Mirror Self-Recognition: What the Mark Test Reveals About Animal Self-Awareness
Topic: Animal Behaviour & Ethology Subject: Mirror Self-Recognition, the Mark Test, and Self-Awareness Last verified: 2026-09-23
Most animals that meet their reflection for the first time treat it as a stranger. They threaten it, court it, or look behind the mirror for the animal they think is there. A few species go further and appear to work out that the image is of themselves. In 1970 the American psychologist Gordon Gallup Jr. turned this observation into an experiment. He showed that chimpanzees given long exposure to mirrors would use their reflections to inspect dye marks on parts of their bodies they could not otherwise see, while monkeys did not [Gallup 1970]. The "mark test" that grew out of this study has since been applied to species ranging from elephants and dolphins to magpies and small reef fish. It is still the most widely used experimental measure of self-recognition in animals, and one of the most argued about [Anderson & Gallup 2015; de Waal 2019].
This article explains how the mark test works and the stages animals pass through in front of a mirror. It then surveys the main positive results: great apes, bottlenose dolphins, an Asian elephant, Eurasian magpies and the cleaner wrasse. It covers the replication problems and interpretive disputes that followed, the criticism that the test is biased towards vision, and what a pass or a fail can and cannot tell us. It closes with why these questions matter for animal welfare and for how we think about sentience.
Gallup's Experiment and the Mark Test
Gallup's original study, published in Science, placed individually housed chimpanzees (Pan troglodytes) in front of a mirror. After prolonged exposure to their reflections, the chimpanzees were marked with red dye and showed evidence of being able to recognise their own reflections. The monkeys tested in the same study did not appear to have this capacity [Gallup 1970]. In the procedure, each animal is anaesthetised and a mark is placed on a part of the body it cannot see directly, such as the brow ridge or ear. The mark is made with a non-irritant, odourless dye, so that the animal has no cue to its presence other than the mirror. The key question is whether, once the mirror is returned, the animal spontaneously touches or investigates the mark on its own body rather than the image in the glass [Gallup 1970; Anderson & Gallup 2015].
Two controls make the test informative. First, mark-directed behaviour is compared with behaviour when no mirror is present, to show that the animal is responding to what it sees in the reflection and not to a feeling on its skin. Second, many studies apply a "sham" mark, which is transparent or otherwise invisible, to rule out the possibility that the marking procedure itself draws attention to the area [Plotnik et al. 2006; Reiss & Marino 2001]. A comparable test was developed independently in developmental psychology, in which a spot of rouge is placed on an infant's nose [Amsterdam 1972]. Most human children pass a mirror mark test by about two years of age [Chang et al. 2015].
Stages of Mirror Behaviour
Animals that eventually recognise themselves do not do so at once. Joshua Plotnik, Frans de Waal and Diana Reiss described a typical progression through four stages when an animal faces a mirror: (i) social responses, as if to another individual; (ii) physical inspection, such as looking behind the mirror; (iii) repetitive mirror-testing behaviour, in which the animal makes movements and watches the reflection move in step; and (iv) realisation of seeing itself, shown by self-directed behaviour [Plotnik et al. 2006]. The third stage, often called contingency testing, is considered especially telling. An animal that repeatedly moves and checks whether the image copies it appears to be exploring the link between its own actions and what it sees [Plotnik et al. 2006; Anderson & Gallup 2015].
Even within a species that passes, not every individual does. Daniel Povinelli and colleagues observed 105 chimpanzees aged from 10 months to 40 years across five days of continuous mirror exposure, and followed up with mark tests and a longitudinal study of young animals [Povinelli et al. 1993]. Their results suggested that self-recognition typically emerges between 4.5 and 8 years of age, that at the population level the capacity declines in adulthood, and that in group settings it typically appears within minutes of an animal's exposure to a mirror [Povinelli et al. 1993].
