Demonic Possession: What the Documented Cases Actually Show

Summary

The Catholic Church has published a list of things a possessed person can do that a sick person cannot.  That list is the reason this question can be investigated at all.  This article takes the best-documented cases of the last century, holds each one against that list, and reports what survives.

• The Church's own test is a good one, and it is empirical. The Roman Ritual tells the priest not to believe too easily, and gives three signs: speaking or understanding an unknown language, disclosing distant or hidden things, and strength beyond one's age or condition.¹ These are not feelings. They are claims about what can be recorded, transcribed, sealed in an envelope, or measured.
• No case has been found in which any of the three was documented under conditions that exclude an ordinary explanation. Repeated searching turned up no possession case with a recorded, transcribed, linguist-examined language sample; none with a hidden fact written down before it was disclosed; none with a measured force. Absence of a find is not proof of absence, and this article says so — but a claim this old, this common, and this consequential should not be this hard to document.
• The decisive experiment is already routine practice, and nobody writes down the result. Exorcists describe testing subjects with identical bottles of water, one secretly blessed. Adam Blai, who screens cases for a diocese: "They can always pick out the blessed one every time."² If that is right, twenty properly blinded trials would be the most important experiment ever run in the study of religion, and it costs a supermarket trip. No trial count, no protocol and no tally has ever been published, and in the version described only the subject is blind. France ran the controlled version once, on Marthe Brossier in 1599, with a design better than today's. A royal commission under the physician Michel Marescot used sham holy objects — unconsecrated water in a holy-water bottle, unconsecrated wafers from a box that normally held consecrated ones — over a forty-day trial, and concluded that she "reacted similarly when exposed to both genuine and sham religious objects."³ The reaction followed the label, not the blessing. That sham arm is the feature no modern protocol has.
• The spectacular claims never terminate in a first-person witness. Gallagher: "I have not witnessed a levitation myself, but half a dozen people I work with vow that they've seen it." Blai: "I haven't seen it yet, but I know a number of people that have." Rossetti: "someone said they did, but I have never seen it." Fr. Gary Thomas has had people spit stones, and knows a priest who saw nails — which then "all disappeared except one." The one exorcist who does report a first-hand levitation, Fr. Vincent Lampert, volunteers that "it could have been just something that I believe that I was seeing," and adds that the other priest in the room did not see it.⁴
• The one exorcist who kept a written case log produced the field's only real dataset, and a Catholic exorcist in it makes this article's own central criticism. An Italian diocesan exorcist recorded 1,075 consultations over ten years in a 200,000-word diary and handed it to sociologists, who published it in Current Sociology. Exorcism was performed in 5.1% of cases; of those 55 people, 40% appear to have recovered, 25% were still being exorcised, 22% left or were dropped as hopeless, and one man had received the ritual 354 times.⁵ He refused to use the diagnostic exorcism at all, on the grounds that "if you perform the exorcism, a person could act as though they're possessed, even become convinced."⁵
• The experience is reproducible in a laboratory, and the state has no physiological signature. Hypnotic suggestion plus fMRI at King's College London deliberately modelled spirit possession in fifteen volunteers.⁶ Thirteen trance channels alternating between channelling and not showed "no substantive differences… in EEG frequency power, ECG measures, GSR and respiration."⁷ There is no photograph or instrumented record of a levitation anywhere in the modern literature, and the psychiatrist who is the practice's best-credentialled defender concedes he has never seen one.⁸
• The most experienced exorcist in the world reported that the signs only ever appear after the ritual starts. Fr. Gabriele Amorth, exorcist of Rome from 1986: "these signs always surfaced during an exorcism, never before. It would be unrealistic to expect the manifestation of those signs before proceeding with the exorcism."⁹ He also held that "only through the exorcism itself can we determine with certainty whether there is a satanic influence."⁹ A test that can only be run by administering the treatment is not a test.
• Amorth's own tally is the strongest single number in this field, and it runs against possession. In nine years he "exorcised over thirty thousand people," and recorded the name of every one he judged possessed: "ninety-three so far."⁹ That is 0.3%. To Vanity Fair he put it as "one or two at the most" per hundred.¹⁰
• The famous Gary, Indiana wall-walk was tested on the spot, minutes later, and failed. The state's own case file says the attending physician "did not believe it so he asked if the child can perform this act again." They went back to the room. The boy said "I did not walk up a wall. I cannot walk up a wall," and "when he tried to walk up the wall again he could not get his balance nor place his feet properly."¹¹ The same report says his grandmother was holding his hand throughout the original event and "he never let go."¹¹ This is the single most-cited modern American case, and its one attempt at replication is almost never mentioned.
• The one exorcism ever filmed and shown to specialists produced a split verdict, and no anomaly. William Friedkin filmed Amorth in 2016 and screened it for two UCLA neurosurgeons and four Columbia psychiatrists. The neurosurgeons were startled — "This goes beyond anything I've ever experienced."¹⁰ The psychiatrists were not: "this is unconvincing as to anything that could be supernatural," and "It fits recognized psychiatric syndromes that have been defined. It's classic."¹⁰ All four agreed it was not fraud.
• Where possession reports cluster is where possession is expected. Pew asked ten countries whether they had ever seen the devil driven out of a person. Kenya 61%, Nigeria 57%, Brazil 34%, United States 11%, South Korea 6%.¹² Anthropologists have found belief in spirit possession in about 70% of 488 societies.¹³ The entity's identity, behaviour and exit always match local expectation.
• The harm is real, documented, and sometimes lethal. Anneliese Michel died of starvation in 1976 after nine months of exorcisms and no doctor. Eight-year-old Terrance Cottrell Jr. died of "mechanical asphyxia" while a minister lay across his chest for two hours.¹⁴ A Romanian nun died bound to a cross in 2005. In each case a person with a treatable illness was treated as a battlefield instead.
• None of this touches whether demons exist, and the New Testament evidence stands on its own footing. That Jesus was known in his own lifetime as an exorcist is accepted across the scholarly spectrum.¹⁵ What fails is not the doctrine. What fails is the modern evidential claim that possession can be identified today by observable signs.

Verdict:  the phenomenon is real — people genuinely lose control, in a state that is not faked and not ordinary.  Every serious observer agrees on that much, including the Columbia psychiatrists who watched Amorth work.  What is not established is the cause.  After a century in which the claim could have been settled by one recording, one sealed envelope, or one properly blinded bottle of holy water, none of those exists — and the bottle test, remarkably, is run all the time and simply never written down.  The honest position for a Christian is that Scripture teaches demons are real and that no modern case has been shown to be one.  Those are two different claims, and apologetics that runs them together is trading a doctrine anyone can hold for an evidential claim nobody can support.

What is being claimed, and how it could be settled

The Church wrote the test itself

The 1614 Roman Ritual, Title XI, opens its instructions to the exorcist with a warning, not a method.  In the Latin:  In primis, ne facile credat, aliquem a daemonio esse obsessum — "first of all, he should not easily believe that anyone is possessed by a demon" — but should know the signs distinguishing the possessed "from those suffering from some illness."¹  Then the signs:

ignota lingua loqui pluribus verbis, vel loquentem intellegere;  distantia et occulta patefacere;  vires supra aetatis seu conditionis naturam ostendere

Speaking many words of an unknown language, or understanding one who speaks it.  Disclosing distant and hidden things.  Showing strength beyond the nature of one's age or condition.  The ritual adds that these count for more the more of them coincide.

This is a remarkable document to have on the other side of the table.  A church that wanted an unfalsifiable claim would have written "the exorcist shall discern by prayer."  Instead it wrote down three things that can be checked by a tape recorder, a sealed envelope, and a force gauge.  The rest of this article takes that list at face value.

Two later documents tightened the frame.  In 1985 Cardinal Ratzinger's Congregation for the Doctrine of the Faith wrote to bishops that canon 1172 permits exorcism only with "special and express permission from the local Ordinary," granted "only to priests who are endowed with piety, knowledge, prudence and integrity of life," and that lay groups must not question demons directly or seek their identity.¹⁶  In 1999 a revised rite, De exorcismis et supplicationibus quibusdam, replaced the 1614 text;  it was approved by John Paul II on 1 October 1998 and released on 26 January 1999, eighty-four pages, in Latin.¹⁷  Cardinal Medina Estévez told reporters the rite "has not been greatly changed," that there are "very few cases" in which it is used, and that belief in the devil "is not a matter of opinion which one can accept or reject."

What would count as evidence

Four things, any one of which would settle the question in an afternoon:

1. A recording. Twenty seconds of a subject holding a two-way conversation in a language they have never learned, transcribed and checked by a linguist who did not know what the recording was.
2. A sealed envelope. A fact written down and sealed before the session — the contents of a stranger's pocket, a number chosen by someone in another building — then stated by the subject.
3. A measurement. A force plate, a load cell, a strain gauge. Emergency departments restrain violently agitated patients every day and know what human maximum looks like; a number outside that range would be decisive.
4. A blinded bottle. Identical containers, some holy water and some tap, labelled by someone who then leaves the room. This costs nothing and takes ten minutes.

None of these has been done and published.  That absence is the central fact of this article, and everything below is an attempt to explain it fairly.

