People can see, hear, or sense a distant loved one at the exact moment of that person's death
Where the evidence lands: UnresolvedThat some people receive a real, non-physical signal from a person who is dying or in grave danger at a distance, experienced as an apparition, a voice, a touch, or an overwhelming sense of alarm, and that this coincides with the death too often and too specifically to be explained by chance, suggestion, or faulty recollection; and that this points to telepathy between living minds, or to the survival of consciousness after death.
Believed by: A broad and varied public: bereaved people who describe a personal experience they cannot explain away, along with parapsychologists and survival researchers who read the pattern as evidence of telepathy or an afterlife. Surveys of hallucinatory and bereavement experiences suggest a meaningful minority of the general population reports something of this kind.
The full story
What is documented
Start with what is not in dispute. For as long as records exist, people have reported a particular kind of experience: a sudden, vivid sense of a loved one, a figure seen at the foot of the bed, a voice calling a name, an inexplicable stab of dread, that later lined up with the death or grave danger of that person somewhere far away. These are the reports we call crisis apparitions.
What sets the subject apart from ordinary ghost lore is that it was, unusually early, treated as something to be counted rather than merely retold. In 1882 a group including the Cambridge philosopher Henry Sidgwick and the classicist Frederic W. H. Myers founded the Society for Psychical Research (SPR) to bring evidence-gathering methods to claims like these. In 1886 Edmund Gurney, Myers, and Frank Podmore published Phantasms of the Living, several hundred first-hand cases assembled with an argument attached: that many apparitions were not ghosts of the dead but impressions sent, somehow, from a living person in extremity.
Then, from 1889, the SPR did something more ambitious. Its Census of Hallucinations asked roughly 17,000 people whether they had ever had a vivid impression of seeing, hearing, or being touched by something that was not physically there. The 1894 report sorted the well-attested replies and concluded that the cases coinciding with a death occurred far more often than chance seemed to allow, a result often quoted as around 440 times the expected rate. That the reports exist, that they were gathered carefully, and that the early numbers looked anomalous: all of this is part of the record. The question is what it means.
The case for something real
The serious version of the claim deserves a fair hearing, because it was built by people who were not credulous. The founders of the SPR included working scientists and philosophers who expected to find fraud and error, and often did, and who nonetheless kept returning to a residue of cases they could not dismiss.
The strongest cases share a shape. Someone has a sharp, specific experience, a relative appearing, a voice, a sudden conviction that a particular person is in danger, at a time when there is no known reason to expect it. Later, word arrives that the person died or was in mortal peril at close to that moment. When such an account is written down or told to a witness before the news comes back, coincidence gets much harder to invoke, because the record predates the event it seems to anticipate. A handful of the old cases have exactly that structure.
The census result is the other pillar. If death coincidences really did occur hundreds of times more often than the base rate predicts, that is not the kind of thing a single anecdote can produce; it is a statistical shadow, the sort of pattern that in other fields would prompt a search for a cause. Add the sheer breadth of the reports, recurring across cultures and centuries in much the same form, and the case for something genuine is not that any mechanism has been proven, but that a real effect may be hiding inside data too easily waved away.
The reports are sincere, the pattern is old, and the best cases were recorded before the news arrived. That is not proof of anything. It is a reason not to close the file.
Where the paranormal claim breaks down
The distance between this is a real experience and this is a signal from the dying is where the evidence thins out. Nothing below denies that people have these experiences. The question is only whether they require a new force in the world, and the honest answer is that they do not, yet.
The central problem is retrospective testimony. Most accounts are recorded after the death is known, and human memory is not a recording; it reshapes. A vague unease the night before, once it is followed by terrible news, can be remembered as a sharp premonition timed to the moment of death. The detail that seems to have been received in advance can be, unconsciously, filled in from what was learned afterward. The rare cases with a genuine pre-notification record are the ones worth studying, and they are a small minority of the corpus.
Then there is base rate and selective memory. Vast numbers of people love vast numbers of others, think of them often, and occasionally feel a stab of worry. People also die constantly. With numbers that large, striking coincidences are not just possible; they are guaranteed. We remember the premonition that hit and forget the thousand uneasy nights that led nowhere. The census excess, impressive as it looks, was drawn from self-reported memories under assumptions the committee had to choose, and later reviewers flagged exactly these weaknesses.
Finally, the experiences themselves are well within the range of ordinary neurology. Grief, stress, and the edges of sleep reliably produce vivid perceptions with no outside cause; bereavement hallucinations of a lost person are common and, for most people, unremarkable in medical terms. A perception can be intense, detailed, and utterly convincing and still originate entirely inside the skull. When a mundane account is sufficient, an extraordinary one has to earn its place, and here it has not.
The reach for meaning
It is worth understanding why the paranormal reading is so compelling, because the pull is not weakness of mind; it is the mind working normally under the heaviest load it ever carries.
