Groundbreaking effort launched to decode whale language - National Geographic

Marina: Welcome to Deep Voices — I'm so glad you're here. I'm Marina, a marine biologist who has spent the better part of two decades completely, hopelessly in love with whales and dolphins, and this podcast is my excuse to talk about them every single week. Marina: Before we dive in, a warm thank you to the sponsors who make this show possible. Ada, who helps businesses integrate A I into their workflows in smart, practical ways. Ago Consulting — pronounced "Ahh-Go" — silicon development experts working across everything from A I to SoC design. And Zen Semiconductor, an umbrella company with a remarkable portfolio spanning the silicon industry. We are so grateful for all three of them. Marina: Today's episode is a big one. We are talking about sperm whales — their clicks, their complexity, and a genuinely landmark moment in the science of animal communication. We're going to meet the researchers who think we may be on the edge of understanding what these animals are actually saying to each other. And we'll end with a story about humans and dolphins that I think you'll carry with you for a while. Marina: Let's go deep. Marina: So — voices of the deep. Let's start with sound, because with sperm whales, sound is everything. These animals live in a world of almost total darkness. Below a few hundred meters, sunlight just doesn't reach. And yet sperm whales navigate, hunt, find each other, and communicate across vast stretches of ocean. They do all of it with sound. Marina: The sounds they make are called codas — short, rhythmic patterns of clicks. And for a long time, scientists knew these codas existed, knew they varied between groups, knew they seemed to carry some kind of social meaning. But the deeper structure? That was still a mystery. Marina: Then came some findings that genuinely made me catch my breath. Researchers studying sperm whale clicks have found what appear to be vowel-like patterns embedded in the structure of those codas. Let me say that again, because it's remarkable. Vowel-like patterns. That's the kind of feature we associate with human language — the idea that the internal shape of a sound unit carries meaning, not just its rhythm or timing. Marina: And it doesn't stop there. Scientists have also documented something called ornamentation — subtle variations layered onto the core click pattern, a bit like the way a singer might add a flourish to a note. And rubato — which is a musical term for the expressive stretching or compressing of timing. Jazz musicians use rubato. And apparently, sperm whales do too. Marina: What this tells us is that sperm whale communication has combinatorial properties. Meaning: they're not just using a fixed set of signals with fixed meanings. They may be combining elements — rhythm, ornamentation, timing, internal click structure — in ways that generate a much larger expressive range. That is a hallmark of complex communication systems. It is the kind of architecture that language is built on. Marina: And then there's the sheer physical scale of it. Male sperm whales produce the loudest communication sounds ever recorded in any mammal. We're talking about clicks that can reach around two hundred and thirty decibels underwater. To put that in context — that is louder than a jet engine at close range. These animals are not whispering to each other. They are broadcasting across miles of open ocean, and the messages they're sending appear to be far more nuanced than we ever imagined. Marina: Now let's talk about the science — the extraordinary, ambitious, A I-powered science — that is trying to actually decode what sperm whales are saying. Marina: Project CETI — that stands for Cetacean Translation Initiative — is probably the most ambitious animal communication project in history. It launched with a simple but staggering goal: to listen to sperm whales long enough, carefully enough, and with powerful enough tools, to understand the structure of their communication. And potentially, eventually, to figure out what it means. Marina: Here's how they're doing it. The project deploys underwater hydrophones — sensitive microphones — near sperm whale families off the island of Dominica in the eastern Caribbean. These whales have been studied for years by researchers like Shane Gero, who leads the biological side of the project. He knows individual animals. He knows their family histories. And now, the team is building a dataset of whale vocalizations that is orders of magnitude larger than anything that existed before. Marina: Why does scale matter? Because of the A I. The machine learning models that Project CETI is using — some of them inspired by the same large language models behind tools like the ones you might use for text — need enormous amounts of data to find patterns. The idea is that if you give the A I enough examples of sperm whale codas, recorded in known social contexts, the system might start to identify structure. Grammatical-like structure. Contextual variation. The kinds of patterns that, in human language, carry meaning. Marina: And the early results are genuinely exciting. The discovery of those vowel-like features I mentioned — that came out of this kind of deep computational analysis. Researchers found that sperm whale clicks aren't just varying in number and rhythm. They're varying in internal acoustic qualities in systematic ways. Ways that look, mathematically, like the way phonemes — the building blocks of spoken language — vary in human speech. Marina: Now, I want to be careful here, because the scientists themselves are careful. This does not mean we've translated whale language. It doesn't mean we know what any