How Humpback Whale Song Dialects Shift Over Decades and What Cultural Transmission Tells Us

In 1998, marine biologist Michael Noad and his colleagues at the University of Queensland documented something that had no clean explanation at the time. Two humpback whales off the east coast of Australia were singing a song-type that belonged to the west coast population, a dialect from the other side of the continent. Within two years, the entire east coast population had abandoned its own song and adopted the new one. All 82 males monitored had switched. The old dialect was gone.
This is not metaphor. Humpback whale males (Megaptera novaeangliae) produce complex, structured vocalizations that last anywhere from six to thirty minutes and repeat for hours. Every male in a given ocean basin sings the same song at any one time, but that song changes continuously. New phrases get introduced. Old ones drop out. The direction of change is not random, it spreads, like a cultural trend, from individual to individual across thousands of kilometres of open ocean.
The Noad study, published in Nature, was the first hard evidence that this spread was social learning, not convergent evolution or genetic drift. The whales were copying each other.
A whale song is built in layers: units combine into phrases, phrases repeat into themes, themes sequence into a full song. Researchers tracking the Pacific humpback population over decades have found that themes are added and dropped in a consistent order, always forward, never backward. No population has ever been recorded reverting to an older version of its song. This directional drift is one of the clearest markers of cumulative cultural change in any non-human animal.

Cetacean biologist Ellen Garland, working with colleagues across the Pacific, mapped song transmission across six distinct humpback populations between Australia and French Polynesia. Her 2011 paper in Current Biology showed that new song types moved eastward across the Pacific in a wave, population by population, over a period of years. The pattern matched human cultural diffusion models more closely than it matched any biological transmission model.
The mechanism behind this is vocal learning, the ability to hear a sound, encode it, and reproduce it. Most mammals cannot do this. Humpbacks can, and they are precise about it. Acoustic analysis shows that whales copy new phrases with high fidelity, often within a single breeding season.

Only male humpbacks produce the full song, and they sing almost exclusively on breeding grounds and during migration. The leading hypothesis is that song functions in male competition and mate attraction, though the exact mechanism is still debated. What is clear is that conformity matters, a male singing last year's dialect on this year's breeding ground is singing the wrong song. The social pressure to update is real and measurable in the speed of adoption.
This conformity pressure may be what drives cultural transmission so efficiently. Sperm whales offer a parallel case. Their clicks, called codas, vary by clan, and calves learn the coda pattern of their maternal group. A 2016 study by Shane Gero and Hal Whitehead in Royal Society Open Science confirmed that coda repertoires are stable within clans across generations but differ sharply between clans occupying the same waters. The whales are not just making sounds. They are signalling group membership through learned vocal identity.

Cultural transmission in animals has been documented in chimpanzees, orangutans, and some bird species. But those cases typically involve tool use or foraging technique, behaviours that are easier to attribute to individual learning or environmental pressure. Whale song is different because it is purely social. There is no foraging advantage to singing a particular phrase. The only reason to sing what your neighbours sing is because you heard them and chose to copy them.
That distinction matters to researchers working on the definition of culture in non-human animals. The standard framework, proposed by primatologist Andrew Whiten and colleagues, requires that a behaviour be socially learned, spread through a population, and persist across time. Humpback song dialects meet all three criteria, and they do so at a population scale and geographic range that no primate behaviour has matched.
The Indian Ocean humpback population, which migrates along the western coast of India and into the Arabian Sea, has received less acoustic monitoring than Pacific populations, but preliminary tagging data from the Wildlife Institute of India and international collaborators suggest the same pattern of dialect variation between northern and southern groups. The research is ongoing.
The standard objection to calling whale song a form of culture is that whales lack the social complexity to sustain it, no persistent groups, no face-to-face teaching. The evidence now suggests that objection had the causality backwards. Humpbacks on breeding grounds are temporary aggregations, not stable communities, yet song spreads through them with the speed and fidelity of a viral trend. The transmission does not require long-term bonds. It requires attention, memory, and the capacity to reproduce what was heard.
Those three things, taken together, are the minimum substrate for culture. The whale does not need to understand why the new phrase is being sung. It needs only to hear it, remember it, and produce it. That is precisely what the acoustic record shows happening, decade after decade, across every humpback population with sufficient monitoring.
The song is not evidence of language, and it would be a mistake to frame it that way. But it is evidence of something the field spent decades assuming was uniquely human: a population-level behaviour that changes not because the environment changed, but because individuals heard something new and passed it on. The ocean, it turns out, has been running this experiment for longer than anyone was listening.