Ethan Muse has spent the past year building one of the most rigorous defences of the miracle case for Catholicism on offer, and the apologetic project he sets up turns on two commitments concerning miracles and divine hiddenness whose joint acceptance, I will argue, produces a calibration problem that has to be addressed before any historical question becomes tractable. The three main instalments on Fátima, the Eucharistic Miracle of Lanciano, and the charisms of St. Pio of Pietrelcina, together with his reply to Maxime Georgel on Lanciano, all set themselves the goal of presenting a case ‘rigorous enough to persuade those that are strongly predisposed towards skepticism’. My goal is not to dispute the historical sources, the medical findings, or the photogrammetric analyses; doing so adequately would require a separate article for each of the three cases, and a level of technical competence in cardiac histology, forensic serology, and applied photogrammetry that I do not possess. My target is the problem that the two commitments produce when taken together; a problem that survives even if Muse wins every empirical question, and therefore ought to be settled on its own terms.
A brief word on the probabilistic apparatus, since the article makes some use of it. The argument is broadly Bayesian: a rational agent assigns numerical credences (degrees of confidence between zero and one) to hypotheses, and updates those credences as evidence comes in, with hypotheses that predict the observed evidence well gaining credence at the expense of hypotheses that predict it less well. The central technical notion is the likelihood, which is just the probability a hypothesis assigns to the evidence. If hypothesis A assigns a higher probability to the evidence than hypothesis B does, the evidence favours A over B, and the strength of favouring is proportional to how much higher A’s probability is. When I say evidence ‘disconfirms’ a hypothesis relative to a rival, I mean the rival assigns the evidence a higher probability than the original does, so that observing it raises credence in the rival at the expense of the original. Priors will occasionally enter the discussion, but the argument turns on likelihoods. Muse himself adopts this when pressed in dialogue, framing the question of Fátima as one of comparing ‘the prior over given that Catholicism were false and that demons were out there doing massive miracles’ with ‘the likelihood we’d have a distribution of miracles that look like this’ on the supposition that Catholicism is true.
To be clear, I am not defending any specific rival hypothesis against Catholicism, and I am not arguing that any particular alternative supernatural agent best explains the data at Fátima, Lanciano, or San Giovanni Rotondo. The argument turns on what follows from Muse’s own joint commitments. Which rival hypotheses are best positioned to exploit the resulting problem is a separate question, and one I will take up in a later quantitative supplement. For now, the claim is that Muse's commitments create a problem that significantly constrains his case, and that the problem opens the door to further criticism, not only of Muse's argument but of similar miracle-based arguments advanced by other theists.
1. The Hiddenness Constraint
Muse’s first commitment appears in §3 of the Lanciano reply, defending the imperfect preservation of the relics against the objection that, were the preservation truly miraculous, it ought to be perfect:
This can be expressed as two claims. The first is positive: God produces miraculous evidence of sufficient strength to vindicate Catholic claims to the receptive observer. The second is restrictive: God does not produce evidence so overwhelming that the obstinate sceptic has no rational pretext for dissent. Together these amount to an upper bound on the strength of miraculous evidence God will permit. Following Schellenberg and his interlocutors, the literature sometimes calls this the epistemic distance God maintains from creatures; the label matters less than the constraint, which I will call simply the hiddenness constraint. Evidence strong enough to clear that upper bound is hiddenness-crossing evidence.
One could read Muse’s clause as concerning the psychological possibility of doubt for reprobates rather than the rational force of the evidence itself. On this reading God might produce evidence of arbitrarily high rational force whilst still leaving the obstinate sceptic with motivational and dispositional routes of resistance, such as selective dismissal of testimony, motivated reframings of the data, or the weighting of priors. The trade-off Muse describes would then be between evidential force and the availability of psychological resistance, not between evidential force and rational warrant for resistance.
