The Person at the Breakfast Table
In 1676, British soldiers arrived in Virginia to suppress Bacon’s Rebellion. Some of them gathered an unfamiliar weed for a salad. The colonial historian Robert Beverley recorded what happened.
For eleven days the men behaved as if a world existed that no one else could see. One sat naked in a corner, making faces at his comrades. Another threw straws at a feather he held aloft. They held conversations. They reacted to a room that wasn’t the room around them. On the eleventh day, they returned to themselves, with no memory of what they had done.
Beverley’s soldiers had eaten jimsonweed. Datura stramonium. Devil’s Trumpet. Hell’s Bells. Thornapple. The Indigenous peoples of the region had known the plant for centuries — its preparation, its dosing, its uses in ritual, its dangers. European folk traditions placed it among the witches’ weeds, alongside belladonna, henbane, and mandrake. The locals knew what it was. The soldiers did not.
—
Jimsonweed is closer to a true hallucinogen than the compounds we more commonly discuss — LSD, DMT, psilocin. Those work through serotonergic systems. They warp surfaces, generate fractals, layer geometry onto a world you still recognize as yours — and recognize as altered. Under a classical psychedelic, generally, you know you are tripping. The distortion is its own signal. Jimsonweed operates on a different system, and produces what psychiatry calls a true hallucination: a perception with no external referent that the experiencer takes for ordinary reality.
David Luke, a psychologist who has interviewed many users of datura and its botanical cousins — belladonna, brugmansia, henbane, all in the nightshade family — describes the signature:
“Everything looks normal, and then you turn around and you’re having a conversation with a werewolf, and that seems normal. And then you turn back and it disappeared.”
The conversation is the thing. Not the werewolf. The hallucination is not an image the experiencer is watching. It is a relationship the experiencer is inside.
People on these compounds spend long stretches in conversation with people who are not in the room. Friends. Family. Often — and this is the part of the experience the popular literature underplays — the dead. A grandmother, a brother, a friend, a lover. Sitting across the table, speaking as they did in life, holding ordinary conversations. The experiencer typically does not realize they have been hallucinating until they look away for a moment, look back, and the person is gone.
This is the phenomenology of ghosts in the older sense: the visitation. The dead walk in unbidden. They stand in the kitchen. They behave as they did when alive. Full-bodied. Quiet. Normal. This is what folk records have called ghosts for centuries — across cultures, in archives older than psychiatry — and the form is consistent: a figure the experiencer recognizes. A friend. A family member. The dead. Sometimes a character from a film, or a creature from a story they grew up with. The werewolf is retrieved rather than invented.
Modern psychedelic reports focus on the serotonergics — DMT especially — where the entities are a different kind of being. Machine elves are alien: geometric, non-human, met in alien space. Ghosts are someone you knew, in your room, doing what they used to do. One is encountered. The other is recognized.
The question is why different brain systems produce different kinds of entities — and why the kind is so consistent. The phenomenology is too steady, across too many cases and centuries, to be noise.
—
Over a decade ago I was prescribed Cogentin — an anticholinergic drug typically given to manage the movement side effects of antipsychotic medications, a second drug to treat the symptoms caused by a first. This was during a period when the list of antidepressants that hadn’t worked left prescribers reaching for whatever might. I was desperate for relief and far less knowledgeable about psychiatric medications and mental health care in general than I am today. I had no idea what Cogentin was, mechanistically. The prescriber, evidently, did not think it worth explaining either.
The effect was specific. Almost everything stayed normal. I could walk, talk, carry on conversations, function in most of the ways I usually did.
I could not read. I could not draw. I could not write.
The page would blur when I looked at it. The blur was at the wrong layer for visual acuity — my eyes were fine, I didn’t wear glasses and had no other cause for poor vision. The marks on the page would not lock in. They wouldn’t settle into letters. Drawing did the same thing in reverse: the line I tried to put down would not stabilize as the shape I intended. The whole symbolic-binding layer of perception had gone soft.
Reading and drawing were the two practices I most relied on for self-regulation. The drug took them, and only them. After three days it cleared, and mercifully they came back.
I later learned that Cogentin, or benztropine, blocks acetylcholine — one of the brain’s signaling molecules. The class of drugs that do this is called anticholinergic. Jimsonweed is in the same class. So is scopolamine, the truth-serum drug. So are the alkaloids in belladonna, brugmansia, and henbane. Different doses, different routes, different intensities. The same gate failing.
—
A woman wakes one morning and her husband is in the kitchen. He fills two cups of coffee. He sits in the chair by the window and turns on the radio. He looks up and asks where the milk is.
He died last spring. She buried him. Her daughter sat with her through the service.
She knows this when asked. She also sees him there, waiting for an answer.
—
The neuroscientist Michael Hasselmo has spent his career mapping that gate.
The hippocampus is the memory encoder. More precisely, it’s the system that binds a moment into something retrievable later — what happened, who was there, when it occurred, what came before and after. Every time you bring a specific person to mind, you’re retrieving a representation the hippocampus encoded. Mannerisms. Voice. Face. Relationship.
Acetylcholine, released at the right rhythm, tells the hippocampus when to encode and when to retrieve. Block the acetylcholine and the gate sticks open in retrieval mode. Stored representations get pulled up and presented to consciousness with no constraint — no check against present sensory input, no source-tag for when this person was last alive, no flag marking the contents as memory rather than perception.
