The Question of Animals and Language… Why Is it Even a Question?
Jeremy Bentham, a philosopher and professor of law, posed this thought in 1789 about animals: “The question is not, can they reason? Nor, can they talk? But, can they suffer?” This notion that suffering is what makes humans consider and care about animals led to hundreds of years of animal welfare. This concern over animal suffering led to the Five Freedoms, generally recognized as a method of assessing animal welfare. The five freedoms are: freedom from hunger and thirst; freedom from pain, injury, and disease; freedom from fear and distress; freedom from discomfort; freedom to display natural behaviors.
Animal welfare has developed further than those simple freedoms, though it has never abandoned these freedoms and the moral concern over animal suffering. But even now, in 2025, animal researchers and scientists often don’t take seriously the idea that animals can talk. Obviously, any animal scientist or really any observant human can see that animals have methods of communication, but the debate that’s existed mostly in recent times is whether that communication is language… Are they talking? What is it that we consider a language and talking? Ludwig Wittgenstein wrote in his Philosophical Investigations, “If a lion could talk, we wouldn’t be able to understand it.” We can’t fully, deeply understand a mind and body that is not our own. We can more closely understand the mind and body of other humans, but we are never actually having the identical experience of anyone else. So then it follows that we cannot simply look at a lion and know whether the lion has a shared language among lions, whether the lion’s thoughts move and flow in the same way that a human’s thoughts often do. But, if we watch lions in the wild for a long enough time, we see them interact with other lions and the other lions are not shocked, confused, panicked, or behaving in erratic and different ways each time they simply look at each other. We also see them interact with other animals. Perhaps the lions aren’t saying much to the antelope they are hunting, chasing, catching, and eating, but the antelope’s behavior demonstrates that it has received a communication that there is a lion, a dangerous predator, hunting it and tries to flee. The lioness group coordinates, working together, to hunt and catch the antelope. If they were not communicating between themselves in some way, they’d interact with each other in a random and meaningless way, akin to pebbles in a jar of water being shaken. If we can, which may be impossible, remove our human anthropomorphizing of the lions at least enough to forget what we think we know and understand about language, our human language, could we discover the lion’s language? Would we then all be able to agree that lions are talking?
Am I talking about philosophy or story? I think it might need to be both, and maybe even neither. There’s a logic and a science to communication and behavior. I always insist that everything is a story, all types and levels of communication, from the mundanely simple to the intricately complex. This is not philosophy. Communication is transferring meaning: whether the meaning is understood or received does not matter. Demonstrating any type of communication, whether we recognize it or not, then requires intent. Every living creature that moves, breathes, eats, or reacts in any way to the world around it must have intent to do any of that. Even plants communicate with each other and the world around them, and move and bend to face the light or shrink away from it. Our human understanding, or lack thereof, does not negate or endorse their intentions and communications, it only impedes us from examining them more closely when we become dismissive of other creatures as less capable than humans, simply because their way of existing in the world is different from ours.
Wilhelm von Osten was a gymnasium mathematics teacher, phrenologist, and amateur horse trainer in Germany who owned a horse named der Kluge Hans, or Clever Hans. The horse was born around 1895 and died 1916. Clever Hans was special, being exhibited around Germany as a horse who von Osten claimed had been taught to do math, tell time, understand a calendar, and read, spell, and understand German. Clever Hans would respond to questions by tapping his hoof. Around this time, interest in animal intelligence was enhanced due to the work of Charles Darwin, so Clever Hans drew crowds and von Osten never charged admission. This interest led to the German Board of Education investigating von Osten’s scientific claims, so philosopher and psychologist Carl Stumpf formed the Hans Commission, which determined in September 1904 that Hans’ performance was not a trick. However, Oskar Pfungst, a German comparative biologist and psychologist, took over, and in 1907, after trials, it was determined that Hans could correctly answer 89% of the time (50 out of 56 questions) when von Osten was visible to Hans, even if others were asking questions. But when Hans could not see von Osten, Hans was only able to answer 6% of questions (2 out of 35) correctly. Pfungst had already known about cases of trained horses and dogs who were able to read their master’s expressions and respond, and that the investigation had not uncovered fraudulent tricks, so he determined that Hans was observing von Osten’s body language to decide when he was to tap his hoof and when to stop.
