clock menu more-arrow no yes mobile
frankieleon / Flickr

Do animals feel empathy? Inside the decades-long quest for an answer.

Throughout her career as a neurobiologist, Peggy Mason has been told over and over that the rats she experiments on are not capable of empathy. Only humans and other primates can understand the emotions of another. Most other animals can't. And certainly not beady-eyed rats.

But what she was witnessing in the lab was telling her something very different. In experiments, Mason and her colleagues at the University of Chicago were finding that when one rat was placed near another jailed rat, the free rat would open the hatch for a rescue — something it wouldn't do for a toy rat. What's more, when given the choice between saving a fellow rat and some delicious chocolate, the free rat would open both cells and then share.

The study, published in Science in 2011, was a breakthrough. If rats were capable of basic forms of empathy, then perhaps empathy was common — or even universal — among mammals. Studying animal empathy could give us insight into how human empathy evolved. ("I consider myself just a fancy rat," Mason told me.)

But almost immediately, Mason's results met intense skepticism.

Alex Kacelnik, a behavioral ecologist, argued that Mason was simply projecting humanlike feelings and emotions onto these rat "rescues" — a tendency known as anthropomorphism.

"We don’t have evidence that there is an internal first-person experience that leads the animal to do it," Kacelnik tells me on a Skype call from his office in Oxford. "Do they experience any emotion when helping a partner? It may as well be, but we don't know."

In a response to Mason's study, Kacelnik and his colleagues wrote that tiny ants can also appear as if they are "rescuing" other ants in some situations. But ants are nearly brainless, and few consider them capable of empathy.

Alan Silberberg, a psychologist at American University, also wondered if Mason was inferring too much from the data. In his view, the free rat opened the cage for selfish reasons: It liked playing with the other rat. He published a paper replicating Mason’s design, but with a twist. He showed when the rats couldn’t play with each other after the trap opened, the free rat wasn’t as interested in conducting the jailbreak.

These criticisms don't mean Mason’s findings are incorrect. But they do illustrate a core difficulty in studying animal empathy: While it's easy to observe animal behavior, it's near impossible to confirm the motivations behind that behavior.

The quest to understand animal empathy has been long and fraught, but it isn't trivial. If animals can indeed feel emotions like our own, the revelation may one day lead to treatments for conditions where social bonding is difficult — like in autism — or is nonexistent, like in sociopathy.

We should care whether animals have empathy. It would mean animal brains are not that different from our own.


Nuwandalice / Flickr

For decades, scientists trying to study animal empathy have run into a simple fundamental problem. "There are so many possible interpretations of [empathetic] behavior that all need to be eliminated if you're going to say they're responding because of the emotions of another animal," James Burkett, a neuroscience researcher at Emory, explains. After all, psychologists can't put a rat on a couch and ask it about its feelings.

Early experiments to prove that animals have empathy were all confounded by this limitation. In 1959, Russell Church, who still conducts neuroscience research at Brown, showed that rats conditioned to press a lever for food will stop when another rat is shocked in an adjacent cage. He argued this was a sign of anxiety, and a possible sign of empathy.

Then in 1962, George Rice and Priscilla Gainer at Agnes Scott College went a step further, arguing that rats would try to save other rats in trouble. In their experiment, rats were placed in two adjoining see-through cells. One of the rats was put into a harness and hoisted above ground — an unpleasant experience that caused wailing. The free rat in the other cell could then push a lever to lower the harness. Rats would push the lever to "free" their fellow rats, but wouldn't push the lever if a block of Styrofoam were being hoisted.

Rice and Gainer concluded that "this behavior might be homologous to altruism."

But other psychologists at the time weren't convinced by these conclusions. They easily found alternative explanations for the rat behavior.


A rebuttal article in Science argued that Rice and Gainer's rats were simply responding to the loud noise of the other rat's screams. These critics showed that rats would use the lever to free anything making a loud noise — not just their fellow rats. This wasn't empathy, their argument went; this was wanting their fellow rats to shut up.

