In the early 1990s, Sandia National Laboratories put a panel of experts in a room and handed them a brief nobody had written before: design a sign for the year 12,000 that anyone could decipher.

Half a mile under the desert east of Carlsbad, New Mexico, the government was preparing to bury plutonium-contaminated gloves, tools, and machinery left over from building nuclear weapons, and federal rules required markers that would keep people from digging into the site for ten thousand years. The expert panel held linguists, anthropologists, an astronomer, a science fiction writer. They were told to assume no shared language with whoever showed up, no surviving institution, no continuity of any kind. Every writing system on earth is younger than the hazard.

What was drafted reads like scripture written in a panic. This place is not a place of honor. No highly esteemed deed is commemorated here. Nothing valued is here. What is here was dangerous and repulsive to us. Every line braces against the same fear, that a person standing on that ground would read a marked, engineered, deliberate site the way we read the pyramids: as proof that something precious is buried underneath.


The fear is earned, because we have watched a language go dark before. Egyptians wrote in hieroglyphs for more than three thousand years. The last dated inscription was cut into Hadrian's Gate at Philae in the summer of 394, by a priest of Isis who wrote that he hoped his words would stand "for all time and eternity." Then the reading stopped. Fourteen centuries followed in which those carvings stood in the open on temple walls people walked past every week, perfectly visible as marks and completely mute as language. The stone held up its end. What failed was the last generation of people who knew what the marks were for, and the message only came back in 1822 because a slab dug out of the mud at Rosetta happened to carry the same decree in a script scholars could still read.


Even a sign that stays legible won't hold still. A skull above crossed bones has meant incompatible things inside a single culture: carved on European tombs and ossuary walls it was a memento mori, and in some traditions a promise of resurrection rather than a threat. Two centuries later the same arrangement flew from a pirate mast. By 1971, when the new poison center at Children's Hospital of Pittsburgh tested the symbol on local kids, they read it as adventure, treasure, the baseball team downtown. The center's answer was Mr. Yuk, a scowling green face invented because the internationally standard mark for lethal poison had drifted into a mascot.


Once you accept that the code drifts and the reader keeps changing, the panel's stranger proposals start to look like the sober ones. Thomas Sebeok, a semiotician who worked on an earlier study, argued that no object could carry the warning at all and proposed an atomic priesthood: a standing body charged with holding the knowledge and reissuing the message every generation, wrapped in ritual and taboo so the rule would transmit whether or not anyone still understood the physics. Françoise Bastide and Paolo Fabbri proposed breeding cats that visibly change color near radiation, then seeding songs and proverbs about them, so a rule about a glowing cat would ride down the centuries in folklore long after the reason for it was lost. Ray cats. From an engineering standpoint the suggestions are ridiculous. As semiotics they're precisely right, because both proposals give up on the durable object and hand the message to a living carrier.


Along the Sanriku shore of northeastern Japan stand hundreds of stone markers, some of them six centuries old, set by survivors to record how far the water came. One above the hamlet of Aneyoshi, placed after the tsunami of 1933, carries a flat instruction: High dwellings are the peace and harmony of our descendants. Remember the calamity of the great tsunamis. Do not build any homes below this point. In March 2011 the wave stopped about three hundred feet short of that stone. The eleven households above it were still standing that evening.

Elsewhere on that coast, houses and shops went up below the markers, and the stones became scenery to people who had stopped repeating the sentence. Aneyoshi's stone worked because a small village kept the rule in the mouths of its children, and because the language on the rock is still the language spoken at the bottom of the hill. The transmission happened between people.


We get this backwards nearly every time. When we want something to outlast us we reach for the tougher material: the archival copy, the etched plate, the standard, the document written once and filed. Durability of the object is the cheap half of the problem, and the only half we control. The expensive half is a reader who still holds the code, and that half can't be built, only kept alive one generation at a time.

Most families have a box of photographs where the faces have gone anonymous. The paper is fine. The silver is stable. Someone knew every one of those names and never wrote them down, because at the time the knowing seemed like the permanent part.

So make the object as durable as you like, then do the harder thing and hand the meaning to someone who will keep saying it out loud. Stone keeps its shape on its own. A sentence needs a mouth.