The Nature of Visual Self-Recognition
Total Page:16
File Type:pdf, Size:1020Kb
Review The nature of visual self-recognition Thomas Suddendorf and David L. Butler School of Psychology, University of Queensland, Brisbane, Qld 4072, Australia Visual self-recognition is often controversially cited as an The mirror mark test indicator of self-awareness and assessed with the mirror- Gallup exposed chimpanzees to a mirror before placing an mark test. Great apes and humans, unlike small apes and odorless, tactile-free mark on the uppermost portion of monkeys, have repeatedly passed mirror tests, suggest- their eyebrow ridge and ear while they were anaesthetized. ing that the underlying brain processes are homologous Upon recovery, a mirror was reintroduced and their be- and evolved 14–18 million years ago. However, neurosci- havior observed. Self-directed responses to the marked entific, developmental, and clinical dissociations show areas significantly increased compared to control condi- that the medium used for self-recognition (mirror vs tions without the mirror. Subsequent replications, typical- photograph vs video) significantly alters behavioral and ly using surreptitious marking rather than anesthetics, brain responses, likely due to perceptual differences confirmed that chimpanzees can recognize themselves in among the different media and prior experience. On the mirrors [10]. Many species have subsequently been tested basis of this evidence and evolutionary considerations, (Box 1), but only the closest relatives of chimpanzees we argue that the visual self-recognition skills evident in (humans, gorillas, and orangutans) have so far provided humans and great apes are a byproduct of a general independently replicated evidence of passing the mark test capacity to collate representations, and need not index [11,12]. When using carefully matched criteria, chimpan- other aspects of self-awareness. zee and human infants develop the capacity for mirror self- recognition in a similar manner [13]. Reflecting on visual self-recognition Although some signs of self-recognition, such as famil- Visual self-recognition has long fascinated scholars partly iarity with one’s image, are evident in two- to three-month- because of its intuitive appeal as a potential indicator of self- old infants [14], the mark test is not typically passed until awareness. Many humans regularly spend time in front of over a year later. Using rouge or stickers to surreptitiously mirrors and invest efforts into improving their looks. Vari- mark toddlers, researchers have repeatedly found that ous other species adjust their appearance to impress poten- children begin to pass the test from 15 months of age; tial partners and opponents or to camouflage, but they do not by 24 months most children pass [15–18]. Some diversity in seem to take advantage of reflective surfaces in their efforts pass rates has recently been reported from different cul- to do so. Fish sometimes show aggressive behavior towards tures [19]. In one study on 16–21-month-olds, urban Ger- their mirror image, although their brain responses differ man and Indian children passed the test earlier than rural compared to when they fight a real fish, which suggests that Indian and Nso children, leading the authors to conclude they recognize something unusual about the mirror [1]. that this reflects an autonomy-supporting cultural context Numerous species, including pigs [2] and New Caledonian in the former and a relational cultural context in the latter crows [3], can use mirrors to find hidden objects. Great apes [20]. Nonetheless, the basic ability seems to be universally can even be observed using mirrors to examine body parts, acquired in toddlers. For instance, Bedouin children, even such as their anal region, that they cannot otherwise see and without previous experience with mirrors, were found to their understanding was confirmed through a now widely develop competence in the same way as Israeli children used mirror mark test first developed by Gordon Gallup over familiar with mirrors [21]. 40 years ago [4]. In recent years, significant research efforts have been devoted to understanding the development, What does the mirror mark test measure? neuro-cognitive basis, and disorders of visual self-recogni- In spite of its intuitive appeal, it has been controversial tion [5–9]. Here, we highlight recent conceptual and meth- what the mirror mark test indicates [6,22–25]. Gallup odological issues that are critical to interdisciplinary argued that it measures self-awareness because one needs integration. In particular, it has become evident that to be able to become the object of one’s own attention to behavioral and brain responses can vary considerably pass the test. He went on to argue that the task implies a depending on the medium used to present images of self. self-concept, a capacity for introspection, theory of mind, We argue that these findings, when integrated with devel- and awareness of death [24]. In support of such rich opmental and comparative data, suggest that successful interpretations, researchers have documented associa- performance in self-recognition tasks draws on a broader tions between the onset of mirror self-recognition and other capacity to collate representations, rather than on general, purported indicators of self-awareness, such as use of context-independent self-awareness. personal pronouns, empathy, synchronic imitation, and embarrassment [20,26]. However, there is no direct evi- Corresponding author: Suddendorf, T. ([email protected]). dence that supports links between self-recognition and Keywords: self-awareness; phylogenetic reconstruction; ape; child; mirror; self-recognition. death awareness or introspection. 