• How Can You Know What Your Dog
    is Really Feeling?

    Depressed, Confused, Excited, Surprised... and We're Not Listening

    Every day watching dogs I see emotion in them. At the lab, many of the scenes we create for them are inadvertent emotion provocation. I see curiosity directed at a small robotic “dog” toy that dances and plays a tune. I see surprise when a hidden person appears from behind a door. Dogs may feel anxious when I open an umbrella; disgusted when they sniff a very strong smell; delighted when their owner stops listening to me and turns to pet them again.

    When I watch dogs in the “wild”—out in parks and on sidewalks among people and other dogs—I see regular displays of joy, interest, and affection; of apprehension and fear.

    Still, one of the questions I am most often asked is whether dogs really love us, feel bored, and get angry: a testament to both the ardor of our interest in our dogs, and our uncertainty about the dog’s experience. As our own days may be colored with anxiety, anticipation, or foreboding—are dogs’ days so colored? As we respond to events and people with empathy, sarcasm, or incredulity—do dogs tend toward such sentiments?

    Many of these questions boil down to whether dogs have feelings or emotions at all. But of course they do. Look at it adaptively: emotions are messaging to the muscles and response system to circumvent the closed-door discussions between the sensory organs and brain. I see a tiger; I know that tigers are predators and this one is coming toward me . . . and Hey!, chimes the brain emotively, Be afraid! Run!

    Look at it neurologically: the areas of human brains that are active when we feel, sigh, yearn, and despair are also found in dogs’ brains.

    Look at it behaviorally: though we are not always great at naming which behavior indicates what emotion (as we will shortly see), the wide array of different behaviors and postures of dogs tells us about their internal states.

    Look at it sensibly. The alternative to having emotions—having undifferentiated experience—defies reason, defies Darwin, defies continuity. Human emotions did not emerge mysteriously and fully formed out of unfeeling automata. Keep in mind that the last popular advocate of the latter belief, Descartes, lived in a time when bloodletting was still considered salubrious.

    Martin Seligman’s research on depression’s “learned helplessness” model was first tested on dogs.

    My own dogs, subject to my near-continuous gaze, appear to be great furry balls of emotion, sentiment, and expression: anticipation at a walk, disappointment at being left at home, grumpiness at a friendly cat’s attention. I naturally see Finnegan’s hoisting of an improbably large stick out of a river as pride; the dour look he gives me as I allow the cat to curl on my lap as jealousy; his look when discovered later sneaking mouthfuls of the cat’s food as guilt. There is a coyness in Upton covering his face with a paw; amusement in his self-invented game of mimicking the sounds of trumpet practice; embarrassment at his own ghost-thrusting hips long after his playmate is gone.

    As shorthand, it makes good sense to me to use emotional terms to describe what I’m seeing. In the lab, I would more likely say, The dog’s head extends forward, leading the body by an extra half-step; the ears are perked into their full height (read: curiosity). A dog jumps back, preparing the body for escape; a “rurf” sound slips out (surprise). Retreating, the dog’s body shrinks down and back (anxiety); on approach, a dog pulls away her head, lifts her paw, curls her lip (disgust); with a high, loosely wagging tail, the dog leaps with two or four legs and attempts to lick every nearby face, dog or human (delight).

    I don’t use those shorthand words as my first descriptions of what they are doing—because I hesitate to assume that a dog’s experience of what looks like curiosity or delight is precisely like mine. While the similarities across mammalian brains make it highly likely that all mammals have diverse emotional experiences, we all also have very different lived experiences, based on, for humans, our cultures, where we live, and the people we meet. So, too, for dogs. My own guess is that, planted into a dog’s body, we wouldn’t recognize the feelings we’re flooded with as being just like our own. But that there are feelings, I’ve no doubt.

    In this way, I inhabit the territory between the presumptive granting of subjective experience just like humans—and complete denial of any experience. Not presuming to know the dog’s subjective experience is not at all the same as denying them any experience at all. In fact, though, that denial has been the default model in much of science. Without definitive evidence of an animal’s fear of pain, researchers say, how can we be sure that the animal feels fear—or pain—at all?

