Mixing Memory

By now you’ve probably all heard about the paper published by Plotnik, de Waal, and Reiss in the Proceedings of the National Academy of Sciences in late October titled “Self-recognition in an Asian elephant.” I suspect that for people who study elephants, the results described in that paper come as no surprise. Researchers have been testing elephants on measures of self-awareness for a while, because they seemed like a good candidate. Still, the conclusion, stated in the title of the paper, is probably a bit premature. To understand why, though, it’s important to understand the methods used in the paper, and all of the issues surrounding those methods. So I thought I’d write a post about the mirror test, and end with a discussion of the elephant experiment.

The MIrror Test

The mirror test was first used by Gordon Gallop with chimpanzees in the late 60s1. The procedure in Gallup’s study was pretty simple. The chimps were first familiarized with mirrors by placing mirrors in their enclosures for a period of ten days. At first, when the chimps looked at the mirror, they acted like they were seeing another chimp, and acted accordingly. After a few days, though, they began to use the mirror to guide self-directed behaviors. This led Gallop to believe that they were recognizing themselves in the mirror. In order to test for this possibility, Gallop anesthetized the chimps, and placed red dye marks above their eyes and on their ears. When the chimps woke up, they seemed unaware that anything was different, but when they looked in the mirror, it was clear that they quickly noticed the marks (they kept touching the marks, for example). Gallop and others took this as clear evidence that the chimps were self-aware.

After Gallop’s paper on his mirror test with chimpanzees was published in 1970, researchers were quick to test several other primate species. For years, it was believed that only humans over the age of 2 and the great apes, chimps, gorillas (sometimes), and orangutans. For years, researchers had failed to find mirror self-recognition in any other primates, though a still controversial study published by Hauser et al. in 1995 showed mirror self-recognition in cotton-top tamarins2. As far as I know, other monkey species (e.g., Rhesus monkeys) have yet to show mirror self-recognition. Other promising nonprimate species, like African gray parrots3, have consistently failed the mirror test, though in 2001, Reiss and Marino reported that two botltenosed dolphins had passed it4. There have also been some imaging studies showing that self face recognition takes place in the right hemisphere5, perhaps in the mirror neuron system6. The implication is that self-awareness occurs in the right hemisphere as well.

There are other methods for testing for self-awareness, mostly in the form of various false-belief tests, but mirror self-recognition is by far the most widely used, but it has been seriously challenged on both theoretical and empirical grounds. For one, failure to find self-recognition in some individual members of a species does not mean that other members of the species won’t pass the mirror test. Often when many individuals are tested, a third or fewer pass the test, even in great ape species. So if individuals don’t pass the test, it could just mean that you’ve selected the wrong individuals. Furthermore, there are alternative explanations for passing the mirror test. In an ingenius study published in 1981 by Epstein et al. (B.F. Skinner was one of the et als)7, pigeons were trained to peck red dots in various locations. When a red dot was placed on their body so that they couldn’t see it without a mirror, the pigeons quickly learned to peck at the red dot when they observed it in the mirror. Epstein et al and others have argued that this indicates that animals who are trained with mirrors simply learn to use mirrors to locate objects in their environment, and therefore that passing the mirror test does not provide evidence of self-awareness. Others have argued that the mirror test merely demonstrates bodily awareness, which is at least one step down from awareness of one’s own mental states (which is usually what people mean when they use the phrase “self-awareness”). In sum, then, mirror self-recognition alone is probably not enough to attribute self-awareness to a species.

The Elephant In the Mirror

i-6b245d33867c2aac87febde729b32c24-mirrorelephant.JPG

With all that said, I still think the elephant findings are cool. Previous research had failed to find evidence of mirror self-recognition in Asian elephants (see here), but as I said, negative results on the mirror test are not sufficient to infer that a species is not self-aware, or even that there aren’t at least some members of a species that could pass the mirror test. In the study published last week, Potnik et al8 argue that methodological problems, such as a mirror that was too small, were responsible for failures to find mirror self-recognition in elephants in the past. So they built a giant, “elephant-proof” mirror (2.5 meters tall), and placed it inside the elephant’s enclosure, so that they could touch it. They then placed large, highly visible white X’s on the right side of the elephant’s head (you can see the mark in the photo above, from Figure 2), and subsequently observed one of two elephants touching the mark when looking at herself in the mirror (as in the photo).

The conclusion the authors draw (a conclusion also drawn in the dolphin study mentioned above) is that mirror self-recognition in several different species with fairly divergent evolutionary histories, along with failure to find mirror self-recognition in other species in between those that do demonstrate it, is evidence of convergent evolution, perhaps as a result of the social environment in which these species evolved. I don’t think this conclusion is really warranted, because as I’ve said, failure to find individuals who can pass the mirror self-recognition test is not evidence of a lack of self-awareness or self-recognition in an entire species. Still, I think this opens the door for people to search for better mark test techniques for testing other non-primate species, and I wouldn’t be surprised if these results lead to similar results in a variety of species over the next several years. So in that regard, I think these results are exciting.


1Gallup, G. G., Jr. (1970). Chimpanzees: self-recognition. Science 167, 86-87.
2Hauser, M.D., Kralik, J., Botto-Mahan, C., Garrett, M., & Oser, J. (1995). Self-recognition in primates: Phylogeny and the salience of species-typical features. Proceedings of the National Academy of Sciences, 92, 10811-10814
3Pepperberg, I.M>, Garcia, S.E., Jackson, E.C., & Marconi, S. (1995). Mirror use by African Grey parrots (Psittacus erithacus). Journal of Comparative Psychology, 109(2), 182-195.
4Reiss, D., & Marino, L. (2001). Mirror self-recognition in the bottlenose dolphin: A case of cognitive convergence. Proceedings of the National Academy of Sciences, 98(10), 5937-5942.
5Keenan, J.P., Nelson, A., O’Connor, M., Pascual-Leone, A. (2001). Self-recognition and the right hemisphere. Nature, 409(6818), 305.
6Uddin, L.Q., Kaplan, J.T., Molnar-Szakacs, I., Zaidel, E., Iacoboni, M. (2005). Self-face recognition activates a frontoparietal “mirror” network in the right hemisphere: An event-related fMRI study. NeuroImage, 25, 926-935.
7Epstein R., Lanza R.P. and Skinner B.F. 1981. ”Self-awareness” in the Pigeon. Science, 212, 695-669.