How common is this

Belief is widespread and stable.  YouGov found in 2013 that 51% of Americans believed in possession by evil spirits and 57% in the devil, with born-again Christians at 72%;  among believers in possession, 45% thought it happened "rarely."¹⁸  A 2020 YouGov survey of more than 17,000 US adults found 50% saying demons definitely or probably exist — 66% of Protestants, 58% of Roman Catholics, 10% of atheists.¹⁹

Firsthand experience is far less common, and its distribution is the interesting part.  Pew's ten-country survey asked whether respondents had "experienced or witnessed the devil or evil spirits being driven out of a person":¹²

A Kenyan is ten times more likely to have seen an exorcism than a South Korean.  Within every country, pentecostals report it several times more often than other Christians.  Erika Bourguignon's cross-cultural survey of 488 societies found about 90% with some institutionalised altered state and about 70% associating such states with spirit possession.¹³

Two readings are available.  Either demons work hardest where they are believed in, or possession behaviour is learned.  The rest of the evidence has to decide between them.

Reading the exorcists

Before the cases, it is worth listening to the men who do this work.  They are not a monolith, they are mostly not frauds, and — this is the part that surprises people — they are the strictest gatekeepers in the subject.  The numbers below are theirs, not a critic's.  Several of them speak mainly in long recorded interviews rather than in print, so those are quoted here from the recordings, with video identifiers and timestamps;  the method and its limits are set out in a note.²⁰

How many of the people who come to them are possessed:

The fourth row is the most valuable line in this article, because it is the only number in the field that comes from a written record instead of a memory.  An Italian diocesan exorcist kept a 200,000-word diary of every case he handled — personal details, reason for coming, occupation, age — and gave it to Giuseppe Giordan, a sociologist at the University of Padua, who analysed it with Adam Possamai and published the result in Current Sociology.⁵

The published figures:  1,075 consultations over ten years;  55 exorcism rituals performed, or 5.1%.  Of those 55, 60% were men and 56% were aged forty to forty-nine — the male majority running against the standard anthropological account of possession as a refuge of the powerless.  The exorcised "mostly belong to the working class and lower middle class;  there are only six cases of professionals/graduates."⁵  Fifty-six percent of all consultations were for reasons the authors classify as religious or magical — evil influences, occultism, Satanism, paranormal phenomena — and the rest for family, health or personal problems.

One correction is worth making in the open, because it bears on how any of this should be used.  In a later interview Giordan recalled that among those judged possessed "they were really educated."  His own published table says the opposite.⁵  Where a researcher's recollection and his data disagree, the data wins — and the episode is a small demonstration of why this article keeps insisting on written records over remembered ones, including when the person remembering is a professional with no reason to shade anything.

Lampert says he receives roughly ten emails a day.²¹  On his own figure, a working exorcist could screen for a decade and see a handful of genuine cases.  Whatever else is true, the exorcists are not the credulous party;  the overwhelming majority of what reaches them, they send away.

Where they send it.  Fr. Carlos Martins, whose podcast is among the most listened-to religious shows in the world, describes requiring a full medical and psychiatric workup first — "give me the landscape of this individual" — while insisting the two are not exclusive, since "someone can have mental illness and be possessed."²³  Lampert tells a story about a man he judged not possessed, in a meeting he convened with the man's psychiatrist and caseworker.  The man's reply is the most humane thing in this whole literature:

You, as a psychiatrist, can put a label on me and tell me that I'm schizophrenic, but you can't tell me why.  If it's the devil, I have my why.²¹

That sentence explains more about the persistence of possession belief than any theory of suggestion does.  A diagnosis names a condition.  A demon gives it a reason, an enemy, and a way to fight back.  Any account of this subject that cannot see why that appeals is not paying attention.

And one written record lets that claim be checked.  The Padua diary shows what a careful exorcist actually recommended across 1,075 consultations:  a psychologist in 140 cases (13%), some medical service in 5 (0.5%) — and, far more often, religious remedies:  "religious homework" in 500 cases (46.5%), blessings in 206 (19.2%), prayer in 188 (17.5%), rituals of liberation in 142 (13.2%).⁵  This is the exorcist the authors class as the institutional, cautious type, the one who worked with a psychiatrist and refused the diagnostic exorcism.  Even so, five out of six people who came to him were sent away with a religious practice rather than a referral.  The gatekeeping is real, but it gatekeeps entry to the ritual, not entry to care.

Where the criteria go soft.  Two things in the practitioners' own descriptions should worry a careful reader.  Amorth held that the three signs "always surfaced during an exorcism, never before."⁹  And Martins offers a heuristic that runs the other way from the obvious one:

People who generally are mentally ill, they'll come to you right away and say "I'm not crazy."  But people who are demonically oppressed, they often will say "I think I'm just crazy, I think there's just something mentally wrong with me."²³

Taken as a rule, that makes both answers evidence.  Insisting you are not ill counts against you;  suspecting you are ill counts for you.  A criterion with no possible disconfirming answer is not a criterion, and this one is being offered by a serious practitioner as a mark of the demonic "more often than not."

Numbers at the other end of the scale are worth recording too.  Msgr. Stephen Rossetti has said that perhaps "one out of every four Americans are experiencing demonic obsession" — a lesser category than possession, but a claim about roughly eighty million people, offered without any stated basis.²⁴  The gap between that estimate and Lampert's one-in-five-thousand is not a disagreement about a detail.  It is a field without a measuring instrument.

The documented cases

The cases below are the ones with the best paper behind them.  They are a selection, and a reader is entitled to ask whether it is a fair one, so the full working register — every modern possession and infestation case with at least one named witness from outside the household, nineteen of them, scored on what survived investigation — is published alongside this article rather than kept in a drawer.

Gary, Indiana, 2011–2012: the Ammons family

This is the best-documented modern American case, and it is documented because the state was involved.  Marisa Kwiatkowski of the Indianapolis Star built her January 2014 account from "nearly 800 pages of official records" and more than a dozen interviews with police, child-welfare staff, psychologists, family and a priest.²⁵

Better still, the underlying records are public.  The Star published the Department of Child Services file, the police excerpts and the priest's report to his bishop as scanned documents, so the case can be read in the participants' own contemporaneous words rather than through anyone's retelling.¹¹  What follows is taken from those documents.

What the case file actually says about the wall.  The scene is a small exam room at Northlake Methodist Hospital in Gary on the evening of 19 April 2012.  Present:  family case manager Valerie Washington, psychologists, and the boys' grandmother Rosa Campbell.  Washington's own report:

[He] had the weird grin on his face and began to walk backwards while the grandmother was holding his hand and he walked up the wall backwards while holding the grandmother's hand and he never let go.  He flipped over and landed on his feet in front of the grandmother and sat down in the chair.  A few minutes later he looked up as if he was back to himself.¹¹

Read that slowly.  The grandmother is holding his hand the entire time, and he never lets go.  A boy who walks his feet up a wall while an adult holds his hands and steadies him is doing something most agile nine-year-olds can do;  it is a playground trick, not a suspension of gravity.  The later newspaper phrase "up a wall to the ceiling," and the New York Daily News version in which demons made him "walk on a hospital ceiling," are amplifications of a report that describes a supported climb ending in a flip onto his feet.

And then somebody tested it.  This is the part that almost never survives into the retellings, and it is in the state's own file:

FCM Washington and psychologists left the room immediately and went to the attending doctor, Dr. Richardson, and told him what was just witnessed in the Ammons family room.  Dr. Richardson did not believe it so he asked if the child can perform this act again, the psychologist stated I doubt it seriously, this kid was not himself when he did that.  Dr. Richardson, FCM Washington, and the psychologist went back to the room to ask [if] he can walk up the wall again?  [He] stated "I did not walk up a wall.  I cannot walk up a wall".  When he tried to walk up the wall again he could not get his balance nor place his feet properly to walk up the wall.¹¹

An attending physician heard an extraordinary claim, immediately demanded a repeat, and got a failure — within minutes, in the same room, in front of the same witnesses.  The boy denied he had done it and then could not do it.  That is the only controlled test anywhere in this case, and it is negative.

The rest of the file.  The report that brought DCS in came from a medical professional at the doctor's office, recorded in the file as the Report Source:  "RS states they believe the children are performing for the mother and that she's encouraging the behavior," and "RS states they believe the mother may have mental health concerns."¹¹  The same source noted that "a month ago everything was relatively normal."  Two clinical psychologists then evaluated the children.  Stacy Wright found the youngest boy tended to act possessed "whenever he was challenged, redirected, or asked questions he did not wish to answer," that he was coherent and logical except when talking about demons, and that his demon stories were "bizarre, fragmented and illogical" and changed with each retelling;  her conclusion was "an unfortunate and sad case of a child who has been induced into a delusional system perpetuated by his mother and potentially reinforced" by relatives.²⁵  Joel Schwartz reached a similar conclusion about the other two children.  The family also had a prior DCS finding for educational neglect from 2009, and Ammons told the agency the children missed school because "the spirits would make them sick, or they would be up all night without sleep."²⁶

The occult context came first, not last.  Two clairvoyants told the family the house held more than 200 demons;  on their advice Ammons built an altar in the basement, burned sage and sulfur, and anointed the doors, windows and blinds with oil in the sign of the cross.  The possession behaviour escalated after that.

