A Register of Modern Possession and Infestation Cases

Companion to Demonic Possession:  What the Documented Cases Actually Show.  That article argues from a handful of cases.  This page holds the working register behind it, so a reader can check the sample instead of taking a curated selection on trust.

What is in here, and what is not

Inclusion rule:  at least one named witness from outside the household.  A family's own account, however sincere, cannot be separated from the family's own beliefs, and the corpus of such accounts is effectively unbounded.  Requiring an outside witness — a priest, a police officer, a physician, a teacher, a case manager, a journalist — cuts the field down to something that can be audited, and it selects for exactly the cases apologetics actually cites.

Modern means 1900 onward.  Loudun (1634), Salem (1692) and the Bell Witch (1817) are omitted, not because they are unimportant but because the evidential questions there are historians' questions.  One pre-modern case is worth naming anyway, because it is the only properly controlled test in the whole history of the subject:  Marthe Brossier, France, 1599, whom Bishop Charles Miron examined by giving her real holy water disguised as ordinary water — no reaction — and an ordinary key presented as a relic of the True Cross, together with lines of Virgil presented as the rite — violent reaction.  Eleven physicians appointed by the Parlement of Paris then reported unanimously that there was nothing demonic in the matter.  Every entry below should be read against that precedent, because it shows the test was available four centuries ago.

Two types are tracked together.  Possession claims that a person is inhabited.  Infestation claims that a place is — the poltergeist and haunting cases.  They are kept in one register because the diagnostic problem is identical and because the Catholic taxonomy treats infestation as a category of the same phenomenon.  Where the distinction matters, the table says which is which.

"What survived" is the only column that matters.  It records what is left after investigation, not what was claimed.  A case can be entirely sincere and survive nothing.

This register is incomplete and will stay that way for a while.  Entries marked unaudited are listed with their claim and their best-known source but have not been worked up from primaries;  they are here so that the gaps are visible rather than invisible.  Nothing in the "what survived" column for those rows should be relied on.

The register

Collective outbreaks

A second class of case belongs here, and it is the one with the most witnesses of all:  outbreaks in which dozens or hundreds of people in one institution are affected at once.  These are the strongest available test of the cultural-script model, because they let you watch a possession spread through a population in real time and see what governs its shape.

Three features recur, and all three are hard for the demonic reading and easy for the psychological one.  The outbreaks run along social lines — one class, one dormitory, one workplace — rather than along any line a spirit would be expected to follow.  The content is locally uniform:  every child in Tarapoto saw the same bearded man in black, which is what a shared story produces and not what independent encounters produce.  And they stop when the population is dispersed, which is why closing the school is the standard remedy and why it works.

Historic convent outbreaks — Loudun in the 1630s, Louviers, Auxonne — have exactly this shape, which is why historians treat them as a single phenomenon with the modern school cases rather than as a separate spiritual category.

Fatal cases

Kept as its own table because it is the part of this subject with the firmest evidence and the least attention.  Every row is a court record or a prosecutor's charge, not a testimony.

Two things stand out.  The victims are overwhelmingly children and young women — the least able to leave the room.  And in the cases where the underlying condition is recorded, it was treatable:  autism, epilepsy or convulsions, schizophrenia.  The Guyanese girl who died had convulsions;  Anneliese Michel had epilepsy;  Maricica Cornici had a diagnosis from a psychiatric hospital that was set aside in favour of a monastery.  Whatever else is at stake in this question, that pattern is the reason it is not merely academic.

Notes on the strongest entries

Sauchie, 1960 — the best witnesses anyone has

If a reader wants the single hardest case for a sceptic, it is not any of the famous ones.  It is a Scottish village in the winter of 1960.

Eleven-year-old Virginia Campbell had just been moved from Ireland, separated from her father, her dog and her best friend, and put to share a bed with a nine-year-old cousin.  Over eleven days the household reported knocking, a linen chest that moved, and a pillow that rotated under her head.  What makes the case unusual is who watched.  The local minister, Rev. T.W. Lund, said he saw the linen chest rock, rise slightly, travel about eighteen inches across the floor and return.  Dr W.H. Nisbet, the family doctor, said he watched the chest move about a foot with its lid opening and closing, and described a rippling in the bedclothes as "a puckering, as if due to traction by an invisible agency."  Virginia's teacher, Margaret Stewart, said she saw a desk lid rise and an empty desk lift about an inch, and checked underneath for strings and levers.  Two more physicians corroborated.  A.R.G. Owen, a Cambridge mathematician, interviewed all of them the following month and concluded that trickery could not account for what they described.¹³

Three things a fair reader should hold alongside that.  There is no film, no photograph and no instrument anywhere in the case.  Owen's interviews were conducted about a month after the events, after the witnesses had read each other's accounts in the local newspaper.  And Owen's own conclusion is not a demon:  he explicitly sets aside a discarnate agency and proposes that "certain unknown physical forces operated on matter in the vicinity" as a result of something in the girl's body or mind, noting her upheaval and her rapid puberty.¹³

So the best-witnessed infestation case in the modern literature is one whose own investigator concluded it was about the child.  That is worth sitting with before reaching for the demonic.

Rosenheim, 1967 — the only case with third-party instrument records

A Bavarian law office had its lights fail, its fuses blow and its telephones misbehave.  What lifts this above anecdote is that the phone company's own equipment logged the anomaly:  the speaking-clock number was recorded as dialled far faster and far more often than the office's rotary phones could physically have produced.  Hans Bender brought in two Max Planck Institute physicists, and the investigation identified the effects as occurring only when a nineteen-year-old secretary, Annemarie Schneider, was present;  they stopped when she left the firm.¹⁴

This is the closest the field comes to instrumented evidence, and it is genuinely interesting.  It is also, again, agent-dependent — a fact about a person, not a place — which is the opposite of what an infesting spirit predicts and exactly what the sceptical literature expects.

