Mar 22, 2026 In 2024, a team of researchers from MIT and Project CETI did something no one had done before — they fed over 9,000 sperm whale recordings into an artificial intelligence. What the AI discovered changed everything we thought we knew about animal communication.
Sperm whales aren't just making noise. They have an alphabet. They have vowels. They have dialects that differ from ocean to ocean. They have names. And their communication system might be one of the most sophisticated on the entire planet — besides our own.
This is the story of how scientists finally cracked the code, and what they found is forcing researchers to ask a question that once seemed impossible: Are sperm whales actually speaking a language?
CHAPTERS 0:00 - The Discovery 1:23 - Why Scientists Chose Sperm Whales 4:00 - The Problem With Whale Language 5:51 - The Breakthrough: The Phonetic Alphabet 8:21 - They Have Vowels 10:22 - Clans, Dialects, and Names 12:48 - What Comes Next 14:49 - Conclusion
SOURCES & REFERENCES
This video was made possible by the groundbreaking research from Project CETI (Cetacean Translation Initiative) and the scientists dedicating their lives to understanding sperm whale communication.
Primary Research Papers: Sharma, P., Gero, S., Payne, R. et al. (2024). "Contextual and combinatorial structure in sperm whale vocalisations." Nature Communications. https://doi.org/10.1038/s41467-024-47...
Beguš, G. et al. (2025). "Vowel- and Diphthong-Like Spectral Patterns in Sperm Whale Codas." Open Mind (MIT Press).
Hersh, T. et al. (2022). "Evidence from sperm whale clans of symbolic marking in non-human cultures." Proceedings of the National Academy of Sciences (PNAS). https://doi.org/10.1073/pnas.2201692119
Gero, S., Whitehead, H., Rendell, L. (2016). "Individual, unit and vocal clan level identity cues in sperm whale codas." Royal Society Open Science. https://doi.org/10.1098/rsos.150372
Institutions & Projects: Project CETI (Cetacean Translation Initiative) — https://www.projectceti.org/ MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) Dominica Sperm Whale Project — https://www.thespermwhaleproject.org/ University of California, Berkeley — Department of Linguistics Max Planck Institute for Psycholinguistics Dalhousie University — Department of Biology
Key Researchers Featured: Dr. Shane Gero — Biology Lead, Project CETI; Founder, Dominica Sperm Whale Project Dr. Daniela Rus — Director, MIT CSAIL Dr. David Gruber — Founder & President, Project CETI Dr. Gašper Beguš — Linguistics Lead, Project CETI; UC Berkeley Pratyusha Sharma — PhD Candidate, MIT CSAIL Dr. Taylor Hersh — Max Planck Institute for Psycholinguistics Dr. Hal Whitehead — Dalhousie University Dr. Roger Payne (1935-2023) — Whale biologist, conservationist, MacArthur Fellow
News Coverage & Articles: MIT News: "Exploring the mysterious alphabet of sperm whales" (May 2024) National Geographic: "Sperm whale speech has human-like vowels" (November 2025) Smithsonian Magazine: "Scientists Discover a 'Phonetic Alphabet' Used by Sperm Whales" (May 2024) NPR: "Why sperm whale communication is much more complex than previously thought" (May 2024) Popular Science: "Sperm whales use vowels like humans, new study finds" (November 2025)
⸻
COPYRIGHT & FAIR USE NOTICE
This video was created for educational and documentary purposes under Fair Use (17 U.S.C. § 107). All footage, images, and audio clips from external sources are used for commentary, criticism, and educational illustration only.
We have made every effort to credit all sources appropriately. If you are a rights holder and believe your content has been used in a way that constitutes copyright infringement, or if you would like to be credited differently, please contact us directly at: 📧 [email protected]
We are committed to resolving any concerns promptly and respectfully.
ABOUT THIS CHANNEL; The Bio Files explores the most fascinating mysteries of the natural world — from the creatures of the deep ocean to the science that's changing how we understand life on Earth. Subscribe for weekly deep dives into biology, evolution, and the unknown.
Somewhere in the Caribbean Sea, a mother sperm whale is floating at the surface with her calf.
She's not hunting. She's not sleeping. She's talking. A rapid burst of clicks fires from
deep inside her skull. Click, click, click, click, click. And 400 meters away,
another whale answers. Different pattern. Different rhythm. Almost like a response.
For decades, scientists recorded these sounds. Thousands of hours. Millions of clicks.
They knew whales were communicating. They just couldn't figure out what any of it meant.
It sounded like noise. Random. Primitive. Then, in 2024, a team of researchers did something
no one had done before. They fed 9,000 whale recordings into an artificial intelligence.
And what the AI found changed everything we thought we knew about animal communication.
Sperm whales aren't just making noise. They have an alphabet. They have vowels.
They have dialects that differ from ocean to ocean. They have names. And their communication
system might be one of the most sophisticated on the entire planet, besides our own.
