Bunny the Sheepadoodle looks into a text bubble shaped mirror
Illustration by Cathryn Virginia 

Can Bunny the Talking Dog Really Talk?

Bunny is more than an internet celebrity, she is part of the latest attempt to research the limits of animal cognition.
December 22, 2020, 4:15pm

It's probably the most famous existential crisis since Jean Paul Sartre's novel Nausea: In November, a sheepadoodle named Bunny looked into a mirror and said, "Who this" by pushing buttons on a soundboard designed for her to communicate with her owner, Alexis Devine. Moments after seemingly questioning her existence, Bunny pushed the button, "Help.”

Advertisement

As reckonings with one's existence go, it was more succinct than Sartre's “I am, I exist, I think, therefore I am; I am because I think, why do I think? I don’t want to think any more, I am because I think that I don’t want to be, I think that I… because… ugh!”

With over 5.6 million followers on TikTok and 555,000 followers on Instagram, memes quickly proliferated about Bunny's existential plight. 

“We all jokingly predicted this but I don’t think any one of us actually expected Bunny to cross the line of conscious existentialism,” one person tweeted.  Others made parodies of Bunny using the buttons to ask questions, “What point of Bunny? Bunny die one day?” 

“What would Bunny[’s] answer to the trolly problem be oh my god,” one TikTok user wrote. Daily Show correspondent Jaboukie Young-White tweeted that “Bunny is a button away from needing an emotional support dog.” 

Is Bunny on her way to grappling with ethical thought experiments? Is she using human language to share the inner workings of her mind? To what extent, if at all, can animals use language? These are difficult questions, and ones that researchers have long tried to investigate in one way or another.

Advertisement

Over the decades these inquiries have led to in-fighting between those who believe that some animals can be trained to converse with humans, and those who think that human bias and misinterpretation underlie any such communication skills. 

Bunny is more than a TikTok celebrity, she is a participant in the latest attempt to research these issues, through her involvement in a large study called They Can Talk, based at the Comparative Cognition Lab at University of California San Diego (UCSD). Her custom Augmentative Alternative Communication (AAC) soundboard has over 70 buttons that produce pre-recorded words in six categories; the buttons are organized using the Fitzgerald Key, originally developed to teach deaf children sentence structure. 

“Right now, if you ask me, ‘Does Bunny talk?’ the answer is, I don’t know.”

Bunny is the most well-known of the participants, but there are currently over 900 animals involved. One of the lead researchers, UCSD's Federico Rossano, said he wants to prove that studies that probe the limits of animal communication can be done rigorously, with big sample sizes that collect troves of data, and are followed up with targeted experiments in the lab. 

Yet, to figure out if Bunny can "talk," let alone wonder "who this," we have to grapple with defining what language is in the first place, and ask how our interactions and biases about animals might influence our conclusions. 

Advertisement

“Right now, if you ask me, ‘Does Bunny talk?’ the answer is, I don’t know,” Rossano said. "What I can tell you is that Bunny clearly seems to be engaging in some sort of communicative exchange.”

Outside of her existential pondering, Bunny uses her buttons to ask a lot of questions about poop. At times, Bunny also appears to ask for what she wants, whether it's "scritches," or to be let outside. In one recent video, Bunny pushes the “help” button. Devine asks, “How can I help? What help?” Bunny pushes “friend,” then “play.” “Who do you want to play with?” Devine asks. Bunny walks over to the sliding glass door, and when Devine follows she sees a group of neighborhood dogs waiting patiently behind the fence.

Once it seemed as though Bunny could tell Devine that she had something caught in her paw. Bunny pushes the button, "Mad." 

“Why mad?” Devine asks.

“Ouch,” Bunny pushes. 

“Where is your ouch? Where ouch?” Devine says.

“Stranger, paw,” Bunny pushes. 

“Let me see your paw,” Devine says, and then finds an embedded thorn.

Devine said these interactions have led many of her followers to think of Bunny less as a dog, and more as a human in a dog suit. She sees comments all the time in which people write things like, “Oh my God, this dog is a genius. She's talking. She's just like a toddler,” Devine said. On the flip side, there are people who call Devine a fraud, and accuse her of selectively editing or cherry-picking her videos. 


Watch more from VICE:


“It's really an emotional roller coaster for me,” Devine said. “There are some days where I am so frustrated and I believe it's all random and the skepticism in me overrides everything else. Then there are days when almost every single utterance is clean and concise and contextually appropriate and I’m like, ‘There's definitely something going on here and it's not random.’”

Bunny is a little over a year old. Before Devine brought her home, she came across Christina Hunger, a speech pathologist on Instagram, who was teaching her dog Stella to communicate using similar buttons. “I was astounded by the possibilities,” Devine said. “It’s sort of everyone's childhood dream to be able to talk to their animals. I devoured her blog and I had some buttons waiting for Bunny.”

