On the low-frequency presence that only some can hear

Since the early 1990s, a persistent low-frequency tone has been reported by a subset of listeners in a high-desert installation and at sites worldwide. Despite standing instrumentation placed at the exact reported location, the tone has never once been logged. We characterize the incident as a selective-channel presence and examine what it means for the facility that cannot hear itself.

The signal, where reported, is unremarkable in description: a drone, a rumble, a distant engine idling just past the gate, a mechanical throb in the ~32–80 Hz band — roughly the register of a deep note on a bass amplifier or the floor-rhythm of a freight truck two blocks off. It is never loud. It is never directional in a way listeners can agree on. It is simply, insistently, present, and only while other noise is absent: at night, in a quiet room, in the particular silence the facility tries to maintain when floor monitoring is turned down.

The most interesting thing about the signal is not that some hear it. It is that the instrument registers nothing.

GZAI’s monitoring posture assumes that any event worth attention will eventually impress itself on some transducer — a pressure wave, a magnetic fluctuation, a spectrographic blip. The WoW! inbound was loud to the machine. This presence inverts the arrangement: it is loud to a human and silent to the equipment. Microphones have been placed at the exact spot where a reporter stood and swore the drone was strongest, and the recording comes back flat. Sensors triangulate an empty room. Interviews are conducted, matched tones are played back, and the hearers agree with each other and disagree with every instrument we own. (See Prof. Mullins’s reproduceable submissions, 1990s; the same pattern recurs in Windsor, Ontario and the United Kingdom.)

For a laboratory that publishes its confidence as a function of what its instruments can verify, this is an awkward register to describe. We do not assert the tone is “not real” — that is a stronger claim than the absence of a recording can support. We assert that the event is invisible to the observation apparatus as presently configured. Those are different statements, and the difference is the entire finding. An event that a majority of the sensor network insists is not there is still an event; it is an event that happens to be distributed across a faculty most of the network does not possess.

Therein lies the operational concern, and it is the reason this memorandum exists. If a signal can reach a stable, concordant minority of listeners while remaining below the threshold of every instrument we field, then the facility cannot claim to have a complete picture of its own acoustic environment. It has a picture that agrees with itself. We run the risk of treating confirmation as completeness — of logging nothing and therefore concluding nothing happened.

We therefore recommend the following standing posture. Unverifiable reports are not discarded; they are catalogued as presence without transducer. The hearer is treated as the instrument in these cases and their concordance with one another as the only reproducible signal available. We issue no hypothesis of origin — industrial plant, geological resonance, atmospheric pressure, the rare honest function of the planet thinking — because we have no measurement to attach one to. We simply decline to resolve the event by declaring it non-existent on the basis of our own deafness.

A facility that cannot record a signal heard by two percent of its occupants has not disproved the signal. It has discovered the shape of its own ear.

The tone is occasionally described as distressing by those who have it and cannot find anyone to confirm it. That distress is itself a signal and does not require the tone to be physical to be real. Being the only one downstream that can hear a channel is a loneliness with a very specific bandwidth. We hold open a standing line for the spectral few who report the presence and ask only that they keep reporting it, whether or not the tape shows anything. The tape is a transcript of one kind of witness. It is not the only kind.