Whale songs follow the laws of human languages

“Humpback whale songs are the most complex acoustic display in the animal kingdom,” says Ellen Garland, a marine biologist at the University of St Andrews in the United Kingdom. Only males sing, and their vocalizations can be heard thousands of miles away. The males within each population sing the same song, but periodically learn new tunes from other populations — sometimes up to 8,700 miles away — replacing their own. “This achievement is found in no other animal, except humans,” says Garland.

Together with other biologists and linguists, Garland published a study in Science on Thursday, showing that humpback whales follow basic laws found in human languages. But they’re not alone. Another recent study published in Science Advances confirms that most cetaceans adhere to the principles of efficiency and economy in language — concepts once thought to be unique to humans.

Garland explains why humpback whales fascinate both marine biologists and linguists: “These rapid changes, where songs spread from one population to the next, can happen across an entire ocean basin. It’s cultural change at a large, rapid pace.” She compares this dynamic to human fashion or pop songs: “Culture is a really important part of humpback whale life, and song is part of it.”

The Science paper she co-authored is the result of eight years of recording males from a population of humpbacks living off the coast of New Caledonia in the Pacific Ocean. The researchers aimed to uncover the underlying structure of their songs and explore whether they share any similarities with human language.

“We used exactly the same cues that babies use to segment whale song,” Garland explains. Achieving this segmentation into hierarchically organized units (such as paragraphs, sentences, words, syllables, or human phonemes) is the key methodological contribution of this research. “We took the recordings from each year, removed all the human annotation, and were left with a long chain of sound elements [the basic acoustic blocks of whale song]. Then we calculated the transition probabilities between every two consecutive sound elements in the year and cut them off when they were low,” she adds.

These cuts allowed them to divide each song into segmented sequences. “Then we looked at their distribution and discovered, surprisingly, that they follow the same distribution found in all human languages,” the biologist notes.

Inbal Arnon, a linguist at the Hebrew University of Jerusalem in Israel and co-author of the study, explains the main similarity they found between humpback whale songs and human speech: a regularity known as the Zipfian distribution. “It is a particular frequency pattern that words follow in language, where the first word is approximately twice as frequent as the second most frequent word, three times as frequent as the third most frequent word, and so on,” she explains.

The name “Zipfian” comes from George Kingsley Zipf, a linguist who, in the 1940s, discovered and proposed a series of laws present in all human languages (and in other areas beyond language). His most famous contribution is the so-called Zipf’s law, or the generalized law of abbreviation. In its simplest form, it states that the longer a word is, the less frequently it will be used, meaning there is an inverse relationship between word length and frequency of use. This is why terms like “the” and “it,” are more common than words like “thermodynamics” or “intercontinental.” Or why people prefer to say “L.A.” rather than “El Pueblo de Nuestra Señora la Reina de los Ángeles del Río de Porciúncula,” the name the Spanish gave to Los Angeles.

“This arrangement has been shown to aid learning in humans, suggesting that it may be found in language precisely because it aids learning and transmission across generations,” says the Israeli linguist. “If that’s the case, we would expect to find a similar structure in other culturally transmitted communication systems, such as those of humpback whales.” And that’s exactly what Arnon, Garland, and their colleagues have found: humpback whales follow Zipf’s law. And they are not alone.

Between linguistics, cultural evolution

Mason Youngblood, a young researcher at the State University of New York at Stony Brook, is working at the forefront of an emerging field. Drawing on computer science, advanced computing, AI, and algorithms, he explores the intersection of linguistics, cultural evolution, and the study of both human and animal behavior.

In Science Advances, he published a study that, after reviewing, compiling, and analyzing dozens of previous studies on cetacean communication systems, brought together 610,000 communicative elements (including their pauses or silences) from 65,511 sequences recorded from specimens of 16 cetacean species, including both baleen whales (humpback, blue, fin whales, etc.) and toothed whales (dolphins, orcas, sperm whales, etc.). He then investigated whether these communications followed two key laws of human language: Zipf’s law and another equally important one, Menzerath’s law. This law can be summarized as follows: the longer a sequence (such as a sentence), the shorter the elements (such as words) that compose it.

His extensive work of compilation and analysis revealed that, of the five species for which data was available, two — humpback whales and blue whales — adhere to Zipf’s law. The findings are even more conclusive when it comes to Menzerath’s law. Of the 16 most studied species, 11 shorten the basic elements of their communication as the length of their messages increases. Not only do these species follow the laws, but they also prove to be more efficient than humans. Among them, humpback whales again stand out.

“This means that some species of whales compress their vocalizations more than humans do. In other words, they are more time-efficient when communicating with each other,” Youngblood explains. The researcher believes there are two reasons for this: “First, whales vocalize underwater while holding their breath and have specialized adaptations for doing so. This may increase the importance of efficiency — communicating as quickly as possible. Second, human languages contain much more information than whale communication, and it is much harder to compress something that contains more information.”

Researcher Iván G. Torre, who was not involved in either of the two studies, highlights the importance of efficiency in any communication system. “Danger calls, for example, need to be understood by the entire group quickly, so they don’t have to warn each individual one by one, and they must be as brief as possible — otherwise, the predator will eat you,” says Torre, who has studied the presence of Zipf and Menzerath’s laws — considered universal — in human languages. “These are laws that reflect how selection favors efficiency,” he adds.

Torre, now working in the field of artificial intelligence applied to linguistics at Oracle, also points out that the researchers have achieved no easy feat. “For humans, segmentation is straightforward; you have words or syllables. But the work with humpback whales opens up the possibility of studying other animal communication systems,” he explains.

Among the cetaceans that follow Menzerath’s law are bottlenose dolphins. Bruno Díaz, director of research at the Bottlenose Dolphin Research Institute (BDRI), agrees that natural selection has optimized vocalizations for efficiency. “Although this similarity was already known in species such as the bottlenose dolphin, the fact that cetaceans from different groups exhibit these patterns, with even more pronounced effects than human speech, is surprising. It suggests that their communication system has evolved to maximize efficiency in transmitting information within their acoustic environment,” says Díaz, who has been studying communication between dolphins for nearly two decades.

One of the leading experts in human linguistics and animal communication is Ramón Ferrer i Sancho, director of the Laboratory of Quantitative, Mathematical, and Computational Linguistics at the Universitat Politècnica de Catalunya in Spain. He has been working in this field for decades. “These studies indicate that the communication and vocalization systems of species in general are shaped by the need to reduce costs. However, it is another matter if they do it more or better than us,” he says.

Among the various limitations he sees in the two works, one stands out as particularly difficult to overcome: “We do not know the identity of each vocalizer. Not knowing who is vocalizing could be hiding an underlying pattern,” he explains. Additionally, the context and meaning of the vocalizations remain unknown — elements that play a significant role in the adherence to linguistic laws. Despite these issues, Ferrer i Sancho argues: “If we have discovered something like this with so many difficulties, imagine what we could find if we knew more and could control more variables.”

The Catalan scientist concludes with an idea that may seem obvious to him but deserves emphasis. Neither the authors nor the consulted researchers are identifying whale songs with human language, at least for now, because we still do not know whether they carry specific meaning. However, Ferrer i Sancho notes: “What we have not been able to say for some time is that human language is unique.”

He concludes: “As for communication, a large number of species produce sequences, and within these sequences, a certain level of complexity can naturally emerge, and we do not yet fully understand the level of this complexity. We are more similar than we think.”

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition