A couple of weeks ago, an editor at The Guardian tweeted an image of a bald eagle staring at its reflection in a body of water. “This photo of an eagle taking a hard look at itself is not a metaphor for anything that's been in the news recently,” he wrote.

At the time of this writing, the image has been retweeted 62,000 times.

And it prompted one of my colleagues at The Atlantic to ask: “Are eagles intelligent enough to recognize their own reflections?”


In March 1838, a young and little-known biologist named Charles Darwin asked the same question. On a visit to London Zoo, he stepped into a cage with an orangutan named Jenny, and marveled as she played with a mirror. He noted that she was “astonished beyond measure” at the glass. She examined it, kissed it, made faces at it, and contorted her body as she approached it. What did she see in the mirror? Did she recognize herself? And perhaps most importantly, how could you even tell?

Psychologist Gordon Gallup Jr. came up with a way, over a century later. In 1970, he got four captive chimps accustomed to a mirror, and anesthetized them. While they were out, he dabbed red dye on their eyebrows. When they came to and caught sight of their reflections, they did exactly what humans would do—they stared at their faces and touched their own eyebrows. Monkeys, by contrast, made no moves to examine their own red-marked faces. They couldn’t recognize themselves in the mirror, Gallup concluded. But chimps could.

“Recognition of one’s own reflection,” he wrote, “would seem to require a rather advanced form of intellect… Moreover, insofar as self-recognition of one’s mirror image implies a concept of self, these data would seem to qualify as the first experimental demonstration of a self-concept in a subhuman form.”

Gallup’s mirror test is one of the most famous and controversial techniques in the study of animal intelligence. It has been administered to dozens of species, in much the same way. If an animal inspects the mark on its body (and not on the reflection), and if it interacts with that area more often than usual, it passes the test. At the very least, that’s a sign of self-recognition. It suggests that the animal has a sense of self—a mental representation of its own body or appearance. Perhaps, as Gallup has repeatedly argued, it indicates self-awareness.

Some animal species perform inconsistently on the test. Chimps and orang-utans have unambiguously passed, but they don’t always. An Asian elephant named Happy passed in 2006 after staring into a really big mirror, and examining a mark on her head with her trunk—but two other elephants failed.

Killer whales and bottlenose dolphins have allegedly passed; they’re obviously less dextrous than apes or elephants, but they seemed to go to special lengths to inspect marked body parts by twisting, turning, and stretching. Two captive manta rays reportedly did something similar, but as with many of these studies, it’s hard to interpret the creatures’ behavior.

That’s especially true for the many species that have failed the mirror test—which is to say, most of them. Pandas failed. Dogs failed. Sea-lions failed. They’ll ignore, fight, or play with their reflections, but won’t act as if they recognize themselves. Gorillas have been inconsistent. For a few decades, they were thought to be the only great apes that failed the mirror test; a few individuals have since passed, but all had lived in enriched environments with extensive human contact.

Monkeys also had a terrible track record, beginning with Gallup’s seminal study and continuing for several decades after (including a prominent case of fraudulent monkey business). But in 2010, Abigail Rajala from the University of Wisconsin-Madison noticed that some of her lab monkeys, which had been fitted with head implants as part of an unrelated experiment, were checking themselves out in a mirror. Those same monkeys had previously failed a mark test, but now, they were investigating their weird skull adornments. Later, they even started examining unseen parts of their bodies, like their genitals (as in the video below). “We cannot objectively claim that these animals are self-aware, all the pieces are there to suggest that, in some form, they are,” Rajala and her colleagues wrote.

Gallup wasn’t convinced. He argued that even when monkeys fail the mark tests, they can use mirrors to locate hidden objects. They understand reflections, even if they can’t recognize themselves. So given that they had an implant stuck to their heads, they may just have been trying to examine the strange heavy thing that they could feel but not see. As ever, interpreting the mirror test isn’t easy.

The vast majority of birds have failed, including the omnipresent pigeon, and the linguistically capable African grey parrot. That losing streak broke in 2008, when Helmut Prior from Goethe University showed that two European magpies named Goldie and Gerti would try to wipe off yellow dots that had been painted on their necks. It was the first time that any non-mammal had passed the mirror test—and it made sense that it would be magpies. They are corvids—part of a group that includes ravens, crows, and jays. Known for their intelligence, they’ve been described as “feathered apes.” If any bird would pass the mirror test, safe money would go on a corvid.

