated from the broader sceptical constraints that originally motivated the problem (Boyce Reference Boyce2026, 15–16, 18).

This is the right way to understand what I mean by residual theistic expectations. These are not expectations generated by direct a priori theology alone, but expectations generated for generic theism by D as admissible background information. Unless Boyce’s appeal to D yields some revised expectation of this general sort, the argument loses its probative force. Merely noting that a life-permitting universe is compatible with theism, or not ruled out by sceptical theism, does not yet establish that the datum confirms theism. Confirmation requires more than mere consistency. It requires some reason to think that the evidence is less surprising under the relevant theistic hypothesis than under the relevant rival hypothesis. Boyce’s proposal changes the source of theistic expectations rather than their necessity.

This should not be confused with the cruder complaint that Boyce still relies on divine psychology. The present point is more discriminating. Boyce need not be read as covertly reinstating the very same premise used in standard versions of the fine-tuning argument. His proposal is revisionary in a genuine way. But a revised epistemic basis is not the same thing as no epistemic basis. If the revised argument is to do more than establish bare compatibility, it must still support some evidentially relevant judgement about what would be relatively unsurprising if generic theism is true and D is part of the admissible background information. Once that much is granted, the central question becomes whether the expectations licensed by D can really be quarantined from the broader sceptical-theist constraints that gave rise to the original difficulty.

This is also the point at which the distinction between target hypothesis and evidential background becomes crucial. Boyce’s proposal should not be read as replacing generic theism with a more specific hypothesis that simply entails, or nearly entails, the relevant fine-tuning datum. That would make the argument far less interesting. Rather, Boyce aims to retain generic theism as the hypothesis confirmed by fine-tuning while allowing D to inform the conditional probability of the evidence on that hypothesis. That move is philosophically intelligible, and it may even be unavoidable if sceptical theists are to endorse any positive evidential argument for theism at all. But it is not dialectically neutral. Sceptical theism’s broad anti-evil role was standardly deployed where divine-intention judgements remained comparatively opaque. Once D is allowed to make such judgements sufficiently scrutable for positive confirmation, the debate no longer concerns generic theism considered against an evidentially thin background. It concerns generic theism assessed relative to a more informative background (Boyce Reference Boyce2026, 8–10). The question, then, is not whether Boyce has eliminated theistic expectations, but whether he has shown that background-enriched theistic expectations can do positive evidential work without affecting the broader sceptical posture that originally made fine-tuning problematic for sceptical theists in the first place.

That is the pressure point that must now be addressed directly. If the appeal to D is substantial enough to generate intention-involving expectations in the fine-tuning case, then the evidential profile under which generic theism is assessed has been altered. The next question is whether that altered evidential profile can support positive inference without narrowing sceptical theism’s broader anti-evil role. The following section argues that it cannot.

The dialectical cost of evidential enrichment

The decisive issue is no longer whether sceptical theists can coherently endorse some revised fine-tuning argument. Boyce has given a serious reason to think that they can. The more important question is what exactly follows from that result, and at what evidential level it is achieved. Here the crucial point must be stated precisely: Boyce’s proposal remains a fine-tuning argument for generic theism. It does not preserve the argument by substituting a more specific theistic hypothesis for generic theism. Instead, it preserves the argument, if at all, by allowing D, including alleged instances of divine disclosure, to function as background information that helps generate expectations on generic theism. Boyce’s formulation therefore relocates the epistemic basis of the relevant theistic expectations, not the target hypothesis itself (Boyce Reference Boyce2026, 15–18).

That relocation matters because it changes what sceptical theism can plausibly be used to do. Sceptical theism has typically mattered dialectically not merely because it blocks a few isolated inferences, but because it sustains a broad undercutting posture towards arguments that move from observed suffering to claims about what would be unlikely if God existed. That posture is most naturally available when generic theism is assessed against a comparatively thin background, where the evidential profile of divine intentions remains indeterminate. Once D is introduced in order to support a positive evidential argument, however, the relevant question is no longer simply what follows from generic theism in evidential abstraction. It is what follows from generic theism when assessed relative to D. If Boyce is right that D makes certain claims about divine intentions sufficiently scrutable for confirmatory purposes, then sceptical theism can no longer operate unchanged as if the relevant evidential setting remained maximally thin and opaque.

The governing methodological principle should therefore be stated with care. The argument does not require any strong symmetry thesis according to which positive and negative evidence must be treated in exactly the same way. What it requires is only a weaker principle of minimal admissibility: if D is epistemically admissible for the purpose of underwriting intention-involving judgements in a confirmatory context, then D cannot be excluded in limine from potentially disconfirmatory assessment of generic theism relative to that same background. This does not mean that D must yield parallel conclusions in the two cases. It means only that once D is admitted for the sake of positive inference, one cannot revert without argument to the opacity characteristic of a thinner evidential setting when adverse evidence is considered. What is ruled out, in other words, is not asymmetry of result, but selective evidential enrichment without principled restriction.