The oil, which looked unexplained and is not.  Officers wiped an oil-like substance from bedroom blinds with paper towels, sealed the room for half an hour, and found more oil on their return.  That is a real observation and it sounds decisive — until you read the priest's own report to the bishop, where the explanation sits three sentences away from the claim.  Fr. Michael Maginot noticed the oil on Latoya's blinds and asked about it:  "she mentioned that it was oil and that she made the sign of the cross on all the windows, doors and blinds about a month ago where it had dried."  He then reports that oil "seem[ed] to condense and drip unto the steps from the crosses she made above," and that when the furnace kicked on he could see its vent moving the blind cords, and when it kicked off they stopped.²⁷  Oil applied a month earlier, soaked into slats and cords, warmed by a heating vent directly below the window, will keep migrating out after a wipe.  Maginot wrote down both the cause and the effect and did not connect them.

Everything else deflated too, mostly by the newspaper itself.  Capt. Charles Austin's self-moving car seat:  his mechanic found the driver's seat motor was simply broken.  The photograph of a figure in a window that circulated captioned "Photo by Hammond Police" was not a police photograph;  Joe Nickell, who went to Gary for the Committee for Skeptical Inquiry, was told so by the Hammond police chief, whose department had no involvement in the case.  The "EVP" recording of a voice saying hey was an aspiration close to the microphone.  Nickell also found that no haunting was reported in the house before the Ammons family moved in or after they left — the landlord, Charles Reed, was firm on that — and that the later tenant's only trouble was sightseers.²⁶

What was done with it afterwards.  Fr. Maginot performed the exorcisms on Latoya Ammons — not on the children who were said to be possessed — and then signed with Zak Bagans of Ghost Adventures and with the production company behind The Conjuring.  Capt. Austin told the Star he "expected notoriety and figured a movie would come of this."  Ammons signed releases opening her family's medical, psychological and social records, which is the only reason any of this can be checked.²⁶  The children were returned to her in November 2012;  the wardship dismissal notes that "The family is no longer fixated solely on religion to explain or cope with the children's behavior issues."

Where that leaves the case.  A sincere family in genuine distress, three frightened children, a state file, and a claim that was tested once and failed.  Nothing in the record requires a demon, and the one document usually cited as proof — Washington's report — contains both the wall-walk and the failed repeat, four sentences apart.

Klingenberg, Bavaria, 1975–1976: Anneliese Michel

Anneliese Michel matters because two priests, both her parents, a bishop's authorisation, and tape recordings of the sessions ended in a criminal trial where the evidence was cross-examined.

The medical history came first.  Her seizures began in 1968, around her sixteenth birthday;  the first episode was preceded by a night in which she "felt as if something was pressing down on her.  She couldn't move, couldn't breathe or speak, and lost control of her bladder."²⁸  Her neurologist judged it "probably a case of cerebral seizures of the nocturnal type, with the symptoms of a grand mal epilepsy," though her EEGs showed nothing pathological.  She was prescribed anticonvulsants, including Tegretol.  Depression followed, and statements — "someone else is manipulating me," "My will is not my own" — that a psychiatrist reads as depersonalisation and a priest reads as occupation.  Her interests narrowed to religion;  she withdrew from friends;  she reported smells nobody else could detect.  Hyperreligiosity, reduced sexual interest, and social withdrawal are a recognised behavioural cluster in temporal-lobe epilepsy.

Bishop Josef Stangl authorised the major rite in September 1975.  He did so on the strength of Fr. Ernst Alt's letters;  he never requested her psychiatric records or a second opinion from her neurologist.²⁸  Sessions ran over the next nine months — the number is usually given as sixty-seven, though no source read for this article states it.  She stopped eating.  On 30 June 1976 she weighed between sixty-eight and seventy-two pounds;  she died the next morning.  The autopsy found advanced emaciation from malnutrition and dehydration, along with fractured teeth, bruising, and knees broken by genuflecting an estimated four to six hundred times a day.

Two tests were attempted during the exorcisms, and both belong in this article.

The language test failed.  Fr. Arnold Renz questioned the "demon" in Chinese.  The reply:  "If you want to ask something, ask it in German" — followed by "But I did understand that."  No glossolalia was recorded across sixty-seven sessions.  The demons who identified themselves as Judas, Nero, Cain and Hitler spoke German, and the Latin was a language Anneliese knew from church and school.²⁸

The holy-water test was run, and not controlled.  Renz filled five unmarked bottles, some with tap water and some with holy water, and she reacted only to the holy.  No record exists of any control against her learning which was which — the ordering, the handling, whether the person sprinkling knew, whether anyone tested her when the answer was known to nobody present.²⁸  This is the single easiest experiment in the whole subject, someone did attempt it, and the protocol was not written down.  That is the story of this field in one sentence.

The demons' opinions are the other tell, and here the primary record is available:  the tapes were transcribed and published, with a bishop's imprimatur, as Satan's Answers to the Exorcists.²⁹  What the damned turn out to want is a traditionalist Catholic policy programme.  "The Communion in the hand should be abolished.  This is my work.  The Bishop should now forbid Communion in the hands, if he had permitted it before."  "Yes, you have to follow the message of Fatima!"  "The Encyclical Humanae Vitae is decisive, there is no other way!  The whole Humanae Vitae!"  "One must pay obedience to the Pope."  There is a long passage defending Archbishop Marcel Lefebvre, then under Roman discipline, for recognising in the liturgical changes "a departure from tradition and an adaptation to the Protestant cult."²⁹

Two things follow.  Theologically, this is incoherent:  demons campaigning for orthodoxy, Fatima devotion and papal obedience are not doing the devil's work as any Christian tradition understands it.  Evidentially, it is exactly what the alternative explanation predicts — the positions are those of Anneliese's own traditionalist circle in mid-1970s Bavaria, down to the factional detail of the Lefebvre dispute.  A demon that shares its host's stance on communion in the hand is not evidence of a demon.

At trial in 1978 the two priests and both parents were convicted of negligent homicide and given suspended sentences.  Fr. Renz's defence has been quoted since:  "The exorcism ritual expressly states that the clergymen should not burden themselves with medical matters."²⁸  He was reading the 1614 rite correctly.  It told the exorcist to leave medicine to physicians.  In this case there were no physicians left.

The anthropologist Felicitas Goodman argued in 1981 that Anneliese was not sick at all, and that what looked like disease was religious experience.²⁸  Her case rests on EEGs and an autopsy that found no temporal-lobe abnormality — but in roughly a quarter of temporal-lobe epilepsy cases no structural cause is ever found, so a clean scan is not a clean bill of health.  Goodman also records, without reconciling it, both that Anneliese "continued taking the drugs conscientiously" and that she "often took less than the three tablets per day (of Tegretol) when her prescription was beginning to run out."

Cottage City, Maryland, 1949: the boy behind The Exorcist

Every retelling of the 1949 case descends from a diary kept by Fr. Raymond Bishop, and from a Washington Post story of 20 August 1949 headlined "Priest Frees Mt. Rainier Boy Reported Held in Devil's Grip."

Both are less than they look.  Mark Opsasnick's 1999 investigation for Strange Magazine established that the boy never lived in Mount Rainier at all;  the family home was at 3807 40th Avenue, Cottage City.  Working from a Fate magazine birth date, a 1954 Gonzaga High School yearbook, Maryland birth records and interviews with long-time residents, Opsasnick identified the boy;  a later chain of confirmations — Sergio Rueda's 1990 interview with the Lutheran minister first consulted, Ancestry records, and Fr. Bishop's own unredacted cover letter — settled it as Ronald E. Hunkeler, born 1 June 1935, died 10 May 2020.³⁰  Opsasnick, asked directly after Hunkeler's death:  "The Haunted Boy is Ronald Hunkeler.  That's a fact."

The decisive point for evidence is smaller and duller.  Fr. Bishop was not there for the first two months.  He was brought in when the decision to exorcise was made on 9 March 1949, so the "Background of the Case" section of the primary document — the rapping, the moving bed, the flying objects — is somebody else's account written down by a man who arrived later.³⁰  The onset itself is a familiar shape:  a boy whose recently dead aunt, a spiritualist, had taught him the Ouija board.

Fr. Hughes reportedly told a Mount Rainier neighbour that the boy "went on to graduate from Gonzaga High and turned out fine."³⁰  He worked a long career and died at eighty-four.  Long-time residents of the actual street, interviewed fifty years later, recalled fire-department call-outs to a "commotion" that had always ended by the time the truck arrived, and a boy who struck them as unremarkable.³¹

The psychiatrist's case: Richard Gallagher

Richard Gallagher is the strongest witness the affirmative side has.  He is a board-certified psychiatrist, Princeton and Yale trained, with Columbia psychoanalytic training and a faculty post at New York Medical College, and he has spent decades screening referrals for exorcists.  His July 2016 essay in the Washington Post is the most-read modern statement of the case.

His account of the woman he called Julia:

So I was inclined to skepticism.  But my subject's behavior exceeded what I could explain with my training.  She could tell some people their secret weaknesses, such as undue pride.  She knew how individuals she'd never known had died, including my mother and her fatal case of ovarian cancer.  Six people later vouched to me that, during her exorcisms, they heard her speaking multiple languages, including Latin, completely unfamiliar to her outside of her trances.  This was not psychosis;  it was what I can only describe as paranormal ability.  I concluded that she was possessed.⁸

Take the three claims in order.  "Undue pride" is a weakness most adults will accept as their own;  it is the standard currency of cold reading.  Knowing how a stranger's mother died is striking, but the essay does not say when Gallagher learned that the statement was correct, whether the woman had any route to that information, or whether anything was written down beforehand.  And the languages are not his observation at all — "six people later vouched to me."  Nothing was recorded.