Columbus, 1984 — how these cases usually end when a camera is present

Tina Resch's flying telephone ran on the front page of the Columbus Dispatch and went out on the Associated Press wire.  Terence Hines describes how the photograph was obtained:  the photographer "found that if he watched an object, it stubbornly refused to budge.  So he would hold up his camera and look away."¹⁵  Later a visiting television crew left a camera running and recorded Resch pulling a lamp over her shoulder and then screaming as though frightened.  Confronted, she admitted it.

Two lessons generalise.  First, the requirement for an unobserved moment is itself data.  Second, the fact that a person was caught faking once does not prove they always faked — the Skeptical Inquirer's own treatment makes that point carefully — but it does put the burden where it belongs.¹⁵

Enfield, 1977–79 — the honest contested case

Enfield deserves its reputation as the case that will not resolve.  A police constable signed a statement about a moving chair.  Hundreds of hours of tape exist.  And the two investigators conceded that the girls sometimes tried to trick them, arguing only that the attempts were few and quickly caught;  Janet Hodgson later put it at cheating "about two percent of the time."  Within the Society for Psychical Research itself, Anita Gregory reviewed the case sceptically and argued the girls enjoyed play-acting.¹⁶

An honest register records this as unresolved rather than claiming it for either side.

The cases that collapsed, and why they still matter

Amityville and the Snedeker case are in the register precisely because they are the two that produced the most cultural output and the least evidence.  Amityville's collapse is on the record from the DeFeo defence lawyer, William Weber, who is reported to have admitted that he and the Lutzes "created this horror story over many bottles of wine."¹⁷  The Snedeker case collapsed from the inside — Ray Garton, hired to write it as non-fiction, says the family's accounts did not agree, and that when he raised it with Ed Warren:  "'They're crazy,' he said.  'All the people who come to us are crazy, that's why they come to us.  Just use what you can and make the rest up.  You write scary books, right?  Well, make it up and make it scary.  That's why we hired you.'"¹⁸  Garton has said the same thing in public for thirty years and has not profited by it;  readers should also know that his critics accuse him of settling scores, and that Warren, who died in 2006, never answered the charge.

These two are the reason the register exists.  Both were, in their day, presented exactly as the surviving cases are presented now:  sincere family, outside investigators, a book, a film.

What the register shows in aggregate

Twenty-six cases, and some patterns that no single case would reveal.

1. Where the documentation is best, the claim is weakest. The two cases with real institutional paper — Gary's state file and Klingenberg's tapes and trial — are the two where the supernatural reading fares worst. The cases that look strongest in retelling are the ones documented by a single sympathetic author.
2. Every replication attempt in the register failed. There are exactly three: the Chinese-language question at Klingenberg, the wall-walk repeat demanded by Dr. Richardson at Gary, and the holy-water bottles at Klingenberg — of which only the first two were even attempted under anything like control. Two failed outright; the third has no recorded protocol.
3. The phenomena avoid instruments, not people. Witnesses are plentiful. Recordings of the decisive moment are absent everywhere except Venafro, where the camera showed a recognised clinical picture, and Columbus, where it showed fakery.
4. Agent-dependence keeps turning up. Rosenheim, Sauchie, Enfield and Columbus all centre on one adolescent or young adult, and in each case the effects travel with the person or stop when their circumstances change. That is a strong pattern and it does not point at places or spirits.
5. Five of the nineteen ended in a courtroom, and four of those in a death. Michael Taylor, Terrance Cottrell, Maricica Cornici, Kristy Bamu, and Arne Johnson's victim Alan Bono. No court has ever admitted possession as an explanation; in the one case where it was offered as a defence, the judge excluded it as unprovable.
6. The one controlled test in the history of the subject predates the whole register by three centuries. Miron's 1599 examination of Marthe Brossier had a decoy arm — a fake relic presented as genuine — that no modern protocol includes. Nothing since has matched it.
7. Nothing in the register requires a demon, and nothing in it disproves one. The register's job is to show what the evidence is, and the honest summary is that the evidence is testimony — sincere, sometimes professional, and never captured.

Not yet audited

Listed so the gaps are visible.  Each has at least one outside witness and none has been worked up from primaries for this register.

• Runcorn, Cheshire, 1952.
• The Smurl family, West Pittston, Pennsylvania, 1974–89 — Warren case.
• M. Scott Peck's two cases, United States, 1980s — a psychiatrist's own accounts, in People of the Lie and Glimpses of the Devil.
• Malachi Martin's five cases, United States, before 1976 — listed for completeness only. Names, dates and dioceses were all changed, so no entry in this register can ever be made from them.
• Craig Keener's Global South reports, collected in Miracles (2011), Appendix B — the largest body of first-hand possession and exorcism testimony assembled by a serious scholar, and the most important omission on this page. A copy has not yet been obtained.

Sources

Cases in the register that are sourced in the main article rather than here — Earling, Cottage City, Ossett, Klingenberg, Milwaukee, Tanacu, Newham, Gary, Venafro and Gallagher's "Julia" — carry their citations there.  The Brookfield entry rests on the Connecticut Superior Court proceedings of 1981, in which Judge Robert J. Callahan declined to admit demonic possession as a defence;  that entry has been checked only against secondary reporting and needs a primary source before publication.