Is That Dog Talking or Are We Fooling Ourselves?

This article originally appeared on VICE Netherlands.

“Outside! Now!” Silke has had enough of the conversation between me and her owner Karlijn Koning. After pushing down on the two buttons corresponding to those words, the dog puts a frisbee on the ground in front of Koning, sticking her tongue out and wagging her tail while looking expectantly at her.

“Soon, outside, Silke!” says Koning, 31, from the Dutch town of Rosmalen in the South of the Netherlands. Her dog Silke is a cross between a Croatian sheepdog and a short-haired border collie. Koning laughs at my look of surprise. “We have this conversation almost every day,” she explains. From my vantage point on her couch, I can see that her living room has been taken over by colourful foam mats and about 30 sound buttons, each corresponding to a different word, organised in different coded sections.

You might have come across videos of similar scenes on social media – cute pups (and even cats) telling their owners to give them scritches or to go to the beach. Bunny the sheepadoodle broke the internet with her ability to master 90 buttons and her alleged capacity for philosophical thinking, like, for instance, looking into a mirror and pressing the message: “What? Bunny.”

The whole phenomenon started with another dog, a Catahoula-Blue Heeler mix named Stella, and her owner Christina Hunger, a speech pathologist. “When I brought my new puppy, Stella, home [in 2018], I realised she demonstrated many of the same prelinguistic communication skills as toddlers do right before they start talking,” Hunger writes on her website. 

This gave her an idea – in her practice working with children with language development delays, she uses Augmentative and Alternative Communication (AAC) devices, which helps people communicate without speaking or writing. “Could Stella use an AAC device to express herself the same way my patients did?” Hunger wondered.

Today, five years after they began training, Stella apparently knows 45 words and can string up to five together to make a sentence, according to her owner. In 2021, Hunger published a book about their experience, How Stella Learned to Talk.

This newly observed animal behaviour has also piqued the interest of researchers, who launched the largest study into animal communication to date in March 2020. Researchers at the University of California in San Diego teamed up with the sound button company FluentPet to analyse footage of thousands of pets appearing to speak with their owners, recorded off home cameras across the world.

The aim of the research is to see whether dogs and cats can actually communicate with people with words, whether they really understand what the buttons mean and whether they consciously press them. If that turns out to be the case, this research could expand our understanding of animal intelligence and open the door to a whole new world of interaction between humans and animals.

Silke seems to completely understand what the buttons are for, Koning explains to me. “Her favourite buttons are 'food', 'water' and 'poop',” she says. “When she's in greater need, she presses the corresponding button repeatedly.” Koning believes Silke finds the buttons useful. “Otherwise, she would have to stand by the door [if she needs something]. You're more likely to ignore something like that.” 

Koning has even noticed that Silke might be taking advantage of her new communication tools to manipulate her owner. “Sometimes, she uses 'poop' as an excuse to go for a walk because she knows I won't wait to take her outside,” she laughs. 

According to Koning, Silke might even be able to make up new words. “Last summer something fascinating happened,” says Koning, talking about a video she posted on her Instagram page. “She had a sore tooth, so she often got an ice cube from me for the pain. At one point she pressed ‘bone’ and ‘water’. ‘Water bone? What is that?’ we wondered. And then it suddenly clicked. She meant ice!” 

Not all buttons denote actions – some are used to express emotions: angry, happy, busy, calm. “When it comes to 'busy,' I'm not sure if she means she's bothered by the hustle and bustle or if she presses it when she is busy,” Koning adds. “Sometimes it feels like she uses words just to describe what she sees or experiences. But, of course, it's all just guesswork.”

This is one of the biggest criticisms levelled against the proponents of this theory – how much of this communication is real and how much are owners simply projecting onto their pets?

The field of animal-human communication is plagued by a grim history of failed experiments and malpractice. The decades between the 50s and 70s were a golden age of research in this field: Two schools of thought arose within the linguistics community. Structuralists like Charles Hockett believed they had evidence that animals could be trained in human language. Generalists like Noam Chomsky thought of language as a human phenomenon. Animals could surely be taught tricks, but they didn’t have the mental capacity to understand our complex system of communication.

Heated debates and feuds broke out between the two camps. Then, in 1973, structuralist Herbert Terrace began a landmark experiment that defined the field for years to come – he picked up a baby chimpanzee from a lab and sent him to live with a human family on the Upper West Side of Manhattan, New York. Pettily named Nim Chimpsky, the baby chimp was also taught American Sign Language (ASL) at Columbia University – where Terrace still teaches psychology and psychiatry – by the professor’s team.

Over the next four years, Project Nim became an unmitigated disaster. While the chimp did manage to learn over 120 signs, he would only use them to get rewards, thus failing to prove he understood the concept of language. Members of his team were also accused of crossing boundaries with the animal, including smoking weed with him.

Nim became aggressive with his host family and shuffled from one caretaker to the next. In 1977, the experiment was abandoned – and so was the chimp. Nim began showing signs of depression, and was sent to a medical facility where he spent the rest of his days until his death in 2000. 