On his most spectacular category, Gallagher is candid, and the candour is the point:

(I have not witnessed a levitation myself, but half a dozen people I work with vow that they've seen it in the course of their exorcisms.)⁸

His book gives the levitation in full, and it is worth reading closely, because the detail undercuts the claim.  Gallagher was not present;  the exorcist told him over the telephone, "Julia levitated right out of her chair.  For half an hour."  Gallagher's own narration continues:  "Julia visibly rose about a foot off the chair and, in the clear impression of all attendees, would have ascended higher if it hadn't been for several of those present, including all the men, laboring mightily to hold her down."  And a few lines earlier:  "Fortunately, her arms and hands had been restrained."³²  So the scene is a restrained woman with several men pressing down on her, for half an hour, and the levitation is an inference from upward force felt through their arms — "the clear impression of all attendees" — not an observation of a body hanging unsupported in the air.  A person straining upward against restraint produces exactly that impression.  This is the best-attested levitation in the modern literature, and it is a second-hand report of a woman being held down.

And on why none of it is on tape:

One cannot force these creatures to undergo lab studies or submit to scientific manipulation;  they will also hardly allow themselves to be easily recorded by video equipment, as skeptics sometimes demand.⁸

Three clinical psychologists writing in Skeptical Inquirer — Dean McKay, Rachel Ammirati and the late Scott Lilienfeld — made the obvious replies:  cold reading and hot reading are not exotic hypotheses;  the eyewitness-memory literature is not kind to emotionally charged group recollection;  and "can be explained no other way" is a description of the observer, not the event.³³  Steven Novella added the empirical objection that costs nothing to check:  "I have literally watched dozens of hours of exorcisms on video.  They are all incredibly boring.  Nothing interesting happens.  No levitations."³⁴

Gallagher's position deserves one more sentence of fairness.  He is not credulous about individual referrals — he says the overwhelming majority are psychiatric — and he is not making money from this.  What he is doing is treating a convergent body of second-hand testimony as though it were evidence of a different kind than it is.

Venafro, Italy, 2016: the only exorcism filmed for scientists

In May 2016 William Friedkin, who directed The Exorcist, filmed Fr. Amorth performing the ninth exorcism of an Italian woman, then showed the footage to specialists.  Amorth's own screening rule, as Friedkin reports it, is the base rate this article keeps returning to:  "Out of a hundred people who seek my help," he explained, "one or two at the most may be possessed."¹⁰

Neil Martin, chief of neurosurgery at UCLA, with more than 5,000 brain surgeries behind him:

Absolutely amazing.  There's a major force at work within her somehow… I've done thousands of surgeries, on brain tumors, traumatic brain injuries, ruptured brain aneurysms, infections affecting the brain, and I haven't seen this kind of consequence from any of those disorders.  This goes beyond anything I've ever experienced — that's for certain.¹⁰

Itzhak Fried, an epilepsy surgeon at UCLA and Tel-Aviv Sourasky, was more specific and less impressed:

It looks like something authentic… I believe everything originates in the brain.  So which part of the brain could serve this type of behavior?  The limbic system… and the temporal lobe.  I don't see this as epilepsy.  It's not necessarily a lesion.  It's a physiological state.  It seems to be associated with religious things.  In the temporal lobe there's something called hyper-religiosity.  You probably won't have this in somebody who has no religious background.  Can I characterize it?  Maybe.  Can I treat it?  No.¹⁰

Then four Columbia psychiatrists watched the same tape for ninety minutes.  Jeffrey Lieberman, director of the New York State Psychiatric Institute:  "To be perfectly blunt, this is unconvincing as to anything that could be supernatural or excused from the laws of nature as we know them."  Asked whether it was fraud, all four said no.  Michael First:  "It fits recognized psychiatric syndromes that have been defined.  It's classic.  I would say she fits into the pattern that we call Dissociative Trance and Possession Disorder.  There is no obvious known psychopathology.  Exorcism as a therapeutic technique could work."¹⁰

That is the whole modern case in one room.  A camera was finally present.  Two brain surgeons had never seen anything like it.  Four psychiatrists recognised it immediately.  Nobody thought it was faked.  And the tape contains no unknown language, no hidden knowledge, no measured strength, and no levitation.

Older cases, and how legends are built

Two frequently cited cases are worth naming for what their evidence chains look like rather than for what they contain.

The Loudun possessions in France (1632–1637) ended with Urbain Grandier burned at the stake on 18 August 1634 on the testimony of possessed Ursuline nuns.  A contemporary physician, Marc Duncan, published a dissenting analysis in 1634.  The superior of the convent, Jeanne des Anges, later toured France displaying sacred names that had appeared on her hand, which contemporaries accused her of painting.  The exorcist Jean-Joseph Surin subsequently reported being possessed himself for two decades.  Whatever else Loudun was, it was a legal proceeding in which possession testimony killed a man.

The Earling, Iowa case of 1928 — levitation to the ceiling, hidden sins revealed — descends from a single devotional pamphlet, Carl Vogl's Begone Satan!, published in 1935.  Every later retelling traces back to it.  When a spectacular claim has exactly one source, and that source is a pamphlet written to edify, it is not evidence of the event;  it is evidence that the pamphlet exists.

Malachi Martin's Hostage to the Devil (1976), still the most-read possession book in English, presents five American cases with names, dates and dioceses changed.  That choice makes every claim in it unfalsifiable by design, and Martin's own biography was disputed during his lifetime — including by Robert Blair Kaiser, who covered the Second Vatican Council and later published an account hostile to him.  The book should be read as testimony about a genre, not as a case file.

Auditing the three signs

Speaking an unknown language

This is the sign most often claimed and least often documented.  It also has a real research literature, because Ian Stevenson at the University of Virginia spent years on it — not in possession cases, but in claimed past-life cases, where he distinguished recitative xenoglossy (reciting memorised fragments) from responsive xenoglossy (genuinely conversing).

The linguist Sarah Thomason examined Stevenson's best cases.  Her conclusion:  "although fraud can probably be discounted in all these cases, the linguistic evidence is too weak to provide support for the claims of xenoglossy."³⁵  In the "Jensen" case, the subject's answers "are one- or two-word utterances, with no complex sentences at all."  The vocabulary ran to roughly a hundred words, against the thousands any native speaker commands.

There is one further datum, and it is the most honest number available.  In eleven hundred cases over ten years, the Padua exorcist recorded three or four occasions on which someone spoke "in perfect Latin, without any reason to know it."⁵  Take it at face value:  that is a rate of about 0.3%, in a Catholic country, in a rite conducted in Latin, with no recording, no transcript and no linguist.  Giordan's own summary of the whole diary is "nothing so exceptional."

That is the best-investigated corpus in the entire field, and it fails.  Possession claims are weaker still, because they have produced no corpus at all:  no transcript, no recording submitted to a linguist, no test in which someone addressed the subject in a language chosen by an outsider.  The one time that test was actually run — Fr. Renz's Chinese question to Anneliese Michel — the answer came back in German.²⁸

Knowing hidden things

Every claim under this heading in the modern literature shares one structural feature:  the "hidden fact" is identified as hidden only after the subject states it.  Gallagher's mother's cause of death, the guilty secrets named in front of a room, the "you had a fender bender on Tuesday" anecdote that circulates in exorcist lectures — in each, the record begins with the utterance and the verification comes afterwards, from people who were present, invested, and frightened.

The protocol that fixes this is trivial:  write the fact down first, seal it, and open it after.  Stage mentalists have made a living off the difference for two centuries, and the reason is that cold reading, hot reading, and generous scoring of near-misses reliably produce exactly this pattern.  No published possession case has run the sealed-envelope version.

Strength beyond one's age or condition

Emergency departments restrain agitated, delirious, and intoxicated patients daily, and clinicians know what an adrenalised human can do:  it is a great deal more than a calm one and never more than a human.  The claim under investigation is that some subjects exceed that range.  Testing it requires a number.

No possession case in the modern literature reports one.  What the literature reports instead is impressions formed by people already straining to hold someone down.  Novella's observation about the video record applies here as well — televised exorcisms show subjects restrained by unremarkable people who do not appear to be struggling unusually hard.

Levitation, and the video problem

Levitation is not in the Roman Ritual's list, but it dominates the popular claim, and it is the cleanest test imaginable:  it either happens on camera or it does not.  There is no photograph or instrumented record of a possession levitation in the modern literature.

What there is instead is a chain of testimony that never terminates.  Ask the practitioners directly and the answer keeps taking one step backwards:

• Richard Gallagher, the psychiatrist: "I have not witnessed a levitation myself, but half a dozen people I work with vow that they've seen it."⁸
• Adam Blai, who screens cases for a diocese: "when we talk about levitation and exorcism, which is really rare — I haven't seen it yet, but I know a number of people that have seen it, that, you know, exorcists I know well."²
• Msgr. Stephen Rossetti, until 2026 the chief exorcist of Washington: "people do levitate… Now, I've never seen someone's head turn 360, although someone said they did, but I have never seen it."²⁴
• Fr. Chad Ripperger, on demonic infestation of animals: "I haven't seen them actually infest the animals… But I have heard of other exorcists having that."³⁶
• Fr. Gary Thomas, the American exorcist whose training was the basis of the book and film The Rite: "I've never had somebody… spit nails, but I have had people spit stones. But I know of a priest who experienced the vomiting of nails and who told me that himself — and then the nails all disappeared except one."³⁷

Everyone has a colleague who saw it.  Notice also the shape of Thomas's account:  the ordinary version happened to him, the extraordinary version happened to someone else, and the physical evidence for the extraordinary version disappeared afterwards — except for one item, which is not said to have been examined by anybody.  In a field with an international association, an annual course in Rome and hundreds of appointed exorcists, that structure should be surprising.  It is what a rumour looks like from the inside.

