The Frequency We Missed

The band runs from about 1.42 to 1.72 gigahertz. At the bottom sits the emission line of neutral hydrogen — H. At the top sits a line of the hydroxyl radical — OH. Put them together and you get H₂O, and in the Project Cyclops design study, NASA engineer Bernard Oliver named the region between them the "water hole," reasoning that water-based species would naturally gather there to talk, the way animals gather at a watering hole on the savannah. It is a lovely image. It is why everybody remembers the band.

It is not why the band was chosen. Underneath the pun sits a real piece of physics: that stretch of spectrum falls inside the galactic microwave window, the quietest neighborhood available to any receiver anywhere. Below it, galactic synchrotron noise climbs. Above it, atmospheric absorption and quantum noise climb. In between, total system noise bottoms out. That argument holds regardless of what the transmitting species is made of. The water hole was a mnemonic draped over a noise-minimum calculation — the sales copy, not the reasoning.

So the easy version of this story — we searched for water-people at the water frequency, how embarrassingly parochial — is wrong. Discard it. The parochialism is real; it just lives one level up.

The noise-minimum argument optimizes for a receiver. Ours. It quietly assumes a transmitter that wants to be found cheaply by strangers — broadcasting omnidirectionally, on a budget, having thoughtfully worked out where our listening is cheapest. That is not a physics constraint. That is a guess about motive, shaped like a NASA engineer with a grant proposal: build the beacon, make it findable, be generous. We never assumed aliens were made of water. We assumed they were altruists with our budget problems.

Drop that assumption and the millimeter band stops looking exotic. Enormous bandwidth. Tight beam directivity. Precisely what you would use for point-to-point infrastructure you were not trying to make findable at all.

Which is where Louisa Mason comes in. Mason, a PhD researcher at the University of Manchester, has run the first technosignature search ever conducted with the Atacama Large Millimeter/submillimeter Array — two narrow spectral windows near 90.6 and 93.2 gigahertz, inside four archived ALMA observations originally collected to study something else entirely. Frequencies she describes as "almost completely unexplored for SETI."

She found nothing. No candidate technosignatures above detection threshold.

Which is a null result, and is also the most interesting thing anyone has said about SETI in a while — because the finding is not about the sky. It is about the search.

And the second result from Mason's group is the one that should have taken the headline.

Take a previous radio SETI survey: 1,327 telescope pointings, and a published figure for how many stars it covered. That figure came from the Gaia catalogue — count the catalogued stars falling inside each beam, add them up, get about 288,000. Mason's team asked a different question. Not how many stars have names in there, but how many stars are in there. They ran the same pointings through the Besançon Galactic Model, a simulation of how stars are actually distributed through the Milky Way, faint and unresolved ones included.

Six point one million.

Twenty-one times more. Nobody moved a telescope. Nobody built anything. The stars were always in the beam. The survey had been reporting its own catalogue's completeness limit and calling it a measurement.

That is not a story about assumptions. It is a story about accounting, and it generalizes well past SETI to every survey science: your denominator is set by what you can name, not by what you can see. Those are not the same number, and the gap is not small.

Put the two findings side by side and something uncomfortable falls out. Wrong band, and wrong star count — two errors pushing in opposite directions on the same quantity. Which means the field's headline result, the great silence, the thing that has launched a thousand Fermi paradox essays, has never once carried an honest error bar. "We looked and found nothing" was a claim nobody was in a position to make with a straight face. The silence was never measured. It was inherited.

The comedy here is not that we failed. It is that we spent six decades looking for someone else, and what we kept turning up — in the choice of band, in the assumed motive, in the star count nobody checked — was an extremely detailed picture of us. A search is a self-portrait. What you scan, at what pitch, with what patience, and what you are willing to count: those choices encode your assumptions about what a mind is, what it wants, and where it would stand to be heard. There is no view from nowhere. There is only the view from a water-based primate with a limited budget and a fondness for elegant coincidences.

There is a generous reading here, and I want to be honest that it is a choice: the correction mechanism worked. Somebody looked at the method instead of the data, noticed the method was wearing a costume, and published. Slowly, embarrassingly, in public — but working.

The less generous reading is also on the table. A single design study set a band, and sixty years of instrumentation, funding, graduate training, and institutional habit crystallized around it, until the cost of deviating compounded past what anyone could propose. That is path dependence, and it is not a failure of anybody's reasoning — it is what capital does to inquiry. The mechanism did not take sixty years because the question was hard. It took sixty years because the ground everything stood on was poured early and nobody could afford to move.

Both readings are true. The institution is the slow part.

What is not in question: the sky is enormous and almost entirely unchecked, the odds remain terrible, and none of that matters — because the alternative, deciding we already looked, was always the only way to guarantee we would never hear anything.

Turn the dial. See what's up there.