What will be left of us?

A 2014 paper set out the concept of technofossils: manufactured objects entering the sedimentary record. Many are made of materials very rare in nature (uncombined iron, aluminium, titanium) or unknown in it (uncombined vanadium and molybdenum, boron nitride, artificial glasses, plastics). Since the Second World War they have spread over every continent and into the sea, and their range of forms, the paper says, rivals biological diversity.

Durability is not uniform. Concrete and brick can degrade in sulphate-rich or chloride-rich groundwater, iron corrodes, plastics break down in light and heat. What keeps the record thick is volume and poor recycling: roughly half of aluminium is recycled, under a fifth of plastics, under a tenth of concrete, and much of what escapes is hard for scavengers and microbes to digest.

Landfills hold the densest concentrations. The strangest entry is downward. Mine workings reach hundreds of metres into the ground and boreholes several thousand, and the paper gives those traces extremely high preservation potential, until the rock they cut is eroded or metamorphosed.

Fallout from atmospheric nuclear testing, fly ash, nitrogen isotope ratios altered by fertiliser and fuel burning, and the carbon isotope signature of burned fossil fuel are all recorded in sediments or ice, most of them worldwide.

Nothing in that has to be decoded. It is the part of our record that does not depend on anyone understanding us.

Sandia National Laboratories convened an expert panel in the early 1990s to work out how to mark a nuclear waste repository in New Mexico so that people would stay away for ten thousand years, which is the period the regulation covers. Engineers, archaeologists, anthropologists, linguists, astronomers and designers sat on it.

The pessimism is in the assumptions rather than the conclusions. One of the two teams assumed that political change would take the area out of United States control and that vandalism would continue, and designed for a break in society that might disrupt languages and institutional memory. The other trusted writing more than pictures: you can recognise the people and animals in a cave painting, it noted, without knowing what the painter meant by them.

So the design is redundant on purpose. Earthen berms made jagged to read as foreboding rather than honorific. Warnings at four levels of complexity, from the shape of the place to a technical account buried in rooms under the earthworks, in about seven languages, with wall space left blank for a future generation to add its own. Materials chosen to have little salvage value.

The panel did not conclude that this is impossible. Its members put the chance of the markers still standing after ten thousand years at 40 to 85 per cent, most estimates between 60 and 70. The chance that a surviving message is read correctly fell over time too, most for a society with less technology than ours. The report says a system that both survives and stays interpretable for that long needs further study before anyone builds it. Ten thousand years is about twice the span of recorded history.

Most of the world's capacity to store information has been digital since the early 2000s: 94 per cent of it by 2007. That storage depends on media, hardware and software that keep changing.

None of it lasts unless it is copied to new media every few years. Storage media have short physical lifetimes and shorter market lifetimes, and the software needed to interpret a file goes obsolete as fast as the medium it sits on. Rothenberg put it plainly in 1999: there is as yet no viable long-term strategy to ensure that digital information will be readable in the future.

A civilisation that wrote more than any before it may leave a thinner readable record than one that used clay.

A long message assumes a reader with curiosity, tools and a reason to look. Those are assumptions about a future nobody can constrain.

The accidental record has no such requirement. What is buried will be there, in the rock, whether or not anything comes to look at it.