The Catholic Church, whatever else one thinks of it, is not an institution given to understatement about its own supernatural claims. It has spent seventeen centuries cataloguing saints, compiling martyrologies, and processing reports of divine intervention with the methodical seriousness of a bureaucracy that regards the miraculous as central to its institutional identity. When the Church certifies a miracle, it does so after a process that is, by the standards of confessional bodies, unusually demanding: medical boards, canonical investigations, repeated peer review, and a formal declaration from the relevant dicastery in Rome. The bar is genuinely high, higher than most critics give it credit for. This matters, because the argument that follows rests entirely on the Church’s own published figures rather than on any sceptical inflation of them. The numbers are striking enough without embellishment.
Since 1858, when Bernadette Soubirous reported her apparitions of the Virgin Mary at Massabielle, the Sanctuary of Our Lady of Lourdes in the French Pyrenees has become the single most visited pilgrimage site in the Catholic world. The Sanctuary’s own published figures place cumulative visitor numbers at somewhere around 200 million people across those 160-plus years. Many of those visitors have been sick. The baths at Lourdes, the springs, the processions by candlelight: they were not designed merely as a backdrop to spiritual tourism. The explicit and repeated purpose of the pilgrimage has been healing, the restoration of physical health that medicine has been unable to provide, through the direct intercession of the Virgin and, through her, of God. That is the promise embedded in the founding narrative. It is the promise that fills the coaches and the trains and the aeroplanes that carry the sick and the suffering to the foot of the Pyrenees every year.
The Church has recognised approximately 70 of those claimed healings as genuine miracles. The most recent, as of the time of writing, was the case of Sister Bernadette Moriau, a French nun who had suffered from a refractory cauda equina syndrome for decades before reporting a sudden and complete recovery in 2008. After a nine-year investigation involving the Lourdes International Medical Committee, the Vatican’s Dicastery for the Causes of Saints, and a canonical inquiry within her diocese, her case was formally declared a miracle in 2018. Seventy certifications across 160 years and 200 million visitors: let us think carefully about what that figure means, because the implications run considerably deeper than most defences of the pilgrimage are willing to follow.
1. The Machinery of Certification
Before drawing any statistical inference, it is necessary to understand precisely what the Church means when it certifies a Lourdes miracle, because the process is more rigorous than the popular imagination tends to assume, and the rigour is itself part of the argument. The Lourdes International Medical Committee, known by its French acronym CMIL, was established in its modern form in 1947 and currently draws its membership from physicians across multiple countries and medical specialities. It is not a committee of theologians. When a reported cure is submitted for review, the committee applies criteria that were formalised in 1978 by the then Bureau des Constatations Médicales and which require, among other things, that the original diagnosis be certain and well-documented, that the cure be sudden and complete, that it not be attributable to medical treatment already under way, that it not coincide with any natural resolution of the condition, that it be stable and lasting, and that it be scientifically inexplicable given current medical knowledge. Cases that pass this first filter, a very small fraction of those submitted, are then referred to the diocesan bishop of the claimant, who may initiate a canonical inquiry. If the bishop concludes that a miracle has occurred, the case goes to Rome, where the final declaration is issued or withheld.
This is not a casual process, and deliberately representing it as one would weaken rather than strengthen the sceptical case. The CMIL has, in the past several decades, been at pains to note publicly that it does not itself declare miracles; it only declares that a given cure is currently inexplicable by natural means. The theological step, the attribution of the cure to divine intervention, belongs to the ecclesiastical hierarchy, not to the physicians. That distinction is honest and should be acknowledged. It is also, however, precisely where the argument becomes philosophically interesting, because the gap between “medically inexplicable at present” and “therefore caused by God through the intercession of the Virgin Mary” is not a small gap. It is precisely the gap that David Hume identified in 1748 and that no one has successfully closed since.
Hume’s formulation in “Of Miracles” remains the cleanest statement of the problem. As he put it: “When anyone tells me, that he saw a dead man restored to life, I immediately consider with myself, whether it be more probable, that this person should either deceive or be deceived, or that the fact, which he relates, should really have happened. I weigh the one miracle against the other; and according to the superiority, which I discover, I pronounce my decision, and always reject the greater miracle. If the falsehood of his testimony would be more miraculous, than the event which he relates; then, and not till then, can he pretend to command my belief or opinion.” The same logic applies at Lourdes. The gap between “medically unexplained” and “supernaturally caused” is not bridged by eliminating natural explanations one at a time; it is bridged only when the supernatural explanation becomes more probable than all remaining natural ones, including those we have not yet identified. The Church’s process, for all its rigour, does not attempt that probability calculation. It cannot, because no methodology currently exists for assigning a prior probability to divine intervention.
