Posted by Humpback whale communication is the closest thing SETI has to a field test. The whales are here, on this planet, breathing the same air we do, and after fifty-odd years of listening to them we still cannot say with confidence what their songs mean. That is not a failure. It is data about how hard the problem is, and it comes from an animal that shares our biochemistry, our vertebrate nervous system and tens of millions of years of common ancestry. A radio signal from a star four hundred light-years away would offer none of that.
The idea has a formal home. A loose collaboration calling itself Whale-SETI brings together Laurance Doyle of the SETI Institute, whale biologist Fred Sharpe of the Alaska Whale Foundation, and Brenda McCowan’s animal behavior lab at UC Davis. Doyle is not a marine biologist. He spent years on the Kepler mission and led the team that found Kepler-16b, the first transiting planet confirmed to orbit two stars at once. He came to whales through information theory, which turns out to be the shortest road between the two subjects.
A Song With Rules Nobody Wrote Down
When Roger Payne and Scott McVay published “Songs of Humpback Whales” in Science in 1971, the surprise was not that whales made noise. It was that the noise repeated. Their spectrograms showed the same units recurring in the same order, units building into phrases, phrases repeating into themes, themes running in a fixed sequence, and then the whole cycle starting over, for hours at a stretch. The field still uses Payne’s vocabulary. Ryuji Suzuki, John Buck and Peter Tyack later put the recordings through information-theoretic analysis and found the hierarchy running several levels deep, which is not something random noise does.
Only males sing, mostly on the breeding grounds, and in any given season the males of a population sing essentially the same song.
What the song is actually for remains unsettled. Mate attraction is the oldest hypothesis and still the most popular. Male-male spacing, an acoustic display of fitness, coordination of movement, even some sonar-like function have all been argued in the literature. Half a century in, the honest answer is that nobody has closed the case. That is worth sitting with, because it is precisely the position SETI would occupy the morning after a confirmed detection: an obviously structured signal in hand, and no agreed account of what it is doing.
Song Revolutions, Rolling East
Here is the part that makes humpback whale communication a genuinely strange object rather than merely a complicated one. The songs change, and the changes spread. Michael Noad’s team documented the cleanest case in Nature in 2000: two humpbacks off eastern Australia showed up singing the song of the west Australian population, from the far side of the continent. Within two seasons the entire east coast population had dropped its own song and taken up theirs. Not drifted toward it. Replaced it.
Ellen Garland widened the picture across the South Pacific using eleven years of recordings from six populations, published in Current Biology in 2011. Song types moved eastward in waves, from eastern Australia toward New Caledonia, Tonga and on to French Polynesia, taking roughly two years to make the crossing. A cultural sweep in whales, with a measurable direction and speed.
Whatever that is, it is not genetics and it is not individual invention. Something is being learned and handed on at the scale of an entire population.
The Whale Who Answered
In August 2021 the Whale-SETI group was working in Frederick Sound, in southeast Alaska, with a hydrophone and an underwater speaker. The day before, they had recorded a “whup” from a group of whales, a contact call humpbacks use in ordinary social situations, roughly the acoustic equivalent of a nudge. They played it back into the water. A humpback the researchers know as Twain came in, circled the boat, and answered.
She answered thirty-six times over about twenty minutes. More interesting than the count is what she did with the timing: when the team shortened the gap between playbacks her replies came faster, and when they stretched it out she waited. The encounter was written up in PeerJ in 2023 by McCowan, Sharpe, Doyle and colleagues as a demonstration of interactive bioacoustic playback, which means deliberately provoking a response instead of eavesdropping on one.
Worth being careful here, because the story traveled a good deal further than the paper did. One whale, one afternoon. Nobody on that team claims Twain said anything, understood anything, or grasped that her interlocutor was a boat rather than a whale she could not see. What the encounter establishes about humpback whale communication is narrower and still useful: a nonhuman animal took turns with an unfamiliar caller, on the caller’s schedule, and stayed with it for twenty minutes. Sharpe has put the rationale plainly enough. If a signal ever does arrive, we will badly want to have practiced.
What Information Theory Can and Cannot Do
Doyle’s contribution to humpback whale communication research is a toolkit that does not require you to know what anything means. Rank the signal units of a communication system by how often they occur, plot frequency against rank on log axes, and human languages, all of them, regardless of family, yield a line with a slope near minus one. That is Zipf’s law. Doyle and McCowan ran the same analysis on bottlenose dolphin whistles and on humpback song units and landed in the same neighborhood. Infant babble gives a much flatter line, and it steepens as the child acquires language.
A second measure, conditional entropy, asks how far back the dependencies in a sequence reach. Does knowing the last unit tell you anything about the next one, and the one after that, and how deep does the rule structure go before the sequence stops being predictable?
Neither measure hands you a translation. That is both their point and their ceiling. They can establish that a system carries language-like statistical organization without revealing one word of its content, which is exactly where a real SETI detection would strand us. The Wow! signal of 1977 lasted seventy-two seconds and never came back. BLC1, the Breakthrough Listen candidate that appeared to come from the direction of Proxima Centauri, survived long enough to be thrilling and then resolved into terrestrial interference. Neither one ever reached the interpretation problem. Twain’s thirty-six answers did.
Where the Analogy Gives Out
It gives out fast, and being honest about where matters more than keeping the comparison tidy. We share an ocean planet with humpbacks: air-breathing, warm blood, mothers and calves, hunger, sound as the obvious medium. The overlap is enormous and we still cannot read them. A transmitting civilization would share physics with us and possibly nothing else. Whale researchers at least get context. They can see who called, who was nearby, what happened next, who moved. Interstellar distance deletes every bit of that.
Which is why mathematics keeps being nominated as the common ground: prime numbers, the hydrogen line, atomic constants, the sort of scaffolding Frank Drake built into the 1974 Arecibo message. It may be the only vocabulary two civilizations can count on sharing. It is also thin. You can establish a mutual counting system and learn nothing whatsoever about who is doing the counting.
There is a detail in all this I find hard to shake off. The Voyager Golden Record, launched in 1977 and now well past the heliopause, carries greetings in dozens of human languages and a recording of humpback song. We put a message aboard that we cannot ourselves translate, and sent it toward whoever might be listening. If it is ever found, the finders will hit the same wall with our whale track that we have been standing at since Payne’s spectrograms, and presumably with the human greetings too.
Listening as a Discipline
The useful transfer between the two fields is not a theory. It is a habit: record first, describe the structure before assigning meaning, keep the null hypothesis alive far longer than feels comfortable, and treat a beautiful pattern as a reason to look harder rather than as an answer. Radio astronomers learned this the slow way, through decades of pulsars, satellites, microwave ovens and BLC1. Whale researchers are learning it in Frederick Sound, one playback at a time, and their subject has the enormous advantage of answering back.
SETIworld follows both threads as they develop, from hydrophones in Alaskan channels to technosignature surveys and the arguments about what should count as evidence. If you want the whale-song papers and the radio results in the same place, and people to argue with about both, come read along with us.
