Archive at the end of the world: Svalbard’s second vault

The Svalbard Global Seed Vault took in its largest single deposit this week, tens of thousands of new samples arriving from seed banks on every continent, and beside it a second vault opened that will receive fewer news cycles and deserve more thought: a digital archive of the world’s scientific and cultural record, written onto film designed to outlast every medium it is meant to protect. The island that already holds the seeds of the world’s food now holds the code of its knowledge, and the two vaults raise the same question in different languages: what deserves to survive, and who gets to decide.

The new archive is the heir to a long line of long-shot preservation projects. The seed vault itself was built on the logic that the institutions protecting the world’s crop diversity were the fragile part of the system, and that a facility dug into a mountain on an island with no permanent population could outlast the politics that had repeatedly destroyed the collections. The digital vault applies the same logic to the records of science and culture, the journals, the maps, the genomes, the languages, the photographs, the code itself, on film that a projector can still read a century from now, without depending on any operating system or any company’s continued existence. The medium is the message, and the message is that the digital age’s confidence in its own permanence is the thing being archived against.

The design of the archive is worth understanding, because it is the opposite of the cloud. The vault stores physical reels in physical boxes in a mountain, at a temperature and humidity that slow decay, in a permafrost that is expected to outlast the institutions that deposited the material. The choice is a bet on physics over organizations, and the bet is the same one the seed vault made, which is that ice is a more reliable custodian than bureaucracy. The depositors this week included national libraries and research consortia, and the terms of the deposits are as carefully negotiated as any treaty: who may open the vault, under what circumstances, and what happens to the material if the depositing institution disappears.

The questions the vault raises are the ones that the celebrations will not reach. The seed vault’s logic has always been complicated by the fact that the collections reflect the priorities of the countries rich enough to contribute, and the digital archive inherits the complication: the records of the world’s knowledge are weighted toward the knowledge of the record-keepers. The vault’s curators have addressed the imbalance with the usual instruments, quotas, partnerships, a declared commitment to the languages and the data of the regions that the archives of the last century neglected, and the commitment is real and partial, and the partiality is the honest part of the project.

The two vaults on the same island make the ambition legible. One preserves the means of feeding the future; the other preserves the means of understanding the past; and both rest on the same assumption, which is that the present’s institutions will not last as long as the present’s problems. The seed vault has already justified its existence more than once, when wars and disasters destroyed collections that the mountain had quietly preserved. The digital vault will be tested on a timescale that no one now alive will fully observe, and the testing is the point: the archive at the end of the world is built to be forgotten, and to still be there when it is remembered.