And then there is the exception, which is more valuable than all of it.  Fr. Vincent Lampert, exorcist of the Archdiocese of Indianapolis and the most-quoted exorcist in America, says he personally witnessed a levitation.  Asked by an interviewer what a natural explanation might be, he gave one himself, unprompted:

It could have been just something that I believe that I was seeing, because at the time the person was shaking very violently and then I saw them rise up out of the chair, and then that's when the priest reached over and put his hand on the head of the person and pushed them back in the chair.  Now, there was another priest in the room, and we talked about it later, but he didn't see it.  So I saw it, he didn't.⁴

A violently convulsing person, in a chair, momentarily rising — and the other priest in the room did not see it.  That is the best first-person levitation testimony this article could find anywhere, offered by a man with every reason to state it more strongly, and he states it honestly.  Credit is due for that, and the answer it gives is clear.

The standard explanation for the absence of recordings — that demons will not permit it — has to be weighed honestly.  As theology it is coherent:  intelligent agents can decline to perform.  As evidence it is worthless, because it predicts the same observations as the hypothesis that nothing is there.  But it is worth noting that the practitioners themselves mostly do not use it.  Blai gives a mundane reason instead:  "There's no video recordings because you're not allowed to videotape exorcisms.  You're not allowed to audio record them," and the reason he gives is the sufferer's privacy — you do not record a person's worst hours for your own benefit.²

There is also a precedent that cuts through the whole debate about whether recording is permitted, because it was permitted once, at the highest level, for exactly the reason a believer would want.  In October 1990 the Diocese of Palm Beach authorised ABC's 20/20 to film a full rite, after a six-month investigation of the case, in order — as the Associated Press put it — to dramatise "the reality of evil in the modern world."  It aired in April 1991 to about 29 million viewers.³⁸

What the tape showed is the point.  A sixteen-year-old girl, called Gina, strapped to a chair, "barked and babbled gibberish like 'sanka dali' and 'booga' while Father A pressed a cross to her face" — sounds, not a language.  The priest to the supposed demon:  "You want pain?  I'll give you pain."  Newsweek's verdict on the broadcast was that the ceremony "seemed at times like little more than the gratuitous torment of a deeply disturbed young girl."  Within days of being sent home she was rehospitalised and given antipsychotic drugs.  Her psychiatrist at Miami Children's Hospital, Dr. Warren Schlanger, who treated her before and after, said the ceremony had been "a significant risk" to her health.  A Boston College ethicist asked the obvious question about a hospitalised psychotic teenager on national television:  "Was she capable of informed consent?"³⁹

So the one time a camera was invited in with a bishop's blessing, it recorded no unknown language, no hidden knowledge and no preternatural strength — and the aftermath was a psychiatric readmission.  That is a data point about what the ritual looks like when it is documented, and it did not need a demon to explain it.

Blai's privacy answer is a respectable one, and it is the right instinct about a vulnerable person.  It is also a rule the Church made and could change tomorrow for one consenting adult.  Medicine records patients constantly with consent, faces blurred, for far less.  If a person really were rising off a chair, the case for a single consented recording — which would end a two-thousand-year argument — is overwhelming, and the person on the chair is described as a victim being freed, not a patient being exposed.  The privacy rule explains the absence of footage.  It does not explain why nobody has sought an exception.

The one cheap experiment — which turns out to be standard practice

Aversion to concealed sacred objects is the sign that is trivially testable, and it is the most interesting thing in this whole subject.  Identical bottles.  A neutral party labels them and leaves.  Nobody in the room knows which is which.  Twenty trials.

Exorcists say they run this test routinely, and that it works every single time.

Adam Blai is a lay expert approved by the Diocese of Pittsburgh who holds a master's in psychology and does intake screening;  he is describing standard practice, not a stunt.  In one interview:

Detecting the holy — knowing whether something's blessed or not.  You put out 10 bottles of water, all identical, one of them was blessed before the person arrived.  Every single time they can pick up "that's the one I don't like, I can't touch that one," and all the others are fine.  There's many variations on detecting the holy.⁴⁰

In another, with a different host and a different number of bottles:

You can line up six bottles of water, one of which was blessed in the other room out of the person's sight.  They're all identical.  They can always pick out the blessed one every time.  They can tell you whether there's a relic in that box or not.  I've had cases where they've been able to describe the saint whose relic it is without even seeing it.²

This changes the shape of the argument, and it deserves to be stated plainly rather than explained away.  The claim is not that the phenomenon is untestable.  The claim is that the decisive experiment is already part of the working method, has been run many times, and produces a perfect hit rate.

Which makes one question the whole ballgame:  where is the record?

Not a story about it.  A record:  how many bottles, how many trials, who prepared them, who carried them into the room, whether that person knew which was which, whether the order was randomised, how many hits and how many misses, and what the subject was told beforehand.  Twenty trials of a six-bottle test at chance gives about three hits.  Twenty out of twenty would be a result of roughly one in a hundred million, obtained in an afternoon, with materials from a supermarket.  It would be the most important experiment ever conducted in the study of religion, and any exorcist could run it tomorrow.

No such record exists in the published literature.  This is a search result and it should be auditable, so:  PubMed and OpenAlex were both queried for a blinded or sham-controlled test of holy-water or sacred-object discrimination in possession, under several phrasings, and returned nothing on the subject at all.⁴¹  What exists is the practice, described in interviews, with a remembered result.

And the protocol as Blai describes it has a specific hole.  The water is "blessed in the other room out of the person's sight," which controls what the subject knows and nothing else.  Everyone in the room knows which bottle is the blessed one.  In a setting where the subject is being intently watched by people who already believe, an unblinded experimenter is the oldest confound in the book — the same one that made a horse in Berlin look as if he could do arithmetic.  Blinding the experimenter costs nothing:  a third party numbers the bottles, records the key on paper, seals it, and leaves the building.

The historical attempt has the same defect.  Fr. Renz's five unmarked bottles at Klingenberg is a real experiment, and Anneliese reportedly reacted only to the holy water — but nothing was recorded about who knew what, in what order, or how the bottles were handled.²⁸

It was done properly once, in 1599, and it failed

The blinded version of this experiment is not a modern sceptic's invention.  France ran it, on the most famous possession case of the age, and the design was better than the one in use today.

Marthe Brossier was a weaver's daughter from Romorantin who by 1598 was drawing crowds across France as a possessed woman, at a moment when the claim had sharp political value against Protestants.  Henri IV appointed a commission led by the physician Michel Marescot — the man who named the larynx, pharynx and hyoid bone — and what it carried out is described by medical historians in exactly these words:  "what was, in effect, a placebo-controlled trial."³

The design turned on the rationale of the rite itself.  If a demon cannot bear contact with holy things, then holy things should produce reactions and ordinary things should not.  So the commission brought both, disguised as each other:

these might include using unconsecrated water in a bottle normally used for holy water, or unconsecrated bread (wafers, or hosts) drawn from a box that usually contained only consecrated bread.³

It ran for forty days.  The verdict:

After a 40-day trial, the physicians concluded that Marthe could not have been genuinely possessed by a demon as she reacted similarly when exposed to both genuine and sham religious objects.³

An earlier examination by Bishop Charles Miron had gone the same way from the other direction:  he gave her real holy water presented as ordinary, which did nothing, and an ordinary key wrapped in red silk described as a relic of the True Cross together with lines of Virgil presented as the rite — both of which "agitated her immensely."⁴²  Her reactions tracked what she believed she was being given, never what she was actually being given.

Notice what these men had that the modern protocol does not.  Hiding which bottle is blessed removes one channel.  Marescot added the decisive second arm — presenting a sham object as genuine — and ran it for forty days rather than one afternoon.  If reactions follow the label rather than the blessing, the question is answered in a sitting.  That test costs nothing, needs no laboratory, and has apparently not been run on a possession case in four hundred years.

There is a further irony worth recording.  The intellectual ancestor of this design, on the same historians' account, is Jean Gerson, Chancellor of the University of Paris around 1400, who wrote treatises on how to test claimed revelations and warned against adoring a wafer "blindly" on the basis of appearance rather than condition.³  The demand for a control in matters of the spirit is not a modern hostility to faith.  It came out of the Church's own faculty of theology, six centuries ago, and it was aimed at protecting the genuine from the counterfeit.

So the honest state of this sign is not "untested."  It is worse than that, and more interesting:  tested constantly in an uncontrolled form, believed to succeed perfectly, never once recorded — and the last time anyone controlled it properly, over forty days, the reaction followed the label.  A community that ran twenty blinded trials with a sham arm and published the tally would either hand the world the strongest evidence for the supernatural ever obtained, or learn what Marescot's physicians learned in 1599.  That nobody has done it, in the decades since the informal test became routine, is the single most telling fact in this article.