Dialectical retrenchment occurs when a position that previously functioned as a broad undercutter of a class of anti-theistic inferences can, after the relevant evidential background has been enriched, undercut only a narrower subclass of those inferences. The missing step is therefore dialectical rather than merely methodological. Sceptical theism mattered apologetically not simply because it blocked a few isolated evidential inferences, but because it sustained a broad anti-evil undercutting posture towards arguments that moved from observed suffering to claims about what would be unlikely if God existed. Once D is admitted as a source of intention-involving positive assessment, that broad anti-evil undercutter cannot simply be carried over intact. What follows is not that all sceptical resistance disappears, nor that adverse evidence now defeats generic theism relative to D. What follows is that the original broad anti-evil undercutter has been replaced by a narrower one. That is the precise sense in which evidential enrichment yields dialectical retrenchment.

The result can be put as a dilemma. Either D is not epistemically rich enough to make the relevant intention-involving expectations meaningfully scrutable, in which case Boyce has not yet shown enough to secure a robustly confirmatory fine-tuning argument for sceptical theists. Or D is rich enough to make such expectations scrutable, in which case generic theism, assessed relative to D, becomes the proper object of evidential assessment more generally. On that second horn, sceptical theism may remain available in some form, but it no longer functions in its original broad way as a general undercutter of expectation-guided arguments from evil. What Boyce has then shown is not that sceptical theism and fine-tuning are broadly reconciled without cost, but that a sceptical theist may endorse a fine-tuning argument for generic theism once the evidential background is sufficiently enriched and the scope of sceptical restraint correspondingly narrowed.

Why the enriched problem is not merely formal

This conclusion would be less significant if the enriched problem from evil were merely formal: a possible objection that follows from evidential bookkeeping but has little substantive force. That is not the case. Once D is allowed to inform claims about divine purposes in the fine-tuning argument, several adverse data become serious enough to demand independent attention. The point is not that they defeat generic theism relative to D. The point is that they cannot be dismissed by the same broad sceptical manoeuvre that was available when generic theism was assessed against a thinner background.

Firstly, consider divine hiddenness. If D includes alleged divine disclosure, covenantal address, revelation, or claims about God’s purposive relation to rational creatures, then D does more than support the claim that God intended a life-permitting universe. It also appears to support expectations about divine self-disclosure. A God whose disclosed purposes include relating to rational creatures, eliciting trust, or making divine reality known would not obviously be expected to permit persistent patterns of non-resistant nonbelief, deep religious ambiguity, or radically uneven access to the relevant disclosure. These data are not simply generic evils added to the ledger. They bear directly on the very class of intention-involving judgements that D was introduced to make scrutable (Howard-Snyder and Green Reference Howard-Snyder, Green, Zalta and Nodelman2025).

Secondly, consider suffering that obstructs rather than merely accompanies the goods allegedly disclosed in D. If D is invoked to support the claim that God values life, rational agency, or relationship with creatures, then some forms of suffering appear especially problematic: suffering that destroys the psychological conditions for trust, damages moral agency, severs relational capacities, or makes religious response less rather than more accessible. The problem is not merely that suffering exists. It is that certain kinds of suffering seem to undermine precisely the goods that the enriched theistic background is supposed to identify as divinely intended. At that level, the adverse datum is no longer simply suffering-as-such, but suffering that appears misaligned with the very intentions probabilified by D.

Thirdly, consider distributional and historical evil. If D contains claims about universal divine concern, providential governance, or the availability of divine disclosure to human beings, then the uneven distribution of both suffering and religious access becomes evidentially relevant. The concern is not exhausted by the existence of horrors. It also includes their distribution across history, geography, vulnerability, and epistemic access. A world containing severe suffering, religious fragmentation, and highly uneven access to alleged disclosure is not obviously what one would expect once D is allowed to inform judgements about divine purposes. Again, the claim is not a decisive defeat. It is that generic theism assessed relative to D faces adverse evidential pressure at the same level of specificity at which Boyce’s revised fine-tuning argument operates.

These three cases show why the strengthened conclusion is not a mere verbal upgrade from ‘reopened’ to ‘serious’. Enrichment changes the relevant evidential setting. Where generic theism is assessed against a thin background, the sceptical theist may say that our ignorance of divine reasons blocks confident inferences from evil to divine nonexistence. Where generic theism is assessed relative to D, however, D may itself supply content about divine aims. If that content is specific enough to help fine-tuning, then it is also specific enough to generate prima facie expectations about disclosure, creaturely relationship, and the distribution of suffering. The burden on Boyce is therefore not merely to show that sceptical theists can still utter a fine-tuning argument. It is to explain why the background information that enables the positive argument does not also require a substantive response to these adverse data.

Four Boycean replies

Boyce is aware that his appeal to D may invite evil-based objections once D is admitted into the evidential background, and he offers several replies (Boyce Reference Boyce2026, 8, 17–18). None is negligible. But none preserves sceptical theism’s original broad anti-evil role.

