Animals are talking — will the legal system listen?

The character of Dr. Dolittle, who "walks with the animals, talks with the animals," was the central figure in an early 20th century series of children's books written by Hugh Lofting, acting as a kind of personal antidote to his trauma of experiencing the worst of humanity in the trenches of World War I. "A fox has his rights, the same as you and I have," the eponymous physician-turned-animal doctor said in one book. But Dr. Doolittle is eccentric and, of course, fiction.

Increasingly, however, talking with the animals is becoming less of a fantasy. This doesn’t just invite some fantastic conversations, but recontexualizes the rights we allot to non-human creatures. The past decade has brought a slew of new findings that are changing our understanding of animal consciousness and communication. A new study on complex communication in chimps and another about bonobos' surprising language abilities, and recent findings about how sperm whale communication works, all suggest some animals have complex communication abilities that actually resemble human language in important ways. As AI and bioacoustics, as well as other clever new techniques, bring us ever closer to talking with animals or at least listening in on their conversations — should we prepare to update the way legal systems deal with animals?

"Compositionality, in a nutshell: It's the fact that we can combine meaningful units into larger structures whose meaning is derived from the meaning of the units," Mélissa Berthet, a researcher in primatologist and linguistics at the University of Zurich, who studied bonobos in Democratic Republic of Congo (they are only found in the wild in forests in DRC, south of the Congo river) over an 8-month period, explained to Salon in a video interview. Combining units of language in meaningful ways is one of the hallmarks of what human language does, and it’s evidence of this in a number of animals that suggests we might not be able to use signs of language as a filter to separate humans, who get certain rights, from animals, who don’t.

The University of Zurich team observed a group of bonobos from a project by local biologists that had habituated the animals over 10 to 15 years, depending on the group. This meant that the animals were used to having humans around them, but Berthet's group, like the local scientists, avoided interfering with their behavior and did their best to avoid any disturbance. With this work, the hope is that the animals use essentially the same "dictionary" as they would use if there were no humans present.

"I really think the meaning is not really impacted," Berthet said, though she allows that they may perhaps talk — gossip? — about humans more than they would otherwise do.

There are two kinds of compositionality. In the first, and most common, individual units (like words, for humans), are combined to produce meaning. This is called trivial compositionality, and it has been observed in animals in the past. In the second kind, thought to only exist in humans, when you put two units together, the meaning of one unit will modify the meaning of the other. For example, in the phrase "a blue dress," the word blue has actually modified the word dress, giving it a different meaning than it had before. This kind is called non-trivial compositionality, and it allows us to create complex utterances.

"So far, it was thought that in animals, when they could do compositionally, it was mostly restricted to trivial compositionally ... we wanted to see whether bonobos had compositionally, whether it was something that was common in their vocal system. So here we took an approach that is a bit different from former studies, because usually people would investigate one combination, and here we managed to investigated all the combinations of their vocal system." This was possible, Berthet said, because they succeeded in adapting a linguistic method used in humans to the study of animal communications.

"The idea was to see whether compositionally in other species can be as extensive as in humans. Because, like I said, humans use compositionally all the time. We create very long sentences. And so we did it with bonobos. And what was interesting is that we found that four of their combinations are compositional, and all the core types that they have, so all the units can be combined in a compositional structure. So it's basically like all their words can occur in sentences, just like in humans, we can put any words in a sentence, and they can really do the same. And so for us, this is really amazing, because it shows that even though they have a very limited vocabulary, they can really use them and combine them extensively," Berthet explained.

But it got better.

"Among those four combinations that were compositional," Berthet went on, "we found that three of them were non-trivial. And again, this is very big, because for us, it's the first time that we really managed to show that animals can have non-trivial compositionality. So I think what it really means, in general, is that, first, humans are not the only species that can do non-trivial compositionally. Second, other animals can also use compositionally a lot. It's not just about one combination in their own system, that could be a coincidence."

All this suggests that the last common ancestor of chimpanzees, bonobos and humans likely also produced compositional structures.

"Compositionality is very common in human language. We use it all the time, and for some people, it's really a hallmark of human language. And there are some species that do a bit of compositionality, but it's always very limited," said Berthet.  

The finding that among combination types, four are compositional, means they can convey information in a flexible way just like us. 

"So it's basically like all their words can occur in sentences, just like in humans, we can put any words in a sentence, and they can really do the same."

About those whales

We have indeed known for a while that whales are intelligent creatures, with their own eerie and complex form of sound-based communication. Since the 1970s, humans have been entranced by the songs of humpback whales, for example. But another kind of whale communication, among the very social if less musical sperm whales, has more recently become a focus of research into exactly how whale communication, or language, might work. 

Sperm whales use a series of clicks, called codas, to communicate. Until last year, though, we couldn't really say how this form of communication — distinctly different, even alien, to our own — actually works to convey information. We had already used machine intelligence to analyze the codas, recorded using bioacoustics, and determined that both the clicks and the intervals between clicks are important. The question has been, though, how can a system of clicks and spaces, with some 150 individual coda types, defined by their characteristic sequences of inter-click intervals found globally (with different clans or groups of whales using a far smaller number), account for the social and behavioral complexity seen in these animals? Surely there must be some other factor dividing up the use of different codas to allow for more complex ways to combine them?

In recent years, our technological ability to listen in on whales has grown in leaps and bounds. So too, as a result, has our understanding of their unique forms of communication, which turns out, in the sperm whale at least, to be structured, complex, and expressive.

Last year, a group of researchers applied machine learning to the largest existing data set of sperm whale codas, ​​from The Dominica Sperm Whale Project, to demonstrate that sperm whale codas can be sensitive to how they are used — that is, the context created by surrounding codas may change a given coda's intended meaning, just the way words in human sentences create a context that indicate, or change, the intended meaning of a given word. And they also showed that sperm whale codas are indeed combinatorial, similar to the compositionally shown by bonobos and chimps.

The researchers, members of Project CETI, used 8719 codas recorded over 13 years ending in 2018 from the Eastern Caribbean 1 (EC-1) clan, a group of 400 sperm whales. Clans are known to have their own culture, developing unique behaviors and repertoires of coda types. Over those years, Dominica Sperm Whale Project researchers manually recorded codas, while over the last four, Project CETI scientists studied EC-1’s coda repertoire using aerial drones, small computers attached by suction cups (called D-tags or acoustic biologging tags), synced underwater microphones in their hundreds, and swimming robots. The data collected in this way was then analyzed using machine learning, allowing for the sort of complex analysis needed to find patterns indiscernible to the human ear, and in no way similar to what we might expect from our knowledge of human or even primate communication.

"When analysing codas exchanged between whales, we observe fine-grained modulation of inter-click intervals relative to preceding codas, as well as modification of standard coda types via the addition of an extra click. We term these contextual features rubato and ornamentation. Next, we show that the coda repertoire has combinatorial structure: in addition to rubato and ornamentation, codas’ rhythms and tempos can independently be discretised into a small number of categories or types," the researchers wrote in Nature Communications this time last year. Essentially, these different features allow for coda repertoire to act as a kind of phonetic alphabet.

Project CETI doesn’t actually aim to talk with whales, but to listen in and translate what they’re saying — ideally with permission, if they reach a point where they can understand if permission is given or withheld. The project itself is a non-profit organization funded by partners that include Amazon Web Services, Google Research, and the MOTH (More-Than-Human-Life) Program at NYU Law, which advises on the ethics and permission part, as well as universities in Canada, the U.S., the U.K. and, notably, given some of the technology involved, in Israel. 

"We show that all four features are sensed and acted upon by participants in the vocal exchanges, and thus constitute deliberate components of the whale communication system rather than unconscious variation. Rhythm, tempo, rubato and ornamentation can be freely combined, together enabling whales to systematically synthesize an enormous repertoire of distinguishable codas," they explain. The whales could put together long, finely detailed sequences of all of these features.