The natural explanation, stated at full strength

It is a recognised clinical entity

Psychiatry does not deny that possession states occur.  It classifies them.  The World Health Organization's ICD-11 carries "trance and possession disorder," and DSM-5's criteria for dissociative identity disorder explicitly note that the disruption of identity "may be described in some cultures as an experience of possession."  Michael First, who helped write DSM-5 and watched Amorth's tape, named the syndrome on sight.¹⁰

The empirical work is thin but consistent.  Ferracuti, Sacco and Lazzari studied ten people undergoing exorcism in Rome using a structured dissociative-disorders interview and the Rorschach:  they "had many traits in common with dissociative identity disorder patients," most had "severe impairment of reality testing," and — importantly — most maintained normal social functioning.⁴³  A Ugandan study of 119 spirit-possessed patients found their local symptoms — altered consciousness, shaking, speaking in a voice attributed to a spirit — mapped onto the DSM criteria.⁴⁴

The pattern is learned, and the evidence for that is specific

The strongest argument is not that possession looks like illness.  It is that possession looks like the local expectation of possession, in detail, every time.

• The demons at Klingenberg were Catholic traditionalists worried about Vatican II, in a household of Catholic traditionalists worried about Vatican II.²⁸
• The entity's identity varies by culture — demons in Christian settings, jinn in Muslim ones, ancestral spirits across much of Africa — and so do the exit conditions. Bourguignon's survey found the belief in about 70% of societies, with each society's version internally consistent and mutually incompatible with the others.¹³
• Reported experience tracks the local script rather than the local population: 86% of Kenyan pentecostals, 6% of South Koreans overall.¹²
• The same thing happens in a different religion, in a Western clinic, and it lands on a different entity. A Dutch transcultural psychiatric outpatient service screened 551 patients and interviewed 49 with a Muslim background: 43% were positive that their psychiatric symptoms were caused by jinn, 27% thought not, and 31% were unsure; 87.2% had experienced hallucinations at some point.⁴⁵ These are people under psychiatric care, in the Netherlands, attributing the same class of experience to the spirits their own tradition supplies. The authors also note that most eligible patients declined to be interviewed at all, "many express[ing] a fear for stigmatization or metaphysical repercussions if they spoke about jinn" — which means the real rate is probably higher, and which is exactly the reticence Western Christians describe about telling a doctor they think they are oppressed.
• The demon answers the questions it is asked. It names itself when told to name itself. It leaves when commanded. Every one of those behaviours is prescribed in advance by the ritual, which is precisely why the 1985 CDF letter forbade lay groups from questioning demons directly or seeking their identity.¹⁶

And in the largest branch of the practice, the cue is spoken aloud.  Catholic exorcism is rare, slow and gatekept.  Protestant deliverance ministry is none of those things, and it is where most of the world's exorcisms now happen.  At a mass deliverance service, hundreds of people are worked through renunciations together, and the manifestation is requested from the platform.  Pastor Greg Locke, whose Tennessee congregation runs these regularly, opens one by addressing the room:  "all over this room, if you're not already manifesting…"⁴⁶  Later he commands, item by item, through a list:  "Spirit of witchcraft, now manifest and go, I command it — this is not a suggestion";  "every Masonic demon, manifest and go now."⁴⁶

Set aside whether demons are present.  Consider only what a psychologist would predict from that setting:  a crowd, an authority figure, an explicit instruction to manifest, a named list of what to manifest, and social proof from everyone else in the room already doing it.  The prediction is that manifestations will follow, in the forms named, in proportion to how suggestible the person is — and they do.  On the Protestant side there is also no screening step at all;  one minister interviewed on a deliverance channel is introduced as having performed some 25,000 deliverances and says he no longer keeps count.⁴⁷  Amorth's careful 93-in-30,000 has no counterpart here.

A working exorcist made this exact criticism first.  The Padua exorcist whose diary was analysed refused to use the diagnostic exorcism at all.  Giordan's summary of his reasoning:  "Some people use a 'diagnostic exorcism' to see if you're possessed or not.  He didn't use that because he said it was dangerous.  If you perform the exorcism, a person could act as though they're possessed, even become convinced."⁵  That is the objection this article has been building, stated by a Catholic exorcist about his own colleagues' method, and reported in a peer-reviewed sociology journal.  It is not an outsider's complaint.

The rite itself contains an amplification instruction.  Norm 4 of the traditional Roman Ritual tells the exorcist what to do after a session:

let him inquire of the person possessed, following upon one or the other act of exorcism, what the latter experienced in his body or soul while the exorcism was being performed, and to learn also what particular words in the form had a more intimidating effect upon the devil, so that hereafter these words may be employed with greater stress and frequency.¹

Read that as a psychologist would.  Debrief the subject about which stimulus produced the strongest response, then deliver that stimulus harder and more often.  Whatever its intent, that is a procedure for selecting and amplifying whichever cues a particular person reacts to — and it is written into the ritual as a duty, not as an abuse of it.  Over a course of weekly sessions lasting months, a well-executed rite will therefore converge on a highly reliable, highly specific set of triggers for that individual.  It will look exactly like a demon learning to fear the words, and exactly like a person learning what is expected.

And then the admission that no skeptic had to extract.  Amorth, describing the Ritual's three signs:

it is my significant experience, and that of all the other exorcists whom I questioned, that these signs always surfaced during an exorcism, never before.  It would be unrealistic to expect the manifestation of those signs before proceeding with the exorcism.⁹

The diagnostic signs appear only once the diagnostic procedure has begun.  That is the signature of a produced phenomenon, and it is reported as settled fact by the most experienced practitioner in the field, on behalf of every colleague he asked.

Amorth completed the circle himself.  He held that "only through the exorcism itself can we determine with certainty whether there is a satanic influence," that "even few and uncertain signs are sufficient to proceed with the exorcism," and that "An unnecessary exorcism never harmed anyone;  all the exorcists whom I questioned agree with me."⁹  Read together:  the threshold for testing is near zero, the test is the treatment, and the treatment is believed to be harmless.  Under those rules the practice cannot generate a negative result.

What has actually been measured

"Nobody has ever studied this scientifically" is a claim made confidently on both sides, and it is false.  Possession and possession-like trance have been instrumented several times, and the results are worth stating in both directions rather than cherry-picked.

Suggestion alone reproduces the experience.  The most direct experiment is Deeley and colleagues at King's College London, published in Cortex in 2014.  They used hypnotic suggestion plus fMRI in fifteen highly suggestible volunteers to model, deliberately, the two things this article is about:  "Suggestions of external personal control and internal personal control over involuntary movements modelled delusions of control and spirit possession respectively."⁶  The volunteers experienced involuntary movement attributed to an outside agent, and the brain networks recruited differed by what the movement was attributed to rather than by anything happening to them.  A laboratory can produce the core experience of possession in an ordinary person, on request, in an afternoon.  That does not prove any given case is suggestion, but it removes the claim that nothing natural could produce the experience.

The trance state itself has no physiological signature.  Wahbeh and colleagues took thirteen practising trance channels and recorded EEG, ECG, galvanic skin response and respiration across alternating five-minute blocks of channelling and not-channelling, on two separate days.  Their result:  "Despite subjective perceptions of distinctly different states, no substantive differences were seen in EEG frequency power, ECG measures, GSR and respiration."⁷  Only the voice changed.  People who sincerely report being occupied by another agent, under controlled alternation, look physiologically like themselves.

And these subjects are not ill.  A Brazilian team compared sixteen mediums with sixteen controls on pineal and pituitary volumes, melatonin metabolites and stress reactivity.  Their conclusion:  "The normal neuroimaging and stress reactivity findings… contrast with the abnormal results usually observed in subjects with psychotic and dissociative disorders."⁴⁸  This cuts against the lazy version of the sceptical case.  Whatever these people are doing, it is not straightforwardly pathology.

The one instrumented result pointing the other way, stated fairly.  Peres and colleagues scanned ten Brazilian psychographers with SPECT while they produced automatic writing in trance and in normal states.  The experienced mediums showed lower activity in several regions — left hippocampus, left anterior cingulate, right precentral gyrus among them — while the complexity of what they wrote was higher than in their control writing.  The authors' own comment:  the production of complex content while those areas are underactive "deserves further investigation both in terms of replication and explanatory hypotheses."⁴⁹  Ten subjects, one small study, no replication located — but it is a real result in a real journal, and an article that suppressed it would deserve none of the trust it is asking for.

The shape of the measured evidence, then, is this:  the experience is reproducible by suggestion, the state carries no autonomic or EEG signature, the practitioners are not measurably disordered, and one small imaging study found something odd that nobody has yet repeated.  Notably, none of these studies is of a Christian exorcism.  Nobody has instrumented one.

Medical conditions that produce this picture

Several are worth naming, because a possession attribution can delay their treatment past the point of recovery.

• Anti-NMDA receptor encephalitis. An autoimmune brain inflammation producing psychosis, seizures, abnormal movements, autonomic instability and catatonia, mostly in young women, and treatable. Its resemblance to historical possession accounts is close enough that a neurologist published the point in Annals of Neurology under the title "In search of lost time from 'Demonic Possession' to anti-N-methyl-D-aspartate receptor encephalitis."⁵⁰
• Temporal-lobe epilepsy, with its associated hyperreligiosity, reduced sexual interest and social withdrawal — the cluster Anneliese Michel displayed.
• Sleep paralysis. A pooled analysis of thirty-five studies covering 36,533 people found lifetime prevalence of 7.6% in the general population, 28.3% among students and 31.9% among psychiatric patients.⁵¹ Its phenomenology is a felt malevolent presence, chest pressure, and inability to move or speak — which is exactly how Anneliese Michel's first night is described.
• Schizophrenia with religious delusions. A Polish study of four such patients found that in every case the possession attribution "was supported by the local environment and media, led to seeking spiritual help, and delayed diagnostic assessment and treatment."⁵²

What the natural explanation does not do well

Three things, stated plainly.