Marina de Veer and colleagues later retested all 12 chimpanzees from the 8–15-year-old group in the earlier cross-sectional work, eight years on [de Veer et al. 2003]. In the original sample, 75% of animals in this age group had been classified as positive, compared with only 26% of animals aged 16–39. On retesting with the same criteria, 67% of the animals were classified the same as before, and 50% of these now-adult chimpanzees showed self-recognition. Every change in status, however, was in the negative direction [de Veer et al. 2003]. The authors concluded that mirror self-recognition is a highly stable trait in many chimpanzees but may decline with age [de Veer et al. 2003]. A failed test for one individual on one occasion is therefore weak evidence that the species lacks the capacity.
Beyond the Great Apes: Dolphins and Elephants
For three decades after Gallup's study, convincing evidence of mirror self-recognition was limited to humans and great apes [Reiss & Marino 2001]. Among primates, James Anderson and Gordon Gallup's review of attempts to extend the result concluded that there was still no compelling evidence that prosimians, monkeys or the lesser apes (gibbons and siamangs) are capable of mirror self-recognition [Anderson & Gallup 2015]. They were particularly critical of efforts to "engineer" positive results in monkeys by modifying the test or by conditioning [Anderson & Gallup 2015]. In one such study, rhesus macaques (Macaca mulatta) were trained in front of a mirror to touch a laser light spot on their faces that produced an irritant sensation. After two to five weeks of training they learned to touch non-irritant marks seen in the mirror. In their home cages, five of the seven trained monkeys showed mirror-induced self-directed behaviours, while four untrained control monkeys passed none of the mark tests [Chang et al. 2015]. Whether behaviour acquired through such training is equivalent to spontaneous self-recognition remains disputed [Anderson & Gallup 2015].
Diana Reiss and Lori Marino exposed two bottlenose dolphins (Tursiops truncatus) to reflective surfaces. Dolphins have no hands, so a pass had to be judged from how they positioned their bodies. Both animals showed responses consistent with using the mirror to investigate marked parts of their bodies [Reiss & Marino 2001]. The authors described this as a striking example of evolutionary convergence with great apes and humans, given how distantly related dolphins are to primates [Reiss & Marino 2001].
Plotnik, de Waal and Reiss then tested three Asian elephants (Elephas maximus) at the Bronx Zoo with a large mirror, applying both visible marks and invisible sham marks to their heads [Plotnik et al. 2006]. The elephants progressed through the same four stages of mirror behaviour reported in apes and dolphins, and the study reported a successful mark test [Plotnik et al. 2006]. Only one of the three elephants, named Happy, touched the visible mark with her trunk while ignoring the sham mark; the other two did not pass [Plotnik et al. 2006]. The authors suggested that the parallels between apes, dolphins and elephants reflect convergent cognitive evolution, most likely linked to complex sociality and cooperation [Plotnik et al. 2006].
Magpies and the Replication Problem
In 2008 Helmut Prior, Ariane Schwarz and Onur Güntürkün reported mirror-induced behaviour in the Eurasian magpie (Pica pica), a member of the crow family [Prior et al. 2008]. As in apes, some individuals behaved in front of the mirror as if testing behavioural contingencies. When given a mark, magpies showed spontaneous mark-directed behaviour [Prior et al. 2008]. The authors presented this as the first evidence of mirror self-recognition in a non-mammal. They argued that essential components of self-recognition had evolved independently in vertebrate lineages with separate evolutionary histories [Prior et al. 2008].
The result rested on a small sample: two of the five magpies tested passed the mark test [Soler et al. 2020]. Manuel Soler and colleagues carried out a close replication with a larger sample and a very similar protocol [Soler et al. 2020]. As in the original study, their magpies showed both social and self-directed behaviour more often in front of a mirror than in front of a cardboard control. During the mark test itself, however, self-directed behaviour was more frequent in front of the cardboard than in front of the mirror, so the replication failed to confirm the earlier findings [Soler et al. 2020]. The authors noted that a failed replication does not disprove the original study. It does mean that the result should be treated with caution until further work shows that magpies can pass the test consistently [Soler et al. 2020].
Cleaner Wrasse and the Fish Debate
The most contentious claim came from a fish. Masanori Kohda and colleagues exposed the bluestreak cleaner wrasse (Labroides dimidiatus), a small reef fish that removes parasites from other fish, to mirrors [Kohda et al. 2019]. The fish showed behaviour that could reasonably be interpreted as passing through all the phases of the mark test: social reactions towards the reflection, repeated idiosyncratic behaviours towards the mirror, and frequent observation of the reflection [Kohda et al. 2019]. When given a coloured mark, the fish tried to remove it by scraping their bodies against the substrate, but only when a mirror was present. They did not respond in this way to transparent marks, or to coloured marks when no mirror was available [Kohda et al. 2019]. The authors put the dilemma directly: either fish that behave like this are self-aware, or the mark test does not measure what it has been taken to measure [Kohda et al. 2019].
The paper received both positive and negative reviews, and the journal published it alongside a commentary by Frans de Waal [Kohda et al. 2019; de Waal 2019]. De Waal argued that the mirror test had encouraged a binary view in which a few species possess self-awareness and all others lack it, and that evolution calls instead for a gradualist model [de Waal 2019]. In his view, the ambiguous behaviour of the fish and the use of physically irritating marks meant the results fell short of mirror self-recognition. He suggested they indicated an intermediate level of mirror understanding, closer to that of monkeys than of hominids [de Waal 2019]. Gallup and Anderson were more sceptical still. They argued that, on rigorous and reproducible evidence, only humans and some great apes had shown clear, consistent and convincing self-recognition, and they criticised the recent claims for other species, including the cleaner wrasse [Gallup & Anderson 2020].
Kohda's group responded with further experiments [Kohda et al. 2022]. All 14 new fish scraped their throats when given a brown mark, but only in the presence of a mirror. Blue and green marks did not elicit scraping. Injecting the mark deeper beneath the skin caused scraping even without a mirror, which confirmed that a mark can be felt if applied carelessly. Mirror-naive fish scraped less than mirror-experienced ones, and seeing another fish with the same mark did not trigger scraping [Kohda et al. 2022]. The authors concluded that cleaner wrasse pass the mark test when it is presented in an ecologically relevant context, such as a mark resembling a parasite [Kohda et al. 2022].
A follow-up study asked how the fish recognise themselves [Kohda et al. 2023]. Mirror-naive fish attacked photographs of both themselves and unfamiliar individuals. After passing the mark test, they no longer attacked motionless photographs of themselves but still frequently attacked photographs of strangers. With composite photographs, the fish did not attack an image combining their own face with a stranger's body, but did attack a stranger's face on their own body. Shown a photograph of themselves with a mark on the throat, unmarked mirror-experienced fish scraped their throats [Kohda et al. 2023]. The authors argued that the fish recognise their own facial characteristics and hold a mental image of their own face, rather than relying on matching their movements to the image [Kohda et al. 2023]. How far these findings should change the interpretation of the mark test remains an open debate [de Waal 2019; Gallup & Anderson 2020].
Critiques: Vision, Motivation and the Meaning of Failure
The mark test assumes that the animal relies on vision, cares about marks on its body, and can reach or rub the marked area. Alexandra Horowitz pointed out that domestic dogs (Canis familiaris) had not passed the mirror test despite their skill at many social cognitive tasks, and that motivational and sensory barriers might prevent success before self-recognition even comes into question [Horowitz 2017]. Because dogs rely heavily on smell, she devised an "olfactory mirror" test in which dogs investigated canisters of odour, including their own urine [Horowitz 2017]. Dogs investigated their own odour for longer when an additional odour had been added to it than when it was unmodified, and they spent longer investigating other dogs' odours than their own. A second experiment showed that novelty alone did not explain the result [Horowitz 2017]. The study does not show that dogs are self-aware in the sense the mirror test aims at, but it illustrates how a vision-based test can underestimate species whose main sense is not vision [Horowitz 2017].
Other critiques concern what a result actually means. The same cleaner wrasse findings have been read as evidence of self-awareness in fish, as evidence of a partial or intermediate capacity, and as a reason to doubt the claim altogether [Kohda et al. 2019; de Waal 2019; Gallup & Anderson 2020]. Training studies in monkeys raise the question of whether a learned association between a mirror image and the body is the same capacity as spontaneous self-recognition [Chang et al. 2015; Anderson & Gallup 2015]. The chimpanzee data show that individuals of a "passing" species often fail, and that performance varies with age [Povinelli et al. 1993; de Veer et al. 2003].