Undaunted, other scientists kept searching for animal empathy. A few years later, in 1964, scientists found that "rhesus monkeys will consistently suffer hunger rather than secure food at the expense of electroshock to a [another]." Jules Masserman, who once served as a president of the American Psychiatric Association, also noted that rhesus monkeys were more likely to go hungry for those they were familiar with.

Most psychologists, however, refused to believe that animal empathy was possible. The field was dominated by behaviorism — which held that animals don't have observable mental states, just actions. Behavior could be understood in terms of rewards, punishment, competition, and consequences — but not emotion. Whenever animals displayed behavior that looked like empathy, behavioralists could easily attribute it to stimuli response.

And for a long time, the skeptics prevailed. Scientists simply didn't have the tools to demonstrate that animal empathy was real. That is, until Jeffrey Mogil, a geneticist at McGill University, noticed his mice behaving oddly in the early 2000s.


La Tarte au Citron / Flickr

Though he hurts mice for a living, Mogil says it's not for a lack of his own empathy. His research is about understanding why two people may experience pain in different ways, with the hope of developing better pharmaceuticals for managing it. "My sympathy is mostly reserved for chronic pain patients," Mogil tells me.

For decades, Mogil and his team have conducted experiments testing the pain threshold of mice in order to see what role genetics play in pain tolerance.

In one experiment, the researchers would take a mouse from a cage, restrain it, and then dip its tail in a hot 120°F bath. They wanted to see how long it would take for the mouse to yank its tail out — a measure of how well the animal withstood pain.

In the early 2000s, Mogil and his team began analyzing data from thousands of tail-dip trials. And they happened upon a bizarre pattern. The order in which the mice were tested in each trial seemed to matter a lot for pain tolerance. That is, on average, the first mouse taken from a communal cage and dipped in hot water experienced the least amount of discomfort. The second one felt a little more. And so on.

That "was really, really surprising to us," Mogil tells me. It was as if the first mouse had somehow told its cage mates, "That really hurt." And upon hearing that, the other mice winced.

Mogil's was the first evidence that animals other than primates could experience "emotion contagion." This is considered by many to be the most basic form of empathy. When I see a broken arm, I wince. When one mouse is in pain, others feel it.

Mogil staged several follow-up experiments to solidify his hunch. First he repeated the tail-in-water test, but instead of putting the mice back into the same cage, he put them into a different one. When that happened, he says, "this effect goes away completely." Another test showed that mice feel greater pain when they are around other mice in pain. This is especially true if the mice know one another (i.e., they were cage mates beforehand).

That was just the beginning. Mogil later showed that mice were capable of consolation — another sign of empathy. In one experiment, researchers "jailed" certain mice and injected them with vinegar, a painful experience. Other mice were then placed in the "jail block" and were free to visit the imprisoned mice or ignore them completely. When the mice were familiar with one another, the free mouse would visit with the caged ones.

"If you think about that for a second, that's crazy," Mogil says. It doesn't make any sense for one mouse to approach another that is in pain — after all, that pained mouse could be dangerous. "The only reason you would do it is if you were trying to help."



Zach Johnson / Science

Mogil's work inspired Mason and her team, who took it a step further. Mason's study was evidence of "pro-social behavior": a form of empathy where animals engage in actions that benefit the group rather than an individual.

That said, there do seem to be limits to animal empathy. The highest level of empathy is called cognitive empathy. This is the ability to think through feelings and weigh options: "Emily seems upset right now, does she want me to ask what's wrong?" No researcher I spoke with for this article would go so far as to argue that animals have cognitive empathy. That's distinctly human.

But scientists have made extensive progress studying the lower, gut-level forms of empathy, in order to understand how this capacity might have evolved in animals. That, in turn, might shed some insight into how the more advanced forms of empathy seen in primates and then humans developed.

"We should think of ourselves as a part of a continuum," says Larry Young, a neuroscientist at Emory University. "These animals have some basic fundamental underlying neural mechanism that cause them to engage in a behavior similar to what we do."