    Weirdly, most of the history of medical and psychiatric research has also seemed not to doubt the reality of animals’ feelings. In fact, it presumes feelings in its very premise. To prove the efficacy of an anti-anxiety drug for humans, the drug first has to be roundly vetted on an “animal model”: essentially, lab animals have to be made anxious, then given the test, and have their anxiety dissipate (while no other ill effects arise). A history of this kind of thinking is written between the lines of every medical study using animals: they are so similar to us, thus they are a good model for humans.

    Dogs—the same dogs who express alarm at a deflating balloon loping down the sidewalk; the same dogs who, having lived in your home for even only one day, gleefully greet you at the front door—have not been immune from this manner of research.

    Should someone make the claim to me that a dog definitely can’t be “depressed,” or benefit from anti-depression medication, I’ll take their hand and walk them back in time. Several decades ago, depression research took a step forward with the development of the “learned helplessness” model, made famous by Martin Seligman. He and his colleague came up with a scheme to see if helplessness could be induced by circumstance. Brace yourself: it involved dogs.

    To watch struggling animals without working to relieve their struggle demonstrates the great dissociation we condone with animals.

    I was born at the oldest hospital in Philadelphia, about a mile from the mid-century building where Seligman’s experiment was likely conducted two years before. I met my first dog, an Irish setter named Trevor—long-haired, graceful, more fluid in movement than my toddler self—at my grandparents’ house a few miles to the north. Twenty years on, in the fall, I walked the same paths at Penn that Seligman did, with leaves flamboyantly littering the ground and the air scented with the new season.

    His dogs never smelled that air. They lived in the laboratory, the 32 “adult mongrel dogs” who were his first compulsory subjects. I don’t know where they came from, but if they were mongrels they might have come from the city’s shelters. Like the shelter that, two years after I arrived at Penn, I walked out of with my first mongrel dog, Pumpernickel, whose soft black form and gentle pacing as we exited the building into the sunlight I still remember. Pump walked those paths and frolicked in those leaves.

    One day, these 32 adult mongrel dogs were strapped into rubber harnesses in a small cubicle, with their legs hanging through leg holes. Their heads were yoked in a fixed position by panels restricting head and neck movement. A 70 dB noise—about the loudness of a nearby vacuum cleaner—was played in the box. Seligman, or his assistant, taped brass plate electrodes to their hind feet. Then they sent from 64 to 640 electric shocks through the plates.

    The strength of the shocks, six milliamps, is described in the literature with adult humans as “painful”; “muscular control is lost” when experienced for one second. The dogs felt these shocks from five seconds to two minutes each time, dozens or hundreds of times that day. When other control dogs felt what the study describes as this “severe, pulsating shock,” the researchers noted that they “barked, yelped, ran, and jumped until (they) escaped.”

    In some conditions, the dogs could stop the shocks by pushing their head against the panels, if they discovered this in their struggle; in another group, there was no escaping the shock. It went on and on, with no sign of ending, and despite their attempts to move, and despite their cries. Finally it did abruptly end. The next day both groups were put in a different cage with a metal grid on the floor and an obstacle separating their cage from an adjoining one. The grid was electrified. Dogs who had learned to stop the shock the previous day quickly jumped over the obstacle and escaped the shock. Those who were previously exposed to the inescapable shock sat passively, unmoving, unescaping. This is what excited the researchers. These dogs had learned that they were powerless: what the researchers called “learned helplessness.”

    So, dogs were shocked, driven to depression and passivity and impotence, to prove that we could feel passivity and impotent in depression. Dogs are still widely used in medical research, make no mistake: this is happening now. Also now. And again.

    For learned helplessness studies, though, subsequent researchers found it hard to stomach using dogs, and rodents have taken their place. You might not feel, on first pass, that the fact of a mouse going through the study is as dramatically upsetting as the dogs’ fate. I’d hazard if you knew any mice for more than a few hours you might feel differently. Or maybe if you heard about the current test of choice, a hugely common test called the “forced swim” test—which is just as it sounds. The test is also officially called the “despair” test.

    It is, as one paper writes, “probably the most widely used screening test of antidepressant potential of novel compounds”: a good way to test if antidepressants work at “rendering or preventing depressive-like states” in rodents. A mouse (or a rat) is placed in a tank or bucket filled with water, from which they cannot escape. Researchers watch them, for many minutes. The amount of what is described as the animals’ “struggling” is measured. After a while, the mice lose their wills, they lose energy, and they become passive. Their feet still stir the water, but their heads are barely above the surface, just holding on to life. But, hey, tested antidepressants “reduce the immobility time,” that is, cause the mouse to keep struggling.

