Are dogs self-aware? Do they recognize themselves as individuals, distinct from others?Other Animals Have It: Although rather tricky topics of study, animal self-recognition, self-awareness and consciousness have been examined by scientists for decades. Animal consciousness is neither a new idea, nor is it a radical way of thinking. Lucky for us, we no longer live in the age of Descartes when animals other than those of the human variety were viewed as non-thinking automatons who lacked both consciousness and the ability to feel emotions. (Though, personally I can think of a few humans who may fit that description).
Evidence for at least a rudimentary sense of self-awareness is available in a wide range of non-human animal species. A leading theory of the evolutionary benefits of this trait is that the ability to distinguish self from other helps social animals (including humans) to recognize their place within their social group, to cooperate successfully with others, and to identify individuals who are outside of their group. Dogs, also members of a highly social species, are now known to have much more complex inner lives than we once gave them credit for. They readily follow the gaze of another dog or person, understand pointing, attend to the emotional states of others, and demonstrate rudimentary aspects of perspective taking (knowing what someone else can see or know). Having a sense of self as distinct from others is an additional cognitive talent that dogs may possess given their highly social nature and the functional benefits of self-recognition and self-awareness.
Mirror, Mirror: The classic test used to study self-recognition has been the mirror self-recognition (MSR) test. Using this method, the subject animal examines her image in a mirror after an area of her body has been surreptitiously marked with a spot of dye. The animal’s reaction to this alteration is observed and if the subject uses the mirror to examine the spot on her body, this attention is interpreted as evidence for recognizing the image in the mirror as oneself rather than simply an image of a like-looking animal with a funny spot on her head. Species that regularly pass the MSR test include the great apes (chimpanzees, bonobos, orangutans), dolphins, a single elephant, and even some bird species, such as the Magpie. Oh yeah, and most humans pass, as long as they are older than 2-years of age.
What about Dogs? Dogs however, have routinely failed this test. Dog folks are now certainly muttering, “Well of course, dogs do not use vision as their primary special sense – they use olfaction – their noses“. This difference is significant, since dogs believe what their nose tells them first and foremost, compared with primate species such as ourselves, who perceive the world primarily through vision. Additionally, because of anatomical and social differences, dogs do not regularly self-groom in the same manner that primates do, so are not as apt to care about an unexpected spot that suddenly shows up on the top of their head. For those who study dogs, clearly, another type of test was needed.
Enter Alexandra Horowitz and her team of dog pee researchers at Barnard College in New York City.
The Significance of Pee: Dogs regularly investigate the urine scent of other dogs. There is evidence that they spend more time investigating the urine markings of other dogs and less time sniffing their own urine, which suggests that dogs distinguish their own scent from that of others. Using this knowledge, Horowitz devised a new type of mirror test for dogs – this one based upon their primary sense – smell. She reasoned that just as a chimpanzee notices the sudden change in appearance when a spot of dye shows up on her head, if dogs recognize their own scent, then they too should be surprised to find an unexpected change in that smell and attend to it (sniff it) for a longer period of time. She devised a pair of controlled experiments that asked, using their sense of smell – “Do dogs recognize themselves?”
The Study First, the team of researchers collected the pee of a group of volunteer dogs (well, okay, the owners volunteered their dogs’ pee. We are not really sure how the dogs felt about that part). The author also collected urine from her own dog, who would serve as the “unfamiliar dog” sample. Each dog was tested individually with a set of three scent canisters for three separate trials and comparisons. One canister contained water only (decoy sample), one contained the subject dog’s urine (self), and the third contained either (1) the subject dog’s adulterated urine (marker self), (2) the urine of an unfamiliar dog (other), or (3) the scent of the adulteration substance alone (marker). Two experiments were conducted, with the only difference being the way in which the subject dog’s urine was altered. In Exp. 1, a tissue sample of dog spleen was added to the urine. In Exp. 2, a small amount of anise essential oil was added.
Results: Similar to mirror tests, the researchers expected dogs to pay more attention to a scent of themselves that was unexpectedly altered compared with their reaction to their unaltered urine scent. Here is what they found:
- Who’s this guy? As earlier research has shown, dogs spent more time investigating the urine of an unfamiliar dog compared with the time that they spent sniffing their own urine. (“Hmm…. Smells like I was here earlier……whoa…..hello….who is this new dude who peed here too?)