While the manner of this investigation and assessment have had some debate, like that Pfungst should have done further experiments in mimicking von Osten’s body language with von Osten out of sight, Pfungst’s determinations led to the observer-expectancy effect being referred to as the Clever Hans effect. It is still relevant in the way animal studies are conducted today. It is even seen in our human interactions with artificial intelligence. Our human bias infiltrates our observations and we anthropomorphize even without intention. The Clever Hans effect is often the first method of dismissal regarding animal cognitive abilities and communication. Animals speaking is seen as something that we as humans simply project. But, this branch of research generally ignores something very important. Clever Hans was able to determine the subtleties of von Osten’s body language and facial expressions to a high degree of accuracy, all while von Osten was convinced he had taught the horse math and language. Clever Hans was communicating and he certainly understood that it was his turn to respond in some way when tapping his hoof. A language can’t be formed without the understanding that it is a sharing of information between more than one individual. Was Hans truly understanding German? From the accounting of von Osten’s shows and the commission’s determinations, it would be wild to claim that Hans was communicating using the German language. But though the observer-expectation effect clearly comes into play here, it would also be wild to claim that no communication between an animal and a human occurred.
Koko the gorilla was a captive gorilla trained in American Sign Language by Francine Patterson for The Gorilla Foundation’s ape language experiments. Her language skills have been debated, with the common scientific consensus that Koko was not demonstrating the syntax or grammar required of true language. Koko, still an incredible ambassador for her endangered species, may or may not have demonstrated language acknowledged by experts, but even after her death, science publications still attribute the experiments with her as having contributed to the human understanding of animal emotions and intelligence. Koko, as a gorilla with hands shaped differently than human hands, had many of the ASL signs she was taught adapted for her. While the interpretation of Koko’s behavior and communication may have been incorrectly interpreted by her trainer, the primary researcher caring for Koko, who had Koko with her and isolated from other gorillas for most of Koko’s life, video of Koko interacting with humans is at least akin to Clever Hans: it would be wild to claim that Koko is not communicating something.
In episode 1727 of Mister Rogers’ Neighborhood, aired in 1998 and titled You and I Together, Fred Rogers visited with Koko, who had been watching his television show for some time. When Mister Rogers arrived, Koko grabbed his hands, hugged him, and even took his shoes off, which is what Mister Rogers does when he begins the show. Whether Koko is using ASL correctly, or demonstrating true language or not, these behaviors are specific and performed with intent. Koko isn’t receiving treats or praise for performing every behavior, either. This means that, whatever can be said of whether she is using “language,” Koko is demonstrating the intention to communicate something. Simply because we don’t understand it and can’t align it with our human understanding does not mean that it is not there.
Clever Hans and Koko aren’t the only examples of animals in the world of humans trying to use human language to communicate, though. Very recently, companies have popped up that make pet buttons, like FluentPet, and have made available the purchase of buttons for use with your regular, every day companion pets. These buttons are laid out on a slip-resistant mat and can be recorded by owners with a word or very short phrase for the animal to use a paw to press. Once pressed, the recording plays, and thus the animal has communicated in a human language… or so the concept goes. While influencers on social media mostly consist of dogs using these buttons, there are even some cats. One of the largest “dogfluencers” on the internet is Bunny, Alex Devine’s sheepadoodle dog, whose interactions with the buttons Devine published a book about titled I Am Bunny. I still have only read a small portion of this book, but I have watched many of Bunny’s videos. There are generally two opinions about Bunny and dogs like her using buttons to communicate. One is that, of course, our dogs are always communicating with us and of course they can talk with this Augmentative and Alternative Communication (AAC) mode. The other side believes that this is all fake and essentially a human projection. Watch any Bunny video, and it’s clear that Bunny is intentionally initiating an interaction. What’s less clear is whether Devine is interpreting Bunny’s button presses in a way that is actually true to Bunny’s experience as a creature with the brain of a dog, who thinks like a dog and definitely not like a human.