The distinction between the medically inexplicable and the supernaturally caused is not pedantic. It is the entire question. A physician who cannot explain a recovery has reached the limit of current medical knowledge. A theologian who interprets that limit as evidence of God has made a logical leap that the medical evidence itself does not authorise. The CMIL’s criteria are designed to eliminate alternative natural explanations; they are not designed to establish the probability of a supernatural one. These are different tasks, and conflating them produces the false impression that a residue of unexplained cases constitutes evidence of miraculous causation. What a residue of unexplained cases actually constitutes is a residue of unexplained cases: genuinely interesting from a clinical standpoint, but not sufficient, without further argument, to support the conclusion that omnipotent divine intervention has occurred.
2. The Base Rate Problem
Setting theology to one side for a moment and thinking purely in terms of rates, the figures produce a result that should give any defender of the miraculous serious pause. Across an estimated 200 million visitors, the Church has certified 70 miracles. That is a rate of 3.5 certifications per 10 million visitors, or approximately one certification per 2.86 million pilgrims. Expressed differently, if you have travelled to Lourdes seeking a miraculous cure, the institution most motivated to find and certify such a cure has assessed the rate of success at roughly 0.000035 percent. These are not sceptical numbers. These are the Church’s own numbers, applied to the Church’s own process, evaluated by the Church’s own criteria.
The temporal distribution of certifications compounds the picture considerably. In the first several decades after 1858, certifications came relatively quickly: the Church recognised a number of cases in the late nineteenth and early twentieth centuries as the institutional machinery for processing claims was being established and as Lourdes was being promoted by the French Church with considerable energy. The rate has since declined substantially. Across the last four decades, certifications have become extraordinarily rare, with only a handful of cases declared miraculous since 1980, and with years-long intervals between them. Sister Bernadette Moriau’s case, certified in 2018, was the first declaration in a decade. The CMIL itself has acknowledged publicly that it is receiving fewer cases that pass its preliminary criteria, not because fewer pilgrims are reporting cures, but because the bar for eliminating natural explanations has risen as medical knowledge has expanded. This is a candid admission with significant implications: as the available space for the medically inexplicable shrinks, so does the supply of miracle candidates.
Now set those 70 certifications against what one might expect from spontaneous remission alone. Spontaneous remission, the natural and unexplained regression of a disease without conventional treatment, is a documented phenomenon across a range of conditions. The oncology literature has tracked cases of spontaneous remission in cancers including renal cell carcinoma, melanoma, neuroblastoma, and certain lymphomas for well over a century. A 1966 survey by Tilden Everson and Warren Cole documented 176 cases of apparent spontaneous cancer remission in the medical literature up to that point, and subsequent compilations have added substantially to that number. The remission rates vary dramatically by condition and are in most cases very low, but they are not zero, and they occur in populations that include no religious component whatsoever. Among 200 million people who are, by the nature of the pilgrimage, disproportionately ill and suffering from serious conditions, the number of spontaneous remissions one would expect purely from natural processes across 160 years is not easy to calculate precisely, but it is almost certainly larger than 70.
This is not a point about any individual case. It is a point about the aggregate. If a hospital recorded 70 unexpected recoveries from a population of 200 million patients across 160 years, and if many of those recoveries could not be immediately explained by the treatment administered, no epidemiologist would conclude that a supernatural agent had intervened. They would conclude that spontaneous remission, misdiagnosis, delayed treatment effects, and the natural resolution of self-limiting conditions had operated within their known parameters. The Church’s process attempts to exclude these explanations for each individual case, and it does so with genuine care. But the process evaluates cases one at a time, against the state of medical knowledge at the moment of evaluation, without a control group, without randomisation, and without any mechanism for comparing the recovery rate of Lourdes pilgrims against the recovery rate of equally sick people who stayed at home. Those are not minor methodological gaps; they are the entire methodological architecture that any serious scientific claim requires.
The raw numbers also carry a quietly devastating implication for the theology of intercessory prayer that the Lourdes narrative depends upon. If prayer to the Virgin Mary at this specific site produces miraculous physical healing, one would expect the rate of certified miraculous recovery to be, at the very minimum, higher than the background rate of spontaneous remission in a general population of seriously ill people. The evidence that it exceeds that background rate has never been produced. What has been produced is a set of 70 carefully documented case studies, remarkable in their clinical detail and genuinely puzzling in certain individual instances, but far too small and far too methodologically compromised to establish a healing rate above what natural processes predict. The absence of that comparison, between Lourdes pilgrims and matched non-pilgrims, is not an oversight that future research might correct. It is a structural silence at the centre of the entire enterprise, and it has endured for 160 years.
