Humans believe they understand their dogs. Our research gave us pause.

Washington Post

26-05-2025

Oliver, one of the researcher's dogs, is seen on a black background, as part of an experiment. (Holly Molinaro)
Dog owners don't understand their pet's emotions as well as they think they do.
Clive D.L. Wynne is a professor of psychology and director of the Canine Science Collaboratory at Arizona State University. Holly Molinaro has recently completed her PhD at Arizona State.
Most of us have powerful intuitions about how our dog is feeling — starting with that flag attached to the rear end, the tail. Tail wagging: dog happy. Tail tucked: dog sad or scared. And yet the scientific literature is surprisingly quiet about whether we are actually good at reading a dog's emotions. If people are going to care for dogs, they need to know how their pet is really feeling — so we studied just how well they understand dogs' emotions.
Our work started during the pandemic with one of us, Clive, in Arizona and the other, Holly, in Connecticut. As we struggled to master Zoom, we realized that manipulating video could help us investigate this question.
First, Holly filmed her dog Oliver playing with her father in several situations. Some positive, like giving him a treat …
... and some negative, like showing his nemesis, Saffron.
Holly then edited the videos, so they showed only Oliver against a black backdrop.
The videos were shown to hundreds of people who were asked how Oliver was feeling. A key finding was that people couldn't say how Oliver was feeling without any context.
Holly filmed her (now much-missed) dog Oliver playing with her father, Rich. Some of the time Rich set up situations that would be considered positive; such as playing with Oliver, showing him his leash or giving him a treat. Rich also created negative situations, such as showing Oliver his nemesis in the house, Saffron the cat.
Holly filmed everything, and then, just as Zoom makes it possible to obscure the background, she edited the videos so that viewers only saw Oliver against a black backdrop.
We then showed 400 people these videos and asked them how Oliver was feeling. First, we showed just Oliver on the black background, and then we let people see the same videos with the full context: Oliver, Rich and anything Rich had with him — like a treat or Saffron. No surprise, when given full context, an overwhelming majority of people rated Oliver as happy in positive situations and less happy in negative ones. But in videos without contextual information — no Rich, no leash, no Saffron or anything else — they couldn't tell us how Oliver was feeling.
Story continues below advertisement
Advertisement
This was a shock. Surely people could tell a happy dog from an unhappy one? We delved deeper. Since context seemed so important, what if the context was … wrong?
Holly and her dad (and Oliver!) went back to work. Holly made movies of Rich and Oliver in different scenarios and manipulated some videos to make it appear that Oliver was playing with Rich when the unedited footage was actually of Oliver being reprimanded. In others, Oliver appeared to be responding to a reprimand, when in reality he had been shown his leash which promised a fun walk.
We sent this second survey to 500 people and found that when they saw Rich doing something fun, such as offering Oliver a treat, they responded consistently that Oliver was feeling good, regardless of whether the footage they saw was of Oliver actually reacting to a positive or a negative situation. When people saw Rich doing something a little mean to Oliver, they thought the dog was more sad and anxious, regardless of what Oliver was actually reacting to. Our participants rated how Oliver was feeling based solely on what Rich was doing.
Story continues below advertisement
Advertisement
You might think, 'Okay, well, that's someone else's dog. I surely know my own dog the best.'
Holly showed her dad the edited videos as well. When Rich watched, even he was stumped as to what his dog was really feeling.
'Oh, that video was definitely the one where I showed Oliver some cheese. He loves cheese!'
'Actually, no, Dad. That is the one where he is being shown the cat.'
'Well, then — '
So what is going on here? Are we truly just terrible at understanding if our dog is happy or not?
Can you guess Oliver's emotion?
Happy Sad
The owner is playing with Oliver, asking him to roll over.
Happy Sad
Oliver is being reprimanded, with the owner pointing a finger at him.
Happy Sad
Oliver is facing an enemy – a cat named Saffron.
A pair of studies in Italy a decade ago helps fill out this picture. A team at the Universities of Bari and Trieste put dogs one by one in a wooden box with cameras above them and a window in front. The cameras were trained on the dogs' wagging tails while the researchers presented things to look at through the window. The researchers showed the dogs their owner, an unfamiliar person and an unfamiliar dog. The dogs showed a strong, consistent bias to wag their tails to the right when shown their owner or an unfamiliar human but a left bias toward the unfamiliar dog, indicating that dogs' wagging tails show their emotional state not simply by how much they wag them but also the side of the body they wag their tails toward.
This likely is connected to how the left side of the brain is more specialized for approach and the right side for withdrawal. In the dog these signals cross over on their way from brain to tail, leading to more rightward wagging for something the dog would like to approach and more leftward wagging for something it would rather retreat from. This is a striking finding, because in all the millennia people have been watching dogs and writing about them, nobody had ever noticed that the direction a tail wags makes any difference.
While humans may be blind to this aspect of emotional expression in dogs, our canine friends certainly notice. In a follow-up study, the researchers connected dogs to heart rate monitors and showed them videos of other dogs wagging their tails. If the dogs saw a left-wagging tail, their heart rate revealed they were more anxious than when they watched a right-wagging tail.
People and dogs have been living together for more than 15,000 years. In that time, what have we learned? Our study along with the research from Italy, shows that, despite intense intuitions, people are poor at recognizing the emotional state of dogs. Instead, we look at everything around the dog to guess what our pet must be feeling but fail to look closely at the animal itself.
This might not seem so surprising. After all, we don't have tails to wag, and we don't sniff our friends' backsides to learn how they're feeling. But it's crucial to the success of our lives together because the world we share with our dogs has changed dramatically over recent decades. Our dogs no longer live in kennels in the backyard, as their great-grandparents did. More than three-quarters of dogs in America today curl up each night in bed with people who consider them family members. Highly trained hounds console patients in hospitals, and there are even churches that involve dogs as part of their ministry. This increased intimacy requires us to accurately gauge our dogs' moods.
Story continues below advertisement
Advertisement
Meanwhile, at the other end of the leash, several studies have shown that dogs are remarkably good at recognizing human emotional expressions. They can tell what emotion a human face is showing or respond with empathetic concern to a weeping person. Where our comprehension of dogs' emotions is so weak, their understanding of us is remarkably strong.
We need to confront our biases and be more modest in our assessment of canine emotions. We have to recognize that it isn't easy to know how a dog is feeling, but with careful attention to each individual dog we might be able to learn what their happiness looks like.
Post Opinions wants to know: How did your relationship with your dog evolve over time? Share your responses and they might be published as letters to the editor.