Taken together, the literature supports a few careful conclusions. A spontaneous, well-controlled pass shows that an animal can use a mirror to gain information about its own body. A failure, by contrast, can reflect the animal's senses, motivation, anatomy, age or individual history rather than an absence of self-awareness [Horowitz 2017; de Veer et al. 2003]. And the test addresses one aspect of how an animal represents itself, not whether the animal has feelings or experiences at all [de Waal 2019; Birch et al. 2020].
Conservation and Welfare Relevance
Mirror research has sometimes been used to rank species by how "intelligent" or "self-aware" they are. Jonathan Birch, Alexandra Schnell and Nicola Clayton argued against any single scale of this kind [Birch et al. 2020]. Their framework distinguishes five dimensions along which consciousness may vary: perceptual richness, evaluative richness, integration at a time, integration across time, and self-consciousness. On this view each species has its own distinctive consciousness profile, and none can simply be ranked as more or less conscious than another [Birch et al. 2020]. Mirror self-recognition bears on only one of those dimensions [Birch et al. 2020].
This distinction matters for welfare policy. The capacity that most directly matters for an animal's welfare is sentience, the capacity to have feelings such as pain, pleasure, hunger or comfort. A 2021 review commissioned by the UK government, led by Jonathan Birch at the London School of Economics, drew on over 300 scientific studies to assess the evidence of sentience in cephalopod molluscs and decapod crustaceans. It assessed sentience against criteria centred on the capacity to experience pain, such as nociceptors, integrative brain regions, motivational trade-offs and flexible self-protective behaviour, none of which requires self-recognition [Birch et al. 2021]. Species that fail or cannot take a mirror test should not be assumed to lack morally relevant experiences [Birch et al. 2020; Birch et al. 2021].
Mirror studies also show the value of careful cognitive research on animals in human care. The elephant study was carried out at a zoo, and Plotnik and colleagues later argued that zoological facilities offer an opportunity for cognitive research on elephants, species that cannot easily be kept in laboratories and about whose cognition remarkably little experimental research had been done [Plotnik et al. 2006; Plotnik et al. 2010]. The divergent findings across individuals, species and laboratories point to the same lesson for welfare assessment: an animal's abilities are best judged with tests designed around its own senses and ecology, and one negative result should not be taken as the last word [Horowitz 2017; Kohda et al. 2022; Soler et al. 2020].
References
[Amsterdam 1972] Amsterdam, B. (1972). Mirror self-image reactions before age two. Developmental Psychobiology, 5(4), 297–305. https://doi.org/10.1002/dev.420050403
[Anderson & Gallup 2015] Anderson, J.R. & Gallup, G.G., Jr. (2015). Mirror self-recognition: a review and critique of attempts to promote and engineer self-recognition in primates. Primates, 56(4), 317–326. https://doi.org/10.1007/s10329-015-0488-9
[Birch et al. 2020] Birch, J., Schnell, A.K. & Clayton, N.S. (2020). Dimensions of animal consciousness. Trends in Cognitive Sciences, 24(10), 789–801. https://doi.org/10.1016/j.tics.2020.07.007
[Birch et al. 2021] Birch, J., Burn, C., Schnell, A., Browning, H. & Crump, A. (2021). Review of the Evidence of Sentience in Cephalopod Molluscs and Decapod Crustaceans. London: LSE Consulting, London School of Economics and Political Science. (Report; no DOI.) https://www.lse.ac.uk/business/consulting/reports/review-of-the-evidence-of-sentiences-in-cephalopod-molluscs-and-decapod-crustaceans