3. What the Lourdes Medical Committee Cannot Tell Us
The CMIL is a medical committee operating within a religious framework, and it is worth being precise about what it can and cannot establish. What it can establish, when it applies its criteria carefully, is that a given reported cure is not currently explicable by reference to the medical treatment the patient received or to the known natural history of the condition as documented at the time of evaluation. That is a genuine and sometimes significant finding. What it cannot establish, by any means available to it, is the causal mechanism responsible for the cure. “Inexplicable” and “miraculous” are not synonyms, and the gap between them is not one that any committee, however constituted, can close by inspection of medical records alone.
Consider the role of the placebo effect and its cognates, processes that the CMIL’s criteria do not and cannot fully control for. The conditions under which pilgrims arrive at Lourdes are psychologically extraordinary: intense communal expectation, profound personal faith, the sustained emotional experience of collective prayer and ritual, the symbolic power of water and light, the presence of hundreds or thousands of others in similar states of hope. These are not trivial psychological variables. Psychoneuroimmunological research across the past several decades has established with increasing confidence that psychological states can produce measurable physiological changes, including transient improvements in autoimmune conditions, pain perception, and neurological symptoms. None of this is to say that every claimed improvement at Lourdes is psychosomatic in the pejorative sense of that term. It is to say that the CMIL’s process, which focuses on the documented medical state of the patient and the documented absence of conventional treatment, cannot fully account for the influence of these well-established psychophysiological mechanisms on a given outcome.
The problem of initial diagnosis is also more significant than the Church’s promotional literature tends to acknowledge. A number of the earlier certified miracles have been reassessed in the light of subsequent medical advances, and in several cases the original diagnosis appears, in retrospect, to have been uncertain or incorrect. If a patient was diagnosed with a condition that, at the time, carried a grave prognosis but which modern understanding would classify differently, the subsequent resolution of their symptoms may reflect not a miracle but a misdiagnosis. The CMIL has tightened its diagnostic requirements substantially over the decades, partly in response to exactly this problem, which accounts in part for the declining rate of certifications. This is, again, an honest and creditable acknowledgement. But it also means that some portion of the 70 certified miracles rests on a diagnostic foundation that, by the committee’s own evolving standards, might not today pass the initial filter.
There is a further and rarely discussed difficulty: the absence of data on Lourdes pilgrims who were not healed and who did not recover. The institutional focus of the Sanctuary is necessarily on reported cures, for obvious reasons. The pilgrims who return home without improvement, who continue to decline, who die within weeks or months of their visit, are not tracked in any systematic way by the Church’s apparatus. This is not a criticism of the Sanctuary as a caring institution; it is a structural observation about the limits of the evidence base. Without knowing the full range of serious cases and their subsequent outcomes, it is impossible to assess whether the rate of recovery at Lourdes differs in any statistically meaningful way from the rate of recovery among equally sick people who did not make the journey. The answer to that question, if anyone were to attempt a properly controlled study, would be the most honest measure of Lourdes’ therapeutic efficacy. To date, no such study exists, and the institution best placed to commission it has shown no inclination to do so.
Mark Twain, writing about faith healing in a different context in 1907, observed that the various cures on offer, regardless of their theological trappings, appeared to work through “the patient’s imagination.” He was describing American mind-cure movements of his era, not Lourdes, but the underlying psychophysiological observation retains its relevance. The mechanisms by which belief, expectation, and communal ritual can produce genuine physiological changes are better understood today than they were in Twain’s time, and they are sufficiently well-established to represent a serious competing explanation for any individual outcome that the CMIL evaluates. The committee’s methodology was not designed to control for them, and its published criteria do not attempt to do so. This is not a flaw introduced by bad faith; it is a structural limitation of retrospective case review, which is all the committee can conduct given the nature of the pilgrimage and the voluntary character of case submission.
4. The Theology of a 0.000035 Percent Success Rate
Let us now consider the theological implications of the Church’s own numbers, not as a rhetorical exercise but as a genuine intellectual problem for those who accept both the reality of a benevolent and omnipotent God and the authenticity of Lourdes as a site of divine intervention. If God has chosen Lourdes as a location at which miraculous healing is available, and if the institution responsible for managing that claim has identified 70 verified instances across 200 million visitors, then the God implied by those numbers is a God who selectively heals a statistically negligible fraction of the sick pilgrims who make the journey, while allowing the overwhelming majority to return home uncured or to die of their conditions. The children with terminal cancers who were carried to the baths and were not healed. The young parents with motor neurone disease who knelt in the spring water and felt nothing change. The elderly nuns who prayed in absolute sincerity and declined regardless. These are not invented figures; they are the arithmetic of the 200 million minus 70.
