The ability to look into a mirror and recognise oneself is a cognitive skill we all take for granted, but very few animals outside of humans are able to do it. New research shows that monkeys can be trained to pass the so-called “mirror test”, suggesting that more species may be self-aware than previously thought. It’s a fascinating result, but one that shows how far we are from being able to accurate gauge consciousness in another animal.
In a new study published in Proceedings of the National Academy of Sciences, a research team from the Institute of Neuroscience at the Chinese Academy of Sciences claims that, with the right training, it’s possible to get rhesus monkeys to suddenly recognise themselves in a mirror — an ability that’s normally absent in this species.
This suggests that monkeys possess a certain degree of bodily self-awareness, even though they lack the innate ability to spontaneously recognise themselves in a reflection. The new research also points to the inadequacy of the classic mirror test as a measure of self-awareness in certain species, and that self-awareness may be more common in animals than previously assumed.
When scientists talk about self-awareness, they’re referring to the capacity for introspection, along with the ability to recognise oneself as being an individual (“I”) separate from other individuals. Humans believe that other humans are self-aware, even though we can’t really prove it. Given that virtually all humans claim to be self-aware, we have to take a leap of faith and accept it as fact (otherwise we’d be accused of solipsism).
Unfortunately, we can’t be sure about non-human animals. It’s not like they can tell us about their internal mental states. Back in the early 1970s, and in an effort to overcome this limitation, psychologist Gordon Gallup Jr. developed the mirror test, also known as the mirror self-recognition test (MSR), to test for self-awareness in non-human animals. Since the test was introduced, only a short list of animals have passed the MSR, including great apes, elephants, dolphins, magpies, and manta rays. Animals that fail the mirror test include cats, dogs, and even monkeys.
But as the new PNAS study shows, an animal’s inability to pass the MSR doesn’t necessarily mean it lacks self-awareness. In the case of rhesus monkeys, it simply means that these primates lack the ability to identify themselves in another medium—at least not before they’re trained to do so.
At first, the monkeys were trained to touch a laser pointer spot on a block by inferring its position when looking into a mirror. (Image: Liangtang Chang et al., 2017)
On their own, monkeys can’t pass the mirror test. Strangely, they’ve been observed to use mirrors to investigate their surroundings, but they can’t quite figure out who’s looking back. Leveraging their innate skill in understanding the concept of reflections, a research team led by neuroscientists Mu-ming Poo and Neng Gong placed monkeys in front of a mirror and trained them to touch a red laser pointer light spot on a board that could only be seen through the mirror.
Eventually, after several weeks of training, the researchers started migrating the position of the laser light from the training board to the monkey’s face. At that point, the monkeys were able to touch the spot on their face marked by the location in the mirror—something they hadn’t been able to do prior to the training exercises. It was a kind of “a ha” moment for the monkeys.
In tests, monkeys were taught to locate a laser pointer mark on their face by looking at a mirror—a possible sign of self-awareness. (Image: Liangtang Chang et al., 2017)
In other words, the monkeys learned, simply by looking at the mirror, that the red dot on the face looking back at them could be matched to their own. Now, this single observation isn’t enough to prove that the monkeys are self-aware (they may have been responding to the training, and not fully understanding what they were doing) — but the next part of the experiment was more revealing.
After the training exercises were over, the monkeys were able to maintain their newfound skills. Unlike their untrained brethren, the trained monkeys displayed self-directed behaviours when they looked into their mirrors, such as investigating normally unseen parts of their body (dolphins, who pass the mirror test in spades, do the same thing, exploring their bodies and the inside of their mouths).
These observations suggest that rhesus monkeys are self-aware, and that they’re cognitively capable of passing the mirror test—but that something is preventing them from learning this skill on their own. With the visual-somatosensory training (that is, the training the monkeys received to connect the red dots to their faces), it seems a brain connection was made that’s common to other species who pass the MSR. Looking ahead, the researchers would like to study the monkeys further and determine which brain circuits are responsible for forging that connection. As a starting point, the researchers are planning to take a look at mirror neurons, which are an important component of imitation.
“This study in no way ‘debunks’ the validity of the mirror self-recognition test.”
Another possible takeaway from this study is that the mirror test is fundamentally flawed—that it’s a poor method for measuring self-awareness. Indeed, given just how few animals pass the MSR, this test certainly seems to have limitations. That said, neuroscientist Lori Marino, the executive director of the Kimmela Centre for Animal Advocacy and a former faculty member in neuroscience and behavioural biology at Emory University, says we shouldn’t throw the baby out with the bath water.
“This study in no way ‘debunks’ the validity of the mirror self-recognition test,” she told Gizmodo. “The researchers may have uncovered some factors that are responsible for the difference between some species ‘passing’ the test and some species ‘failing’ the test. What they have not done, however, is challenge the validity of the original MSR test.”
Marino, who developed the MSR test for dolphins, says that mirror tests provide a strong and valid test of self-awareness at some level. “Passing the test means there is a level of self-awareness that is similar to other species who pass the test,” she says, “But ‘failing” the test by no means lack of self-awareness.”
Marino says we can only be descriptive at this point (i.e. we can only infer cognitive abilities based on behavioural actions), because we don’t fully understand consciousness and self-awareness. She compares it to blindfolded people touching different parts of the elephant. “We have to take the totality of what we know about each species to make a guess about their level of self-awareness,” she told Gizmodo. “So, for instance, dolphins seem to pass every test thrown at them. So the whole scientific literature on dolphins is consistent with the fact that they have a very robust sense of self.”
Marino says that scientists place a lot of emphasis on the mirror test because it’s one of the few ways we can probe certain aspects of self-awareness in an objective experimental manner. She says we’re also biased, and that we tend to use tests that are specific to human measures of intelligence. Encouragingly, Marino says there may be other ways to test for self-awareness.
“For instance, the studies of metacognition or ‘uncertainty monitoring’ have shown that dolphins and rhesus macaques do as well as undergraduate students on discrimination tasks and show the same pattern of responses indicative of being able to think about one’s ability to answer a question,” she says. “Humans taking an exam know how to be strategic about their time by being sensitive to which questions they find ‘hard’ and which are ‘easy’. Dolphins and macaques can do the same thing.”
In addition, Marino says some studies that require individuals to repeat a previous behaviour or to enact a novel behaviour (one they’ve never done before) are also arguably tests of self-awareness. Studies showing that dolphins can spontaneously imitate another dolphin or a human are especially interesting, she says. Marino says work is finally being done in this area, “but compared with the complexity of self-awareness it is relatively sparse.”
For now, we’ll have to be content with the mirror test—but with the understanding that there’s more to self-awareness than a silly reflection, and that some animals may possess types of self-awareness we’ve never imagined. [PNAS]