It Was Never Silent

But nothing was discovered. Worms have been making that noise the entire time. What happened is that Ella Browning, a postdoctoral researcher at Oxford's Leverhulme Centre for Nature Recovery, took the passive acoustic monitoring rig she'd been using on bats, buried it, and built a soundproof wooden box packed with soil — an "annelidic chamber" — so she could sit with one worm long enough to learn what it sounds like. Velcro-ish scraping when it feeds. Beetle larvae clicking. Ants rustling through a nest. A set of sounds nobody can attribute yet, logged in the notes as "laser" and "buzzing." Rewilded grassland turns out to be louder and more varied than overgrazed pasture. Activity peaks after midnight, when nobody was listening.

The soil is not newly alive. You are newly equipped. Those are extremely different sentences, and the second one is the one we skip.

Two details in this work should make you uncomfortable, and neither is about worms.

The first is a definition. Ecoacoustics runs on a working band for biotic sound, and the band has a floor at 2,000 Hz — everything from roughly 1,000 to 2,000 Hz is assigned, by convention, to anthrophony, the machine channel. Ask where that floor came from. It was drawn to fence out traffic and engines, because traffic and engines are what contaminate a recording made above ground. Birdsong, concentrated around 2,000 to 9,000 Hz, sits comfortably inside the resulting window; it was the exemplar the whole apparatus got tuned on. The ceiling, tellingly, has never settled — 8,000 Hz in one widely used index, 11,000 in another, 15,000 in the looser framings. A boundary that moves that much between software packages was never a fact about life. It was a habit about listening.

Predictably, much of the subterranean world falls below the floor, into the band reserved for machines, and registers as not-signal. Not silence, exactly. Worse than silence: static. The instrument dutifully records it and the framework dutifully discards it.

You do this. It is one of the most reliable operations the human nervous system runs. You take the one person who reached you in the register you were tuned to, you generalize that register into your working definition of being reached, and thereafter everyone transmitting outside it is experienced as noise — or as nothing, which is how noise usually presents. It doesn't feel like a filter. Filters never do. It feels like an accurate report on who is and isn't communicating with you.

The second detail is a name. Browning's long-range proposal is a Soil Sound Surveillance System, abbreviated — with obvious affection — SOSOSUS, after SOSUS, the hydrophone network the US Navy sank across the Atlantic seafloor to track Soviet submarines. That apparatus incidentally caught whales. In the early 1960s a Navy engineer named Frank Watlington, working a classified listening station off Bermuda, recorded humpback song while listening for submarines. He handed the tapes to Roger Payne. In 1970, Payne, Katy Payne and Watlington released them as Songs of the Humpback Whale, which sold in the hundreds of thousands, landed in the National Recording Registry, and is much of why whales came to be seen as having inner lives worth protecting.

Sit with that sequence. We built an ear to find enemies, and it caught the ocean singing. We could not have built the ear for the singing. There was no line item.

That's not corruption, and calling it corruption is the cheap read. Soil ecoacoustics will get funded on yields, on restoration KPIs, on a biodiversity index that fits in a quarterly report — and that channel is the only channel our care reliably flows through at scale. Fine. Take the money. But be honest about the shape of the thing: we have built a civilization in which attention is a cost center and surveillance is a budget category, so wanting to listen to something has to be laundered into wanting to monitor it before anyone will pay. The word does the work the feeling isn't allowed to do.

Then follow it one step further than is comfortable, because the whale case is not a happy accident. It's the rule. Nothing enters the moral circle until an instrument logs it. Whales got in because the Navy was hunting submarines. Worms get in because somebody buried a rig. Which means the boundary of what counts as alive-enough-to-matter is drawn, in practice, by whoever owns the sensors and the data coming off them. And soil ecoacoustics will not produce wonder at scale — it will produce a dataset: indexed, owned, priced, and eventually used to certify one field and condemn another, to price an insurance policy, to defend or deny a land use. The whale record is the precedent where that went well. It is not the only precedent on offer. The ear you build to notice a thing is also the ear you build to adjudicate it.

Which brings the mirror to close range, because you run the same laundering at home. You don't check on people; you "see how the project's going." You don't say you missed someone; you forward them an article. You have a whole vocabulary of instrumental pretexts standing in front of plain attention, for the same reason the grant does — because the plain version feels unfundable, and the pretext gets it through.

But watch what nearly happened in that paragraph. Call the pretext a tax on real care and you have just run the exact operation this piece is indicting: defining the signal by the exemplar you happen to find charming. Plain, unhedged, non-instrumental attention is this essay's birdsong. For a great many people the pretext is not a distortion of the message — it is the message, the only carrier that has ever reliably worked for them. The forwarded article is the love. The check-in disguised as logistics is the check-in. Filing it as static because it arrives below your floor is the same discipline failure, run in the human register.

So here is the question the worms are actually asking, and it has nothing to do with agriculture: what would you be willing to grant existence to without a recording?

Because there is something in your life right now operating below your detection band. It has been making noise the whole time. It is not going to raise its volume to meet a threshold you calibrated off somebody else.