However, I do not think this reading survives the text, as the passage describes the preservation as ‘not so conspicuous that obstinate sceptics had no pretext for dismissing them’. The key word is ‘conspicuous’, which is a property of the evidence, not of reprobate psychology. The contrast Muse draws is between perfect preservation, which would have been more conspicuous, and imperfect preservation, which is less so. The difference between these is a difference in the rational force of the evidence: a perfectly preserved miraculous artefact provides stronger evidence than an imperfectly preserved one, because the imperfect preservation gives the obstinate sceptic a pretext for natural-cause explanations that the perfect preservation would not. Pretexts are propositions one can offer as reasons for dissent. Their availability is a function of the evidence leaving room for them, which is a property of the evidence and not of the reprobate. So the natural reading is that God calibrates the strength of his evidential output below the point at which pretexts would no longer be available. That calibration is exactly the upper bound on evidential strength I am attributing to him. A defender of the psychological-resistance-only reading has to recast the §3 argument substantially, and at that point the imperfect-preservation datum stops being naturally explained at all.
Nor is this imputed to Muse: he endorses it himself, frames it as a positive account of divine hiddenness, and uses it to do explanatory work. If the constraint is given up, the §3 argument loses its motivation and the imperfect-preservation datum cries out for some other explanation. It must also be non-trivial; a constraint set so high that it never binds explains nothing. And although Muse invokes it in connection with preservation specifically, the rationale generalises naturally to the rest of the case. The trade-off Muse describes is between epistemic force for seekers and the psychological possibility of doubt for reprobates, and both terms of that trade-off are properties of the inquirer’s overall epistemic situation rather than of any specific kind of evidence within it. A reprobate confronted with a perfectly preserved relic, a perfectly executed prophetic apparition, or a perfectly attested healing all face the same problem (no remaining pretext for dismissal), and there is no obvious reason why the calibration God allegedly performs in the relic case would not be performed equally in the apparition and healing cases, or in any aggregation across them. The trade-off applies to apparitions, healings, charisms, and cumulative cases just as much as to physical artefacts; in particular, it applies to the aggregate, and a calibration that governs each case individually but exempts their conjunction would need a motivation the hiddenness rationale does not supply.
What follows assumes Muse means what he says, that the constraint does real work, and that the rationale generalises naturally.
2. The Discernibility Principle
The second commitment comes through most clearly in his recent live debate with Shaun Luke on Capturing Catholicism. Muse defends the claim that miracles must be evidentially sufficient to identify their source independently of the revelation those miracles are supposed to confirm:
I will call this the discernibility principle. Without it, miracles could not confirm revelation without presupposing the very revelation they are meant to confirm: if the only way to tell a true miracle from a counterfeit were to know the true revelation in advance, then miracles could not play the evidential role Catholic apologetics has historically assigned to them. So if miracles are to confirm revelation at all, they must come equipped with markers that let the inquirer identify their source independently of the doctrinal content the source is being invoked to verify. Evidence strong enough to do that is discernibility-sufficient.
Muse develops the principle by appealing to subsidiary criteria, including prior announcement in the name of God the creator (Deuteronomy 13 versus 18), divine foreknowledge of events that exceed rational inference (Isaiah 44 and 48), healing miracles of a kind he takes to exceed the power of demons (John 9), and a sufficiently exalted moral character on the part of those associated with the miraculous data (the good-tree / bad-tree principle of Matthew 7). The underlying intuition for all of these is a more general principle from Réginald Garrigou-Lagrange, which Muse quotes approvingly in the same opening statement:
I will call this the divine non-deception principle: God will not permit his name to be invoked over evidence that would invincibly mislead the rational inquirer into accepting a false revelation. On Muse’s view, discernibility follows from non-deception: if God will not permit invincibly-deceptive miracles in his name, miracles in his name must come equipped with markers that let the discerning observer identify their source independently of the revelation they are invoked to confirm.
Two things to bear in mind about discernibility: it must be non-trivial, since a principle that licenses only weak evidential discrimination would not license the strong conclusions Muse wants to draw, given that the principle is supposed to underwrite confident identification of miracles with the Catholic God; and it must extend beyond demonic counterfeits to alternative supernatural hypotheses generally, since the non-deception principle is supposed to underwrite the rational confidence of the inquirer regardless of which alternative happens to be in view.
3. The Calibration Problem
The two commitments constrain miraculous evidence from opposite directions. Hiddenness imposes an upper bound on its strength, so as to leave room for the reprobate’s rational dissent; discernibility imposes a lower bound, requiring strength enough to identify the producing agent. Both bounds are non-trivial, both are endorsed by Muse, and both have to be satisfied by the same body of evidence for his cases to succeed. The question is whether the two bounds leave any admissible interval of evidential strengths between them, and whether Muse’s case falls inside it.
