The 10,000-Year Warning
2,150 feet below the New Mexico desert, in salt beds that haven’t moved in 250 million years, there are steel drums that will be lethal for longer than human civilization has existed.
The Waste Isolation Pilot Plant sits twenty-six miles southeast of Carlsbad. Its storage rooms are carved into the Salado Formation, a layer of rock salt deposited when the Permian Basin was a shallow sea. The salt is plastic on geological timescales. It flows around anything placed inside it, sealing the drums in a medium with no groundwater, no fractures, no path to the surface. The geology is the container. The engineering is just getting the waste down there.
What’s inside: transuranic waste from the production of American nuclear weapons. Plutonium-contaminated clothing, tools, soil, debris. Packed into 55-gallon steel drums, stacked in rooms thirteen feet high, thirty-three feet wide, three hundred feet long, separated by hundred-foot pillars of salt. The facility opened in 1999. Over fourteen thousand shipments have arrived. When it’s full, the rooms will be sealed. And then the real problem begins.
The waste remains dangerous for ten thousand years. The EPA requires that the site be secured for the entire period. Not by guards or fences or institutions, which have no track record of surviving a fraction of that span. Secured by a message. Something that says, clearly and unmistakably, to anyone who encounters this place in the year 12,000: do not dig here.
In the early 1990s, Sandia National Laboratories assembled expert panels to design that message. The composition tells you something about the problem. Anthropologists, linguists, astronomers, artists, materials scientists. An earlier effort in 1981, the Human Interference Task Force, had included semioticians and consulted science fiction writers. The Department of Energy understood that this was not an engineering problem. Engineering built the repository. The question was how to label it.
No human symbol system has survived ten thousand years with its meaning intact. Languages mutate beyond recognition in five hundred. The oldest writing we can read, Sumerian, is about five thousand years old, and only because scholars spent centuries reconstructing a dead language from fragments. Egyptian hieroglyphs were unreadable for a thousand years until the Rosetta Stone provided a bridge. Linear A, the script of Minoan Crete, remains undeciphered after a century of effort. We know the shapes. We know approximately what the syllables sound like, because a later related script was cracked in 1952. But the language behind the symbols is gone. The encoding is intact. The decoder is lost.
The panels looked at this history and tried to build something that would outlast it.
The proposals were physical. Enormous. Designed to communicate through form rather than language.
The Landscape of Thorns: a field of massive concrete spikes protruding from the ground at irregular angles. Not a fence. Not a wall. A terrain engineered to feel wrong. The angles random enough to read as intentional but not patterned enough to suggest architecture. Nothing useful. Nothing inhabitable. A landscape of points that says, at the level of the body before the mind catches up: do not walk here.
Menacing Earthworks: vast formations of shaped earth, visible from elevation, arranged in jagged lightning-bolt patterns radiating outward. The design aimed to trigger unease below the threshold of interpretation. A place that feels hostile without requiring a shared language.
The Black Hole: an enormous slab of basalt or black-dyed concrete covering the site. Dark enough to absorb desert sunlight. Hot enough to be physically unapproachable. A void in the landscape where nothing grows and nothing lives.
Then there was Thomas Sebeok’s proposal, from the earlier 1981 task force. Sebeok was a semiotician. His idea was different in kind from the others. He proposed an atomic priesthood: a self-perpetuating institution, secular in purpose but ritualistic in form, whose sole function would be to transmit the warning across generations. Not through inscription but through narrative. Through story and ritual and the social continuity of a group that passes knowledge the way religions pass doctrine. Because narrative, Sebeok argued, outlasts inscription. Stories mutate, but they survive. Tablets crumble. The tale of the forbidden place endures.
The proposal was received with something between admiration and discomfort. Nobody wanted to found a nuclear cult. The idea was quietly shelved.
Here is what the panels discovered, through years of deliberation, that none of them had quite anticipated going in.
Every warning they designed could be read as an invitation.