Nim was not an isolated case – it came after a string of failed research projects involving apes and ASL, with Koko the gorilla being another prominent experiment. Researchers working in the field were accused of cherry-picking data, prompting their subjects and over-interpreting results. The sceptical generalists seemed to have been proven right, and all studies in the field came to a halt.

That is, until 2020, when comparative psychologist Federico Rossano from the California University in San Diego started They Can Talk, the research project Koning and Silke are participating in. Rossano, 43, has a PhD in linguistics and studies the behaviour and mental processes of children, apes, goats and wolves, among other things.

When Hunger’s videos first went viral, Rossano’s colleagues reached out to him, but he was sceptical at first. Then, he received a message from Leo Trottier, a former alumni of his department who also happened to be the CEO of PetFluent. “​​What Leo told me is: ‘Look, we already have 1,000 people who ordered buttons. At least half of them said they would be interested in doing a study. Why don’t we do it?” the researcher recalls.

The message came early in the pandemic when Rossano couldn’t continue his regular work on children and apes. Undertaking the study was a risky move, since the linguistics community is still passionately divided on this topic, but it was also an exciting one. What ultimately tipped the scale for him was the scope of the project. “People are gonna start doing this, potentially millions,” he explains. “It’s an opportunity for citizen science.”

Today, over 10,000 pets have entered the study, with about 2,000 currently participating in 27 countries. The project combines live video monitoring with home visits, surveys, interviews, machine learning tools and a number of experiments. Their first paper has been in review for a few months. “We can't discuss the findings,” Rossano says with a twinkle in his eyes.

“Our goal is to start with very basic questions,” Rossano adds. “Can we make the case that they are indeed learning something, or is it completely random? Is it imitation? Or is there something else? To be able to address this, you have to crunch a lot of data.” Luckily, many researchers – including sceptics – are showing an interest in the project.

One of the frameworks Rossano is using to get a better understanding of this phenomenon is the so-called 13 design features of human language first developed by Hockett, which distinguish human language from animal communication. 

Productivity, for instance, refers to the human ability to create new words. One way to do this is by combining two words, an ability observed in young children, but not (yet) in animals. As a possible example of productivity, Rossano mentions a dog pressing the buttons for "squeaky" and "car" every time it hears an ambulance. Silke’s “water bone” could also be evidence of productivity.

Another feature seen as uniquely human is the ability to describe things that are not nearby. This could be illustrated, for instance, by a video of tabby cat Billi apparently missing her dad. Her owner expresses their love for Billi, but the cat taps the button “mad” twice with her paw. But does this really mean that Billi misses “dad” and is upset because of it? “We see that they are pushing the buttons like it’s a sequence,” Rossano explains. “The question is: Is it random, or is it structured?” 

A big criticism of Rossano's project is that the animals might be copying their owners’ behaviour without understanding what they’re actually doing. “It’s fair to be sceptical,” Rossano says. “As a scientist, you need to see data first.” However, “none of the people who criticise this have seen the data,” Rossano continues. “Maybe, you’d see something else if you watched 10,000 dogs for a few years.” (The study includes fewer cats as they’re much harder to train – sorry, cat lovers.)

What seems fairly clear to him at this point is that the animals are trying to communicate with their humans, although the degree of sophistication of this communication is still up for debate. In one video he saw, a dog began pressing the “help” button when his canine buddy became trapped under the couch. In another video, a dog is seen pushing the buttons “stomach” and “ouch” for ten minutes before vomiting. 

“This is fascinating for us,” Rossano observes. “Because: one, [pets] are [apparently] aware of things happening around them; two, they can communicate it; and three, they care.”

Of course, not all animals are interested in the buttons, even if they live in the same household and have similar breeds and training. But this type of individual variation is to be expected, as it occurs naturally in people too, Rossano says. “We want to see to what degree this has to do with temperament and the age of starting [with training],” he adds.

If his evidence holds up, the study could have immense implications. Owners would be able to learn more about their pets’ thoughts and needs. Vets could ask sick animals what might be wrong with them. It could also revolutionise the field of animal welfare – would we be as comfortable with industrial farming if pigs could tell us how they feel?

Ultimately, Rossano concludes, humans have always been fascinated by the mystique surrounding the animal mind, and hoped we’d be able to talk to them some day. From fictional heroes like Dr Doolittle to real-life research projects, like Jane Goodall’s work or the SETI Institute’s study of whale communication, we’ve invested large amounts of time and money into these big ideas. But our passion for this issue could be a double-edged sword – maybe we just want to believe in this possibility a little too much.

Koning truly believes that her dog is talking to her. She has depression, a condition that some studies indicate dogs might be able to sense and even want to provide support for. When she is down, Koning has the impression that Silke tries to cheer her up through the buttons. 

In a video shared on her Instagram page, Silke is seen pressing the buttons “Want”, “Good Job!” and “Karlijn” in quick succession, with “Good Job!” being a phrase Koning always says to Silke in a cheerful state. She then wags her tail and puts a paw on her owner’s shoulder. “This,” Koning writes in her caption, “meant the world to me.”