There are several problems with these buttons, starting with how these buttons appearing “in the wild” with everyday owners can’t constitute a well-constructed experiment, so no matter what appears to be happening in the videos, the only thing for certain is that it isn’t certain. Another issue with the buttons is that humans are interpreting them: a video clip of a dog who saw an ambulance outside shows the dog pressing “squeaker” and “car” and then looking at its human, who had to get up and see outside to think that the dog might have been describing the ambulance. When using any human language, especially one in which we have limited knowledge, we often do end up describing things to create a meaning of something we don’t know how to say. A phrase I learned practically at the same time as “hello” in Spanish long ago was “¿cómo se dice?” which means, “how do you say?” If I had to describe something in Spanish requesting assistance in learning the word, it began with that phrase. I recently learned how to say “train” in Chinese, but I haven’t learned the word for the engine. Fortunately, I do know how to say “the first thing on the train,” which another human would probably understand right away. Putting together a description this way is called “productivity.” But when we aren’t sure what the communication between humans and animals is, we can’t know or expect that they can get past this problem and that we’ll have a clue if they do.
Yet another issue for dogs with buttons is that we have strong evidence that dogs have trouble distinguishing our vocalizations very finely: a dog may hear “stay” or “shay” and know not to move based on the issuer of the command’s body language more than the sound the human made. This muddies the waters even more for dogs trying to learn and communicate with us in our human languages. We know that dogs do have a great capacity for human vocalized words and sometimes can even distinguish pictures. The world’s record holding dog, Chaser, was able to identify over a thousand of her toys by name, and the controlled experiments for her hold up. Chaser would leave the room to retrieve the correct toy from a large pile, being out of sight of her owner or anyone else saying the name of the toy. Chaser, obviously, had a great sense of distinguishing human vocalizations, but since not every dog demonstrates such a huge vocabulary, it seems she is likely among a minority.
So what’s right about Augmentative Interspecies Communication? Dogs have been able to go over to the door and ring a jingle bell to notify humans that the dog needed or wanted to go outside for as long as dogs have lived inside houses with us. Dogs are definitely demonstrating the ability to use these buttons at this level, by matching something with a meaning and even what would happen next. Anyone who has ever owned a dog who loved going for a walk to the point where you’ve run out of ways to mention a “walk” in any language you know and with any spellings of those words because the dog has learned them all and gets excited even when you’re not ready for the walk knows that dogs have a good idea about actions like this, with a short-term cause and effect. The blind tests also prove that dogs can and do use productivity, forming compound words like “squeaker car,” and that they can respond to verbs such as “play” without any human cueing.
Why aren’t the buttons a fantastic thing? What I am really interested in, and would love to have opportunity to study and experiment on my own in the future, is how to design a different kind of language. Humans call creating a language a constructed language or “conlang.” These are typically for a story or as in the case of Esperanto, with the intention to create an easy-to-learn, neutral language for international use. If you can speak Klingon, you speak a conlang. If you have heard someone telling you about the Elvish languages of The Lord of the Rings world and the difference between Quenya and Sindarin, you’ve heard of conlangs. And if you’ve seen the movie Arrival with Amy Adams starring as Louise, the linguist recruited to communicate with aliens, you have an idea about how understanding the other individual and their communication and language is essential to creating meaning between two parties. Conlangs begin their construction through assessing and designing for the target language user. In just the world of The Lord of the Rings alone, many different creatures, including trees that can walk and talk, have their own languages and each creature’s physical and biological aspects are the constraints to how the language develops. A science fiction trilogy from one of my favorite authors, Robert J. Sawyer, is about Neandertals in an alternate version of Earth accidentally breaching into a version of Earth with humans and Sawyer was careful to include that the Neandertal characters were unable to voice the phoneme /ee/ which is the long E sound in English like in “bee” or “fee.”
The science and research into animal communication and language ability was dismissed far too soon, and even now, we have Federico Rossano, a cognitive scientist and professor at the University of California San Diego, who is the leader of the Animal Communication Project, doing research with dogs and their ability to use AAC with some fear for his scientific reputation. Proclaiming that dogs are demonstrating understanding and meaning in their use of these buttons is difficult, but it should not be a risk to one’s reputation. I’ve said several times just in the course of this piece that the behavior of these interesting animals might not demonstrate human language use, but it cannot be ruled out that these animals are communicating in a way that fits them. If Bunny had no sense of language or communication, she would not have the executive function to engage with Devine at all. She’d smash buttons or ignore the mat entirely and only echo back a behavior by accident. Instead of only trying to interpret animal communication through an augmented human language, why wouldn’t we try instead to create an interspecies language between animals and humans that functions with what we do know and understand about that species? Dogs were domesticated by humans thousands of years ago and were domesticated even before horses. They are essentially wolves genetically and can interbreed with wolves, but only dogs can follow a human pointing or even look to a human for assistance in solving tasks, while wolves take no cues from humans, even if they’ve grown up hand raised by them. Only dogs out of all the species on this planet have evolved to naturally be attuned to human facial expressions, use body language and eye tracking to partner with humans, and can exploit their cuteness so brilliantly. We have the biggest wealth of information and data about dog and human interactions, so much so that when Colossal Biosciences wanted to choose the first species to test their genetic manipulations on and call “de-extinct,” they chose to manipulate gray wolf DNA to mimic a dire wolf due to that wealth of information.