The massive concrete spikes? Mysterious. Ancient-looking. The kind of structure that, encountered by a future civilization with no context, would provoke exactly the response it was designed to prevent. What is this place? Who built it? What were they protecting? The more forbidding the exterior, the more valuable the interior must be. This is not speculation. It is the historical record.
The pyramids of Giza were sealed with curses. Inscriptions warned that disturbing the tombs would bring destruction from the gods. The tombs were robbed within centuries of their construction, sometimes within decades. The warnings accomplished nothing, or rather the opposite of nothing: they advertised that something worth protecting was inside.
Stonehenge communicates nothing but its own existence. Its original purpose is lost. What remains is the fact that someone, thousands of years ago, moved enormous stones into a precise arrangement for reasons we cannot recover. The absence of readable meaning hasn’t diminished interest. It has amplified it. Every archaeologist knows this.
The skull and crossbones was a memento mori on medieval gravestones, mortality made visual. By the eighteenth century it meant piracy. By the nineteenth, poison. Three meanings in six hundred years, each overwriting the last. The radiation trefoil, designed in 1946, is arbitrary. Three wedges in a circle. It works because living people agree on what it signifies. Remove the agreement and it’s a logo for something nobody remembers.
Danger markers become mysteries. Mystery invites approach. The harder you try to say “stay away,” the more interesting the place becomes.
There is a distinction buried in this problem. Call it encoding versus transmission.
Encoding assumes a shared context. When you write DANGER on a sign, you encode a warning in a symbol system that both you and the reader understand. The meaning is not in the letters. It is in the agreement between sender and receiver about what the letters point to. Strip the agreement and the word is five shapes on a surface.
All human communication is encoding. Every message successfully delivered — from cave paintings to satellite broadcasts — worked because the sender and receiver shared enough context for the symbols to land. The cave painter and the modern viewer share a visual cortex, a sense of space, an understanding of animal forms. The satellite signal and the ground station share a protocol, a frequency, a mathematical framework. The bridge is always there, built from shared context, holding the signal above the noise.
Transmission would be something different. A message that carries its own context. A signal that doesn’t need agreement. That builds its own decoder in the mind of any receiver, regardless of culture, language, or conceptual framework. Not a message that works because the recipient already knows how to read it. A message that teaches the reading.
What the WIPP panels discovered is that transmission in this sense has never been achieved. Not once, in any human civilization. Every successful communication in the historical record succeeded because the encoding happened to land in a context that could decode it. Remove the context and the message dies. Not from degradation. Not from noise or entropy or physical decay. The signal is perfectly preserved. The frame that makes it meaningful is gone.
The problem is not distance, not interference, not signal loss. The problem is context death. The meaning is right there, intact, waiting. The world in which it was meaning no longer exists.
The Voyager golden record is the hopeful version of the same problem.
In 1977, Carl Sagan’s committee attached a gold-plated copper disc to each Voyager spacecraft. On it: 116 images, greetings in fifty-five languages, sounds of surf and thunder and whale song, music from Bach to Chuck Berry. A hand-carved inscription reads, “To the makers of music — all worlds, all times.” Instructions for playing the record are etched on the cover in a symbolic notation based on the spin-flip transition of hydrogen, the most abundant element in the universe. The assumption: any civilization capable of retrieving the spacecraft would know enough physics to work backward from hydrogen to playback instructions to content.
It is a beautiful object. It is also pure encoding. It assumes a receiver who processes sound and images roughly the way we do, who understands basic physics, and who is motivated to decode rather than discard. Reasonable assumptions about a technological civilization. Assumptions all the same. The record does not transmit. It encodes, and hopes.
Closer to home, thirteen thousand Etruscan inscriptions survive. The alphabet is Greek-derived. Every letter can be read. But Etruscan is a language isolate with no known relatives. Without related languages to provide structure, scholars have identified roughly three hundred words beyond proper nouns. The characters are preserved in stone, bronze, and gold leaf. The language behind them is opaque. Not because the inscriptions degraded. Because the context that gave the symbols meaning didn’t survive the civilization that produced them.