Defenders of Lourdes have several responses available to them, and intellectual honesty requires acknowledging the most serious of these. The first is that miracles are, by definition, exceptional acts of divine grace and not a medical treatment programme with a predictable success rate; God is not obliged to heal everyone who asks, and the pilgrimage has spiritual value independent of physical cure. This is a coherent position within the theological framework, but consider what it concedes: it concedes that one cannot go to Lourdes expecting to be healed, that the statistical expectation of a cure is vanishingly small, and that the pilgrimage’s medical claims rest on an evidentiary base of 70 cases across 160 years. Once that concession is made, the pilgrimage loses its character as a reliable avenue of supernatural healing and becomes something rather different: a site of spiritual consolation at which miraculous healing has occasionally been reported. That may be a perfectly legitimate thing for it to be, but it is not what the brochures and the fundraising appeals have historically communicated to the seriously ill who have spent their limited savings on the journey.
The second standard theological response is that God heals in ways beyond the physical, that the real miracle of Lourdes is the spiritual transformation experienced by pilgrims, the renewal of faith, the experience of community, the acceptance of suffering. Again, this is a position that deserves to be taken seriously on its own terms, and there is plainly something real being described: many pilgrims report profound personal experiences at Lourdes that have nothing to do with physical healing and that have sustained them through illness and loss. The question, however, is whether that description is compatible with the founding claim of the apparitions, which involved the Virgin Mary directing Bernadette Soubirous to a spring with explicit healing properties, and with the institutional infrastructure of baths, processions, and medical committees that the Church has maintained for 160 years. An institution that set up a medical committee to certify physical cures cannot, when the numbers prove inconvenient, retreat entirely to the claim that the real healing was always spiritual. The physical cure was the claim. The medical committee is the Church’s own attempt to verify it. The numbers belong to the Church, and the Church must account for them.
There is a third theological move that appears less often in popular apologetics but surfaces regularly in more sophisticated Catholic writing: the argument that miracles are signs rather than solutions, addressed not to the sick individual but to the broader community of faith, intended to strengthen belief rather than to function as a therapeutic service. This is an intellectually serious position, and it deserves a serious response. If miracles are primarily communicative acts, addressed to the community of believers rather than to the individual in need of healing, then the question of why God communicates so rarely and so selectively becomes acute. A God who has access to 200 million sick pilgrims at a single site over 160 years and chooses to send a confirmatory sign to 70 of them while allowing the rest to suffer is a God with a communications strategy that requires very considerable explaining. The selectivity of the miracle, on this reading, is not incidental; it is theologically central. And the criteria by which the 70 are chosen, while the remaining millions are passed over, are entirely opaque. No theological account of Lourdes has produced a satisfying answer to that question.
5. Hume’s Standard and the Evidence Before Us
Jerry Coyne, summarising the Humean position with characteristic directness, has written: “Hume was right about one thing: to have real confidence in a miracle, one needs evidence, massive, well-documented, and either replicated or independently corroborated evidence from multiple and reliable sources. No religious miracle even comes close to meeting those standards.” The 70 Lourdes certifications, assessed honestly, do not meet those standards, for reasons that have nothing to do with anti-religious prejudice and everything to do with the basic requirements of robust evidence.
Consider what the evidentiary requirements for a genuine miracle claim would look like if applied rigorously. You would need, first, a control group: a population of equally sick people, matched for condition, severity, age, and psychological state, who did not visit Lourdes, whose subsequent outcomes could be compared against those of pilgrims. You would need prospective rather than retrospective case ascertainment: systematic recording of all serious cases at the time of pilgrimage, rather than reliance on the subset who subsequently reported improvement and chose to submit their cases for review. You would need independent replication: verification of the cure by physicians with no institutional stake in the outcome. You would need, above all, some positive evidence for the proposed causal mechanism, rather than simply the absence of a currently available natural explanation. The CMIL’s process, for all its conscientiousness, provides none of these things in any systematic way. It provides detailed retrospective case studies of a self-selected group of reported recoveries, evaluated against the medical knowledge available at the time of assessment. That is valuable clinical documentation, but it is not evidence of supernatural causation.
