Death Came for the Fortune Teller

John Barker opened a bureau to collect warnings of disasters before they happened, hoping that lives might be saved. Then two of its most remarkable contributors began warning the psychiatrist about his own death.

A disaster that seemed foretold

On 21 October 1966, a collapsing colliery spoil tip overwhelmed part of Aberfan in south Wales. Among the buildings in its path was Pantglas Junior School. The disaster killed 144 people, including 116 children, and left a community facing losses for which no explanation could offer consolation.¹

Barker wanted to investigate something beyond the physical causes. People were reporting dreams, impressions and feelings that seemed to anticipate the catastrophe. A psychiatrist working at Shelton Hospital near Shrewsbury, he was interested in whether such experiences could be gathered and examined rather than dismissed or accepted on the strength of a compelling anecdote.

An appeal appeared in the Evening Standard on 28 October, a week after the disaster. A published overview of the initiative records 76 responses. Their timing presented an immediate difficulty: the catastrophe was already known when the accounts reached the investigators. An accurate description received afterwards could not, by itself, establish what somebody had experienced beforehand.²

Barker pursued the question in a paper published by the Society for Psychical Research in 1967. The society’s catalogue describes 35 accounts selected for analysis, with supporting testimony that some experiences had been related before the disaster. Such witnesses mattered. They offered a possible way to distinguish an earlier warning from a memory unconsciously reshaped by the news.³

A witness remembering an earlier conversation is still different from a document made before the event. Both people may remember sincerely while disagreeing about words, sequence or emphasis. A disturbing dream can also contain several images, only one of which later seems important. The investigator needs the whole account, including the details that did not fit.

The correspondence was more extensive than a single appeal. In a 2019 conference abstract, researcher Alan Murdie described another collection prompted by an Oxford appeal, containing letters from 72 correspondents. Copies eventually entered the society’s archive, where they were rediscovered decades later. The letters preserve responses to Aberfan as well as claims about its anticipation.⁴

That distinction is important to the history. These were people trying to explain an appalling event and, in some cases, the guilt of having felt something that they had been unable to interpret or prevent. Their accounts deserve care without requiring an investigator to endorse every conclusion drawn from them.

Barker’s next step addressed the central weakness directly. If he could collect impressions before the next catastrophe, he would no longer have to depend entirely on recollection. A dated prediction could wait for its possible fulfilment. The newspaper could provide both a public invitation and somewhere for the incoming material to be recorded.

The proposal also changed the stakes. A collection of curious experiences might contribute to an argument about the mind. A warning service would have to distinguish a useful alarm from the many disturbing thoughts that never became events. Barker was attempting to move between those two ambitions, with a national newspaper helping to bring the experiment to the public.

A desk for tomorrow

The Premonitions Bureau launched on 4 January 1967, with Barker working alongside the Evening Standard’s science journalist Peter Fairley. Readers were invited to submit experiences that might concern future events. The very name gave an elusive subject an administrative form: a bureau could receive, date and compare material that otherwise remained private.²

Its proposed scoring system shows an attempt to go beyond deciding whether a story felt impressive. As later described by Sky History, a prediction could receive points for unusualness, accuracy and timing. The scheme recognised that a precise detail delivered beforehand should carry more weight than a common anxiety attached to an event afterwards.²

Recording the claim in advance was a real improvement. Once words are fixed, the investigator can return to them instead of accepting an increasingly polished version. A contributor cannot quietly replace a vague description with a place name, and a journalist cannot legitimately omit an inconvenient detail merely because another part of the prediction looks extraordinary.

A fixed record does not settle how the words should be interpreted. Suppose an account mentions a railway, a crowd and a station. Does any railway accident qualify? Must the accident happen at the station, or is a train travelling towards it enough? Is the crowd a detail about the prediction, or something almost any public disaster would supply?

Those questions have to be answered before the news supplies an attractive candidate. Otherwise the record remains fixed while the rules move around it. An experiment can preserve a prediction perfectly and still overestimate its success by accepting generous matches, overlooking failures or allowing an indefinite period in which something suitable might happen.

