Animal Reflections: Do Mirror Images Mirror Self Awareness?

What do self-directed movements guided by a mirror image really mean?

Posted Sep 18, 2018

What are nonhumans reflecting about when they see their own reflections? What do they know about themselves?

" light of Darwin’s principle of evolutionary continuity, we urge abandonment of the anthropocentric view that only big-brained creatures, such as great apes, elephants and cetaceans have sufficient mental capacities for the most complex degree of self-cognizance: self-consciousness. We hope the current conventional wisdom that only a few species are self-conscious will become a historical curiosity and that, in its place, will arise an empirical understanding of where the minds of various social vertebrates and invertebrates lie on a continuum of self-cognizance." (Marc Bekoff and Paul Sherman, "Reflections on animal selves." P. 179)

A recent essay entitled "The Why of Me," written by University of Oxford graduate student Sofia Deleniv in New Scientist magazine, is a must read for people interested in what nonhuman animals (animals) might know about themselves and how self-aware they really are. The online version of Ms. Deleniv's insightful essay is called "The ‘Me’ Illusion: How Your Brain Conjures Up Your Sense of Self" and is only available to subscribers, so below I've included a few snippets to whet your appetite for more. An introductory leader titled "We’re Not Unique – Lots of Species Can Recognize Themselves" (also in New Scientist)  lays out some of her thoughtful arguments. The subtitle for the leader reads, "We should be open to the idea that human intelligence isn't as special as we like to think it is," a point with which I wholeheartedly agree. While we are unique and exceptional in various ways, so too are other animals

I've been thinking about the general topic of self-awareness in animals for a long time, and Deleniv's piece motivated me to revisit a number of issues centering on how we study self-awareness in nonhumans, what different sorts of data mean, the robustness of data collected in what's called "the mirror test," and arguments about why a sense of self has clearly evolved in diverse nonhumans. A recent discovery that a fish called the cleaner wrasse has passed the "mirror test," one of the standard ways in which researchers study self-awareness in other animals, shows that we need to keep an open mind about the taxonomic distribution of different sorts of self-cognizance (for more discussion of the fish study, please see "Is This Fish Self-Aware?" And, for more on the general topic of self-awareness in other animals, please see the essay written by Paul Sherman I, "Reflections on Animal Selves").

For the purpose of this essay, I'm using the phrase "self-awareness" to refer to "self-recognition" or "self-consciousness" to make the discussion less cumbersome. Dr. Sherman and I used the phrase "self cognizance" to call attention to the different perspective for which we were arguing. In essence, we used "self-cognizance' as an umbrella term to cover the continuum from 'self-referencing' to 'self-awareness' to 'self-consciousness.'" We also considered the limitations of the mirror test. Basically, when researchers use the mirror test, they "place a visual marking on an animal’s body, usually with scentless paints, dyes, or stickers. They then observe what happens when the marked animal is placed in front of a mirror. The researchers compare the animal’s reaction to other times when the animal saw itself in the mirror without any markings on its body."

Concerning the mirror test, in the leader to Ms. Deleniv's essay we read, "Admittedly, the mirror test is a questionable way of probing the minds of other animals. But the finding does fit with an emerging idea that the ability to recognise oneself is more related to an animal’s lifestyle than to its brain size." Others (including myself) have been writing about the limitations of the mirror test for some time, and in an email to me (September 10, 2018), renowned conservationist and writer Dr. Carl Safina wrote, "The mirror test does not test for self-awareness. It tests for an understanding of reflection. The whole field is self-confused..." In his book Beyond Words: What Animals Think and Feel he writes, "All the mirror test shows is whether an animal understands reflection of itself and cares about its reflection. Mirrors are extremely primitive tools for understanding the complexities of minds. It’s preposterous to say that animals who don’t understand their reflection don’t have self-awareness." (Pages 277-8)

Moving beyond visual aspects of self-cognizance: Making sense of scents

The mirror test focuses on visual cues; however, there is no reason to think that other animals only use vision to make some assessments of self. The results of studies of self-recognition are a mixed bag, and the most important points are that while some of the subjects (or often only one individual) touch the spot that's been placed on their body, not all individuals of a species show these self-directed movements. Also, just because some animals don’t make self-directed movements, this does not mean they don’t have some sense of self. For example, decades ago, Michael Fox and I tried to do the “mirror test’ on dogs and wolves, and none showed any interest in the spot on their forehead.