- Hey Sally! Interestingly, dogs did not spend more time investigating the urine of a known dog (their housemate) compared with time spent smelling their own urine. (Looks like Sally was visiting at the same time I was. Funny, I don’t remember seeing her here….”)
- Does this smell funny to you? Last, dogs spend significantly more time investigating the canisters that contained their altered urine scent compared with how long they investigated their unadulterated urine. This difference occurred with both types of marker substance – spleen tissue and anise oil. Dogs also returned to the canisters more often when their urine was compared with their adulterated urine. (“Wowza. This is weird. Did I eat something odd last night? Maybe I am getting a cold? What the heck IS that smell on me???”)
The authors conclude that these results support the use of their newly designed (and quite ingenious, if I may add) “smell test” as species-relevant analog to the MSR test. The fact that the dogs spent more time investigating their own urine when it had been unexpectedly changed supports some level of recognition of their own odor and by extension, perhaps a rudimentary “sense of self”. Similarly, dogs were highly interested in the scent of unfamiliar dogs (Hey! Who’s this guy??) but not to the odor of their housemate.
Yeah, I have an opinion on this one. First though, I have to say that this is one of the most creative and clever studies that I have read in some time. (Not to mention it being ripe for witticisms and puns……).
The results of this study suggest that dogs may possess one of the cognitive traits, self-recognition, that humans have historically co-opted for our species and our species alone. In past, we have worked diligently to make clear cognitive distinctions between human animals (us) and non-human animals (everyone else). A wide range of traits have been used for this purpose, many of which have fallen like a house of cards as they are discovered to exist in other animals. Examples include the expression of emotions, perspective taking, tool use and tool making, existence of culture, ability to reason, and the demonstration of altruism. We also know that humans do not hold exclusive rights to the expression of self-awareness and consciousness and are not the only species capable of complex thought, internal representations of the world, planning, intention and deception. Yeah, we do have language and we are capable of “meta-thinking” (thinking about thinking), but many types of cognition and complex thought have been demonstrated to exist in some form in a host of other animals, including dogs. So what is the big deal? Is there really anything to argue about here? Well, yeah, as a dog trainer (a clicker trainer, I must emphasize), I think that there is an important point to be made.
It is this. Behaviorism alone can no longer be enough. The science of behaviorism and its application in dog training no longer can adequately capture and address all that is dog. Sorry to all of you purists out there, but there it is. (And remember, I am a clicker trainer).
Here is my argument: Although dogs respond well to the laws of behaviorism (just as humans do), the fact that we successfully use operant and classical conditioning to train dogs should not be confused for evidence that dogs are lacking in a host of mental skills that fall higher on the cognitive complexity scale. Behaviorism and social cognition are not mutually exclusive sciences (though to listen to some trainers and some scientists, you would think they were disciplines existing on different planets).
The reason that I bring up this particular issue in this particular essay is because self-recognition and self-awareness seem to be a current “hot spot” in this debate between behaviorism and cognitive science. Pure behaviorism has its benefits – mainly it works great when applied as a training technique. However, given the boatloads of research published by cognitive scientists that demonstrate the social complexity of the domestic dog (and now – self-recognition!), we cannot discount as trainers evidence showing that dogs pay attention to the social cues of humans and of other dogs, that they possess some level of perspective taking, that they regularly learn through observation of others, that they can recognize one another and understand intent by the sound of their barks, and that they can recognize one another and themselves through smell. It is time for trainers to embrace both of these important and enlightening bodies of science. We should support and use behaviorism because it provides simple and elegant rules for training that work, and we must also encourage studies of canine social cognition because they continue to teach us more about the internal lives, experiences and perceptions of our canine best friends.
Off of soap box. Back to pee jokes.
Cited Study: Horowitz, A. Smelling themselves: Dogs investigate their own odours longer when modified in an “olfactory mirror” test. Behavioural Processes, 2017; 143:17-24.
8 thoughts on “Does this Smell Funny to You?”
Pingback: Does this Smell Funny to You? — The Science Dog | Train Positive Dog!