The practical ambition created another problem. An impression that seems specific afterwards may offer no safe instruction beforehand. Closing a station, grounding an aircraft or evacuating a hospital requires a location and a reason to identify that place rather than hundreds of alternatives. An investigator needs to know how many false alarms would accompany a successful warning.

Barker’s interest should not erase his medical career. A contemporary bibliographic record places his letter about overcrowding in psychiatric hospitals in the British Medical Journal in May 1968. His professional world included ordinary institutional problems alongside his investigation of extraordinary claims. The two concerns would acquire a terrible proximity during the Bureau’s short founding period.⁵

The experiment also depended on the people who kept sending material. Two became particularly prominent: Alan Hencher, a telephone operator, and Kathleen Middleton, usually known as Lorna, a piano and ballet teacher. Their contributions gave the Bureau recurring correspondents whose impressions could be followed over time, rather than a succession of unconnected accounts arriving after disasters.²

The man on the switchboard

Hencher’s association with telephones made him an apt figure for an enterprise built around incoming warnings. Later accounts describe a man who reported disturbing impressions of events beyond his immediate life. The most frequently discussed example concerned an aircraft and an apparently remarkable number of deaths.²⁶

According to the account reproduced by paranormal researcher Leo Ruickbie, Hencher supplied a prediction on 21 March 1967 involving an aircraft disaster and 123 or 124 people. The detail helped the case survive. A number looks harder to accommodate than a general feeling that an aeroplane will crash, and a warning recorded weeks beforehand appears to answer the objection of hindsight.⁶

In April, a British passenger aircraft crashed near Nicosia in Cyprus. Early reporting gave a death toll of 124, creating the sort of correspondence that could make the Bureau seem to be working. The subsequent official record is more exact, and more difficult for the prediction’s reputation.⁷⁸

The aircraft was a Bristol Britannia 313, registration HB-ITB. It was carrying 130 people, of whom 126 died. The accident report describes an aircraft approaching Nicosia, overshooting and then flying a circuit too low before striking rising ground. Its chronology records the impact at 23.13 GMT on 19 April, which was already 20 April locally.⁷

The final tally

These particulars matter because the version of Hencher’s warning examined by the writer Karl Sabbagh specified a Caravelle jet and an accident shortly after take-off. The Britannia was a turboprop, and the accident occurred during the approach and attempted landing sequence. The striking early total also changed as the casualties were established.⁸

There is a real difference between saying that a number resembled an initial news report and saying that it correctly predicted the final death toll. The early figure belongs in the history of why people were impressed. The final figure belongs in any assessment of what happened. Retaining both makes the case more informative than either a triumphant retelling or a dismissive summary.

The mismatches also reveal how much work can be done by a broad category. Aircraft disasters share passengers, crews, airports and phases of flight. If an investigator accepts the general event and a near numerical match while overlooking aircraft type and flight phase, the test has become less demanding than its opening description suggests.

None of this requires Hencher to have been dishonest. A person can report an impression faithfully, and an investigator can preserve it accurately, without the later event being the one that the impression supposedly described. Sincerity, accurate recording and successful prediction are separate questions, each requiring its own evidence.

What would have made the case stronger? A securely dated complete account identifying the aircraft, place and relevant period, followed by an event matching those particulars under rules already agreed. The public retellings examined here do not provide that demonstration. Their most memorable detail is inseparable from the parts that failed to correspond.

For Barker, however, these were accounts arriving while the experiment was unfolding. Hencher was a person on the other end of a telephone, with further impressions to report. The relationship could become more intimate than the comparison of a prediction card with a newspaper column, especially when the subject of a warning ceased to be an anonymous passenger.

