Scientists have accomplished a whale of a feat. They’ve identified previously unknown complexity in whale communication by analyzing thousands of recorded sequences of sperm whale clicks with artificial intelligence.
Variations in tempo, rhythm and length of the whales’ click sequences, called codas, weave a rich acoustic tapestry. These variables hint that whales can combine click patterns in multiple ways, mixing and matching phrases to convey a broad range of information to one another.
What sperm whales are saying with their clicks remains a mystery to human ears. Still, uncovering the scope of whales’ vocal exchanges is an important step toward linking whale calls to specific messages or social behaviors, the scientists reported May 7 in the journal Nature Communications.
“This work builds on a lot of prior work focused on understanding the calls of sperm whales. However, this is the first work that has started to look at sperm whale calls in their wider communicative context and in the context of exchanges between whales, which has made some of the findings possible,” said study coauthor Dr. Daniela Rus, director of the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT, in an email.
“Understanding what aspects of their codas they can control and vary helps us understand how they can encode information in their calls,” Rus said.
Related article Great apes tease each other just like humans do, says new study
The researchers dubbed their catalog of sound combinations a “phonetic alphabet” for sperm whales, comparing variations in the whales’ click sequences to the production of different phonetic sounds in human speech.
But while the team’s findings are interesting, that term offers a misleading perspective on whales’ vocal interactions, said Dr. Luke Rendell, a researcher at the University of St. Andrews in the United Kingdom whose work focuses on communication in marine mammals, in an email.
“The presentation of the ‘phonetic alphabet’ — it’s nothing of the sort,” said Rendell, who was not involved in the research.
“The way the tempo variation is used is completely different to how, say, we use elements of an alphabet to construct linguistic expression,” he said. “There’s no evidence of that, and it’s not a super helpful interpretation because it forces everything into a restricted and somewhat over-sold perspective of ‘is it like human language or not,’ when there are a much broader range of interpretations available.”
Pattern recognition
Sperm whales produce their clicks by forcing air through an organ in their heads called the spermaceti, and these sounds can be as loud as 230 decibels — louder than a rocket launch and capable of rupturing human eardrums — another team of scientists previously reported in the journal Scientific Reports.
For the new study, the researchers used machine learning to detect patterns in audio data collected by The Dominica Sperm Whale Project, a repository for observations of sperm whales that inhabit the Caribbean Sea. The recordings represented the voices of approximately 60 sperm whales — a subset of a group of about 400 whales known as the Eastern Caribbean clan — and the vocalizations were recorded between 2005 and 2018.
Prior research had identified 150 types of codas in sperm whales worldwide, but the Caribbean whales used just 21 of those codas.
The scientists examined the timing and frequency of 8,719 coda sequences — in solitary whale utterances, in choruses and in call-and-response exchanges between whales. When visualized with artificial intelligence, previously unseen coda patterns emerged.
Related article These snakes not only fake their own deaths, they use gory special effects to do it
The study authors defined four features in codas: rhythm, tempo, rubato and ornamentation. Rhythm describes the sequence of intervals between clicks. Tempo is the duration of the entire coda. Rubato refers to variations in duration across adjacent codas of the same rhythm and tempo. And ornamentation is an “extra click” added at the end of a coda in a group of shorter codas, Rus explained.
These so-called ornament clicks “occur more towards the beginning and end of turns” during vocal exchanges between whales, “behaving like discourse markers,” Rus said.
The discovery that whales could synchronize variations in coda tempo was “a really interesting observation,” Rendell said.
“I am less convinced by the ‘ornamentation,’” he added. “It occurs very rarely, and I think we need more evidence that they aren’t just production glitches,” or filler sounds, “like when we say ‘um’ or ‘err.’”
In all, the program detected 18 types of rhythm, five types of tempo, three types of rubato and two types of ornamentation. These coda features could all be mixed and matched to form an “enormous repertoire” of phrases, the study authors reported. What’s more, meaning could be tweaked even further depending on the placement of a coda — following or overlapping other codas — within an exchange or chorus involving two or more whales.
Interactive experimentation
“Actually, many of us have been waiting for advanced technology to allow us to do something like this for decades!” said Dr. Brenda McCowan, a professor at the University of California Davis School of Veterinary Medicine, in an email.
McCowan, who was not involved in the research, was part of another team that, in 2021, conducted an interactive “conversation” with a humpback whale in waters near Alaska. For about 20 minutes, a curious whale repeatedly responded to a recording of a humpback song transmitted from the scientists’ boat.
“This particular playback (with the humpback in 2021) was an opportunistic experiment with an inquisitive whale engaging us both behaviorally and vocally, and completely at her own volition,” McCowan said.
Related article A lone orca killed a great white in less than two minutes. Scientists say it could signal an ecological shift
Such interactive experimentation with whales, along with observations of whale behavior, could be an important part of unraveling the syntax of sperm whale click sequences, the authors wrote in the study.
Their machine learning method may also prove useful for studying other types of animal vocalizations, McCowan added.
“Tempo, rhythm, rubato and ornamentation are likely to be found in other species of whales,” McCowan said. “We already know this is true of humpback song. But there is also evidence for this type of patterning in other aquatic, terrestrial and arboreal species to which this approach could be applied.”
But although this technique is helpful for identifying certain aspects of communication, it’s no Rosetta stone, Rendell cautioned.
“Machine learning is great for finding patterns in large datasets,” he said, “but it doesn’t create meaning.”
Mindy Weisberger is a science writer and media producer whose work has appeared in Live Science, Scientific American and How It Works magazine.