The point about self-selection is worth dwelling on at some length, because it is the most fundamental methodological problem with the entire enterprise. The 70 certified cases were not identified by scanning all 200 million pilgrims for health changes. They were identified because individual pilgrims, having experienced what they believed to be an improvement, chose to report it and chose to submit their cases for review, and their physicians and ultimately their bishops chose to pursue the certification process. At every stage, the pathway to certification required motivated agents making positive choices to advance the claim. The pilgrims who improved but attributed their improvement to conventional treatment did not submit their cases. The pilgrims who improved but were not interested in pursuing ecclesiastical certification did not submit their cases. The pilgrims who experienced genuine but transient improvements that later resolved did not achieve certification. The resulting 70 cases are not a random sample of outcomes among Lourdes pilgrims. They are the cases that survived an institutional filter designed by an institution with a vested interest in the existence of miracles. Even granting that the CMIL’s medical evaluation is conducted in good faith, the cases that reach it have already been selected by processes that are anything but random.
This selection bias cannot be corrected after the fact by examining the 70 certified cases more closely. No amount of additional scrutiny of the surviving cases can restore the information lost by the systematic non-recording of the cases that did not survive the filter. This is not a technical quibble; it is a fundamental epistemological obstacle. The Church cannot, on the basis of its existing process, tell us whether Lourdes pilgrims recover from serious illness at a rate above the background rate of spontaneous remission, because the Church’s process was not designed to answer that question. It was designed to identify and authenticate individual cases of apparent miraculous healing, which is a related but entirely different enterprise. The distinction between these two questions, whether miraculous healing occurs at Lourdes at all, and whether individual cases can be certified as miraculous, is precisely the distinction that the Church’s institutional response has consistently collapsed, and the collapse is not innocent. Conflating these two questions allows the institution to point to the rigour of its individual case evaluation as though that rigour answered the population-level question, when in fact it does not address that question at all.
6. The Numbers in International Context
Lourdes does not exist in isolation. The Catholic Church maintains a global infrastructure for the processing of miracle claims, most notably in the context of canonisation, where evidence of miraculous healing through the intercession of a proposed saint is required for beatification and canonisation under the procedures formalised by Pope Urban VIII in the seventeenth century and substantially revised by John Paul II in 1983. The global rate of certified miracles, across all Marian apparition sites, all canonisation proceedings, and all other channels of reported supernatural healing, is not published in any single consolidated form, but the Church is simultaneously the world’s most motivated and most methodical collector of miracle claims, and the world’s most conservative certifier of them. Taken together, those two facts suggest something worth reflecting on: the institution that has the greatest incentive to find miracles and the most elaborate machinery for detecting them produces, by its own account, a rate of confirmed supernatural events that requires a denominator of millions to express in percentage terms.
Lourdes is frequently contrasted, in apologetic literature, with secular healing sites and with the claimed healing rates of other religious traditions, to argue that its record is uniquely impressive. The comparison is not straightforward. Lourdes is exceptional not because it has an unusually high rate of certified miraculous healing, but because it has an unusually rigorous institutional process for evaluating claims. Other sites produce far larger numbers of claimed cures, but without anything resembling the CMIL’s evidentiary standards. The relative rigour of the Lourdes process is, from a sceptical standpoint, one of its genuinely admirable features: it has said no far more often than it has said yes, and the nos deserve as much attention as the yeses. The question is whether the yeses, evaluated against the full context of 200 million visitors and the background rate of spontaneous remission, constitute evidence of anything beyond what natural processes alone would predict.
The international context also raises the question of comparative pilgrimage efficacy. Fátima in Portugal, Guadalupe in Mexico, and Medjugorje in Bosnia-Herzegovina each attract tens of millions of visitors and each generate substantial numbers of claimed miraculous healings. The Church has been considerably more cautious about certifying miracles associated with these sites than with Lourdes, though the psychological and social conditions of the pilgrimages are broadly similar. If divine healing is the mechanism, one would expect the outcomes to track the sincerity and desperation of the pilgrims rather than the institutional certification process of the particular site. The fact that certification rates vary so dramatically across sites, and that the most rigorous site produces the lowest certification rate, is more consistent with the hypothesis that the certification rate reflects the stringency of the evaluation process than with the hypothesis that it reflects the frequency of divine intervention.
There is an additional observation that the international context makes possible. The conditions that have attracted certified miracle status at Lourdes have, with very few exceptions, been conditions in which spontaneous remission, however rare, is at least a documented biological possibility. The list of certified miracles, available on the Sanctuary’s own website and through the CMIL’s published archives, includes cases of tuberculosis of the bone, multiple sclerosis, Addison’s disease, and the cauda equina syndrome of Sister Bernadette Moriau. These are serious conditions, and apparent recoveries from them are genuinely surprising by clinical standards. They are not, however, conditions that place complete recovery entirely outside the range of known natural biological processes. The catalogue of certified miracles does not include a single case of verified limb regeneration, and it is worth asking why not. If divine intervention is available at Lourdes, the same God who apparently resolved Sister Bernadette Moriau’s spinal condition could, in principle, restore an amputated limb. The systematic absence of such cases from the certified list is not a small observation. It suggests that whatever is producing the 70 certified outcomes operates within the constraints of known biological possibility, which is precisely what natural explanations, including spontaneous remission, would predict. The miracles, in other words, are curiously respectful of the limits of current biology.