Charing Cross in a vision

Middleton’s reports offered a different kind of apparent correspondence. Among the cases discussed by Ruickbie is an impression recorded on 1 November 1967 involving a crowd, a railway setting and Charing Cross. Four days later, a train heading for the London terminus was involved in one of Britain’s worst post-war railway accidents.⁶

On the evening of 5 November, the 19.43 Hastings to Charing Cross service derailed near Hither Green. The Ministry of Transport’s report describes a 12-coach train travelling at about 70 miles an hour. A broken rail derailed the third coach’s leading wheels; about a mile further on, those wheels struck intersecting railway tracks, causing the much larger derailment. Forty-nine people died and 78 were injured.⁹

Charing Cross was the train’s destination, giving Middleton’s reported impression an identifiable connection with the event. It was not the place where the train derailed. The distinction does not make the report worthless, but it changes the question from whether she named the accident site to how much latitude a destination should receive as a match.

A passenger reading the warning beforehand would also have faced a practical difficulty. A London terminus serves many trains. Without a particular service, time or defect, a reference to Charing Cross cannot isolate the vulnerable stretch of track. The operational explanation supplied by the inquiry was a broken rail, something that could be investigated through engineering evidence.

The railway report and the premonition account are therefore evidence of different kinds. One reconstructs a physical sequence from an accident investigation. The other records a reported experience later associated with that accident. A historian can place them beside each other without treating the second as a competing explanation for why the train left the rails.

Middleton’s ordinary occupation is worth retaining in a story so easily dominated by the word “seer”. A teacher’s work and a contributor’s unsettling experiences could coexist without every aspect of her life becoming an omen. The warning accounts describe what she reported; they do not entitle a later writer to diagnose her or invent private motives.²

Sam Knight, whose later research brought the Bureau to a new readership, has emphasised the human lives behind the experiment. In an interview about his work, he described investigating letters and records while trying to understand the people involved. Their experiences were part of a social history as well as a dispute about paranormal claims.¹⁰

That approach helps explain why the Bureau remains interesting even to readers unconvinced by precognition. A psychiatrist and a journalist created a place where unusual experiences would be heard. Contributors responded with accounts they considered important, and the resulting relationships could carry sympathy, expectation and anxiety alongside the original attempt at measurement.

A public experiment can also change the behaviour it observes. If contributors know which details attract attention, later reports may become more specific or more frequent. If investigators begin to regard particular people as unusually gifted, their submissions may receive closer scrutiny and more imaginative interpretation. Those are possibilities a testing procedure should control, rather than accusations about any individual’s conduct.

What counts as a warning?

The Bureau’s most durable question is methodological. How can an investigator tell an exceptional prediction from a correspondence found by searching a large collection of impressions against an even larger collection of events? The answer cannot come from the vividness of a single story, however sincere its participants or disturbing its subject.

A vague prediction of smoke in a large building could fit a factory fire, a hospital fire or a warehouse explosion. Such hypothetical matches show why a prediction needs fixed criteria before the search begins.

The investigator would need to decide how long the prediction remains open, which buildings qualify and what would count as failure. A description that never expires has indefinitely many opportunities to acquire a match. A description allowed to change its meaning after an event is even harder to distinguish from ordinary association.

A complete register would help. It would retain every original submission, its receipt date, amendments and unsuccessful candidates. Independent judges could compare predictions with several possible events without knowing which pairing the investigator preferred. A scoring rule fixed beforehand could then be tested against predictions made without any claimed special ability.

Later research has explored some of these difficulties. A 2015 paper by Caroline Watt and colleagues compared prospectively reported dreams with retrospective reports, using a registry to collect material before possible fulfilments were identified. The authors treated the label “precognitive” as describing the experiences being reported, rather than assuming that a paranormal ability had been established.¹¹

Their discussion found reasons to take reporting and interpretation seriously. The prospective material did not disclose a distinctive feature that reliably marked out the relevant dreams, and independent judges could disagree over the correspondences. A record made in advance improved the evidence about timing without eliminating the problem of deciding what constituted a match.¹¹

That later study does not retrospectively test Barker’s correspondents. It identifies difficulties that also bear on their story. Remembering a dream differently after a catastrophe and interpreting an unchanged dream generously are separate processes. Preventing the first leaves the second to be examined, so a date stamp is a beginning rather than a complete experimental design.

