Ghosts and the Paranormal: Mostly Debunked Except the Hard Problem

The list of things that will make a person see a ghost is embarrassingly long. Carbon monoxide poisoning causes hallucinations, paranoia, a creeping sense of presence in empty rooms. Infrasound at 18.98 Hz, inaudible to human ears but felt in the chest and the visual cortex, produces feelings of dread, peripheral motion, and the distinct sensation of being watched. Sleep paralysis generates hypnagogic hallucinations vivid enough that cultures across human history independently invented demons to explain them: the Old Hag of Newfoundland, the Kanashibari of Japan, the Bakhtak of Iran. Pattern recognition, the same cognitive machinery that reads faces in clouds and finds meaning in noise, runs on our visual system constantly and cannot be turned off. We are built to find intention in darkness.

A 2003 investigation by Vic Tandy of the University of Coventry found that a supposedly haunted Victorian location in Coventry had an infrasound resonance near 19 Hz produced by an industrial fan in the basement. Remove the fan: feelings of dread disappear. A significant proportion of haunted house reports turn out on inspection to involve elevated CO from faulty heating systems. The Winchester Mystery House is a fascinating piece of architecture built by someone who was, by most assessments, genuinely mentally ill and probably suffering from grief. The famous Enfield Poltergeist has been comprehensively documented as involving children who admitted to hoaxing significant portions of it and investigators who wanted to believe.

This is the debunking section, and it was supposed to go on longer. The problem is that it bottoms out too quickly against something the debunkers cannot touch.

Most Ghost Experiences Have Mundane Explanations, and the Mundane Explanations Are Genuinely Good

The infrasound mechanism alone accounts for a remarkable percentage of “haunted location” reports. Todd Disotell, a primatologist and skeptic who spent years investigating paranormal claims, noted that almost every case he investigated had either a physical or psychological explanation within reach. The human perceptual system is not a neutral recorder of events. It is a prediction machine that constructs a model of the world and only updates that model when the discrepancy with incoming data becomes too large to ignore. Under the right conditions; low light, emotional stress, CO exposure, social priming (“this place is supposed to be haunted”), sleep deprivation; the prediction machine produces confident false outputs.

The grief hallucination research is particularly relevant here. Studies published in the British Medical Journal and elsewhere have found that somewhere between 40% and 80% of widowed people report experiences of the deceased: seeing them, hearing their voice, feeling their presence. These are not metaphors. They are perceptual experiences that the people having them cannot distinguish from ordinary perception. And they are, by virtually all clinical accounts, normal. Not pathological. A feature of grief, not a bug. The brain, which modeled the presence of the deceased person continuously for years or decades, keeps generating that presence for some time after the person dies. The model hasn’t been updated yet. The ghost is a lag.

Electromagnetic field fluctuations, which haunted location enthusiasts measure with devices they buy on Amazon, are genuinely able to produce unusual perceptual effects under laboratory conditions. Michael Persinger at Laurentian University spent years demonstrating that weak EM fields applied to the temporal lobes could generate feelings of presence, spiritual encounters, and what subjects described as communication with entities. His work has replication problems and has been disputed, but the core phenomenon, that brains exposed to unusual inputs produce unusual outputs, is not in doubt.

The role of suggestion in haunted location experiences is probably larger than most people account for. Researchers at Hertfordshire University led by Richard Wiseman and Matthew Smith ran a controlled study at Hampton Court Palace, a site with long-documented ghost reports, and found that people reported anomalous experiences at locations with no objective unusual features but which had been pre-designated as “active.” The expectation of experience shapes the experience. People touring a location they’ve been told is haunted are primed to interpret ambiguous sensory data (a draft, a creak, a shadow at the edge of vision) as paranormal. The exact same data, in the exact same building, without the priming, gets filed as “old building does old building things.”

The psychology of testimony is the final piece of the debunking. Ghost stories are social. People share them, elaborate them, and calibrate them to the expectations of the listener. The original experience, which might have been mild (a feeling of unease, a shadow), grows in the retelling. Memory is constructive, not archival. Each time a ghost story is retold, the memory is retrieved and reconstructed, and each reconstruction incorporates a little more of what the teller expects, what the listener responds to, what fits the genre conventions of a ghost story. By the tenth telling, an uncomfortable feeling in an old house has become a full apparition with period clothing.

So the debunk is real. Most ghost experiences are pattern recognition, grief lag, physiological interference, or hoax. The field guide rating for “ghosts as evidence of dead people persisting in some detectable form” is Copper: plausible-sounding, fails on contact with investigation, the explanatory work done by mundane causes is nearly complete.

Nearly.

The Dismissal Rests on an Assumption Nobody Has Actually Checked

The full debunking argument requires one more premise that skeptics tend to deploy without examination: when the body dies, consciousness ends, because consciousness is produced by the brain, and when the brain stops, the production stops. This is presented as established fact. It is not established fact. It is the dominant working assumption of neuroscience, and working assumptions are not the same as solved problems.