Dog looking into a mirror
Source: Dreamstime

However, a couple of years ago, when I was talking with a group of students, Arianna Schlumbohm told a story about her dog, Honey. "One day a few years ago, Honey had been lying with me on my bed. I was wearing these truly awful purple fuzzy socks, and she got some fuzz on her forehead at some point. It was adorable. After a little bit of this, she caught a glance of herself in my mirror and almost immediately reacted. She batted at the fuzz with her front paws until it caught, then sat on my stomach until I pulled the fuzz off her paw. Then she went back to the foot of the bed for a few more hours. Honey was really upset, but calmed down as soon as she saw the purple was off. I always just thought of it as a cute, dopey dog story, but I really hope that it will help out your research!" Concerning dogs and mirrors, a research paper by Megumi Fuzuzawa and Ayano Hasha called “Can we estimate dogs’ recognition of objects in mirrors from their behavior and response time?” shows that dogs can learn to use mirrors to locate food without humans helping them. (For more discussion of the cognitive and emotional lives of dogs, please see Canine Confidential: Why Dogs Do What They Do.) 

Moving beyond the visual aspects of self-cognizance, it's highly likely that some animals use olfactory or perhaps auditory cues to make some assessments of self. For example, the “yellow snow” test shows that a sense of self may be related to assessments of olfactory rather than visual cues. In one study I learned that my dog Jethro spent less time sniffing his own urine than that of other males or females, and that while his interest in his own urine waned with time, it remained relatively constant for other individuals’ urine. Jethro infrequently urinated over or sniffed and then immediately urinated over his own urine. He marked over the urine of other males more frequently than he marked over females’ urine. I concluded that Jethro clearly had some olfactory sense of “self”, a sense of “mine-ness” but not necessarily of “I-ness.”

Drs. Alexandra Horowitz and Roberto Cazzolla Gatti have followed up on my study. In her book Being a Dog: Following the Dog into a World of Smell, Dr. Horowitz writes about the results of a more systematic study of self-recognition with dogs (for more discussion, please see "Dogs: When They Smell Their Pee They Know It's 'Me'" for an interview I did with Dr. Horowitz). She notes that the dogs “peed only on other dogs’ containers, not their own. They saw themselves" (page 28). Neither Dr. Horowitz nor I are sure what these studies tell us about self-awareness, but they do say something about identity. Dr. Roberto Cazzolla Gatti also also studied olfactory aspects of self-recognition in dogs and presents his findings in an essay called "Self-consciousness: beyond the looking-glass and what dogs found there." 

More on "the 'me' illusion"

"Self-awareness isn't the pinnacle of consciousnessit's just an accidental byproduct of evolution, and a figment of our minds."

The above quote is the subtitle from the online version of Ms. Deleniv's essay. In the print edition, it reads as a question: "Is having a sense of self really the hallmark of a sophisticated brain, or simply an accident of evolution?" I prefer keeping the door open on either possibility, and available data suggest that this is the best way to go at the moment. We really don't know if there has been direct selection for self-awareness, or if it is a coincidental/accidental byproduct of selection for another trait(s). In his classic book The Nature of Selection: Evolutionary Theory in Philosophical Focus, University of Wisconsin philosopher, Dr. Elliott Sober, distinguishes between two evolutionary processes—namely, selection for and section of different traits. Basically, when a trait is selected for, it is targeted directly, but when a trait appears coincidentally/accidentally as a byproduct, there has been selection of that characteristic. So, for example, it remains possible that some sense of self evolved because of its adaptiveness—it was selected for—or it hitchhiked onto another as yet unidentified trait as a byproduct—there was selection of self-awareness. (For more discussion about different types of selection, please see "Animals Feel Pain Because Something Hurts."1)

In her essay, Ms. Deleniv clarifies much of what we know and don't know about self-awareness/self-cognizance in other animals. Here are some of her thoughts on the wide-ranging matters at hand. She begins by stating, "Self-awareness is one of the biggest mysteries of the mind." She also writes that self-awareness isn't displayed by only bright animals (whatever that means) and that we need to "fundamentally rethink our ideas about mirrors and minds." She also notes that the "self-aware elite contains some bizarre anomalies, including pigeons, manta rays, and ants, and even a robot." And, we can now add a fish to the self-awareness club.

Furthermore, rhesus monkeys who have been thought not to have the cognitive capacity for self-recognition have recently been observed displaying some sort of self-awareness. Ms. Deleniv writes, "Last year, Liangtang Chang and colleagues at the Shanghai Institutes for Biological Sciences, China, released video footage [in an essay in the prestigious Proceedings of the National Academy of Sciences] of a small group of rhesus macaques interacting with a mirror. It shows the monkeys contorting their bodies, tugging at their facial hair, inspecting their fingertips and making flashy displays of their genitals, all the while keeping their eyes on their reflections. They are captivated, leaving little doubt they recognise themselves. Yet, rhesus macaques have consistently failed the mirror test." She goes on, "Chang’s team wondered whether the monkeys genuinely lacked self-awareness, or whether they were being held back by a lack of coordination – an inability to link what they saw with internal signals generated by their muscle movements. To test this, they taught the monkeys to link vision and movement by giving them a food reward for touching a projected laser dot. At first, the researchers shone the laser where the monkeys could easily see it, then gradually worked up to shining it in places only visible in the mirror. Fast-forward a few weeks of practice, and they passed the face-mark test with flying colors."

