Unleashing the Truth: The Case of the Talking Dog
Keira Moore, PH-D, BCBA-D
If you've scrolled through social media recently, chances are you've seen the talking dog videos. Bunny, a famous black and white Doodle, is featured on Netflix's The Hidden Lives of Pets. Bunny and other dogs are typically shown having a verbal dialogue with their human "parents" by pressing speech buttons associated with words or phrases. On the surface, these lovely dogs seem to be contemplating their self-identity, showing emotions by saying, "I love you," and understanding abstract concepts like now or later. Being the sucker for pet videos that I am, I want this to be true. Who doesn't want to know what our little friends really think?! However, the scientist in me struggles to see this as anything more than entertainment.
What Do We Know?
There's plenty of content written by critics debunking talking dogs. One relevant point is the "Clever Hans" effect when owners subtly cue their dogs which button to press. Others point to carefully crafted clips only showing us times the dogs were successful. Some state the videos demonstrate a phenomenon psychologists call "apophenia," a tendency for humans to make connections and meaning of unrelated things.
A current study conducted by a group called They Can Talk claims to be reviewing owner-submitted videos with a "rigorous scientific approach to determine whether, and if so, how and how much non-humans can express themselves in language-like ways." However, as their name suggests, the group's efforts seem biased and void of real scientific "rigor." The study is sponsored by the CEO of Clever Pet, the same company that sells the talking buttons in the videos. Participants are encouraged to buy Fluent Pet products to enroll in the study. In addition, Bunny, the famous dog, is a Fluent Pet "brand influencer," and the owner gets paid for generated sales.
The dogs' speech buttons are considered augmented assistive communication (AAC) devices. These devices are used by individuals with speech or other physical disabilities to understand and communicate using expressive language. The idea is that specific buttons, pictures, or symbols are conditioned to represent phrases, words, and letters without requiring the individual to speak themselves. As we know from decades of research teaching humans to use these devices, they take painstaking, deliberate, and careful work to get right.
Who is Right?
So far, all we have is conjecture. Bunny's owner and many other pet owners are convinced their animals can talk like humans. Skeptics (like me) are convinced there are more plausible alternative explanations for what we see in these viral videos. Watch one of the videos. What do you think?
In this heavily shared video, Bunny appears to say her paw hurts due to a "stranger," later identified as foxtail, in her foot. Bunny initiated "mad," then after a pause and owner-led cueing, pushed the button for "ouch." The video then jumps to a scene where Bunny accidentally presses the "stranger" button when scratching her ear. When asked, "where stranger?" Bunny then hits the button for her paw.
The edited clips and questionable inferences are on display. The owner even notes Bunny has been using "ouch" in different contexts, suggesting she hits this button often for other things. The owner "cueing" Bunny how to respond in this clip is most evident at the end when the owner calls her over and subtly puts out her hand prompting Bunny to give a paw. The video jumps ahead again to show a foxtail pulled out of matted fur on that paw.
To Bunny's owner, her dog communicates her feelings, wants, and needs. However, this video appears to be just a dog hitting conveniently located buttons, then repeating this behavior when rewarded with attention by her owner. This is a textbook example of operant conditioning - behavior that gets rewarded is repeated.
The Science of Talking
Is there a way to really tell if dogs can communicate with human language? First, we need to understand what language is and how we learn it. In behavior analysis, we use the label verbal behavior to describe all forms of communication, including sign language, body language, writing, and speech buttons. Individual words or communicative responses are termed verbal operants.
There are four basic verbal operants: echoics, mands, tacts, and intraverbals. Its function or purpose defines each. Echoics are imitations and have exact point-to-point correspondence. For example, I say "apple," and you repeat "apple." Mands are requests and serve a particular purpose, to get something you want. For example, I might say, "get me a drink," when my friend opens the fridge, and I am thirsty. Tacts are labels and names for things used to convey information to others. If I see a cat and say, "Kitty!" I am tacting a cat. Intraverbals are more complex verbal operants controlled by other language, like a back-and-forth exchange. For instance, if I asked, "What did you have for lunch today?" and you responded, "a sandwich," your response would be classified as an intraverbal.