Firstly, Boyce may reply that sceptical theism still undermines those versions of the problem of evil that rely on a priori reasoning about God’s intentions. This is true, and it is a genuine dialectical gain. But it is also a narrower gain than the sceptical theist originally wanted. If the debate now concerns generic theism assessed relative to D, then defeating only the a priori version of the problem of evil leaves the most relevant version untouched. A victory over generic, a priori divine-psychology arguments becomes technically correct but dialectically incomplete if a background-enriched problem from evil remains serious. The sceptical theist has not preserved the old broad undercutter; she has preserved a more limited undercutter against a narrower family of arguments.

Secondly, Boyce may reply that D itself already entails a decent magnitude of evil and suffering, thereby reducing the surprise of the evils we observe. This reply is too quick. Even if D includes claims or narratives involving suffering, it does not follow that D entails the type, scale, intensity, and distribution of observed suffering. Nor is it clear that scriptural or historically mediated revelation, considered as D, must be read as straightforwardly entailing that every narrated instance of suffering occurred literally rather than being symbolic, liturgical, mythological, parabolic, or theologically shaped. More importantly, what matters for the enriched problem is not merely whether D predicts some evil. It is whether D makes unsurprising the actual pattern of evil, including horrors, hiddenness, religious ambiguity, and suffering that appears to obstruct the goods allegedly identified by D. The move from ‘D contains suffering’ to ‘D substantially blunts the evidential force of observed suffering’ therefore requires further argument.

Thirdly, Boyce may reply that various purported revelations contain theological resources that increase the likelihood of evil given theism. This is the strongest reply. For example, some Christian doctrinal claims – that the purpose of life is not happiness but knowledge of God, that humanity has rebelled against God, that divine purposes extend into an afterlife, and that knowledge of God is an incommensurable good – may reduce the evidential surprise of some suffering (Craig Reference Craig2026, 295–301). This point should be conceded. But the concession confirms rather than removes the present argument. Once the sceptical theist relies on such resources, the response is no longer supplied by sceptical theism alone. It is supplied by a more substantive, tradition-specific theology. Moreover, these resources do not automatically address the cases just identified: hiddenness among non-resistant persons, suffering that destroys rather than facilitates knowledge of God, and historically uneven access to divine disclosure. They may improve the theist’s evidential position, but only by moving the debate into detailed theological assessment. That is precisely the relocation of evidential background, not a preservation of sceptical theism’s original broad role.

Fourthly, Boyce may reply that sceptical theism can still undercut inferences to the effect that God is failing to realize intentions probabilified by D in morally optimal ways (Boyce Reference Boyce2026, 8). This reply also has some force, but it cannot be treated as a free-standing solution. If it means only that we are not well placed to know whether God realizes divine intentions in the morally optimal manner, then it introduces a further form of optimal-implementation scepticism. But that is an additional thesis, not a result already secured by Boyce’s appeal to D. It also threatens to reintroduce the original problem at a higher level: the more sceptical we become about whether intentions probabilified by D are implemented in ways we can recognize, the less clear it becomes how D was able to support the fine-tuning inference in the first place. If D can make divine intentions sufficiently scrutable for positive confirmation but not sufficiently scrutable for adverse assessment of implementation, a principled account of that asymmetry is needed.

The upshot is not that Boyce has no replies. The upshot is that his replies confirm the restricted character of the result. The sceptical theist can still resist some a priori problems of evil; D may include some evil; doctrinal resources may lower the surprise of some suffering; and further sceptical claims about optimal implementation may be developed. But these moves do not preserve sceptical theism’s original broad anti-evil role. They replace it with a local, background-sensitive, and tradition-sensitive strategy that must be defended case by case.

This is not merely an external complaint. Boyce effectively acknowledges the relevant structure of the problem. He does not claim that appeal to D removes all room for evil-based objections. Rather, he allows that such objections may reappear once one moves to more specific religious claims about divine intentions, and he suggests that the sceptical theist may be content with a more limited ambition in responding to the problem of evil (Boyce Reference Boyce2026, 17–18). That concession matters. It shows that Boyce’s proposal is compatible with the present paper’s main claim: even if coherence is preserved, breadth of dialectical function is not. The difference is that the present argument makes explicit why the cost is substantial. The enriched problem from evil is not merely reopened in principle; it becomes serious enough to require independent philosophical and theological treatment.

The result can now be stated precisely. Boyce establishes, at most, that sceptical theists can coherently endorse a fine-tuning argument for generic theism by appealing to D as enriched background information. What he does not establish is that they can do so while preserving the broader dialectical role that sceptical theism has typically played in undercutting expectation-guided arguments from evil. The cost of his proposal is therefore not inconsistency but evidential relocation together with dialectical retrenchment. The debate is not shifted from generic theism to a different target hypothesis. It is shifted from an evidentially thin assessment of generic theism to a background-enriched assessment of generic theism, and sceptical theism survives only in a correspondingly narrower form.