    As wrongheaded as it is to presume dogs to be unfeeling, it is no more correct to presumptively grant them a humanlike emotional life.

    To watch struggling animals without working to relieve their struggle demonstrates the great dissociation we condone with animals. Our society’s attitude toward animals is thus mismatched. We grant them feelings when it suits our testing needs, but grant them no feelings when it would not suit our testing needs. The human behavior in these test settings—electrocuting; near-drowning—is considered animal cruelty anywhere outside of the test setting.

    So why is the question of animal emotions still posed? We are trapped on the far reaches of the pendulum’s swing: either assuming dogs are entirely unlike us or assuming dogs are just like us. As wrongheaded as it is to presume dogs to be unfeeling, it is no more correct to presumptively grant them a humanlike emotional life. (Nor must it be somewhere in-between: for all we know, dogs’ emotional experience is far more elaborate than ours.) We glance at dogs and conclude we know what they’re feeling, but our haste to make such conjecture on little evidence—and inability to read a dog’s emotions when they are displayed—is profound.

    There is little place better to see this than in the movies. Dogs are in films not because they are great actors but because they are part of our lives. The actual dogs who trot through most films are methodically managed, like everything that appears on screen. They are situated as companions, caring about what happens. But their body language easily betrays their indifference. If you zoom in on the dogs at the side of the shot, they are often doing something quite unlike what the scene requires.

    Dorothy arrives in Oz with Toto, a cairn terrier. As she gazes around her at the colorful, fantastic new world, a radiant bubble appears in the sky: Glinda the Good Witch is about to make her appearance. Dorothy is struck with apprehension and amazement, staring at the approaching bubble. Indeed, even after a trip in a tornado, one would be surprised at this new weather phenomenon. But look at the little dog at her feet. Toto, apparently not sharing her anticipation, is the picture of insouciance. She gives a small shake, turns away, and casually exits the scene.

    “What is happening” for the dog (for Terry, as she was known) is defined by the presence of a trainer off-screen. A keen-eyed movie viewer can readily spot the dog’s participation in a parallel experience, their attention directed toward an unseen presence. They are, of course, unwitting actors: their performance is only training to “act” in a certain way on cue. What film directors know, though, is that the audience’s willingness to read, say, a dog’s covering her eyes with a paw (a behavior not particularly common in the species) as modesty or apprehension will trump our interest in seeing what the dog is really doing (performing a behavior for a reward).

    The dog is acting—for the trainer; the actors are acting—for the director; and we are acting—like humans, who will suspend not just disbelief but all common sense. Movie dogs are meant to be shy of their nakedness, to be covetous, to be uncertain—in short, they are meant to be quadrupedal versions of us. They are not meant to be honest-to-goodness dogs. Movies with talking dogs are beyond the pale: dogs nearly stop being canid at all. While they are still dog-shaped, and do a handful of banal doglike things (bark, sniff butts, scratch an ear), they are just furry mannequins on which to drape our own concerns and feelings. We willfully disregard what a dog’s posture or expression might indicate about what they are actually saying or feeling.


    Our Dogs, Ourselves

    From Our Dogs, Ourselves by Alexandra Horowitz. Featured with permission of the publisher, Scribner. Copyright © 2019 by Alexandra Horowitz.

    Alexandra Horowitz
    Alexandra Horowitz
    Alexandra Horowitz is the author of the #1 New York Times bestseller Inside of a Dog: What Dogs See, Smell, and Know and On Looking: A Walker’s Guide to the Art of Observation. She teaches at Barnard College, where she runs the Dog Cognition Lab. She lives with her family and two large, highly sniffy dogs in New York City. Most recently, she is the author of Our Dogs, Ourselves.

    More Story
    The Woman Who Beat the Nazis in Europe's Deadliest Horse Race Amidst growing enthusiasm for horse racing among Czechoslovak patriots, President Tomáš Masaryk himself was persuaded to visit...
  • Become a Lit Hub Supporting Member: Because Books Matter

    For the past decade, Literary Hub has brought you the best of the book world for free—no paywall. But our future relies on you. In return for a donation, you’ll get an ad-free reading experience, exclusive editors’ picks, book giveaways, and our coveted Joan Didion Lit Hub tote bag. Most importantly, you’ll keep independent book coverage alive and thriving on the internet.