Researchers say they have unlocked a kind of sperm whale “alphabet” with the aid of machine learning, a breakthrough in understanding cetacean communication
The result extends Project CETI’s effort to apply AI to sperm-whale vocalizations and follows broader work using machine learning across animal communication research. It matters because a repeatable structure in whale sounds gives researchers a more concrete unit to test than undifferentiated recordings.
The work also fits the use of self-supervised learning for animal sounds, where models can find patterns in large, weakly labeled acoustic datasets before scientists determine what those patterns mean.
First-order effects
Cetacean-communication researchers gain a machine-learning-derived vocabulary of sound units to validate against behavioral and social observations.
The finding shifts the immediate task from collecting and cataloging calls toward testing whether combinations and contexts carry consistent meaning.
Second-order effects
Projects studying other species can use the result as support for adapting language-model and representation-learning methods to nonhuman vocal data, as in cross-species ML decoding research.
The value of long-term, high-quality underwater audio datasets rises, since model findings require repeated observations and contextual validation rather than audio alone.
Third-order effects
If results are independently replicated, animal-communication research could become more standardized around shared datasets, model benchmarks, and behavioral validation protocols.
The field may increasingly separate pattern discovery from claims of meaning: AI can surface structure at scale, while biological interpretation remains the limiting scientific step.
The trend: Machine learning is moving animal-communication research from acoustic pattern detection toward testable models of structured signaling.
Sperm whale ‘alphabet’ discovered, thanks to machine learning - https://techcrunch.com/... great: soon we will be able to apologise to #whales for hunting them almost to extinction... #linguistics
The authors are CETI team members from @MIT_CSAIL: @jacobandreas, Daniela Rus, @pratyusha_PS and Antonio Torralba. With @sgero, CETI's Biology Lead/Scientist in Residence @CarletonScience, the late Roger Payne, CETI's principal advisor, and @davidfgruber, CETI's Founder.
This revolutionary paper shows that sperm whale communication, known as codas, exhibit contextual and combinatorial structure, revealing sophisticated structures akin to human phonetics and communication systems in other animal species.
Fascinating new article by the amazing @carlzimmer discussing new findings from @pratyusha_PS and @ProjectCETI on coda ‘alphabets’ in Sperm Whale songs https://www.nytimes.com/...
They show that coda types are not arbitrary but that they form a combinatorial coding system, which gives rise to a large inventory of distinguishable codas which lay the framework for creating a sperm whale phonetic alphabet, and ultimately translating sperm whale communication
And for something completely different: our paper on the combinatorial structure of sperm whale vocalizations (led by @pratyusha_PS, in collab w @ProjectCETI) is out in Nature Comms today! https://www.nature.com/... https://www.nytimes.com/...
Exciting: Scientists including @pratyusha_PS @sgero say sperm whales “use a much richer set of sounds than previously known, which they called a ‘sperm whale phonetic alphabet.’” #scicomm by @CarlZimmer https://www.nytimes.com/... #whales #cetaceans #animalcommunication #animals
Whale, hello there, nice to sea you! Scientists studying the sperm whales that live around the Caribbean island of Dominica have described for the first time the basic elements of how they might be talking to each other. https://apnews.com/...
If sperm whales could talk, what would they say? New research on their communication reveals a complex combinatorial system that challenges our understanding of animal vocalizations. By analyzing 8,719 whale click sounds called “codas”, using machine learning, MIT CSAIL &... [vid…
CETI's science team has used machine learning to uncover the “sperm whale phonetic alphabet”!🐳 Read the paper at the link in our bio! #machinelearning #CETI #Dominica #spermwhales #interspeciescommunication #MIT #CSAIL #Carleton #DSWP [image]