This is what was happening when the marks on the page wouldn’t lock in. The encoder couldn’t hold stable contents long enough for reading to work. Reading is sustained binding of marks to referents. Drawing is sustained binding of intended forms to motor output. Both require the encoder to hold for several seconds at a stretch. With the gate disrupted, the contents wouldn’t stay still.
Push the same disruption further, or hit it during retrieval rather than encoding, and you get what Beverley’s soldiers got. The person is still in there. Fully. The death-tag is what’s missing.
—
In 1971, the BMJ published a study by W. Dewi Rees of 293 widowed people in Wales. Nearly half — 46.7 percent — reported experiencing the deceased after death. Seeing them. Hearing them. Feeling their presence. Sometimes holding conversations with them. The average duration of the phenomenon was about ten years.
Most had told no one.
The medical profession hadn’t asked. The bereaved had learned not to volunteer their experiences. Half a population was carrying an ordinary feature of grief as a private secret, because the language available to them coded it as pathology.
Grief, sleep loss, and stress all suppress cholinergic function temporarily. What the bereaved hippocampus does for weeks or months, Lewy body pathology does permanently. In Lewy Body dementia, the rate of “phantom boarder” experience — the felt presence of someone who isn’t there, often a deceased relative — is roughly 41 percent. In Alzheimer’s, where the cholinergic pathology is different, it’s about 10 percent. The marker is mechanism-specific. Donepezil, a drug that raises acetylcholine levels, partially reverses the phenomenology in Lewy Body patients. The mechanism predicts the reversal. The reversal confirms the mechanism.
Three populations, three different routes in — accidental poisoning, neurodegeneration, ordinary grief — produce the same characteristic: a present-feeling person, often someone who has died, behaving in ways consistent with their living self, indistinguishable from a real perception until something breaks the spell.
The route in doesn’t determine the output. The system that fails does.
—
The previous piece in this series, The Code on the Wall, made an analogous case for a different layer. When subjects look at a diffracted laser on DMT, the angular characters they see are the geometric primitives the primary visual cortex extracts from every scene — junctions of edges, normally invisible because the higher levels of vision integrate them into objects and faces. Lower the integration, and the primitives show themselves. This piece makes the same argument one layer up. When the cholinergic gating relaxes, what becomes perceptible is the hippocampal encoder’s output without source-monitoring constraint. The encoder’s job is to hold the people who matter to you, fully formed, retrievable on demand. Push retrieval into a mode with no flag for present-vs-past, alive-vs-dead, memory-vs-perception, and the people walk back into the room.
—
The brain has several such layers. Each, when its integration loosens, surfaces what it normally does silently.
The geometry seen under the influence of drugs like DMT comes from V1. The phantom person comes from the hippocampus. The felt presence behind you comes from another system. The voice that speaks from no body comes from another. The cosmic intelligence that addresses you from above comes from another still.
What we call entity encounters are, when read carefully, a small number of brain systems each becoming briefly visible. The entity you encounter is a map of which layer has been exposed.
This is more claim than this single piece can prove. The hippocampus, this piece has shown. The other systems wait for their own pieces. The mechanism is one question. The meaning is another.
—
Beverley’s soldiers came back to themselves on the eleventh day with no memory of what had happened. That detail matters. The hippocampus had been disrupted hard enough that no episodic record of the eleven days was ever laid down. Their semantic, motor, and emotional systems had been running fine. The encoder was offline. The result: eleven days of behavior, conversation, possibly conversations with people who weren’t there, and then nothing left to remember, because the part of the brain that would have made the days rememberable wasn’t working.
The encoder holds the people who shaped us. The gate stays open and they come back. The person at the breakfast table is the encoder, doing what it has always done — holding them where the world no longer can.
This is the second piece in a series on what becomes visible when the brain’s integration loosens.
The first, The Code on the Wall, looked at the geometry behind vision.
The next pieces look at felt presence, the voice from no body, and the intelligence that addresses from above.
Subscribe at Inspirited In Sight to follow the rest.
The Rees finding has a second layer. Half a population carried an ordinary feature of grief as a private secret because the language available to them coded it as pathology. The mechanism was biological. The silence was linguistic. Firetongue is where I work the second layer — how vocabulary captures experience and how to test what a word is doing to the body it lands in. Subscribe to read the audits.
References
Aarsland, D., Ballard, C., Larsen, J. P., & McKeith, I. (2001). A comparative study of psychiatric symptoms in dementia with Lewy bodies and Parkinson’s disease with and without dementia. International Journal of Geriatric Psychiatry, 16(5), 528–536.
Beverley, R. (1705/1947). The History and Present State of Virginia. L. B. Wright (Ed.). University of North Carolina Press.
Hasselmo, M. E. (2006). The role of acetylcholine in learning and memory. Current Opinion in Neurobiology, 16(6), 710–715.
Luke, D. (2017, June 14). Synesthesia, hallucinations, & telepathy with DMT [Audio podcast episode]. In The Third Wave Podcast, episode 80.
Luke, D. (2019). Otherworlds: Psychedelics and Exceptional Human Experience (2nd ed.). Aeon Books.
Mori, E., Ikeda, M., Kosaka, K., & the Donepezil-DLB Study Investigators (2012). Donepezil for dementia with Lewy bodies: A randomized, placebo-controlled trial. Annals of Neurology, 72(1), 41–52.
Rees, W. D. (1971). The hallucinations of widowhood. British Medical Journal, 4(5778), 37–41.