Sceptical investigator Joe Nickell has argued that selective reporting, memory and retrospective fitting offer ordinary explanations for accounts of apparent foresight. Those possibilities deserve testing in individual cases. They should not become a licence to claim that every contributor lied, or that a complete explanation of each surviving account has been demonstrated.¹²

Supporters can fairly reply that an advance record is better evidence than a story first told afterwards. Ruickbie’s treatment places the Bureau within an argument for taking unusual experiences seriously.⁶ The strength of that starting point depends on preserving the full prediction and applying demanding comparisons. Removing the mismatches makes the material harder to assess, whatever position the writer favours.

The warning turns inward

The premise became painfully personal in 1967. Knight’s account of the surviving material describes Hencher warning Barker about danger to his own life in April. Details included gas, although Barker’s home had no gas supply. Barker recorded his concern in a memorandum, later shared with Middleton. Knight located the document in the Cambridge archive. The sequence matters because the two contributors’ later warnings cannot simply be assumed to be independent.¹³

The warning was no longer something that could be filed while the investigator waited for a newspaper report. It concerned the man recording it, who could alter his behaviour in response. Every subsequent accident, illness or alarming coincidence might appear relevant, and there was no clear point at which the prediction could safely be declared to have failed.

This is a logical difficulty for testing, quite apart from any emotional effect. If survival confirms that a warning was heeded and death confirms that it was accurate, neither outcome challenges the claim. A useful assessment needs a possible result that would count against the prediction, rather than an explanation available for every eventuality.

According to Ruickbie’s later account, Middleton reported a further disturbing impression in February 1968, involving Barker alongside his dead parents. The report belongs to the history of concern about his safety. It is not a medical finding, and its imagery does not by itself identify a disease, date or means of preventing his death.⁶

The popular version of the story often compresses these events into two gifted strangers independently foreseeing the psychiatrist’s fate. The documented sequence, as reconstructed by later writers, is more complicated. Warnings accumulated over time among people who were in contact, and Barker himself helped circulate material relating to the danger he was said to face.

Contact does not prove that one contributor copied another. It does remove the justification for assuming that their agreement arose without any exchange of information. To establish independence, an investigator would need dated accounts and evidence about what each person knew when submitting them. Two reports cannot automatically be counted as two separate confirmations.

There is a human asymmetry here. Researchers can debate a warning’s specificity from a comfortable distance. Its recipient must decide whether to ignore a person whose earlier reports he has taken seriously. Barker’s experiment had created a relationship in which a troubling message could carry more authority than it would from a stranger making an unsolicited prediction.

No surviving evidence examined for this article establishes that fear determined what happened to him. The relationship between an ominous warning and a later death is precisely what needs explanation. It would be circular to assume that the warning was powerful because the death occurred, then use that assumed power to explain the death.

An appointment he could not cancel

Barker had already been studying the possibility that fear could have grave physical consequences. His 1968 book, Scared to Death, investigated that subject. Historian David Clarke places the work within a longer history of medical and folkloric accounts in which a frightening encounter, curse or expectation of death was said to have become fatal.¹⁴

Such stories cross an awkward boundary. Emotional states can be accompanied by physical changes, but a narrative about someone dying after a fright does not establish the precise cause. Medical evidence, circumstances and alternative explanations matter. A memorable sequence can encourage a causal claim long before there is enough information to support it.

That caution applies especially strongly when a writer becomes part of a story resembling his subject. Barker’s interest in death through fear gives the ending an unsettling symmetry. It does not establish that he died through fear, that a prediction made him ill, or that his medical work had somehow supplied the script for his own death.