Advertisement

When Bunny arrived at her new home, there was an “outside” button by the door. “I had zero expectation and quite frankly, I didn't think I would have much success because I don't have any knowledge in any field even remotely related to teaching my dog how to talk.”

But Bunny took to the button within just a few weeks—she was using the button consistently to request going outside.

“At that point, it was game on,” Devine said. 

Rossano’s background is in linguistics, but because of his expertise in nonverbal communication, he has dedicated a large part of his career to animal cognition research. He has worked with goats, horses, wolves, chimpanzees, bonobos, gorillas, orangutans, gibbons, marmosets, and Japanese macaques—and now, dogs.

He said that dogs provide a new angle through which to study animal communication because rather than being trained in a lab, removed from their natural setting, dogs can participate from their own homes, and researchers get to take advantage of a strong bond that naturally exists. 

“We know that dogs care about humans enormously,” Rossano said. “We know that dogs are much better than any other animal in understanding human communicative signals.” As Yale psychologist Paul Bloom wrote in Science in 2004, “For psychologists, dogs may be the new chimpanzees.”

Rossano acknowledged that it's very possible Bunny could be pushing buttons by accident, leading to random presses that occasionally make sense to us. It's also possible  she could be memorizing certain combinations that produce positive responses from Devine, while not fully understanding what the sounds mean.

“It is absolutely true that we need to be super careful,” Rossano said. “It’s why we have not put out a paper yet.”

Advertisement

Devine has had three cameras on Bunny’s soundboard that capture video for 24 hours a day for the last several months, so Rossano and the research team don’t get just the highlight reel that social media does—they see everything, including mistakes and button presses that don't make sense.

Rossano said that after examining Bunny and other canine subjects around the clock and analyzing thousands of hours of data, he and his colleagues will create experiments where the dogs will be tested on the limits of their understanding. This would include having someone who isn’t Devine communicating with Bunny using the soundboard, or testing her on words with no human in sight. Or removing buttons and seeing how she might try and communicate without a button she usually uses. Could she combine other buttons to get the same point across? 

“We are not taking what we see on those clips as evidence of abilities,” Rossano said. “We have more work to do. But we do think that it is possible.”

Humans have an endless curiosity about the inner lives of animals. Call it a desire that stems from the loneliness of human sentience—we want to know if the animals we share the planet with have thoughts like ours, or if they could talk to us if we provided them the right tools. 

When animals appear to show a flicker of cognitive achievement, we are enthralled. As Margaret Talbot wrote in The New Yorker in 2008, when Alex the talking African Grey parrot died, his owner and trainer Irene Pepperberg received around 6000 condolence emails.

Advertisement

But seconds after answering the phone, Pepperberg, a research associate and lecturer at Harvard University, cautioned heavily against using the word “language” when describing what Bunny, and other animals in the study, could do with their soundboards. 

People have misinterpreted animal communication before, she advised. In one famous example, a horse named Hans seemed to be able to answer basic math questions like "What is 2+2?” through tapping his hoof. It was revealed that Hans wasn’t actually a math whiz, but able to intuit the correct response by watching for subtle changes in body language of the people around him. 

It was a clear warning that animals might have remarkable skills of perception, but are not always intelligent in the ways humans first think they are. A misinterpretation of an animal’s behavior as higher-level cognition is now called the “Clever Hans Effect.” 

Pepperberg said that in-fighting has plagued the field of animal cognition regarding the credibility of the attempts to teach animals language. Starting the 1960s with a chimpanzee named Washoe, researchers have utilized American Sign Language to teach various animals the signs for words, and documented their developing vocabularies. In the 1970s and 1980s, these efforts continued with well-known research subjects like Koko the gorilla, Kanzi the bonobo, and Nim Chimpsky. Kanzi, for example, appeared to understand more than 300 English words. He would listen to the words through headphones—to avoid behavioral cues from the researchers—and then point to a symbol on his keyboard that represented the word that he heard. He was also able to respond to various commands. 

Advertisement

But in 1979, Herbert Terrace, one of the lead researchers on the Nim Chimsky project published an article in Science called,“Can an Ape Create a Sentence?” His answer was, resolutely, no. He claimed that Nim hadn’t learned to sign at all, but was prompted unwittingly by over-enthusiastic human teachers. He concluded that chimpanzees couldn’t even learn words, let alone grammar. 

Animals might have remarkable skills of perception, but are not always intelligent in the ways humans first think they are.

When Terrace wrote that the project with Nim had essentially been an accidental fraud, the implication was that other animal language studies were prey to similar biases and blindspots. In a New York Times article from 1995, Terrace said that Kanzi, “like the disappointing Nim Chimpsky, is simply ‘going through a bag of tricks in order to get things.’”