7. The Pilgrims and the Institution
Nothing in the foregoing analysis should be read as an indictment of the people who make the journey to Lourdes. They are, in the overwhelming majority of cases, people in serious distress who are seeking consolation, community, and hope. These are entirely legitimate human needs, and the experience of pilgrimage, the shared ritual, the beauty of the location, the encounter with others in similar circumstances, the sense of being held within a larger narrative, can provide genuine comfort. The psychological and social dimensions of the Lourdes experience are real even if the supernatural dimension is not, and it would be both wrong and uncharitable to dismiss them. Many pilgrims, including those who are not healed and who return knowing they are not healed, report that the experience was profoundly valuable. That testimony deserves to be taken seriously on its own terms.
The institution, however, is not the same thing as the pilgrims it serves, and the institution’s claims are not immune from scrutiny simply because the people who make the journey are sympathetic. The Sanctuary of Lourdes is a substantial commercial and institutional operation with an annual budget running to tens of millions of euros, a marketing apparatus that reaches the sick and the desperate across the Catholic world, and an ongoing responsibility to represent its offerings accurately. An institution that implicitly or explicitly suggests to terminally ill patients that miraculous healing is available at its site, while knowing that its own certifying body has identified 70 such healings across 200 million visitors, has a serious obligation to be honest about those numbers. The gap between the implicit promise of the pilgrimage, as it is marketed to the seriously ill, and the rate that the Church’s own process has established, is not a minor discrepancy. It is a matter of moral consequence, because people make life-altering decisions, including decisions about medical treatment, on the basis of religious expectations that the evidence does not support.
The concern that the promotion of miraculous healing at sites like Lourdes may lead vulnerable people to abandon or delay conventional medical treatment has been raised repeatedly by medical ethicists and secular critics for well over a century, and the evidence that this concern has material consequences is patchy but real. There are documented cases in which individuals have forgone chemotherapy, declined surgery, or refused blood transfusions in the expectation of miraculous healing. These cases are not the norm, and the Church does not formally instruct pilgrims to abandon medical treatment. But the implicit logic of the pilgrimage, the suggestion that God can and does heal at this site, carries a corollary that is dangerous in the hands of those who take it literally. The institution cannot entirely disclaim responsibility for that corollary simply by adding a small-print disclaimer to its promotional materials, particularly when those materials have, for 160 years, placed extraordinary physical recovery at the centre of the Lourdes story.
There is a further dimension to the institutional question that deserves direct treatment. The Sanctuary of Lourdes has, at various points in its history, been promoted with language that substantially overstated the evidential basis for miraculous healing. The phrase “miraculous cures” appears regularly in Catholic promotional literature about Lourdes without the qualification that the Church’s own medical committee has certified only 70 such cures across 160 years and 200 million pilgrims. The absence of that qualification is not accidental; it is structurally necessary to the marketing of the pilgrimage to a desperate clientele. A brochure that accurately described the certified cure rate would have a considerably harder task attracting pilgrims than one that dwells on the specific cases of apparent healing without providing the denominators. This is not necessarily deliberate deception in any individual instance; it is the institutional logic of an organisation whose identity and revenue depend on maintaining the plausibility of a supernatural claim that its own evidence struggles to support. Institutional logic of that kind is, in many ways, more insidious than individual dishonesty precisely because it requires no individual bad actor to perpetuate it.
8. Science, Scepticism, and the Honest Appraisal
There is a version of the Lourdes discussion in which the sceptical case is made by ridicule: the image of peasant farmers kneeling in cold water, the credulous faithful waving crutches, the medieval atmosphere of candles and incense imposed on the modern world. That version of the argument is both easy and wrong. It is easy because the visual contrasts are vivid. It is wrong because it substitutes mockery for analysis and because it disrespects people in genuine suffering in order to score a rhetorical point. The argument from the numbers requires none of that, and it is far more powerful for its absence.