Chimp talk

Catherine Crockford, of the Max Planck Institute for Evolutionary Anthropology, an author of the new study, sees this combinatorial system as perhaps representing a transitional or intermediate phase between animal calls and language as seen in humans. Whether the chimps’ bigrams should be considered like human units of communication, similar to words, or simple phrases that can be combined in longer sentences, or even like the ideograms used in some human writing is, Crockford told Salon in an email interview, the million-dollar question.

“They do differ from words and sentences, but exactly how they differ is what we are trying to figure out — and involves knotty philosophical issues, like are the signals fully arbitrary, like words are; are they intentionally produced to convey a specific message to others; are they learned; what meaning is really conveyed — these are all questions several teams of scientists are trying to address in animals,” she said.

But complex communication involving compositional structures and even a degree of syntax (or word order that plays a role in determining meaning, as in human language) isn't just a mammal thing. Just as we've had to gradually acknowledge that it's not only humans, then not only primates, that are capable of complex "sentence" structures that flexibly convey meaning — whether this represents something like a transitional evolutionary phase between animal calls and human-type language, or something complex and sophisticated enough to be considered a language of its own — we now have to further expand our understanding. As it turns out, the only living descendants of the dinosaurs also seem to combine in complex ways to produce an array of meanings. Among birds, the Japanese tit offers the first experimental evidence for compositional syntax in any non-human species. 

Meanwhile, the southern pied babbler does not, in fact, merely babble. By playing back so-called "mobbing sequences", combinations of calls by which the birds seem to recruit other group members in response to danger and watching the reaction, scientists were able determine that the combination of two calls used for other purposes communicates information about the context — danger — and the requested action — "let's get together and fight" — in a form of rudimentary compositionally. These findings of two birds that combine individual "alarm" and "recruitment" calls to  express a more complex idea using the two idea units suggests that syntax may in fact be more widespread than we thought.

OK, so what do they talk about?

This, Crockford said, is the most fascinating question. 

“Chimps are highly social and intelligent. They have close friends and family that they like to spend the day with and nest together with at night. But they live in dense forest where it’s incredibly easy to lose each other without using sound. So their calls have become quite context-specific to firstly, tell others what they are doing: ‘I'm resting’, 'I'm travelling’, so others know how to find them. Then they have specific calls to recruit others — to help in a hunt, to join them at food, to help them chase off a leopard or nasty neighbours — each call different. Then they have a bunch of specific calls they give when interacting with others to keep the interaction going longer, such as laughing during play, or teeth clacking during grooming,” Crockford said.

And as they work on those bigrams, the researchers are beginning to understand call combinations. “It seems they can combine two meanings together into an utterance, for example, 'rest' and 'food', which may translate into something like 'I'm staying longer at this food patch'. Telling others what you are doing, helps others decide if they want to stay and do it with you or move on. Such utterances probably mainly help bond partners to coordinate their activities in the dense forest vegetation so that they can stay together for longer,” Crockford added.

By contrast, Berthet notes that bonobo communication focuses mostly on getting the group together. This makes sense, as the highly social, matriarchal primates live in fission-fusion societies, which means that they split up into smaller groups to forage during the day, then get back together again as a larger group. 

And whales? We don’t know enough yet to say, but they seem to be gossips, the matriarchal and matrilineal society of sperm whales apparently consisting of a bunch of yentas who chit-chat all day long, talking over each other about anything and nothing for the sake of being together, while the adult men, though stereotyped as loners, may hang out together, perhaps sharing their own feelings with close friends, on the high seas. Shane Gero, the Canadian whale biologist who founded the Dominica Sperm Whale Project twenty years ago and is now biology lead for Project CETI, told NPR last year that “it's hard not to see cousins playing while chatting … to not see moms hand over to a babysitter and exchange a few words before walking out the door, so to speak, to go eat in the deep ocean."