Grief demands meaning. To have felt a lost person one final time, and to believe the feeling was truly them, is a form of contact and comfort that no statistical correction can offer. Told that the experience was coincidence or a trick of a grieving brain, many people feel the explanation does not fit the force of what happened, and they are not wrong that it fails to capture how real it felt. The disagreement is about the cause, not the sincerity.
The mind is also built to find pattern and to resist chance. A hit that lands almost to the minute feels like an insult to call random, because coincidence seems far too flimsy to produce something so precise. That intuition is powerful and, in this case, misleading: rare, exact coincidences are precisely what enormous numbers of people and events throw off. The correction is true but invisible from the inside, while the experience arrives whole, immediate, and overwhelming.
A coincidence that lands to the minute does not feel like a coincidence. That is the whole difficulty, and it is a fact about minds, not about the dead.
Where the evidence lands
Hold the two things apart. That people have crisis experiences, and that early researchers found more death coincidences than they expected, is documented and real. The rated claim is different: that these experiences are caused by a genuine transfer of information from a dying mind, by telepathy or by consciousness surviving death. On that claim the verdict is Unproven.
It is unproven in the exact sense the word should carry. No mechanism has been demonstrated. The historical statistics rest on retrospective, self-reported testimony that later reviewers found too soft to bear the weight placed on it. Coincidence, selective memory, confabulation, and the ordinary hallucinations of grief are, together, sufficient to account for the reports without invoking anything new. That is enough to withhold the paranormal conclusion.
It is unproven rather than debunked because a genuine residue resists the easy dismissal too. The best documented cases, the ones recorded before any news arrived, are rare but not obviously reducible to chance, and the census anomaly was never cleanly reproduced nor cleanly explained away. The intellectually honest posture is to grant the mundane explanation its strength, to insist that an extraordinary cause be shown rather than assumed, and to leave the door open a crack for the small number of cases that still do not sit still.
Above all this is a subject about love and death, and the file treats it that way. To conclude that the cause is unproven is not to tell anyone that what they felt was worthless, or that their grief invented a lie. It is only to say that a real, sincere, and moving experience is not, by itself, evidence of a signal from beyond, and that the difference between those two things is the whole of the case.
What's still unexplained
- The best cases, those with a contemporaneous written or witnessed record made before the death was known, are rare but not obviously reducible to coincidence. How many genuinely well-documented, pre-notification cases exist, and what would a modern, pre-registered study of them find?
- The Census of Hallucinations reported an excess of death coincidences over chance that has never been cleanly explained away or reproduced under tighter conditions. Was the excess an artifact of the survey's design, or a real signal buried in unreliable data?
- Bereavement hallucinations are common and largely benign, and crisis apparitions overlap with them. Where, if anywhere, is the line between an ordinary grief experience and a report that would demand a new explanation?
- Why the specific form recurs, a figure at the moment of death, a voice, a sudden dread, across very different cultures and eras, is itself unsettled: shared neurology, shared storytelling, or shared experience of something real.
Point by point
The claim: The Census of Hallucinations found death coincidences hundreds of times more often than chance, which cannot be an accident.
What the record shows: The 1894 census did report a striking excess of death-coincidence cases, but the number rests on retrospective, self-reported testimony and on assumptions the committee itself had to make. Critics from the period onward, and later historians reviewing the census, noted that older cases were harder to verify, that memories of a coincidence can sharpen after the death is learned, and that the base rate calculation depended on choices about which cases counted. A large excess over chance in a survey of this design is suggestive, not decisive, and it has never been reproduced under conditions that close those gaps.
The claim: The experiences are too vivid and specific to be ordinary imagination.
What the record shows: Vividness is not the same as veridicality. The brain readily produces intense, convincing perceptions with no external cause, in states of stress, grief, exhaustion, and half-sleep, and bereavement hallucinations of a lost loved one are well documented and common. A sharply detailed sense of a person says a great deal about how the mind renders emotion and does not, by itself, establish that any signal arrived from outside.
The claim: The apparition often carries accurate information the percipient could not have known, such as the fact or manner of the death.
What the record shows: This is the strongest form of the claim and also the hardest to verify. In practice most accounts are recorded after the death became known, so the detail that seems to have been received in advance can be shaped, unconsciously, by later knowledge. Cases with a contemporaneous written record made before news arrived are rare and are the ones worth scrutiny, but the surviving corpus is dominated by testimony reconstructed after the fact, which cannot rule out normal transmission of the news, coincidence, or memory conforming to events.
The claim: So many sincere people across cultures and centuries report the same thing that it must reflect a real phenomenon.