The hard problem of consciousness is the name philosopher David Chalmers gave to a specific explanatory gap in 1995. Not the “easy problems,” which are genuinely hard in terms of complexity but which everyone agrees are in principle solvable: how does the brain process sensory input, integrate information, produce behavior, regulate attention. These are mechanistic questions, and mechanistic questions yield to investigation. The hard problem is different. It is the question of why any of this processing is accompanied by subjective experience at all. Why is there something it is like to be you. Why, when your visual cortex processes wavelengths in the 700nm range, do you experience the redness of red rather than simply responding to red-wavelength information the way a thermostat responds to temperature.

No one has an answer to this. That is not an exaggeration for rhetorical effect. There is no scientific paper that explains why physical processes give rise to felt experience. There are many theories, and we’ll look at them in more detail in a later entry in this series. But none of them close the explanatory gap. The gap is real.

This matters for ghosts because the entire “consciousness ends at death” argument is downstream of an assumption about what consciousness is. If consciousness is simply identical to information processing in neurons, then yes, it ends when the neurons stop firing. But if consciousness is something else, something that arises from information processing but is not reducible to it, or something that the brain mediates but does not generate, then the picture gets more complicated and the “of course it ends at death” conclusion becomes less obvious.

The Steel Man Is Not That Ghosts Are Real; It Is That We Cannot Fully Exclude Persistence

The credulous version of the ghost argument is easy to dismiss: the spirit leaves the body, becomes a floating presence, throws things around, makes cold spots, wants something. This requires ghosts to have physical causal power despite having no mass or energy. It requires some substrate for the personality and memory to persist in. It hasn’t been established. Copper.

The philosophically honest version is harder to dismiss, and it goes like this: we have no complete theory of consciousness. We do not know what it is, where precisely it comes from, or what its relationship to physical processes actually is. In the absence of a complete theory, we cannot construct a complete proof that it terminates at death. We can say it is likely, and the weight of available evidence does point in that direction. But “likely” is different from “proven,” and in a domain where we cannot define the thing we are making claims about, epistemic humility is the correct position.

Philosopher David Chalmers has said directly that given genuine uncertainty about the nature of consciousness, various forms of persistence after death cannot be ruled out on purely scientific grounds. He is not a ghost enthusiast. He is applying the same standards of rigor to this question that he applies to everything else.

The near-death experience literature is relevant here, not because NDEs prove anything but because they constitute a data set that resists full explanation. The AWARE study (Sam Parnia, Southampton University) attempted to verify claims of out-of-body perception during cardiac arrest by placing visual targets in cardiac resuscitation areas that would only be visible from above. The results were inconclusive (one potential hit, not enough to be statistically meaningful). But the broader NDE phenomenology, the tunnel, the light, the life review, the sense of peace, recurring cross-culturally, in blind subjects who report visual experience during periods of verified cardiac arrest, is not fully explained by brain anoxia, though anoxia is the leading candidate. The data keeps being weirder than the clean debunk would prefer.

Pim van Lommel, a Dutch cardiologist who published a large prospective NDE study in The Lancet in 2001, collected reports from over 340 cardiac arrest survivors. Eighteen percent reported some form of NDE, and a smaller subset reported experiences during periods when their brains showed flat EEG, indicating no measurable electrical activity. Van Lommel concluded that current neuroscience cannot fully account for these experiences and called for a “non-local” theory of consciousness. Most cardiologists find this conclusion premature. The flat EEG argument is not as clean as it sounds: EEG measures surface cortical activity and does not capture all brain activity, and timing of NDE experience relative to cardiac events is difficult to verify retrospectively. But van Lommel’s study was prospective, methodologically serious, and published in one of the highest-tier medical journals in the world. It is not a paper by someone who believes the Earth is hollow. The honest debunker acknowledges it without overclaiming it.

The Consciousness Problem Means We Can’t Fully Close the Book

Here is where the honest rating gets complicated. The experiences people attribute to ghosts: feelings of presence, voices, apparitions, inexplicable cold spots, are nearly all explicable through infrasound, electromagnetic fields, CO, pattern recognition, grief, suggestion, and in some cases deliberate fraud. On the direct “ghosts-as-persistent-dead-people” question, the debunking is substantive.

But the debunking depends on a theory of consciousness that is incomplete in ways that are not incidental. The hard problem is not a gap we expect to fill shortly. It is a structural mystery about why physical processes generate felt experience at all, and nobody has a satisfying answer. Not even close.

What this means for ghosts is not that ghosts are real. It means that “consciousness ends at death” is an inference, not a proof, and the inference rests on premises that are under genuine philosophical dispute. The question of what happens to subjective experience when the system that generates it (or mediates it, or whatever the right verb turns out to be) shuts down is not actually settled.

That’s uncomfortable if you want the whole thing tidily debunked. The uncomfortable version is more accurate. Most ghost experiences: explained, Copper. The underlying question about consciousness those experiences are gesturing at: genuinely open, and connected to the deepest unsolved problem in science. The debunkers who say “of course it ends at death, it’s just brain activity” are making a metaphysical claim they cannot fully defend.

The honest position is: almost certainly nothing persists after death in any form that would manifest as a ghost. But “almost certainly” is not “certainly,” and the reason we can’t close it to certainty is because we don’t actually know what consciousness is. That’s not a concession to the paranormal. It’s a concession to the genuine state of knowledge, which is messier and more interesting than either side wants to admit.

The ghost is probably the lag. But nobody has fully proved that the thing creating the lag doesn’t have somewhere else to go.