Muse’s hiddenness account says the Catholic God does not produce hiddenness-crossing evidence: evidence above the upper bound. His apologetic, however, needs the cumulative case to be strong enough to identify the Catholic God as its source. If the strength required for that identification is at or above the upper bound, as I argue in §4 is the natural reading of the two commitments, then the apologetic requires the cumulative case to be exactly the kind of evidence Muse’s own account of the Catholic God says that God does not produce. Catholicism, as Muse construes it, then assigns the very evidence the apologetic relies on a low likelihood once that evidence is strong enough to do the work he assigns it, and the question is not whether some other hypothesis predicts the evidence better but whether Muse’s own hypothesis predicts it at all once it exceeds the upper bound his own account erects.
This makes evidential strength a non-monotonic asset for Muse: more of it helps only up to a point, after which more of it hurts. Below the upper bound, additional strength is pure gain, since it makes the miracle harder for the obstinate sceptic to dismiss. At the upper bound the sign of the effect reverses. Past that point, additional strength is a liability, because the evidence is now of the kind Muse’s own God is said not to produce, so each further increment lowers, rather than raises, the likelihood Catholicism-as-Muse-construes-it confers on the total body of evidence. The strength that satisfies the lower bound is therefore, on the natural reading argued in §4, the very strength that carries the case past the upper bound, at which point the evidence is more surprising on the hypothesis it was assembled to support than it would have been had it been weaker. A body of evidence that its own target hypothesis renders surprising cannot straightforwardly confirm that hypothesis. This does not say Catholicism is false. It says the cumulative case cannot perform the confirming function Muse assigns it once it exceeds the upper bound his own hiddenness account erects; it could reach confirming strength only by becoming evidence the confirmed view does not expect.
Lowering the case below the upper bound does not escape the problem, because the lower bound is still in force. If Muse reduces the case below the strength discernibility requires, the case no longer identifies the Catholic God without presupposing the revelation it is meant to confirm, and a case that cannot identify its source in that way cannot perform the evidential role Muse assigns to miracles in the first place. His own objection to miracle apologetics that lack a strong discernibility condition is that such an argument ‘may prove both ways’ and is therefore useless in the confirmation of revelation; a case held below the lower bound leaves the inquirer with nothing that singles out the Catholic God rather than some other supernatural source, which is precisely that uselessness. So the move that takes the case below the upper bound, if carried far enough to matter, carries it below the lower bound as well. A conclusion that I doubt Muse would be eager to accept.
A defender will try to relax the lower bound by leaning on priors rather than the evidence itself, and the Garrigou-Lagrange thesis is where he will look for the resource: God will not permit invincibly-deceptive miracles in his name, so, the defender presses, no rival capable of producing Catholicism-confirming evidence whilst evading detection is theologically permitted. But non-deception, naturally read, rules out only invincible deception, deception the inquirer reasoning correctly from the evidence could not detect; it does not rule out a rival that is evidentially compatible with correct reasoning but dispreferred on other grounds. It governs the all-things-considered judgment, where priors enter, not the bare likelihood comparison.
And Muse explicitly rejects presuppositionalism: he wants the cumulative case to move an inquirer who has not already accepted Catholicism, which bars any prior credence in Catholic doctrine brought to the inquiry in advance. The only priors left to him are general and non-doctrinal, and those are silent between Catholicism and any non-orthodox rival that is equally non-deceptive and equally simple. A defender may answer that Catholicism inherits prior support from a larger natural theology that no equally-simple rival inherits, but that turns on the existence of such a rival and the priors it attracts, a comparative question developed elsewhere; it does not relieve the present pressure, since insofar as there is any rival the general priors do not already disfavour, the discrimination must come from the likelihood ratio, and that ratio is exactly what hiddenness suppresses. Non-deception cannot relax the lower bound without reintroducing, at the level of priors, the strong evidential discrimination the upper bound denies.
4. The Dilemma, and the Calibration Burden
The calibration problem can now be stated as a dilemma, and, on a more generous reading of Muse’s commitments, as a burden he has not discharged. The question is whether Muse’s cumulative case is of the strength discernibility requires, namely strong enough to identify the Catholic God as the producing agent independently of revelation-specific priors, or whether it is not.