What the record shows: The sincerity and the cross-cultural recurrence are genuine and are not in dispute. But a widespread, consistent experience can have a common psychological cause rather than a common paranormal one. Humans everywhere grieve, everywhere form intense mental images of the people they love, and everywhere notice and remember the coincidences while forgetting the far larger number of premonitions that came to nothing. Universality points as easily to shared human wiring as to a shared signal.
The claim: Coincidence is not enough, because the timing is often exact, within minutes of the death.
What the record shows: Exact-timing cases are the most arresting, but they are also the most vulnerable to selective memory. When a vague unease or a dream is followed by news of a death, the mind tends to align the two and to round the timing toward the moment of dying; the many uneasy nights not followed by any death leave no trace. Without an independent, time-stamped record made before the news, an apparently precise coincidence cannot be distinguished from a remembered one, and such records are the exception, not the rule.
Timeline
- Pre-1880sAccounts of sensing a distant death, a relative appearing at the moment of dying, a clock stopping, a feeling of dread, circulate for centuries as folklore, family story, and religious testimony, long before anyone tries to study them as a class of event.
- 1882The Society for Psychical Research (SPR) is founded in London by figures including Henry Sidgwick, Frederic W. H. Myers, and Edmund Gurney, as the first body to apply systematic, evidence-gathering methods to reports of apparitions, telepathy, and related claims.
- 1886Gurney, Myers, and Frank Podmore publish Phantasms of the Living, a two-volume collection of several hundred first-hand cases. They argue that many apparitions are not ghosts of the dead but telepathic impressions from a living person in crisis, coining the framing of a phantasm sent at the moment of extremity.
- 1889The SPR launches the Census of Hallucinations, a large survey asking whether respondents had ever had a vivid impression of seeing, hearing, or being touched by something not physically present. Interviewers eventually question roughly 17,000 people across Britain and beyond.
- 1894The census committee, chaired by Henry Sidgwick and including Eleanor Sidgwick, Myers, Podmore, and Alice Johnson, publishes its report. It identifies a set of well-attested death-coincidence cases and calculates that such coincidences occurred far more often than chance alone would predict, a figure often cited as roughly 440 times the expected rate.
- 1903Myers's posthumous Human Personality and Its Survival of Bodily Death gathers the case material into a broader theory of a subliminal self, pressing the argument from telepathy toward the larger question of whether consciousness survives death.
- 20th centuryLater researchers revisit and re-analyze the older collections, while critics press methodological objections: reliance on retrospective testimony, the difficulty of ruling out coincidence and selective memory, and the impossibility of verifying decades-old accounts. Interest waxes and wanes but the core reports keep arriving.
- 21st centuryCrisis apparitions are discussed alongside bereavement hallucinations and end-of-life experiences in psychology and palliative-care literature, which documents that sensing a dead or dying loved one is common in grief, while parapsychologists continue to argue the cases point to something the mundane account misses.
Unresolved. The documented record is not in dispute: for well over a century people have reported a vivid sense of a loved one, a figure at the bedside, a voice, a feeling of dread, that later turned out to coincide with that person's death or grave danger far away. Britain's Society for Psychical Research began cataloguing such accounts in the 1880s and, in its 1894 Census of Hallucinations, reported that death-coincidence cases occurred far more often than chance alone seemed to predict. The rated claim is narrower and stronger: that these experiences are caused by a real transfer of information from the dying person, by telepathy or a surviving consciousness, rather than by coincidence, faulty memory, or the ordinary hallucinations of a grieving mind. That claim is unproven. No mechanism has been demonstrated, the old statistics rest on self-reported and retrospective testimony, and mundane explanations remain sufficient. Equally, the reports are real, sincere, and unexplained to the satisfaction of everyone, so the file leaves the question open while treating the bereaved and the dead with care.
Sources
- 1.8. Spontaneous Apparitions and NDEs, Society for Psychical Research (2024)
- 2.Ghosts and Apparitions in Psi Research (Overview), Psi Encyclopedia, Society for Psychical Research (2023)
- 3.Our History, Society for Psychical Research (2024)
- 4.Census of Hallucinations, Encyclopedia.com (2019)
- 5.Apparitional experience, Wikipedia (2025)
- 6.Frederic W. H. Myers, Wikipedia (2025)
- 7.Death-coincidences or wishful thinking? The Society for Psychical Research and the 1894 Census of Hallucinations, History of Psychiatry (T. R. Dening), SAGE Journals (1994)
- 8.Phantasms of the Living, Volume 1 (1886), Internet Archive (Gurney, Myers, and Podmore) (1886)
- 9.Encounters at the Time of Death, Psychology Today (2022)
Help us investigate
This is a living case file. If you spot an error or know evidence we missed, tell us, and weigh in on where you land.
Where do you land?
Cast your read on this one.
Comments
Add your take. Comments are read and approved by a human before they appear, so keep it on topic and civil. Please do not accuse named, living people of crimes.