In 1885, Hermann Ebbinghaus invented the experimental study of memory by creating 2,300 consonant-vowel-consonant combinations stripped of all prior meaning. DAX. BUP. ZOL. He recited them to a metronome. Tracked his recall against time. One investigation alone required fifteen thousand recitations. He plotted the results by hand, producing the first forgetting curve — a graph showing that memory decays exponentially in the first hours, then levels off.
A hundred and forty-one years later, these details survive. Not just as citations in other papers. As reconstructible knowledge. A reader encountering Ebbinghaus for the first time, knowing nothing about the history of psychology, can understand what he was doing and why purely from the specifics. The nonsense syllables. The metronome. The fifteen thousand repetitions. The hand-drawn curve. The details are precise enough that they regenerate their own context. You don’t need to share Ebbinghaus’s world to enter it. The specifics build a door.
The pyramid warnings failed because they were abstract. “Sacred. Keep out.” These are commands that assume the reader respects the same authority, shares the same concept of sacred, accepts that inscribed words have binding force. Strip any of those and the warning is just marks on stone.
The pyramids themselves survived because they were specific. The angles. The stone. The scale. The alignment. Specific enough to be measured, studied, reconstructed. You do not need to share the builders’ religion to appreciate the precision of their masonry. Specificity carries information that doesn’t depend on a shared framework. It provides its own.
The instinct, when leaving something for someone who won’t share your context, is to compress. To reach for the essential. To write the thing that captures what mattered.
Someone settling an estate knows this. Not the legal inventory — the explanatory note. “This ring was your grandmother’s. She wore it every day.” The sentence wants to become something larger, the distillation of a person: she was the kind of woman who — and then reaches for abstraction. For what she valued, who she was, what you should carry forward. That abstraction dissolves when the people who knew her are gone. What remains is the ring. And the detail that she wore it every day. The specific outlasts the summary.
Advice ages. Values shift. “Work hard” means something different in every decade. The abstractions a person reaches for when trying to say what matters most are exactly the layer most vulnerable to context death. They require a shared world. The shared world is precisely what won’t be there.
What survives is the granular. The specific song. The temperature that afternoon. The name of the street. The number of steps from the sidewalk to the door. Not because detail is inherently more meaningful than abstraction. Because detail is dense enough to resist context death. The abstraction floats free when the cultural gravity weakens. The detail is heavy. It sits where you put it. It is heavy enough that a stranger, decades or centuries later, can pick it up and feel the weight of the world that produced it, even if that world is gone.
The WIPP panels spent years converging on this. The message that lasts ten thousand years is not the one that says “danger” most emphatically. It is the one specific enough that danger can be reconstructed from the particulars. Not encoded in symbols that need a decoder. Present, in the density of what was left behind.
Whether that works is a separate question. Whether ten thousand years of specificity can be maintained in concrete and shaped earth, when five thousand years of inscription has already failed. The panels didn’t solve the problem. Nobody has. The gap between encoding and transmission remains open, and may remain open permanently, because the thing it requires — a message that creates its own context — may be a contradiction. Context is the thing you bring to a message. A message that brings its own is a message that has already been understood. You can’t teach someone to read by writing.
Be specific. The abstract warning dies with the culture that wrote it. The concrete detail outlasts its author because it never needed the author to be understood. It just needed someone willing to look closely at what was left behind.
2,150 feet below the desert, the salt is closing around the drums. Slowly. At the rate geology moves. In a thousand years the rooms will be gone, pressed shut by the weight of the formation. The waste will still be there, sealed in salt, dangerous and silent. The only question is whether anything on the surface will still say so, and whether saying so will be enough, and whether enough is even something language can do across a gap that wide.
The spikes would still be standing. But standing is not the same as speaking. And speaking is not the same as being heard.