But #notallcorvids. Jackdaws have failed, as have jungle crows. Surprisingly, so have New Caledonian crows, whose accomplished tool-using skills have given them a reputation as a corvid’s corvid. Even though they could use mirrors to find hidden food, when confronted with their own reflections, they displayed, attacked, or searched behind the mirror, presumably to find the “other” bird.

Among these failures from the intelligentsia of the animal kingdom, perhaps none have been more surprising than humans. Many psychologists had assumed that the vast majority of children pass the mirror test between 18 and 24 months of age. But as with most such studies, people had only worked in Western countries. When Tanya Broesch from Simon Fraser University went further afield, she found that only 2 out of 82 Kenyan children passed the mirror test. Mirrors aren’t as omnipresent a feature in Kenyan homes as they are in American ones, but they still exist, and they’re used regularly. And yet, most of the children—even some as old as 6—just stared “at their image in the mirror, without any attempt at either touching or removing the mark on their forehead.”

Broesch and others have found similar evidence for children in other non-Western nations, like Fiji, Peru, and Zambia. If these people, who are plainly self-aware, are failing the mirror test, then what is the test actually testing? For example, Broesch suggested that the Kenyan kids understood that they were reflected in the mirror, but didn’t know what to do about that. Unlike Western children, who are encouraged to learn by active participation, children in small, rural communities tend to learn mainly by observing and imitating. Confronted by their marked reflections, they “might be reluctant to either touch or remove the mark, assuming that it was surreptitiously placed on him or her by an adult for a ‘purpose.’”

You could offer similar explanations for other species that have failed the mirror test. Some gorillas have failed, but making direct eye contact is an act of aggression in gorilla society. Dogs fail, but dogs live in a world where smell is more important than vision. Two out of three Asian elephants failed, but elephants are large animals that, as Joshua Plotnik told Scientific American, are “used to putting things on, not taking things off of their bodies, like mud and dirt.”

These animals, upon seeing a mark in the mirror, might not perceive the mark at all, might not know what to make of it, might not care about it. What the mirror test tells us is that chimps and orangutans—dextrous, curious, visually oriented, grooming-obsessed species—react to marked reflections in a way that Western scientists can empathize with and can easily interpret. Maybe it says more about the people administering the test than the people (or animals) being tested.

The binary nature of the mirror test—you pass or fail—is also a problem, because it “presupposes self-recognition exists in entirety or not at all,” wrote Debbie Kelly. It’s possible for a species to sort-of-pass. Take Clark’s nutcrackers—a kind of small, black-and-white corvid.  In the wild, these birds bury nuts, but they’ll restrain themselves if they know they’re watched by another nutcracker. That’s what happened last year, when Kelly placed them in front of a clear mirror: they treated the reflection as an onlooker and potential thief, and refrained from burying.

But if Kelly gave them a frosted and blurry mirror, they buried freely. That weird difference persisted when Kelly administered a traditional mark test. The nutcrackers failed when confronted with a regular mirror, but were more likely to pass with a blurry one.

Credit: Clary & Kelly.

She argues that when looking at the blurry mirror, the birds can recognize themselves by tuning into cues of motion and shadow. But the clear mirror provides a kind of sensory overload, bombarding the birds with unfamiliar features about their own bodies. After all, such surfaces are rare in nature. The kinds of reflective surfaces that nutcrackers encounter are more like the blurry mirror: the surfaces of lakes, streams, or puddles.

Which brings us back to the eagle.

Are eagles intelligent enough to recognize their own reflections? We have no earthly idea. To paraphrase Frans de Waal, we’re not quite smart enough to know how smart animals are, and we need better tests of self-awareness and self-recognition. The eagle might be gazing at its own recognized face, or at what it perceives as a rival. When I initially answered my colleague, I said that it’s most likely looking for something to hunt and devour—which is also not a metaphor for anything that’s been in the news.

That was a joke, but it also indicative of the entire debate about the mirror test. It’s a bit about what animals see reflected in the mirror. But it’s also about what we see of ourselves reflected in animals.