I don’t see how we can infer that dogs have self-recognition from this single study, though it is an interesting one. How do we parse out self-recognition from simply something that went from smelling familiar now smells different and therefore requires a lot of investigation. I don’t have any difficulty imagining that that could just as easily elicit the behavior seen here. An example comes to mind. I have a circle of dead tree stumps in the backyard where I eventually plan to have a fire pit. One evening I was sitting on my porch when I noticed that one of the stumps was out of place and possibly split. Not only did I comment to my friend and then later spend quite some time sitting there trying to figure out whether I had just thought they were somewhere else or something had moved and if so, how, but I went outside and moved them around to see if it seemed like they had been sitting in those spots for long or not. I would love to hear thoughts from a neuroscientist of some kind as I imagine that things that are known/routine get stored away in a different part of our brain and don’t get much attention paid to them so that when something IS different, it may be quite noticeable or even alarming.
Second thought – as soon as I read anise, I wondered if the owners had been questioned about whether their dogs participated in scent work since anise is one of the odors used in the sport. In the methods section, I found this:
“The dogs were brought by their owners to a designated room at
the Port Chester Obedience Training Club (PCOTC) in White Plains,
New York. In preliminary observations, dogs whose owners had not
participated with them in nosework, scent games, or scent work did not
reliably approach the stimulus canisters nor were some dogs able to
separate from their owners to do so. Therefore, only dogs whose owners
encouraged their dogs’ sniffing behaviour were used in the study; they
were recruited via PCOTC electronic mailings. All subjects were familiar
with the facility and had participated in at least one class in the
sport of nosework with their owners at PCOTC before.”
This is *very* concerning to me. Even if these dogs hadn’t been trained specifically on anise, they have been trained to locate an odor that they don’t really encounter elsewhere in their life so it makes sense to me to that an unexpected odor would warrant further investigation in a way that an unfamiliar dog smell (which nosework dogs are trained to work past) wouldn’t.
None of this necessarily discounts your overall points about cognition, but to state that we now know dogs are self-aware from this study seems like a really big jump.
I’ve always wondered this and this article certainly gives a full array of opinions and views. That’s like if dogs really miss you as the owner, I watched some news documentary maybe it was 60 minutes where it stated, the dog actually misses your scent not by the visualization of your being. So they did a test where the owners left for work for the day and then came back to an excited dog. When they left the owner’s shirt with his scent, the dog did not notice or care when the owner came back. This kind of reinstates that point, that it’s all about scents and that visually looking in the mirror they wouldn’t recognize themselves.
But — but — one’s humans’ scents are all over the house. The dog could go anywhere and still get their person’s scent, no need for a shirt. I wonder who did the research ansd how they tested for it? In a laboratory, or in the dog’s normal place where it stays when their person is absent?
Nice! Seems like a more rigorous version of the “yellow snow” experiment that ethologist Mark Bekoff and his dog Jethro conducted and that I covered in my blog a few years ago: https://animalwise.org/2011/08/16/the-yellow-snow-test-for-self-recognition/…
Great point, Peter. I agree that there are some limitations to these tests, both the mirror and the urine tests (and the authors do discuss these in their paper). I love your point about mirror learning not being natural (as I am sure you know, the MSR tests typically require pre-test exposure to mirrors), while urine sniffing is a naturally selected-for behavior. Perhaps we will see some additional work with the comparison that you mention!
Intriguing. I haven’t got around the paywall yet, but here is an important difference between the mirror tests and the urine tests – no animal naturally sees its mirror image, but dogs, and mammals in general, are adapted to smelling and recognising their own urine and other scent marks, and so should be naturally good at it. I’m not sure whether that affects the conclusions – certainly it opens the door to an “instinctive” unconcious mental process. A more rigorous control would have been to compare the impact on sniffing time between modified own urine and modified stranger urine – if the dogs recognised self as self they would have shown more extra interest in their own modifed urine than in modified stranger urine.
LikeLiked by 1 person
Ah, now, this makes me think.
Has anyone tested to see if dogs can recognise other individual dogs when seen in a mirror??
I know that at least my own dogs can recognise individuals (humans, other dogs and cats) through glass which would eliminate smell as the identifier.