Young and his colleagues recently published a paper in Science that took an important next step toward proving that the small acts of empathy displayed by rodents are, in fact, like our own.

The study in question involved prairie voles, one of the few species of mammals that mate for life. Young found that prairie vole companions will look out for one another much like human spouses do. When one vole is scared, its counterpart will come over to comfort it by licking. They won't do this for strangers. The vole mates are also emotionally in sync: Their stress hormones will rise and fall with one another when one is shocked and the other is not. (Experiments have shown that this behavior seems to be inborn, not conditioned.)

More importantly, Young and his team discovered that these bonds are facilitated by oxytocin, a neurochemical found in all mammals that is believed to help facilitate social interactions between humans.

"Oxytocin is what makes the brain pay attention to the social cues of the partner," Young explains. In humans, variations on the gene for oxytocin have been shown to predict social problems in childhood and relationship stability later in life. Studies have shown that nasal injections of oxytocin may help autistic people — who often have trouble interacting with others — maintain eye contact.

When Young blocked the oxytocin receptors of prairie voles, they stopped caring for one another.

"We share the neurocircuitry of this with animals," Young says about his results. "We can come to the conclusion that what we have is not necessarily so unique."


Hannah Knutson / Flickr

The wrong conclusion to take away from all these studies is that these rodents are humanlike. The right conclusion is that we're animal-like.

Frans de Waal is one of the world's leading primate behavior researchers. Since the 1970s, he's made thousands of observations of primate communities. He's shown that many primates will console one another after fights. He's seen them hug and kiss. In 2010, he co-authored a paper in the Proceedings of the National Academy of Science compiling data from more than 3,000 observations of chimpanzee fights. The paper found that chimps will commonly console the losers of fights a behavior especially pronounced among chimps with kinship bonds.

De Waal thinks it's wrongheaded for some scientists to dismiss observations of empathy in animals. After all, it makes sense from an evolutionary perspective. If human empathy is so robust and adaptive, it must have evolved from more primitive forms.

"It is hard to imagine that empathy — a characteristic so basic to the human species — came into existence only when our lineage split off from that of the apes," says de Waal. "It must be far older than that."

If human and animal empathy are the same, it means lessons learned from the brains of animals can be applied to heal our own.

Young is hopeful his continued work on prairie voles will yield animal models for psychiatric drugs. With a good animal model he could, in theory, test whether a drug would make the prairie voles more or less likely to comfort one another. He could test to see if there are certain genes responsible for empathetic behavior, and target those for intervention.

"We know that there are many psychiatric disorders where the ability to empathize is deficient," Burkett says, "but we have no treatment for those deficiencies."

Perhaps the fluffy, heartsick prairie vole can open that door.

We still don't know how many animals can actually empathize. But the mounting studies suggest there's a fundamental baseline, at least among mammals. "When we're talking about the ability to sense the emotion of others, and to respond to them in some way, that's probably very widespread," Burkett says.

The capacities may vary depending on how animal societies are structured. Chimps and primates are thought to have the most humanlike emotional capacities. Dogs have evolved to live in packs and are attuned to our emotional needs. Cats, evolved to be solitary hunters, care much less.

Many of the researchers I spoke to say those who discount animals' ability to empathize are making empathy out to be more complicated than it actually is. In Silberberg's view, for instance, empathy requires that an individual knowingly act at a cost to itself to aid another.

But what difference does it make if chimps know why they're helping out a friend? Isn't it enough that they engage in the behavior, just like we do?

"Let's say you are living in a house with children, and you are crying, and your children approach you and touch you," de Waal says. "You're not, at that moment, thinking, 'Are they now altruistic or selfish when they do this?' That would be a silly distinction. Because your children are responding to your emotions."

That's all that really matters.


Editor: Brad Plumer


Down to Earth

Otters are thriving in … Iowa?

Science & Health

A pregnancy turns deadly in an anti-abortion state. What happens next?

Health Care

How Dobbs is affecting abortion care, one week on

View all stories in Science & Health