This is the story of how scientists finally cracked the code. And what they discovered
is making some researchers ask a question that sounds insane. Are sperm whales actually speaking
a language? Before we get into what science has found, you need to understand why they chose sperm
whales in the first place. There are 90 species of whales and dolphins. Humpbacks sing those long,
beautiful songs. Orcas hunt in packs with military precision. Blue whales are the largest animals
that ever lived. So why sperm whales? Three reasons. First, the brain. Sperm whales have
the largest brain of any animal that has ever existed on Earth. Not just the largest today,
the largest ever. Their brain weighs about 8 kilograms. That's roughly six times heavier
than a human brain. Now, bigger doesn't automatically mean smarter. Elephants have
big brains, too. But sperm whale brains aren't just big. They're structured differently. They
have massive regions devoted to processing sound. They have spindle cells, the same type of neurons
linked to emotion and social bonding in humans. These are not simple animals. Second, their social
lives. Sperm whales live in tight-knit family groups led by older females. Grandmothers, mothers,
aunts. These families stay together for decades. They babysit each other's calves. They nurse each
other's young. They grieve when family members die. And here's the strange part. Different families
that live in the same waters refuse to interact with each other. Not because they're enemies,
they just don't associate. It's like invisible cultural boundaries running through the ocean.
Scientists had a theory about why. They thought it had something to do with how the whales
communicate. Different groups seem to click differently, like accents or dialects. But they
couldn't prove it. Not yet. Third reason, the clicks themselves. Unlike humpback whale songs,
which are long and flowing, sperm whale communication is built entirely on clicks,
short bursts of sound, rapid patterns called codas. Click, click, click, pause, click, click. It's almost
like Morse code. Structured, rhythmic, repetitive. And to researchers, that structure hinted at
something bigger. If you wanted to search for meaning in animal communication, clicks seem like
a better place to start than songs. So in 2020, a team of scientists launched one of the most
ambitious projects in the history of animal communication. They called it Project SETI,
Cetacean Translation Initiative. Their goal? Use artificial intelligence to decode what sperm
whales are actually saying. Now, trying to understand what whales are saying sounds simple
enough, right? Just record them, figure out which sounds go with which behaviors, build a dictionary.
Except it's not simple at all. Here's the problem. When scientists first started studying sperm whale
clicks in the 1970s and 80s, they noticed whales made different patterns. Some clicked three times.
Some clicked five times. Some had long pauses between clicks. Some had short pauses. They grouped
these patterns into categories and called them coda types. By the early 2000s, researchers had
identified about 21 different coda types in the Caribbean. 21. That's it. If sperm whales only had
21 different sounds, that's not really a language. That's more like alarm calls, danger, food, hello,
goodbye. Simple signals, not complex communication. And for years, that's what most scientists believed.
Sperm whale codas were probably just basic social signals. Important for whale society, sure, but not
language. There was another problem, too. Even when scientists recorded the same coda type from different
whales, the clicks weren't exactly identical. There were tiny variations, small differences in timing,
subtle changes in rhythm. Most researchers assumed this was just noise, biological sloppiness,
the whale equivalent of mumbling. But a few scientists weren't so sure. What if those tiny
variations weren't noise? What if they were information? The only way to find out was to
analyze whale recordings at a scale no human could do manually. You'd need a compare thousands of
codas, measure every click down to the millisecond, look for patterns humans might list. You'd need
artificial intelligence. In May 2024, Project SETI published a paper in the journal Nature
Communications. The findings sent shockwaves through the scientific community. Using machine
learning, the researchers analyzed 8,719 sperm whale codas recorded off the coast of Dominica
in the Caribbean. These recordings came from about 60 individual whales, tracked over more than a
decade. And what the AI discovered blew apart everything scientists thought they knew. Remember
those 21 coda types? The AI found 156, not 21, 156 distinct codas. But that wasn't even the big
discovery. The AI revealed that those tiny variations scientists had been dismissing as noise,
they weren't noise at all. They were meaningful. The researchers identified four different features
that whales systematically control in their clicks. First, rhythm. The basic pattern of clicks. Three
clicks, five clicks, seven clicks. Second, tempo. How fast or slow the overall coda is delivered.
Third, rubato, a musical term. It means the whale gradually speeds up or slows down during a single
coda, like a singer stretching a note. Fourth, ornamentation. Extra clicks added at the beginning
or end, like punctuation. Here's what's incredible. Sperm whales combine these four features in
different ways to produce different codas. It's combinatorial, like how humans combine a limited
set of sounds to create thousands of words. The researchers called it a sperm whale phonetic
alphabet. Let that sink in for a second. Humans have phonemes, small units of sound that combine
to form words. The sounds B and A and T combine to make fat. Sperm whales appear to have something
similar, basic building blocks that combine to form more complex messages. One of the lead
researchers, Daniela Russ from MIT, put it this way, our findings challenge the prevailing belief
among many linguists that complex communication is unique to humans. But the team wasn't done.
In November 2025, they published another paper, and this one was even stranger. A linguist named
Gaspar Begas was analyzing sperm whale recordings when he noticed something odd. The whale clicks
sounded simple at normal speed, just clicks, binary, on or off. But Begas had an idea.