In 1980, the New York Academy of Sciences hosted a conference organized by Thomas Sebeok, a researcher in zoosemiotics, and Robert Rosenthal, an expert in nonverbal communication. At the conference, people who taught sign language to animals were “accused of cuing their apes by ostensive signals,” and “of consistently overinterpreting the animals’ signs," Pepperberg wrote in a 2016 paper titled, Animal language studies: What happened?

Advertisement

As Science reported at the time of the 1980 conference, “It was amazing that any of the ape language researchers should even have considered stepping into such a lions' den… The very framework of the conference implied that their work fell into the category either of circus tricks or of self-delusion.” In a press conference the next day, Sebeok said that “the alleged language experiments with apes divide into three groups: one, outright fraud; two, self-deception; three, those conducted by Terrace. The largest class by far is the middle one." 

In spite of the attempt to tarnish the reputation of monkey and ape research, some researchers continued their work, and with other animals. Pepperberg's work on Alex showed evidence that he wasn't just "parroting" words, but could understand categories like "same and different," and "bigger and smaller." 

“He even seemed to develop an understanding of absence, something akin to the concept of zero,” Talbot wrote. “If asked what the difference was between two identical blue keys, Alex learned to reply, ‘None.’ (He pronounced it ‘nuh.’)”

Pepperberg told Talbot that once, when rushing into the lab, Alex said, “Calm down.” During long training sessions he could say, “Wanna go back,” meaning, to his cage. He achieved a similar level of fame as Bunny did, appearing on television programs and having countless articles published about him. According to the New York Times, his final words were: “You be good, see you tomorrow, I love you.”

In 2004, a study published in Science documented how a dog named Rico knew over 200 different toys by spoken name. In carefully designed studies, Juliane Kaminski, a Reader in Comparative Psychology at the University of Portsmouth in the U.K., and her colleagues asked Rico to pick out toys by name, but hid themselves out of sight, so that Rico couldn’t pick up on behavioral cues from the humans. Rico could remember items' names for around four weeks.

But what really made Rico famous, Kaminski said, is that he could learn new words by exclusion. In other words, you could present him with a new object that he didn’t know the name of, and ask him for it, and he would attach this new label to the new thing that he had never seen.

Advertisement

Learning by exclusion was seen to be a uniquely human word learning process before Rico. “He was the first dog to be shown on the scientific measures to be able to do this,” Kaminski said.

Kaminski first saw Rico on a German TV show where he was showing off his skills. “I thought that he wasn’t really understanding,” she said. “That the was just responding to nonverbal cues that the owner was giving unconsciously, not really responding to the words.” 

Working with Rico changed her mind. She wouldn't call Rico's ability "language," but she thinks it goes beyond a Clever Hans effect.  “I think it's a big reminder that we need to be open minded," she said. 

Rossano thinks in light of all the bad blood from the past, it's important to refine our line of questioning. Rather than just asking whether animals have language, we first have to ask ourselves more basic questions: What is language? What is the line between communication and language? Is language like a light switch, either on or off, or does it operate on a dimmer,with humans at full brightness and an animal like Bunny only a quarter of the way on? 

Instead of asking whether an animal can learn human language—a question that may always lead to disappointed answers, because of the way it's framed—Rossano said he’s interested in learning whether dogs have the ability to demonstrate one or two of the qualities of language that we’ve so far considered exclusive to humans. 

Advertisement

Language is more than just a vocabulary of nouns, or making associations between a sound and an action or object. “Anybody who has a dog knows that after a while you have to spell out ‘walk’ unless you actually want to take them out,” Pepperberg said. Does this mean they know that the sound and word "walk" symbolically represents an action? This burden of proof will fall on Bunny's research team— to prove to what extent the dogs really understand that the buttons stand for something else, and have meaning. 

It could be that all the buttons elicit a positive response, whether it’s a walk, scratches, or a ball. “In that case, the dog doesn't necessarily need to specifically distinguish,” Kaminski said. “It's just something positive is going to happen and I'm going to be excited either way.” 

Another facet of human language is that it's generative, which means that using the building blocks or letters and words, we can generate an infinite number of new sentences and ideas. Rossano wants to see if dogs can learn to combine buttons in new ways, to reference new things that they don't have a specific button for. He said that both with Bunny and other participants, he’s already seen examples of dogs combining words in this way. One example is that when it rained, a dog pushed the buttons, "water" and "outside." 

When humans speak, we use "displacement," or refer to things that are not present in time or space. It’s something we do all of the time, but there’s not a lot of evidence that animals can do so. Yet, in some of Bunny’s videos, there's a compelling suggestion of it. 

