Posted by Radio missions aimed at finding another civilization do not look for messages. They look for something far cruder and far more tractable: a lump of electromagnetic energy squeezed into a frequency range so narrow that no known astrophysical process can make it. Nature is loud across the radio spectrum — pulsars, masers, synchrotron emission from galaxies — but nature smears its energy out. Technology concentrates it. That single asymmetry is the physical foundation of the entire enterprise.
Whether anyone out there is transmitting is unknown. What is known is what a transmitter would look like arriving here, and that is enough to build instruments around.
Why radio, and why these frequencies
Radio waves cross interstellar space cheaply. They pass through the gas and dust that block visible light, they do not require the transmitter and receiver to be pointed with absurd precision, and the technology to generate them is something a civilization is likely to stumble into early, since we did.
The frequency question has no clean answer, which has not stopped people arguing about it since the 1950s. The classic proposal is the “waterhole,” the quiet band between the 1420 MHz emission line of neutral hydrogen and the 1720 MHz line of hydroxyl — the two dissociation products of water, bracketing a stretch of spectrum where the sky happens to be unusually quiet. It is a nice piece of reasoning that assumes the other party finds it as compelling as we do. Modern radio missions hedge by covering as much bandwidth as the receivers allow, often 1 to 12 GHz in a single campaign.
What the software is actually hunting for
A digital backend chops the incoming spectrum into channels — for narrowband SETI work, channels roughly a hertz wide, which means billions of them across a wide band. The search then runs over frequency and time simultaneously, looking for a thin bright line in what is otherwise a field of noise.
The critical extra dimension is drift. Earth spins and orbits; so does whatever planet holds the transmitter. Those motions Doppler-shift the signal, so a genuinely extraterrestrial narrowband tone should slide steadily in frequency over the course of an observation rather than sitting still. Pipelines search across a whole range of possible drift rates, and a signal with a drift rate of exactly zero is usually a strong hint that the source is sitting on the ground, moving with the telescope.
Human interference is the permanent adversary. Satellite downlinks, aircraft transponders, mobile networks, a badly shielded pump in the observatory basement — all technological by definition, and therefore all capable of producing exactly the narrowband structure SETI is tuned to notice. Most working time in radio missions goes into throwing these out, not into finding anything.
The nodding trick
The main defence is elegantly simple. Point the dish at the target star for five minutes, then swing off to an empty patch a few degrees away for five minutes, then back. Repeat: on, off, on, off, on, off. A signal from that star should appear only in the on-source scans. Interference leaking in through the receiver or the sidelobes does not care where the dish points and shows up in everything.
Breakthrough Listen made this cadence standard practice on the Green Bank 100-metre, and it kills the overwhelming majority of candidates immediately. Multiple telescopes at different sites provide the next layer, since interference in West Virginia has no reason to appear simultaneously in New South Wales at the same sky coordinates.
Then there is the case that survives the first cuts. In 2017 a team led by Abel Méndez recorded an odd repeating signal at Arecibo while observing the red dwarf Ross 128. The pattern was strange enough to attract international attention for a fortnight before follow-up traced it to geostationary satellites transmitting in a nearby band. Nobody had done anything wrong; the system worked exactly as intended, in public and fairly quickly.
Where to point, and how long to stare
Two strategies compete for telescope time. Targeted searches take a list of nearby stars — increasingly stars with confirmed planets, courtesy of Kepler and TESS — and observe each one deeply. Proximity is the whole argument: signal strength falls with the square of distance, so a transmitter at ten light-years needs a hundredth of the power of the same transmitter at a hundred.
Wide surveys sweep the galactic plane instead, accepting less sensitivity per target in exchange for enormous numbers of stars in the beam. Neither approach can be called correct, because nobody knows whether technological civilizations cluster around Sun-like stars or turn up somewhere no astrobiologist would have suggested.
A third idea has crept in more recently: the SETI Ellipsoid. If a civilization within a few hundred light-years watched a conspicuous event — a bright supernova, say — and decided that was a sensible moment to announce itself, the resulting signal would reach Earth along a calculable surface in space and time. Astronomers can work out which stars fall on that surface this year and watch them. It is a guess about behaviour rather than physics, but it produces a finite, checkable target list.
Timing sabotages every strategy. A beacon that runs for ten seconds a year, or a beam that sweeps across the sky pointing at each star in turn, would be missed by almost any observing schedule ever executed. A twenty-minute non-detection is a statement about twenty minutes, and dressing it up as anything larger would be dishonest.
Getting the sky for free
The most consequential recent change in radio missions is that they no longer always need their own telescope time. Commensal observing takes the raw data stream from an instrument booked for entirely different research — pulsar timing, galaxy surveys, transient monitoring — and copies it into a parallel SETI backend.
The COSMIC system on the Karl G. Jansky Very Large Array does exactly this, quietly processing whatever the array happens to be looking at and pushing the number of stars examined into the millions without competing for a single observing proposal. MeerKAT hosts a similar arrangement. The Square Kilometre Array, under construction in Australia and South Africa, is being designed with this kind of piggyback capability in mind from the start.
The cost is data volume. These systems generate more raw spectra than any human will ever review, which pushes the field toward automated candidate ranking and machine-learning classifiers. A computer can flag an anomaly; it cannot tell you what the anomaly is. That determination still requires someone to re-observe, check the archives and try hard to explain the thing away.
What a real detection would need
Start with reappearance. A candidate that never returns, however beautiful, joins the Wow! signal of 1977 in permanent limbo. Then localization: the signal must stay fixed to sky coordinates as the Earth rotates beneath it. Then independent confirmation from a separate observatory on a different continent, running different hardware, with people who were not involved in the original claim.
Structure would be the last and strongest layer — modulation, repeating sequences, information content that no natural source produces. That would probably arrive long after the initial detection, if it arrived at all, and a first technosignature may well be something dull and unmistakably artificial rather than a message anyone can read.
Six decades of radio missions have not answered the question. They have converted it from speculation into a measurement with stated limits: these stars, these frequencies, this sensitivity, this much time, nothing found. The volume of that searched space is still tiny. SETIworld tracks the surveys and the candidates that pass the first filters, which is where this story will break if it ever does.