The real sceptical case against the miracle claim at Lourdes is not that the pilgrims are foolish or that the Church is dishonest. It is that the evidence, evaluated on the Church’s own terms and by the Church’s own process, does not support the conclusion the Church draws from it. Seventy certifications across 200 million visitors across 160 years, drawn from a self-selected pool of reported cures, evaluated retrospectively without a control group, from a list of conditions in which spontaneous remission is biologically possible, produces a result that is entirely consistent with what natural processes alone would predict. The hypothesis that divine intervention is responsible for these outcomes adds nothing that the data requires. It does not improve the explanatory fit. It introduces an entity, an omnipotent God who selectively heals a vanishingly small fraction of the sick pilgrims who approach him, whose properties are entirely unexplained and whose selection criteria are opaque. That is not a parsimonious explanation. It is an elaborate one that does less explanatory work than the simpler alternatives available.
This is, at its core, the same argument that has been made about supernatural claims in general and miracle claims in particular since at least the Enlightenment, and it has not grown less compelling with time. As science has expanded its understanding of spontaneous biological processes, as psychoneuroimmunology has illuminated the mechanisms by which belief and expectation can influence physiological outcomes, and as the evidentiary bar for extraordinary claims has risen in every other domain of human inquiry, the residual space available for supernatural causation has continued to contract. The CMIL’s declining certification rate is not an accident of institutional policy. It is a reflection of precisely that contraction, acknowledged in candid moments by the committee’s own members, and the direction of movement has been consistent across every decade since the early twentieth century.
Historians of science have sometimes described this process as the retreat of the gods: the progressive withdrawal of supernatural explanations from domains that natural investigation has illuminated, leaving a shrinking territory of the unexplained in which divine action is still claimed to operate. Lourdes is a case study in exactly this dynamic. In 1858, the gap between what medicine could explain and what it could not was enormous; spontaneous remission was poorly understood, psychosomatic mechanisms were barely conceptualised, and the diagnostic tools available to clinicians were rudimentary by modern standards. The miracle certifications of the late nineteenth and early twentieth centuries were produced in that context. As medicine has advanced, the CMIL’s standards have tightened and its certifications have become rarer, not because God has chosen to intervene less frequently, but because the space in which medical inexplicability can be established has narrowed. There is no reason to expect that trajectory to reverse, and every reason, given the pace of medical advance, to expect it to continue.
Clarence Darrow, writing about religion and the growth of scientific understanding, observed: “When every event was a miracle, when there was no order or system or law, there was no occasion for studying any subject, or being interested in anything excepting a religion which took care of the soul. As man doubted the primitive conceptions about religion, and no longer accepted the literal, miraculous teachings of ancient books, he set himself to understand nature. We no longer cure disease by casting out devils.” The trajectory Darrow identified in 1929 has continued in precisely the direction he described. The residual territory of medical inexplicability in which the Lourdes miracle claim operates has contracted with each decade, and the certifications have followed it downward. The relationship between the advance of medical science and the decline of Lourdes certifications is not coincidental; the causal connection runs through the progressive narrowing of the diagnostic and therapeutic gaps in which the claim of supernatural healing can be inserted without immediate falsification.
9. What Would Change the Argument
Intellectual honesty requires asking what evidence would, in principle, be sufficient to establish that miraculous healing at Lourdes is genuinely occurring at a rate above what natural processes predict. The question is worth taking seriously, not because the evidence currently approaches the required standard, but because dismissing miracle claims without specifying what would count as adequate evidence is a form of intellectual evasion that the sceptical tradition should not indulge.
The requirements are not impossibly demanding. A prospective study that systematically tracked all pilgrims with serious, documented medical conditions before and after their visit, compared their outcomes against a matched control group of equally ill non-pilgrims, and found a statistically significant excess of unexpected recoveries in the Lourdes group would constitute genuine evidence for a healing effect at the site. It would not by itself establish the supernatural cause of that effect, because the psychophysiological mechanisms described above would need to be excluded as explanations, but it would at least establish that something beyond the background rate of spontaneous remission was occurring. No such study has been conducted, despite the fact that the institutional infrastructure for it exists and the patient population is substantial. For an institution that has maintained a medical committee for over a century and that regards the miraculous healing of the sick as central to its identity, the absence of a serious prospective investigation is a choice that requires explanation, and no explanation that does not involve a reluctance to test the claim under rigorous conditions is particularly satisfying.
A single case of verified limb regeneration, independently confirmed by multiple physicians with no religious affiliation and documented through contemporaneous imaging and clinical records, would be very difficult to explain by any natural mechanism currently known. The absence of any such case from the Lourdes record after 200 million visits is, as previously noted, a striking silence that becomes more deafening as the visitor numbers accumulate. Cases of complete recovery from conditions classified by the CMIL as medically inexplicable at the time of evaluation are, as the 70 certifications demonstrate, within the range of what the Church’s own process can deliver. They are not, on current evidence, outside the range of what natural processes can also deliver. That is the core of the problem. The Church’s most impressive evidence for miraculous healing at its most famous healing site remains, after 160 years, within the territory that natural explanations can already account for.