Advertisement

Bunny appears to be able to ask for her dad when he isn’t there: In one video, Bunny pushes, “Dad, where, come, settle.” Devine responds, “He’s upstairs working.” 

By using the “where” button, in the case of “where dad,” it suggests she might know the meaning of “where.” “That's what we need to make sure as scientists that we test properly, asking where somebody is or where something is,” Rossano said. 

The simple answer is no, animals don’t have "language," Pepperberg said. They can’t communicate at the level that two humans can. Kaminski said that we need to strive for a middle ground—between asserting that Bunny is reaching  a high level of consciousness, or that all of her behavior is random. After all, animal cognition is still a relatively young field. 

“We’re only in the last 30 to 40 years started looking,” she said. “Up until then, the idea was really that there was no way that animals can kind of have any flexible thoughts about anything. Of course without looking, we can’t find anything.” 

Even if the answer to "Do dogs have language?" ends up being not really, there’s still much to explore about dog communication, and how the bonds can be enhanced between human and pet. 

While adhering to a healthy dose of skepticism, we shouldn’t short-change a dog's abilities to communicate with us already. “They're keen observers of our verbal and nonverbal behavior,” said Zach Silver, a Ph.D. student working at the Yale Canine Cognition Center. There’s even evidence that dogs evolved two specialized muscles, that wolves don't have, so that their emotional expressions would be more coherent to humans. 

Advertisement

The Austrian-British philosopher Ludwig Wittgenstein said “If a lion could talk, we could not understand him.” He meant that even if a lion could speak our language, the world that they inhabit is so different from ours, it would be hard to converse. But dogs have lived in our homes and alongside humans, domesticated, for thousands of years. They know where we live, what we do, and when we're getting upset. 

One dog in the study's owner was getting riled up while watching a presidential debate, and the dog pushed the button “settle.” “It basically was like, “calm down,’” Rossano said. “We do believe that one of the fundamental reasons this could be happening is because of the bonding between the owner and the dog and the relationship that they have."

“The debate often seems to take on a quasi-religious tone, as if the issue were whether animals had souls rather than whether they have languages.”

Depending on the outcome of the study, Silver thinks it could show that dogs have a higher capacity for communication with humans that we previously thought. The prevailing view in the field right now is that dogs are very good at interpreting communication from us, but they're not so great at communicating with us through the production of their own communication—this would challenge that assumption. 

The work could also open up a multitude of new methodological possibilities for how scientists, like Silver, test dogs in the lab. “If dogs can reliably tell us what they were thinking in certain situations, this would open up an entirely new world of studies," Silver said. 

Advertisement

Bunny's fans aren't always interested in whether or not she can "talk", per se, but more about what's going on in her mind, the extent to which she is conscious. “The [nonhuman language] debate often seems to take on a quasi-religious tone, as if the issue were whether animals had souls rather than whether they have languages,” one introductory linguistics class lesson from the University of Pennsylvania pointed out

It comes back to a desire the philosopher Thomas Nagel phrased well in his 1974 paper What Is It Like to Be a Bat? “I want to know what it is like for a bat to be a bat," he wrote. "Yet if I try to imagine this, I am restricted to the resources of my own mind, and those resources are inadequate to the task.”

When we look into the eyes of our pets, or even our human companions, we don't have a way to access what they're thinking or feeling—what it's like to be them. With humans, we can hope to trust that what they tell us, but with animals, our insights are blocked. As a result, we can take it too far, or not far enough, in assessing the extent to which they think. 

Devine said she thinks Bunny "challenges something about our own humanness that we're not the only creatures that have complex inner monologues." 

It's another interesting thought to mull over: Is Bunny communicating thoughts she had already? Or when given the buttons "who" and "is," is she undergoing some sort of cognitive revolution?  Silver said he doesn’t think that Bunny is becoming a kind of Dog 2.0, advancing beyond her species through the use of the buttons.

“My instinct would be to say that we're just allowing dogs to better express the version of themselves that we already know,” Silversaid. “There's a chance that maybe dogs all have that question when they look in the mirror, and we just don't have the tools yet to know that.”

Anything that gives an animal more autonomy in its day-to-day life is a good thing, Pepperberg said. “I think it's very exciting that we can establish some level of communication with our pets," she said. "I’m very positive about those kinds of things. I just want people to be very careful in how they're interpreting what's happening.”

That's Devine's bottom line: She cares first and foremost about her relationship with Bunny, buttons or not. If Bunny wasn’t interested in initiating the dialogue with the buttons herself, Devine said she wouldn’t force it on her. 

The vast majority of Devine’s and Bunny’s day is spent away from the board, doing activities that all dogs love to do—playing with toys, going to the park, romping on the beach and cuddling. 

“None of that is necessitated by her pressing buttons, ever,” Devine said.

Follow Shayla Love on Twitter.