particular coda means. What it means is that the complexity we're finding is real, it's systematic, and it is consistent with the idea that these clicks are doing something far more sophisticated than simple signaling. The structure is there. The meaning is still the mystery we're chasing. Marina: National Geographic called the broader effort to decode whale language groundbreaking — and I think that word is exactly right. Not because we've cracked the code, but because for the first time, we have the tools, the data, and the interdisciplinary teams — biologists, linguists, cryptographers, A I researchers — all working together on the same question. What are the whales saying? Marina: I find that coalition beautiful. Because it says something about us, too. That we've reached a moment where we're willing to ask the question seriously. Not as fantasy, not as science fiction. As science. Marina: Okay — our final thread. A cetacean-and-people story. And for this one, I want to take us somewhere a little unexpected. I want to talk about John Lilly. Marina: John Lilly was a neuroscientist and physician who, in the 1950s and 60s, became obsessed with dolphin intelligence. He was the first researcher to seriously and scientifically propose that dolphins might have a language — that their vocalizations might be genuinely complex, meaningful communication — and he dedicated years of his life, and considerable resources, to trying to bridge the gap between human and dolphin minds. Marina: His work was real science, at least at the start. He was among the first to map the brain of a bottlenose dolphin and note its extraordinary size and complexity. He set up the Communication Research Institute in the Virgin Islands, where researchers lived alongside dolphins in partially flooded houses — sharing space with them, interacting with them constantly — in an effort to accelerate communication attempts. Marina: One of his most famous collaborators was a young woman named Margaret Howe Lovatt, who spent months living with a dolphin named Peter in this flooded house. She was trying to teach Peter to mimic human speech sounds. And Peter did make attempts — effortful, recognizable attempts — to shape dolphin vocalizations into something resembling human words. It was imperfect, it was strange, and it was also genuinely remarkable. Marina: Now — here's where the story gets complicated, and I want to be honest about that, because this is where it starts to overlap with a recent documentary. Lilly's later work became increasingly unscientific. He became involved with psychedelics, including LSD, which he administered to dolphins in experiments that would be considered deeply unethical today — and which were controversial even then. He drifted further from mainstream science, and his reputation suffered for it. Marina: And yet. The core of what he was reaching for — the intuition that dolphins are minded, that they communicate in ways we haven't yet understood, that there is something genuinely home inside those large, complex brains — that intuition was not wrong. The science has caught up to the wonder, even if it had to leave some of Lilly's methods far behind. Marina: What I find moving about this story is the longing in it. Here was a scientist so captivated by the idea of talking to another species that he built a house you could swim through just to spend more time with them. And the dolphins — the dolphins engaged. Peter wasn't performing for food. He was curious. He was social. He was trying, in his way, to meet Lilly and Margaret Howe Lovatt halfway. Marina: We now know so much more about dolphin cognition than Lilly did. We know they pass the mirror self-recognition test, suggesting self-awareness. We know they use tools — some populations in Western Australia use marine sponges to protect their rostrums while foraging on the seafloor, and they teach this to their calves. We know they have signature whistles — unique calls that function like names — and that they remember the whistles of individuals they haven't seen in twenty years. Marina: The conversation Lilly was trying to start? It's still going. It's just moved into the water, into the data, into the A I models that are listening now with more patience and more precision than any human researcher could manage alone. Marina: Maybe that's the most beautiful thing of all. That the question — what are you saying? — has never gone away. We just keep finding better ways to ask it. Marina: And that brings us to the end of today's Deep Voices. I want to sit with that for just a second — the fact that right now, as we speak, there are hydrophones in the warm water off Dominica, recording the conversations of sperm whale families. Mothers and daughters. Grandmothers. Calves learning the codas of their clan. And somewhere, an A I is listening to all of it, looking for the shape of meaning. Marina: We are living in an extraordinary moment for this science. And I believe, with everything I have, that what we find is going to change the way we understand minds, language, and our place in the community of life on this planet. Marina: Deep Voices, brought to you by Ada, Ago Consulting, and Zen Semiconductor — thank you, genuinely, for supporting this work and making these conversations possible. Marina: If today's episode moved you, please subscribe, leave a review, and share it with someone who loves the ocean, or who just needs a reminder that the world is more wondrous than the news would have us believe. Links to all the studies and stories we talked about today are in the show notes — the link is right there waiting for you. Marina: Until next time — stay curious, stay humble, and keep listening. The ocean has a lot left to say.