Time to catch up

The structures that determine our relationships with animals and their status in relation to humans were all developed many years before we understood just how much some of them are like us. Like us, that is, in areas that have historically been important in determining these things among humans, such as showing compassion for our fellows, cooperation, strategic use of tools, culture and language. We might also note that plenty of research in human psychology — not to mention our apparent failure as a species to sustainably manage the ecosystems that in turn sustain us and all life on Earth — has demonstrated that our own claims to rationality and consciousness are, well, less impressive than previously advertised.

As Anne Benvenuti, a professor emerita of Psychology and Philosophy at Cerro Coso Community College in California, put it in a 2016 paper in the International Journal of Law and Psychiatry, "While convergent research on animal cognition, emotion, and behavior has increasingly pointed in the direction of animal “personhood,” interdisciplinary research in human cognition has simultaneously confirmed Sigmund Freud's hypothesis that not only are human beings not always self-aware and rational, but also the human unconscious mind motivates much of human behavior; and that human consciousness is fragmented at best." 

Language is a big part of this. Back in the 1870s, Oxford professor Friedrich Max Müller argued that "there is between the whole animal kingdom on the one side, and man, even in his lowest state, on the other, a barrier which no animal has ever crossed, and that barrier is —language." The lack of flexible and open-ended communication, which occurs by means of combinatorial structures or compositionality that allows us to put just about any idea into language, was seen as the great barrier. Increasingly, this barrier is being eroded.

 We still don't have evidence that bonobos want to be able to inherit property or that political participation is important to sperm whales. But we now know that many animals use complex communication systems that allow them to function as groups, or communities, and that they are able to express motivation, desire, fear and harm, all feelings that have political dimensions when they are threatened.

"Understanding the way [bonobos] communicate is also a way to understand what is important to them and what matters and what is relevant. So I think this is actually very important in terms of how we see them under the law. I hope it will help," said Berthet, who noted that when we started to understand that whales actually have a very complex language of their own, interest grew in protecting them. Bonobos, whose only wild populations are all in Democratic Republic of Congo, are highly endangered.

"It's an area of the world that is really under a lot of threats because of all the mining and the logging. So I really hope that if we show that they have complex communication, that somehow they resemble humans in some respect, and that they have this intelligence, I might say, then I hope it will help to have concrete actions to really preserve them," Berthet said.

Meanwhile, the very artificial intelligence that is increasingly allowing us to eavesdrop on animals is increasingly showing us that our own intelligence may one day be truly topped by machines, if not by the often very different but perhaps equally complex and meaningful intelligence of our fellow animals. Add the examples from psychology and the bungle we're making of the Earth we "manage," and humans don't have good evidence to put ourselves at the top of some kind of evolutionary hierarchy after all. And then this accumulating research into compositional language increasingly supports the idea that many animals are complex, sensitive and flexible thinkers.

"This simple fact calls for a rethinking of foundational concepts in law and health sciences," Benvenuti wrote, nearly a decade before these latest findings.

“We now realize,” Crockford said about the chimpanzees, “that to understand the complexity of animal 'languages', we need to study the animals' natural communication — what they communicate to each other. In the wild, chimps for example use specific calls for a number of contexts they never experience in captivity — they have a specific bark to recruit others to hunt, and other calls for territorial defense.” 

While the moratorium that currently exists on using chimpanzees in research in western countries, at least, is a good first step, it’s not enough. And rethinking rights could also allow us to increasingly understand our animal brethren, in a kind of virtuous circle.

“We're just beginning to understand the complexity of chimp communication — as they have become critically endangered. As they are hunted by humans for the pet trade and for meat, and their forests are destroyed, their communities are broken up, risking that they lose their local cultures,” Crockford said. “If we do not find a way to stop their decline due to human disturbance, we may never get to understand the complexities of their communication.”