The question of what would change the argument is also a question about the structure of the burden of proof. The Church has made a specific empirical claim: that miraculous physical healing, caused by divine intervention through the intercession of the Virgin Mary, occurs at Lourdes. That claim is extraordinary by any reasonable standard. The evidence assembled in support of it, 70 retrospectively evaluated cases drawn from a self-selected pool, evaluated without a control group by a committee with an institutional stake in the outcome, is not commensurate with the claim. The burden of proof rests with those who assert the miracle, not with those who decline to accept it, and the evidence produced so far does not come close to discharging that burden. This is not a demand for a standard of evidence that no historical claim could meet; it is a demand for the same standard of evidence that any comparable empirical claim, in medicine, in physics, in any other domain of systematic inquiry, would be required to provide before commanding reasonable assent. As this blog has argued in its broader treatment of the relationship between science and religion, the moment a religious institution makes specific empirical predictions about the physical world, those predictions must be held to the same evidential standards as any other empirical claim. There is no principled reason to exempt them.
10. The Persistence of Lourdes
None of this analysis is likely to empty the coaches and trains bound for the Pyrenees. Lourdes has endured for 160 years not primarily because of its record of certified miracles but because it serves genuine human needs that secular culture has not always found adequate ways to meet: the need for community in illness, for ritual in the face of death, for the sense of being held within a story larger than one’s individual suffering. These needs are real and serious, and they deserve a response that goes beyond pointing to the CMIL’s certification rate. The pilgrim who boards the train to Lourdes in the last weeks of a terminal illness is not primarily engaging in an epistemological exercise. She is doing something recognisably human, reaching for consolation in a form that her tradition makes available. That impulse demands respect even where the supernatural claim that packages it does not.
The question is not whether Lourdes provides something of value to its visitors, because it plainly does for many of them. The question is whether the specific claim at the heart of the Lourdes narrative, the claim that miraculous physical healing through divine intervention is available at this site, is supported by the evidence the Church has itself compiled. Assessed against that evidence, using only the Church’s own published figures and the Church’s own standards, the answer is that it is not. The evidence is consistent with natural processes operating at their known rates. It does not require, and does not establish, supernatural intervention. That conclusion is available to anyone willing to follow the Church’s own numbers to their logical destination, and the destination is uncomfortable precisely because the numbers belong to the institution that draws the opposite conclusion from them.
That conclusion carries consequences for how the pilgrimage should be represented to the seriously ill. Not every pilgrim who boards the train to Lourdes does so under the impression that miraculous healing is likely. Many go in a spirit of prayer, solidarity, and spiritual openness, with no expectation of physical cure. But the institutional framing of Lourdes, reinforced over 160 years of Church promotion, carries the persistent suggestion that healing is available there in a way that it is not available elsewhere, and that suggestion has real effects on the decisions made by sick and desperate people. The responsibility for that suggestion rests with the institution, and the institution’s own numbers are the most efficient instrument for assessing what it has actually delivered.
The secular alternative to Lourdes is not cold indifference to suffering. It is honest medicine, genuine community, and the frank acknowledgement that some questions, including the question of why a particular person must suffer and die, do not have satisfying answers. The secular tradition offers neither the consolation of supernatural intervention nor the false promise of miraculous healing; it offers instead the solidarity of people who are willing to sit with those who suffer without pretending that the situation is other than it is. That is a harder gift to package and a harder journey to sell. It does not fill coaches to the Pyrenees. But it rests on an honest accounting of what the evidence actually shows, and that honesty is itself a form of respect that the sick and the dying deserve at least as much as they deserve hope.
Seventy certifications across 200 million visitors across 160 years. The Church’s own numbers. Evaluated on the Church’s own terms. The argument is complete before a single sceptical assumption is introduced, and it stands on evidence that no defender of Lourdes can reasonably contest, because the evidence is theirs. Those who are interested in how this kind of evidential reasoning applies to religious claims more broadly will find related analysis in this blog’s treatment of the evidence on intercessory prayer and in the wider discussion of what science and religion actually say to each other. The numbers at Lourdes are unusually clean, unusually candid, and unusually damaging to the claim they were assembled to support. That the institution most invested in the outcome is willing to certify a rate of roughly one miraculous healing per 2.86 million sick pilgrims is, in the end, among the most powerful arguments sceptics have. It was handed to us by the Church itself.