Later accounts describe a brain haemorrhage in August 1968. He was 44. The contemporary confirmation of the date is brief and unusually useful: an editorial note in the British Medical Journal records that Barker died on 20 August. Some retellings use 18 August, the date associated with his collapse, as though it were the date of death.¹⁵¹⁶

The difference is small in a dramatic summary and essential in a history of predictions. An experiment concerned with what happened when cannot afford to merge the beginning of an illness with its fatal outcome. Nor should an article convert a later biographical account into a death certificate that it has not actually examined.

Barker’s death was undeniably early and tragic. It also remained a medical death, rather than an official finding that precognition had succeeded. A warning about gas does not become a prediction of a cerebral haemorrhage simply because the recipient later dies. A broad apprehension of danger needs to be assessed in its original terms.¹²

The Bureau did not vanish with him. A later account by Ian Beacock describes its continuation under Jennifer Preston into the 1970s. The familiar ending in which the founder’s death closes the experiment is neater than the institutional history. It makes his fate appear to provide a final verdict when the collecting of predictions actually continued.¹⁵

Knight’s 2022 book returned the experiment to public attention, bringing the participants and their correspondence before a new generation of readers.¹⁷ Its appeal is easy to understand: a scheme intended to make the future manageable became entangled with the vulnerability of the man who helped establish it. The ending is compelling without deciding the paranormal claim.

The warnings already there

Another catastrophe had occurred close to Barker’s professional life that year. In 1968, a fire at Shelton Hospital killed 24 women. Shropshire Fire and Rescue Service’s later account records the scale of the response and the subsequent overhaul of hospital safety. It is a history of patients, staff and emergency services, not an incidental omen in a psychiatrist’s biography.¹⁸

There is no basis here for asserting that Barker should have foreseen the fire, or for assigning him personal responsibility for it. The episode instead exposes the distance between collecting impressions of possible disasters and maintaining the protections needed against identifiable dangers. A warning becomes useful only when somebody can understand it and act on it.

The same distinction leads back to Aberfan. The official inquiry did not need a theory of precognition to explain why the spoil tip collapsed. The tribunal’s conclusions, quoted in Parliament, placed responsibility on the National Coal Board and described failures in policy, communication and the response to known conditions. The danger belonged to a physical landscape and an institutional history.¹⁹

There had been ordinary reasons for concern about tipping above the village, including water and the conditions beneath the spoil. Those could be investigated through records, engineering knowledge and testimony. Their existence did not depend on matching a dream with the eventual disaster, and their significance was not diminished because people also reported extraordinary experiences.¹¹⁹

Barker’s project can be understood as an attempt to discover another source of warning after established protections had failed so completely. That motive gives the experiment moral urgency. It does not transfer the evidential strength of the official disaster investigation to the psychic accounts collected in its aftermath.

The distinction also prevents an unfair burden from falling on people who reported troubling dreams. A child’s nightmare or an adult’s feeling of dread did not make that person responsible for identifying a colliery tip, persuading an authority or preventing the catastrophe. The duties of those managing a dangerous site were already present, with or without an extraordinary warning.

As of November 2026, the material examined for this feature supports a documented history of an ambitious experiment, remarkable reported coincidences and a founder who became the subject of his own correspondents’ fears. It does not establish a dependable ability to predict disasters. The original registers and particular memoranda would require further archival examination before stronger claims about individual entries could be justified.

A fresh investigation could still ask useful questions. How complete is the surviving sequence of submissions? Which matches were proposed at the time, and which acquired their importance later? Were unsuccessful warnings retained? Can independent readers agree on what the strongest predictions actually specified, before being shown the events they supposedly anticipated?

Those questions keep the Bureau open to examination without pretending that every answer remains equally plausible. The Nicosia report fixes an aircraft, a sequence and a final toll. The railway inquiry identifies a broken rail. Barker’s contemporary death notice fixes a date. Each ordinary record places a necessary limit on what can be claimed for an extraordinary story.

The most disturbing object in the case remains a warning addressed to the investigator himself. Barker could file other people’s impressions and wait to see what the news brought. When his own name entered the record, the distance between observer and subject disappeared, while the question of what the warning actually predicted remained unresolved.