It does not explain the felt otherness.  Subjects describe something that is not them, using their voice.  Dissociation is a label for that experience, not a mechanism, and psychiatry knows it.  Calling it "possession trance disorder" renames the puzzle.

It does not explain why exorcism often works.  Both Michael First and Amorth said versions of the same thing from opposite premises:  First that "Exorcism as a therapeutic technique could work,"¹⁰ Amorth that the ritual is only "10 percent of the cure" with the rest coming from the person's own life.⁹  A culturally matched ritual delivered by a trusted authority is a plausible therapy — but "plausible therapy" is a hypothesis, and no controlled outcome study of exorcism exists.  The one real outcome table, from the Padua diary, is discussed below, and it is compatible with a supportive ritual that has no specific effect.  The rest of the outcome evidence runs the other way:  documented cases of retraumatisation, treatment delay, and death.

It does not dispose of every reported detail in advance.  Case files contain odd single incidents that nobody investigated, repeated, or measured, and honesty requires saying so rather than waving at them.  What is striking, though, is how the ones that were chased down have gone.  The Gary oil that reappeared in a sealed room looked like the strongest physical residue in the whole case until the priest's own report turned up the family's month-old oil crosses and the heating vent under the window.  The possessed police car was a broken seat motor.  The police photograph was not a police photograph.  Every item in that case that anyone actually followed up dissolved, and the pattern is worth more than any single unexplained item, because it is a track record rather than an anecdote.

Does it work? The only outcome data anybody has published

No controlled outcome study of exorcism exists — no comparison group, no allocation, no blinding, nothing.  That is a search result, not an assumption:  PubMed and OpenAlex were both queried for trials and outcome studies of exorcism and of religious healing for psychiatric illness, and what comes back is literature on help-seeking, stigma and treatment delay, not on whether the ritual works.  But outcome reporting does exist, in two places, and both were found for this article rather than being drawn from the existing apologetic or sceptical literature.  Neither appeared in any apologetic or sceptical treatment consulted for this article, which is itself worth noting:  both sides have been arguing about this for decades without looking at the two places where somebody wrote the results down.

Fr. Gary Thomas, the exorcist of the Diocese of San Jose and the most publicly visible exorcist in America, told U.S. Catholic in 2011 that he had exorcised five people in five years, and that "more than 80 percent of the people who come to him need therapy, not an exorcist."  The reporting then adds the part the interview did not:

What Thomas does not mention is that four of the people he has exorcized gave up midway through the process — frustrated by their lack of improvement.  The fifth is Frank, who drives to Sacred Heart for exorcism once every five months and has done so for two and a half years.²²

Five cases;  four abandoned for lack of improvement;  one continuing after two and a half years without resolution.  That is not a study and should not be treated as one — it is a single practitioner over a short window, and Thomas himself was not the source of the follow-up.  But it is the closest thing to an outcome tally in the modern literature, and it points the same way as everything else.

The fifth man is worth meeting, because he is not the person the popular picture suggests.  "Frank" holds a doctorate in astrophysics, had struggled for a decade with depression and rage, and had tried medication, therapy and spiritualism without relief.  His wife gave him a book on possession.  Jogging one night he prayed a deliverance prayer, had a violent physical reaction, and concluded he was infested.  During Thomas's prayers his face "contorted into a wide yawn," which was read as an indwelling demon's aversion to the sacred — one of the 1614 criteria.  He now refuses psychological treatment beyond two annual visits for an antidepressant prescription:  he tried that, he says, and it did not help;  exorcism has.  He and Thomas have spent two and a half years trying to get roughly two dozen demons to give up their names.²²

And the Padua diary contains a real outcome table — the only one this article located anywhere.  Of the 55 people exorcised:  22 (40%) "appear[] to have recovered, freed from possession," though the authors immediately qualify that this "does not necessarily mean that the disease, depression, misfortune or the need for psychiatric drugs is over, and it is possible to have relapses";  14 (25%) were still being exorcised, "either because liberation has not occurred or because of a relapse";  and 12 (22%) left, or the exorcist withdrew "considering his own efforts useless."  For the rest the diary does not say.⁵  One man had received the ritual 354 times.

Take the 40% at face value and it is still not what it looks like.  The authors say plainly that recovery here can coexist with continuing illness, depression and psychiatric medication.  What the table actually records is that in a self-selected group who believed themselves possessed and were treated by a careful practitioner, four in ten stopped needing the ritual, a quarter never stopped, and a fifth walked away or were dropped.  That is a plausible profile for a supportive ritual with no specific effect, and nobody has ever compared it against anything.

Nothing about that account requires anyone in it to be foolish or dishonest.  A man in long-term treatment-resistant depression found a framework that gave his suffering a shape, an enemy and a task, and he reports feeling better.  It is entirely possible that he is better.  What it is not is evidence of a demon — and the cost, on the record, is that a man with a decade of depression has stopped seeing a psychologist.

The institutional context should be stated too, because it is not neutral.  The number of appointed exorcists in the United States went from one to nineteen in the second half of John Paul II's papacy, and to about thirty under Benedict XVI;  in November 2010 more than a hundred bishops and priests met in Baltimore to recruit more.²²  Fr. Robert Schreiter, a theologian at Catholic Theological Union, reads the revival this way:  "It's a selective return to the past…  A lot of it, indeed, is intended to create a distinctive identity—a sense of the church as over and against the world." and the same reporting notes that the ministry places a priest "in the role of hero rather than scapegoat" in the years after the abuse scandal.  A historian of the rite, Nancy Caciola, observes that exorcism manuals have historically proliferated at moments when the church felt itself under challenge.  None of this shows that any individual case is false.  It does show that the supply of exorcists is responding to something other than the supply of demons.

The harm

An article that audited the evidence and skipped this would be dishonest.

Terrance Cottrell Jr., Milwaukee, 22 August 2003.  Eight and a half years old, autistic.  His mother brought him to the Faith Temple Church of the Apostolic Faith, where the minister, Ray Hemphill — who "had not received any formal training for the ministry, but rather, was ordained by his brother" — concluded the boy's behaviour was demonic.  On a very hot evening, congregation members including his mother held down the child's arms and legs while Hemphill lay across his chest for one and a half to two hours.  The boy struggled, then stopped.  The medical examiner testified to "mechanical asphyxia due to compression, prolonged compression of his chest that restricted his respirations."  Hemphill first told firefighters the boy had collapsed while running with other boys.  He was convicted of physical abuse of a child recklessly causing great bodily harm and sentenced to two years five months of confinement plus seven years five months of extended supervision.¹⁴

Anneliese Michel, 1 July 1976.  Sixty-seven exorcisms, no doctor, death by starvation and dehydration at roughly seventy pounds.²⁸

Maricica Irina Cornici, Tanacu, Romania, June 2005.  A young nun diagnosed with schizophrenia at a local psychiatric hospital, taken back to the monastery, bound, gagged and left without food;  the autopsy found dehydration, exhaustion and lack of oxygen.  The priest was convicted and imprisoned.

Children accused of witchcraft.  In London, the Metropolitan Police's Project Violet was created specifically for violence against children linked to belief in witchcraft or spirit possession;  recorded referrals rose from ten in 2012 to twenty-one in 2013 to thirty-one in 2014, with sixty identified in 2015 — figures everyone involved treats as a floor, since many forces do not record the category at all.  The worst UK case, the torture and drowning of fifteen-year-old Kristy Bamu in Newham on Christmas Day 2010, produced murder convictions in 2012.

Delay.  This is the quiet harm and the most common one.  The Polish schizophrenia series found possession attribution delaying assessment and treatment in every case;⁵² a follow-up found that receiving a formal possession-trance diagnosis "can reinforce patients' beliefs about supernatural causation of symptoms and discourage professional treatment";⁵³ a Mauritian case study concluded that "exorcisms are potentially retraumatizing by the enactment of painful memories and emotions during this ritual."⁵⁴

Amorth's rule that "an unnecessary exorcism never harmed anyone" is the claim these cases refute.  It is not true, and the counterexamples are in court records.

What the Bible actually commits a Christian to

Jesus was known as an exorcist, and that much is history

This is not a devotional claim;  it is a conclusion critical historians reach on ordinary grounds.  Graham Twelftree:  "to sketch a picture of the historical Jesus without significant reference to his ministry of exorcism is to produce a distortion of the evidence."¹⁵  Gerd Theissen and Annette Merz:  "we must assume not only that the historical Jesus was active as an exorcist but also that exorcism was of great importance for his understanding of himself."⁵⁵  Even Bart Ehrman, who is personally skeptical of the miracle traditions, reports the state of the field:  "So abundantly attested are Jesus' miracle-working abilities that even scholars who are otherwise skeptical of the supernatural biases of our sources sometimes claim that whatever else one can say about him, Jesus was almost certainly a healer and an exorcist."⁵⁶

The reasons are the standard ones.  The tradition is multiply attested;  no messianic expectation called for exorcisms, so there was no template to invent from;  and the earliest opponents conceded the deeds and disputed only their source, which is why the Beelzebul charge exists at all.  Celsus called Jesus a sorcerer;  the Talmud says he "practised sorcery."  Hostile witnesses arguing about the power behind an act are conceding the act.