If the case is of discernibility-licensing strength, then on the natural reading it exceeds the upper bound, and so it cannot straightforwardly confirm Catholicism as Muse construes it; it has become the kind of evidence that view does not expect. Something assigns the observed evidence non-trivial probability, since it is observed; and whatever that something is, on this horn it is not the Catholic God as Muse describes him, whose likelihood for evidence of that strength has by hypothesis shrunk.
The Catholic is formally free to keep Catholicism and abandon the calibration, but the calibration is not something to jettison without cost; it is precisely what Muse’s §3 reply to Georgel uses to explain why the relics are imperfectly rather than perfectly preserved. Drop it to escape this horn and the imperfect-preservation datum is handed back unexplained, the very thing the calibration was introduced to account for. A defender might try a narrower escape: keep the calibration for a single miracle, the relic taken on its own, but deny that it applies to the three cases stacked together, where the horn actually bites. That move does not work, and §1 already said why. The hiddenness rationale is a claim about the inquirer’s whole epistemic situation, the total weight of everything they are confronted with, so there is no principled reason it would govern each miracle on its own yet fall silent the moment they are combined. The rule that covers each case covers the pile of them. So the escape is not cheap: the narrower version is blocked by §1, and the broader one buys its way out only by reopening the very thing the calibration was posited to explain, leaving Muse owing a fresh account of imperfect preservation in its place. The stipulation meant to underwrite the apologetic, discernibility secured by divine non-deception, is on this horn the source of its own undoing, since the very threshold the case aspires to clear is the threshold past which the Catholic God, so calibrated, is said not to produce evidence at all.
If the case is not of discernibility-licensing strength, then by the lower-bound argument of §3 it cannot perform the evidential role Muse assigns to miracles. He stays under the upper bound only by falling below the lower one: the case no longer identifies the Catholic God rather than some other supernatural source, and falls into the very ‘may prove both ways’ uselessness his own discernibility principle was invoked to rule out.
The dilemma proper arises on the reading where the strength discernibility requires is at or above the strength hiddenness forbids, that is, where the lower bound sits at or above the upper bound. On that reading the two bounds leave no admissible interval at all: any case strong enough to satisfy discernibility is by that very fact above the upper bound, and any case weak enough to respect the upper bound is by that very fact below the lower one. There is then no interval to fall into, and the apologetic is self-undermining whichever strength the case actually has. This is, I think, the natural reading: the same conspicuousness that lets an honest inquirer identify the source is what strips the obstinate sceptic of a pretext, since both track how hard the evidence is to explain away, so the strength that satisfies the lower bound is the strength that breaches the upper.
It is available to Muse to deny that the two bounds are that close. He may argue that the strength sufficient for responsible identification of the source and the strength that triggers hiddenness suppression are distinct points on the scale, with the lower bound well beneath the upper one: enough evidence for the honest inquirer, short of the conspicuousness at which the calibration engages, with a non-trivial interval between them. Many Catholic thinkers would say this is the whole point of hiddenness. If that interval is non-empty, the dilemma’s first horn does not bite, and the objection becomes a burden rather than a refutation. However, this introduces a heavy burden that Muse will have to bear. The generous reading establishes at most that some admissible interval exists; it does not locate that interval on the scale, and it does not place Muse’s actual case inside it. The point is not that the two bounds are jointly unsatisfiable; on the generous reading they are not. The point, rather, is that the apologetic requires a worked account of how the same evidence clears the lower bound without breaching the upper, and that account is nowhere to be seen.
The dilemma, then, is this. If the lower bound sits at or above the upper bound, the apologetic is undermined either way: above the upper bound the case cannot confirm the view it was built to confirm, below the lower bound it cannot identify its source. If the two bounds pull apart, Muse faces the calibration burden: he must locate both bounds on the scale and show that the cumulative case lies between them. Until that is done, the apologetic cannot proceed on the mere assertion that the evidence is very strong.