What if whale communication operates on a different time scale than human speech?
Humans speak fast. What if whales speak slow? So he sped up the recordings, and when he did,
he heard something no one had heard before. The clicks weren't just clicks, they had texture,
they had tone, they had qualities that sounded like vowels. Begas and his team analyzed the
recordings more closely. They found two distinct patterns appearing across different whales and
different coda types. They called them the A coda and the I coda, like the vowels A and I in human
language. The whales weren't just varying the timing of their clicks, they were varying the
frequency, the spectral quality, the color of the sound. And just like in human language,
these vowels combined with other features. The team even found diphthongs, sounds that glide
from one vowel to another, like the OI in coin. The researcher described what he found as clean,
razor sharp, precise. I've never seen anything so structured and seemingly intentional, he said.
Now, to be clear, scientists are not saying sperm whales have a language in the human sense.
They're not having conversations about philosophy or telling stories. But they are saying sperm
whale communication has structural features that were previously thought to be unique to
human language. That's a massive deal. Because if you find vowels, phonemes, and combinatorial
structure in an animal communication system, you're looking at something far more complex
than simple alarm calls. You might be looking at the building blocks of language itself.
Here's where things get even more fascinating. Remember how sperm whale families in the same
waters refused to interact with each other? Scientists finally figured out why. It's because
they speak differently. In 2022, an international team pooled whale recordings from 23 locations
across the Pacific Ocean – Canada to New Zealand, Japan to South America. They analyzed
over 23,000 quotas. What they found was extraordinary. There are seven distinct
sperm whale clans in the Pacific Ocean. Each clan has its own dialect, a unique set of quotas they
use to communicate. Whales from different clans can live in the same waters, swim past each other,
encounter each other regularly, but they never interact, they never mix. It's like invisible
cultural barriers running through the ocean. And here's the most fascinating part. When whale
clans overlap geographically, when they share the same territory, their dialects become more distinct,
not less, more. The whales exaggerate their differences. They emphasize what makes their
In humans, we call these ethnic markers, distinctive accents, clothing styles,
or customs that signal group identity. I am one of us, not one of them. Sperm whales appear to
be doing the exact same thing. But the cultural complexity doesn't stop at the clan level.
Within each clan, there are family units. And each family has its own signature quota,
a specific pattern that identifies the group. It's like a family name. And it goes even deeper.
Each individual whale has a personal quota, a click pattern with subtle timing variations that
are unique to that whale and that whale alone. When a sperm whale announces itself, it's essentially
saying, I am this specific whale from this family from this clan. This is me. And just like humans,
sperm whales don't come out of the womb knowing their quotas. They learn them. Baby sperm whales
babble. For the first two to three years of life, they produce messy, incoherent clicks,
random patterns, wrong timing. Slowly, over years, they narrow down their repertoire.
They learn the quotas of their family, their clan, their identity. It's cultural transmission,
passed from mother to calf, generation after generation, just like language.
So what does all this mean? Are sperm whales speaking a language? Are they having conversations?
Do they discuss ideas, share memories, tell stories? Honestly, we don't know.
Scientists are clear about what they found and what they haven't. They found structure,
phonetic building blocks, vowels, dialects, individual identity signals, cultural transmission.
What they haven't found is meaning. They know the whales are saying something.
They just don't know what. The next step is to connect sounds to behaviors. Record a whale
clicking, then watch what happens. Does the calf swim closer? Does the group dive? Does another
whale respond? If you can match specific quotas to specific actions, you might start to build
something like a dictionary. But Project SETI researchers admit this could take decades.
The whales spend most of their lives in darkness, thousands of meters underwater,
doing things we can't see. We're trying to understand an alien intelligence that evolved
in a world completely unlike our own. Some scientists think we'll eventually be able
to communicate with them, play back quotas, see if they respond. Others think we'll never truly
understand what they're saying. Their world is too different. Their minds are too different.
But here's what we do know. For nearly 200 years, humans hunted sperm whales to the edge of
extinction. We killed over a million of them. We boiled their bodies for oil. We carved their teeth
into trinkets. We treated them like resources. And now, for the first time, we're starting to
An animal with a brain six times larger than ours. An animal that lives in complex,
multi-generational societies. An animal with cultural identities, family names, and regional
dialects. An animal that might be speaking a language we're only now learning to hear.
In the Caribbean Sea, somewhere off the coast of Dominica, a mother sperm whale clicks to her calf.
The calf babbles back, wrong rhythm, wrong timing. The mother clicks again, patient, teaching.
Somewhere in a lab thousands of miles away, an AI is listening to a recording of this exact exchange.
Click by click, pattern by pattern, we're getting closer. We may never fully understand what sperm
whales are saying, but for the first time in human history, we're finally starting to listen.
If you watched up to this point, you are one of us, and subscribe to The BioFiles
for more educational and interesting files. We love you, and see you in the next file.