1364-6613/$ – see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tics.2013.01.004 Trends in Cognitive Sciences, March 2013, Vol. 17, No. 3 121 Review Trends in Cognitive Sciences March 2013, Vol. 17, No. 3 Box 1. Mirror self-recognition in diverse species that toddlers are equally capable of recognizing their legs in a mirror as they are at passing the standard task, Many other primate species have been tested with this paradigm, undermining theories that have placed special emphasis but consistently fail [11]. Although monkeys may not confuse their mirror image with a social other [61], they have repeatedly failed the on cognition about faces [28]. In another two conditions, formal mark test, even after extensive exposure to mirrors [11,62]. participants were slipped into baggy tracksuit trousers However, occasional high-profile claims have been made about self- that were attached to the highchair and then presented recognition in animals. First, pigeons were conditioned to peck on with a mirrored view of their legs. In one of these condi- their own body in front of mirrors [63], but this behavior was not spontaneous and required extensive training. Curiously, such tions, before the legs were marked with a sticker, a tray conditioning has not been successful when employed with capuchin that had blocked the direct view of their legs was removed monkeys [64]. Next, cotton-top tamarins were claimed to have for 30 seconds, in the other it was not. Although the passed the test when their distinctive white hair was died red [38]. toddlers saw the same mirrored image of baggy trousers However, these results were heavily critiqued [65] and could not be in both conditions, they performed poorly without direct replicated by the original lead author [66]. Then, two bottlenose dolphins were argued to recognize themselves [67], but, because of exposure and performed as well as in the standard task their lack of hands, the dependent variable was not reaching for the when they had the brief opportunity to view what they mark as required in the standard task. Given their brain size and the were wearing. These results strongly suggest that young frequency with which they naturally see their reflections as they children form a mental expectation of what they look like jump out of the water, it would not be surprising if dolphins were and can do so rapidly [16]. capable of visual self-recognition. Replications of this study would be highly desirable. A few years later one of three Asian elephants The mirror mark test, therefore, measures more than tested was claimed to have passed the mirror mark test by reaching Heyes’ lean account proposes. What children saw in the with its trunk towards the mark [68]. However, in an earlier study, mirror was identical in these last two conditions of the leg elephants had failed [69], which highlights the need for independent recognition study, yet they passed only when they had an replication. This is also required for the claim that two out of six opportunity to update their expectations about their phys- tested magpies passed the mark test [70]. Although assertions about potential competences in various species continue to be made [71], ical appearance. So there is more to passing the task than they frequently fall short when examined critically [62]. Careful distinguishing feedback from other sensory input. In a replications are hence essential. Only the great apes have thus far sense, those who pass may be said to be self-aware about provided repeatedly replicated evidence of passing the mark test what they currently look like. However, this need not mean [11]. Even among the great apes, not all individuals pass. It is possible that some animals do not have the capacity (e.g., because that they are self-aware in other respects, as Gallup’s rich of immaturity or old age) [72] or fail because of differences in account conjectures – nor even that they are entirely attention, affect, or motivation [13]. Furthermore, experiments vary realistic with their expectations (Box 2). Individuals can in terms of methodology and precise criteria [13]. clearly be self-aware about one facet (e.g., an aspect of one’s personality), while being completely ignorant about anoth- er (e.g., some weaknesses). Rich interpretations of the task have been repeatedly In sum, we recommend staying close to the data when challenged by leaner accounts.