I have many ideas about designing to experiment and explore language between humans and animals. Please hit subscribe if you haven’t already so that when that follow-up piece arrives, hopefully next week, you’ll be sure to have it appear directly in your inbox.
The follow up can be found here:
References
Bentham, J. (1823). An introduction to the principles of morals and legislation. W. Pickering. Originally published 1789, quote from page 311. Available at https://www.google.com/books/edition/An_Introduction_to_the_Principles_of_Mor/MtkQAAAAYAAJ
Bogdanovic, Z., Galiher, S. A., Jancer, M., Prada, L., Caramela, S., & Fike, A. (2024, August 9). Is that dog talking or are we fooling ourselves?. VICE. https://www.vice.com/en/article/do-fluentpet-buttons-work/
Bunny (dog). (2025, May 28). In Wikipedia. https://en.wikipedia.org/wiki/Bunny_(dog)
Clever Hans. (2025, June 8). In Wikipedia. https://en.wikipedia.org/wiki/Clever_Hans
Devine, A. (2023). I am bunny: How A “talking” dog taught me everything I need to know about being human. William Morrow, an imprint of HarperCollinsPublishers.
Dogs understand words from soundboard buttons, study reveals. Today. (2024, August 28). https://today.ucsd.edu/story/dogs-understand-words-from-soundboard-buttons-study-reveals
Dogs were domesticated not once, but twice... in different parts of the. University of Oxford. (n.d.). https://www.ox.ac.uk/news/2016-06-02-dogs-were-domesticated-not-once-twice%E2%80%A6-different-parts-world
Fig. 1 by University of California. (2025). Is Bunny the “talking" dog legit? Here’s what science says. Is Bunny the “talking" dog legit? Here’s what science says - YouTube. Retrieved from https://youtu.be/cTnVSJ8V4wk?si=PXUqfySJNdm1HPq3.
The five freedoms for animals. Animal Humane Society. (n.d.). https://www.animalhumanesociety.org/health/five-freedoms-animals
Gorilla Foundation. (2009, June 2). Mission Part 1: Research. Gorilla Foundation / meet koko / research. https://web.archive.org/web/20090602082734/http://www.koko.org/friends/research.koko.html Accessed through the Wayback Machine
Koko (gorilla). (2025, June 21). In Wikipedia. https://en.wikipedia.org/wiki/Koko_(gorilla)
Languages constructed by Tolkien. (2025, July 4). In Wikipedia. https://en.wikipedia.org/wiki/Languages_constructed_by_Tolkien
Living on Earth / World Media Foundation / Public Radio International. (2024, October 11). Conversations with dogs. Living on Earth. https://www.loe.org/shows/segments.html?programID=24-P13-00041&segmentID=4
PBS. (1998, July 28). Mister Rogers’ Neighborhood / You and I Together. You and I Together. episode, Pittsburgh, Pennsylvania. Video at https://www.misterrogers.org/episodes/koko-the-gorilla/
PBS. (2010, November 9). NOVA Dogs Decoded. episode. https://www.pbs.org/wgbh/nova/video/dogs-decoded/
Stephanie Gibeault, Ms. (2024, June 6). Dogs might be able to understand the meaning of words. American Kennel Club. https://www.akc.org/expert-advice/training/dogs-may-understand-the-meaning-of-words/
Tung, L. (2023, January 20). “talking” pets: The citizen science project that could reshape what we know about animals and language. WHYY. https://whyy.org/segments/talking-pets-the-citizen-science-project-that-could-reshape-what-we-know-about-animals-and-language/
Wittgenstein, L., & Anscombe, G. E. M. (1953). Philosophical investigations. Macmillan.