Why does all the boring jargon matter? Decades of research in teaching language show us that just because we can use one type of verbal operant doesn't mean we can use them all. Specific training is needed to use words for different purposes and contexts. Let's say I teach a child to say "juice" when thirsty (a mand). It doesn't mean they will say "juice" when I hold up a Capri Sun and ask what it is (a tact). For these relations to occur, we need a lot of repetition in specifically contrived scenarios.
There is also a distinct difference between speaker behavior (or expressive language) and listener behavior (or receptive language). If you've tried to learn another language, you’ve likely experienced this. You might be able to understand more than you can speak. Teaching language requires careful exposure to both conditions that place the learner as a speaker and a listener.
Testing a Talking Dog
While memorizing the terms above might not be essential, this information is vital to show a deeper understanding of what happens when we talk. A scientific approach to language may be the only way to answer whether dogs can talk definitively.
Based on empirical research, I've put together some simple tests to put a dog's ability to communicate like a human to the task. I want to challenge any of the owners of "talking dogs" to test for these skills and show us how well your dog does. Remember, no help, cueing, or edits!
The Mand Test:
To get a true mand, motivation is necessary. The most objective way to do this is to set up a deprivation condition and wait until your dog’s meal time is approaching. Assuming the dog has a corresponding speech button, bring out a couple kinds of their favorite treats and leave them in sight but out of reach.
Wait to see if your dog independently presses a button corresponding to one of the treats. Then to confirm, present more than one of the treats and see if the dog takes the treat they requested. If so, you can be fairly sure you have an intentional request. Any owner-cueing would nullify the results.
The Tact Test:
Set up a scenario where you show your dog an object or picture corresponding to a button and ask, "what is it?" Similarly, this test could be done with a person the dog identifies by asking, "who is it?" In either case, the dog should not receive the item after pressing the button. If the dog correctly labels the item or person by pressing the corresponding button, they’ve passed the test.
The Intraverbal Test:
In this test you would ask your dog a question that can be objectively answered by one of the buttons. For instance “what is your name?” or “what do you like to play with?” or “what did you do today?” I think it’s highly unlikely any dog could accurately pass this test without explicit training.
The Listener Test:
In this test, you need to use the buttons to communicate with the dog to assess their understanding. Make a simple request using the buttons to get a specific object (e.g., " get rope") and see if the dog can follow the instruction. We know most dogs can follow simple instructions or obey simple commands, but it’s less likely they will be able to follow instructions delivered with the buttons.
We Don't Need No Stinking Buttons
We tend to underestimate the fantastic communication abilities of non-human animals. Research has already demonstrated that some dogs have an incredible capacity for communication. Still, I suspect most dogs' ability to respond accurately as a listener pales compared to Chaser, the border Collie in the video, who can accurately respond to over 100 toy names! There's also a fascinating study on listener behavior called the Genius Dog Challenge worth looking at if you are one of the lucky few with a good listener!
Dogs don’t need to talk like us to be phenomenal communicators. The beauty of having a dog is connecting in ways deeper than words. My dog, Frida, tells me when she needs to go out, eat, play a game of tug, or get me off my darn computer and pet her belly. I can tell when she's happy, scared, and sleepy…and your dog probably tells you these things too! I prefer Frida to tell me she loves me with her wiggly tail, excited yelps, and kisses when I come home, then have her push an “I love you” button.
Why can't we take the time to learn their ways to be better communication partners for them?
Message to Dog Owners
While these buttons can be entertaining, I don't think they are a good investment of time or money for most dog owners. They might be able to tell us more about a dog's intelligence, but much research is needed. I would urge pet owners to carefully consider why they want to use speech buttons before diving into them. If you're hoping your dog will become a social media star, and you can cash in like Bunny's owners, I wouldn't count on it
Additionally, consider whether it's worth spending the time and energy appropriately teaching your dog to use the speech buttons. Think of all the skills your dog needs to have to live a safe, happy, and healthy life, and ensure you've taught your dog all of those things before you even think about speech buttons.
Dr. Keira Moore is a Board Certified Behavior Analyst specializing in assessing and treating sleep problems. She offers teleconsultation to parents of children with sleep problems as well as adults all over the world. For more information, visit www.moorebehaviorconsulting.com and follow Dr. Moore on Facebook @moorebehaviorconsulting. To see more of her & Frida, check out their training journey @frida_pawlo_does_tricks on Instagram.