[Chang et al. 2015] Chang, L., Fang, Q., Zhang, S., Poo, M.-M. & Gong, N. (2015). Mirror-induced self-directed behaviors in rhesus monkeys after visual-somatosensory training. Current Biology, 25(2), 212–217. https://doi.org/10.1016/j.cub.2014.11.016
[de Veer et al. 2003] de Veer, M.W., Gallup, G.G., Jr., Theall, L.A., van den Bos, R. & Povinelli, D.J. (2003). An 8-year longitudinal study of mirror self-recognition in chimpanzees (Pan troglodytes). Neuropsychologia, 41(2), 229–234. https://doi.org/10.1016/S0028-3932(02)00153-7
[de Waal 2019] de Waal, F.B.M. (2019). Fish, mirrors, and a gradualist perspective on self-awareness. PLoS Biology, 17(2), e3000112. https://doi.org/10.1371/journal.pbio.3000112
[Gallup 1970] Gallup, G.G., Jr. (1970). Chimpanzees: self-recognition. Science, 167(3914), 86–87. https://doi.org/10.1126/science.167.3914.86
[Gallup & Anderson 2020] Gallup, G.G., Jr. & Anderson, J.R. (2020). Self-recognition in animals: where do we stand 50 years later? Lessons from cleaner wrasse and other species. Psychology of Consciousness: Theory, Research, and Practice, 7(1), 46–58. https://doi.org/10.1037/cns0000206
[Horowitz 2017] Horowitz, A. (2017). Smelling themselves: dogs investigate their own odours longer when modified in an "olfactory mirror" test. Behavioural Processes, 143, 17–24. https://doi.org/10.1016/j.beproc.2017.08.001
[Kohda et al. 2019] Kohda, M., Hotta, T., Takeyama, T., Awata, S., Tanaka, H., Asai, J. & Jordan, A.L. (2019). If a fish can pass the mark test, what are the implications for consciousness and self-awareness testing in animals? PLoS Biology, 17(2), e3000021. https://doi.org/10.1371/journal.pbio.3000021
[Kohda et al. 2022] Kohda, M., Sogawa, S., Jordan, A.L., Kubo, N., Awata, S., Satoh, S., Kobayashi, T., Fujita, A. & Bshary, R. (2022). Further evidence for the capacity of mirror self-recognition in cleaner fish and the significance of ecologically relevant marks. PLoS Biology, 20(2), e3001529. https://doi.org/10.1371/journal.pbio.3001529
[Kohda et al. 2023] Kohda, M., Bshary, R., Kubo, N., Awata, S., Sowersby, W., Kawasaka, K., Kobayashi, T. & Sogawa, S. (2023). Cleaner fish recognize self in a mirror via self-face recognition like humans. Proceedings of the National Academy of Sciences of the USA, 120(7), e2208420120. https://doi.org/10.1073/pnas.2208420120
[Plotnik et al. 2006] Plotnik, J.M., de Waal, F.B.M. & Reiss, D. (2006). Self-recognition in an Asian elephant. Proceedings of the National Academy of Sciences of the USA, 103(45), 17053–17057. https://doi.org/10.1073/pnas.0608062103
[Plotnik et al. 2010] Plotnik, J.M., de Waal, F.B.M., Moore, D. & Reiss, D. (2010). Self-recognition in the Asian elephant and future directions for cognitive research with elephants in zoological settings. Zoo Biology, 29(2), 179–191. https://doi.org/10.1002/zoo.20257
[Povinelli et al. 1993] Povinelli, D.J., Rulf, A.B., Landau, K.R. & Bierschwale, D.T. (1993). Self-recognition in chimpanzees (Pan troglodytes): distribution, ontogeny, and patterns of emergence. Journal of Comparative Psychology, 107(4), 347–372. https://doi.org/10.1037/0735-7036.107.4.347
[Prior et al. 2008] Prior, H., Schwarz, A. & Güntürkün, O. (2008). Mirror-induced behavior in the magpie (Pica pica): evidence of self-recognition. PLoS Biology, 6(8), e202. https://doi.org/10.1371/journal.pbio.0060202
[Reiss & Marino 2001] Reiss, D. & Marino, L. (2001). Mirror self-recognition in the bottlenose dolphin: a case of cognitive convergence. Proceedings of the National Academy of Sciences of the USA, 98(10), 5937–5942. https://doi.org/10.1073/pnas.101086398
[Soler et al. 2020] Soler, M., Colmenero, J.M., Pérez-Contreras, T. & Peralta-Sánchez, J.M. (2020). Replication of the mirror mark test experiment in the magpie (Pica pica) does not provide evidence of self-recognition. Journal of Comparative Psychology, 134(4), 363–371. https://doi.org/10.1037/com0000223