5. Conclusion
Muse’s miracle apologetic cannot simply maximise evidential force to present a case ‘rigorous enough to persuade those that are strongly predisposed towards skepticism’. His hiddenness rationale imposes an upper bound on the strength of evidence the Catholic God is expected to produce; his discernibility principle imposes a lower bound on the strength a miracle must reach if it is to identify its source without presupposing the revelation it confirms. The cumulative case therefore has to lie in the interval between the two bounds: strong enough to clear the lower one, not so strong as to breach the upper. If the strength needed to clear the lower bound also breaches the upper, the case does not merely weaken; the sign of its evidential force reverses, so that it performs the apologetic role only by becoming evidence Muse’s God is not expected to produce, and the only exit, revising the calibration, is foreclosed at the single-case level by the §1 generalisation and at the aggregate level by the fact that the revision hands the imperfect-preservation datum back unexplained. If instead the case is held beneath the upper bound, it thereby falls beneath the lower one, and lapses into the ‘may prove both ways’ uselessness Muse’s own principle was meant to exclude. The charge that survives either way is not that Muse’s view is inconsistent but that it carries a calibration burden he has left undischarged: miracle evidence strong enough to identify Catholicism’s God yet weak enough to preserve the hiddenness rationale he himself invokes.
This is why the divine hiddenness problem is so underrated, and why it's not just a facet of the problem of evil, but also a distinct argument with independent force. Any general explanation of divine hiddenness necessarily reduces the force of every theistic argument (except maybe for ones made on purely a priori grounds, but those are all pretty terrible anyway), since they make the phenomena appealed to by theistic arguments unexpected on theism - we should expect God to instead hide those phenomena just like he does with all the other evidence for his existence. The theist is free to claim that these phenomena still provide some evidence for theism, since even if divine hiddenness renders them less expected, they would be less likely still on atheism, but an honest theist has to admit that divine hiddenness reduces their force. The only way around this would be if theists had an explanation for divine hiddenness that wasn't fully general - i.e., not just a claim about how much total evidence God would allow, but an explanation that only implies that certain kinds of evidence would be lacking. Such an explanation would have to predict that the specific phenomena that theists point to as arguments for God (in this case, a small collection of Catholic miracles) would still be allowed, while all the other potential ways God could reveal himself would be disallowed.
I also think theists who use, "God doesn't want the evidence for his existence to be too strong," as an explanation for the lack of evidence, and especially as a patch for holes in their own arguments (like the imperfect preservation of the Lanciano relic) don't realize how incredibly silly this looks from the outside. I see no difference between this explanation and a conspiracy theorist dismissing any counterevidence as just part of the conspiracy, or a creationist dismissing evidence for evolution as actually being planted by Satan (indeed, this example is isomorphic to the dismissal of non-Catholic miracles by claiming they were performed by demons). If you add to your hypothesis an auxiliary claim that allows you to dismiss any counterevidence or any weaknesses in the evidence as actually being expected, then your hypothesis immediately loses all predictive power, and therefore all evidence in its favor (it's a simple theorem of probability theory that if no observation would be evidence against your hypothesis, no observation can be evidence for it either).
Good work.
I have some thoughts on Ethan’s methodology regarding miracles, and intend to gather them in a response when he publishes his promised article on it. To my mind, there is a fatal flaw in it even before any considerations about specific miracle cases get off the ground, and even if we grant that all of his miracle cases are factual. In this aspect our approaches to Ethan’s work are strikingly similar.
Preview: Ethan wants me to believe that Catholicism is true based on some historically far-removed anomalous physical phenomena at Fatima. This would include believing, based on these far-removed anomalous physical phenomena, that I and all my relatives will be plausibly eternally tortured in fire unless we convert. IMO, that is a ridiculous epistemic demand in comparison with the evidence provided, and this incredulity touches only the incomparable weight of the evidence relative to the assertion, and not also any priors. As far as priors are considered, Ethan’s case becomes even more implausible. Any moral realist not already wedded to eternal conscious torment, for example, will almost certainly find it to be moral nonsense (and thus impossible given a morally perfect God in whom Ethan believes) with such conviction that Ethan’s case does not even begin to scratch the surface of overcoming these priors (I am saying this as a strong moral realist myself). There are many other philosophical, moral and historical objections to Catholicism, and the list goes on.
Ethan says that his case for Catholicism from miracles is so strong as to unilaterally defeat all of these other objections (for example, he says that the Problem of Evil as a whole is defeated by his miracle cases), but to my mind, that is a strikingly overconfident and baseless assertion which overvalues Ethan’s case to the utmost degree.