Ms. Deleniv also notes that Dr. Robin Dunbar's social-brain hypothesis, "according to which life in tight-knit communities is especially challenging because close relationships hinge on being able to understand what is going on in another individual’s mind," doesn't explain the taxonomic distribution of self-awareness. According to Dr. Dunbar's hypothesis, "Brains needed to evolve from being simply things that experience sensations and thoughts to becoming their observer." The ability to step into the heads and hearts—the thoughts and feelings—of other individuals means that an individual has a Theory of Mind (ToM). Basically, ToM "is the ability to attribute mental states— beliefs, intents, desires, pretending, knowledge, etc.—to oneself and others, and to understand that others have beliefs, desires, intentions, and perspectives that are different from one's own."

Many people think that if any nonhumans have a ToM, great apes are the most likely candidates. While some primatologists have been pretty stingy about accepting which species display a Theory of Mind and which don't, it's clear that limiting the possibility only to nonhuman and human primates is far too narrow a view. Indeed, it's entirely reasonable to argue that fair play among diverse species relies on individuals having a ToM. (For more discussion, please see "Theory of Mind and Play: Ape Exceptionalism Is Too Narrow," "When Dogs Talk About Play They Take Turns Sharing Intentions," and links therein.) There's also evidence that ravens and perhaps other birds also are solid candidates for having a ToM (see, for example, "Ravens Know They're Being Watched: Bird Brain Theory of Mind"). 

Beyond Mirrors: The taxonomic distribution of self in other animals

"Indeed, all this makes clear that the best we can hope for with mirrors is an imperfect glimpse into minds like our own. What’s more, if we proceed under the assumption that such minds are the true pinnacles of complexity, then we will miss out on the most beautiful thing about minds – that they are biological machines for adaptation, with contents that can be sophisticated in so many ways." (Sofia Deleniv)

There's still a lot to learn about which animals might think something like, "Wow, that's me," when they look into a mirror, when they sniff different odors, or when they hear a certain sound. It will be very interesting and important to see what we learn using neuroimaging studies to watch which areas of brains light up when animals are tested to see if they have some sort of sense of self. As Ms. Deleniv makes abundantly clear, we need to be very cautious of attempts to exclude different animals from the self-awareness club because we're constantly learning that the club is not as exclusive as some have made it out to be. I strongly encourage people to read her essay and to think carefully about what we know and don't know about which animals have a sense of self and what it means to know something about oneself. All in all, we really don't know all that much about what nonhumans are reflecting about when they see their own reflections, sniff different odors, or perhaps hear different sounds. And, it's still not clear if self-awareness is an accidental byproduct of evolution, or more directly favored trait. As Ms. Deleniv suggests, brain size doesn't seem to be the trait onto which self-awareness hitchhiked. 

How matters of mind inform matters of welfare

Answers to challenging questions about animal self‐awareness clearly have wide‐ranging significance because they are often used as the litmus test for determining and defending the sorts of treatments to which animals can be morally subjected. Of course, individuals can feel pain even if they're not self-aware, and there's no relationship between intelligence and the capacity to suffer (for more discussion about the capacity for "not so smart" animals to suffer, please see "Do 'Smarter' Dogs Really Suffer More than 'Dumber' Mice?"). Also, the pains of supposedly "smarter" animals are not morally more significant than the pains of supposedly "dumber beings." Dr. Georgia Mason (Animal Welfare, 1994, volume 3, Pp. 57-58) points out that there seems to be no good reason why self-awareness needs to be as a prerequisite for suffering, why "the (self­ aware) feeling 'I am suffering' [should] be considered worse than the (not self-aware) feeling 'Something truly terrible is happening.'"

It's clear that researchers need to develop additional tests that can be used on a wide variety of animals who rely on different sensory modalities to learn what they know about themselves. There's still so much more to learn about what they really know about themselves, using cues when they stare into a mirror, sniff something, or hear a sound. How exciting it will be to learn more about the evolution and the taxonomy of self in nonhuman animals including those in which self-awareness has not yet been detected.

Please stay tuned for more discussion on the fascinating topic of self-awareness in other animals and also in humans. Ms. Deleniv's essay is a great beginning for such discussions and I hope it will make its way into cocktail parties, classrooms, and professional venues. Agree or not with what she and others have written about self-awareness, it's high time to revisit basic ideas about this interesting and daunting topic.  


1Another example would be if a polar bear's coat was warm but not heavy, it would have been selected (the published paper can be seen here), but if it were heavy but not warm, it would not have been selected. Warmth, rather than weight, was selected for

More Posts