The New Testament distinguishes possession from illness — imperfectly

Matthew 4:24 lists categories side by side:  "they brought to him all the sick, those who were afflicted with various diseases and pains, demoniacs, epileptics, and paralytics, and he cured them."⁵⁷  Demoniacs and epileptics are different entries on the list.  Whoever wrote that sentence was not treating every seizure as a demon.

But the boy in Mark 9 looks like a seizure:  "whenever it seizes him, it dashes him down, and he foams and grinds his teeth and becomes rigid," and Matthew's parallel calls him an epileptic.  The honest reading is that the New Testament sorts most illness into ordinary categories and treats a smaller set of cases as possession, while at least one narrative sits on the boundary.

Two further facts a skeptic will raise, and neither should be hidden.  Exorcism was not distinctively Christian:  Josephus reports watching a Jewish exorcist named Eleazar draw a demon out through a man's nostrils with a ring, before Vespasian and the whole army, then command it to overturn a basin of water as proof.⁵⁸  And the Hippocratic treatise On the Sacred Disease had argued around 400 BC that epilepsy was not divine at all — a naturalistic option existed in the ancient world and was known.

The public test the early church offered, and nobody took

Tertullian, around AD 197, made the boldest evidential offer in the history of this subject.  Bring a possessed person before the magistrates, he wrote;  a Christian will command the spirit, and it will confess it is a demon;  and if it does not, "then and there shed the blood of that most impudent follower of Christ.  What clearer than a work like that?  What more trustworthy than such a proof?"⁵⁹

That is a falsifiable public challenge with the challenger's life staked on it.  There is no record of any Roman magistrate accepting it.  A Christian may take that as evidence the claim was credible enough to be left alone;  a skeptic may take it as evidence that a rhetorical offer nobody accepts costs nothing.  Both readings are available, and the article that only tells you one is not worth reading.

What follows

Christian doctrine holds that spiritual evil is real.  Nothing in this article touches that.  What this article tests is a narrower and much more recent claim:  that possession can be identified today by observable signs, and that particular modern cases have been shown to be genuine.  On the evidence, that claim is not established.

The two claims come apart cleanly.  A Christian can hold that demons exist, that Jesus cast them out, that the practice of exorcism is legitimate, and that no case in the last hundred years has produced evidence that would convince a neutral observer.  That combination is not a compromise.  It is what the record supports.

Common objections

"You are demanding proof on your terms.  Spiritual realities are not laboratory objects."  The terms are not the skeptic's;  they are the Roman Ritual's.  The Church chose language, hidden knowledge and strength as its criteria precisely because they are checkable from outside.  Asking for a recording of the first is not importing a foreign standard.  And the controlled test is not a modern imposition either:  a Catholic bishop ran one in 1599, and eleven physicians appointed by the Parlement of Paris reviewed the result.

"Demons will not perform for cameras."  This may be true, and it is unanswerable, which is the problem.  A hypothesis that predicts the same evidence as its own falsity has no evidential content.  It can be believed;  it cannot be argued for.  It also does not explain the sealed envelope or the blinded bottle, neither of which requires the demon's cooperation — only the exorcist's.

"Confidentiality prevents documentation."  Medicine records patients constantly with consent.  If a person were levitating, the case for releasing that recording, with a face blurred, would be the strongest imaginable — and the person, on the account given, is a victim being freed, not a patient being exposed.  The objection also has a precedent problem:  in 1990 a diocese authorised ABC to film an entire rite for national broadcast, so the rule is evidently not absolute.  What that tape showed is discussed above.

"The exorcists say the bottle test works every time — that IS the evidence."  It would be, if a count existed.  A remembered impression of a perfect record, formed by people who prepared the bottles and knew the answer, is the exact thing controls are for;  every practitioner in every field who has ever been tested formally has been surprised by how much worse the number is than the memory.  The remedy is not scepticism about the exorcists' honesty.  It is twenty trials, a decoy arm, and a piece of paper.

"Thousands of exorcists across every culture cannot all be wrong."  Convergent testimony is real evidence for something.  What it is evidence for is the question.  Bourguignon's 70% of societies also converge on possession, and they converge on mutually contradictory versions of it:  demons here, jinn there, ancestors elsewhere, each behaving exactly as its own culture predicts.¹³  Cross-cultural convergence on the form of an experience, with divergence on its content, is what a human phenomenon looks like.

"Psychiatry has no explanation either — it just relabels."  Largely fair, and stated above.  "Dissociative trance" describes rather than explains.  But a description that predicts the cultural patterning, the onset during ritual, the response to suggestion, and the medical mimics is doing more work than a hypothesis that predicts nothing in particular.

"You have not proved possession never happens."  Correct, and no such proof is possible or attempted.  The finding here is about the state of the evidence, not about the state of the universe.  If a single well-controlled case were produced tomorrow, this article would be wrong, and its author would rather be wrong that way than right by refusing to look.

"Exorcism helps people, so what is the harm?"  Sometimes it does, and that is worth saying.  The harm is in the cases where it substitutes for treatment:  Anneliese Michel starving with no doctor;  a Polish patient's schizophrenia going untreated;  an autistic eight-year-old asphyxiated on a church floor.¹⁴  The Church's own 1999 rite and its screening norms exist because the harm is known.

Demons and aliens: one article or two?

The claim that UFO and alien-abduction experiences are demonic in origin is a serious position in some Christian circles, and the two subjects share a structure:  a felt malevolent presence, paralysis, an intelligence that seems to know things, and a reported termination when the name of Jesus is invoked.  A reader might reasonably expect them treated together.

They should be two articles.  Three reasons.

First, the evidence bases barely overlap.  Deciding the UFO question means handling military sensor data, radar and infrared artefacts, the 2021 ODNI assessment and the AARO historical review, the history of the abduction template since the Hill case, hypnotic regression and false memory, and the sociology of contactee religions.  None of that is relevant to whether a woman in Venafro is possessed.  Bolting them together would make both treatments shallow.

Second, the quality of the two evidence bases is different, and merging them would launder the weaker one.  The central testable claim on the alien side — that abductions stop at the name of Jesus — rests on case files gathered by advocates who held the conclusion before collecting, largely from self-selected respondents inside a Christian subculture.  The main collection presents "Over 100 Online Testimonies" as "The Most Powerful Evidence in the World Today," under a statement of faith, with no control group, no blinding, and no criterion set in advance for what counts as termination.⁶⁰  Meanwhile sleep paralysis — felt presence, chest pressure, immobility, hallucinated figures, 7.6% lifetime prevalence in the general population⁵¹ — predicts every element of the report including the termination, since arousal and reorientation are what break an episode and prayer is a standard arousal behaviour.  Until someone tests whether invoking another name performs differently, the demonic hypothesis makes no distinguishing prediction.

Third, the Church has just drawn the same line, publicly and expensively.  On 3 June 2026 Cardinal Robert McElroy removed Msgr. Stephen Rossetti — chief exorcist of the Archdiocese of Washington for more than fifteen years — after Rossetti said in a video that "probably many if not most of these UFO sightings are in fact demons."  The archdiocese said the remarks "gravely undermined" Catholic teaching on demons and exorcism and cut ties with his centre;  Rossetti apologised.⁶¹  When the most prominent exorcist in the United States loses his position for merging these two topics, an encyclopedia that merges them without argument is making a claim it has not defended.

The companion article on UFOs, abduction and the demonic hypothesis will be written separately, and it will state plainly which version of the claim is defensible:  that the extraterrestrial hypothesis is poorly supported, that contactee movements are demonstrably occult in character, and that the strong version — abduction as demonic activity, demonstrated by the Jesus-termination effect — is not established.

Where this leaves it

Something happens to these people.  A brain surgeon with five thousand operations behind him had never seen it.  Four psychiatrists recognised it at once and none thought it was faked.  The subjects are not lying, the exorcists are mostly not frauds, and the families are frightened for cause.

What is missing is any observation that requires a demon.  A hundred years of claims have produced no recording, no sealed envelope, no measurement, no published trial, and no photograph — in an era when the last of those has become nearly unavoidable.  Meanwhile the pattern of the reports tracks culture with a precision that a real external agent should not produce;  the experience can be produced in a laboratory by suggestion;  the state carries no physiological signature;  and the most experienced exorcist in the world reported that the diagnostic signs never appear until after the diagnostic ritual has begun.

The strangest fact in the whole subject is the last one this article found.  The one experiment that would settle everything — identical bottles, one of them blessed, nobody in the room knowing which — is not a sceptic's fantasy.  It is described by practitioners as ordinary practice, with a hit rate they put at one hundred percent.  Somewhere between that claim and a piece of paper with twenty tally marks on it lies the entire difference between a private conviction and a public fact.  Nobody has walked the distance in four hundred years — and the last people who did, a bishop and eleven physicians of the Parlement of Paris, found that the reaction followed the label rather than the blessing.  Until somebody repeats their work, the honest verdict is not that the exorcists are lying.  It is that they have never checked.

The Roman Ritual's authors would recognise this conclusion.  They wrote the warning first and the signs second, in that order, for a reason:  ne facile credat.

Sources