Commentary

Self-recognition in primates: irreverence, irrelevance and irony

Recent mirror studies with two corvid species have reported contrasting findings. Jungle crows, Corvus macrorhynchos, showed no self-contingent behaviour when confronted with mirrors, whereas Eurasian magpies, Pica pica, reportedly passed the ‘mark’ test for self-recognition. We investigated mirror-induced behaviour in wild-caught New Caledonian crows, Corvus moneduloides. We first documented the response of 10 naïve crows to a 50 × 40 cm vertical mirror. The crows responded to their mirror image with social displays and engaged in search and mirror-directed exploratory behaviour. Their agonistic social displays towards the mirror did not decrease in frequency over time. We then gave two of these crows and two naïve ones a mirror-mediated spatial location task with a horizontal mirror. All four crows successfully used the horizontal mirror to locate hidden food. Therefore, they were able to exploit the correlation between an object’s mirror reflection and its location in the real world. This suggests that New Caledonian crows may also have the ability to develop an understanding of how mirrors represent objects in the environment, despite the lack of self-directed behaviour in front of mirrors. Our study fills an important gap in mirror studies on corvids, which are considered to be the primate equivalents of the avian world.

Primates undoubtedly have impressive abilities in perceiving, recognizing, manipulating, and predicting other individuals, but only great apes seem to recognize the cognitive basis of manipulative and cooperative tactics or the concept of self. None of these abilities is unique to primates. We distinguish (1) a package of quantitative advantages in social sophistication, perhaps based on more efficient memory, in which neocortical enlargement is associated with the challenge of social living; from (2) a qualitative difference in understanding, whose taxonomic distribution—including several distantly related species, including birds—does not point to an evolutionary origin in social challenges and may instead relate to a need to acquire novel ways of dealing with the physical world. The ability of great apes to learn new manual routines by parsing action components may have driven their qualitatively greater social skill, suggesting that strict partition of physical and social cognition is likely to be misleading.

In this paper we critically review conceptual and methodological issues of mirror self-exploration research. We conclude that: (1) mirror self-exploration provides evidence for mirror self-recognition; (2) inter- and intraspecies differences in terms of presence or absence of mirror self-recognition are not yet sufficiently clear, that is, whether its absence in species and individuals is methodological (false negative with respect to trait) or ontological (true negative with respect to trait) in nature; and (3) proposals of hypotheses to account phylogenetically for the existence of inter- and intraspecies differences in mirror self-recognition are premature. Before hypotheses such as these are advanced, the impact of rearing conditions on mirror self-recognition should be examined further.

An ethological analysis of emotional behaviour suggests that it consists of somatic and autonomic elements of behaviours that are thwarted or in conflict (unpleasant emotions). Such displays change the behaviour of social partners which may end the thwarting, whereupon there is a relaxation and switch to the intended consummatory behaviour (pleasant emotions). Self-perception through distance exteroceptors allows the individual to respond to its own emotional displays as if they were those of a social partner. This is the basis of the experience of emotion and of self-awareness and self-control. Vocal signals give the most veridical self-perceptions and are produced and shaped by emotional actions. Speech and words may have developed as vocal signals associated with and symbolizing social actions with particular objects. These symbols are self-produced and self-perceived so that sequential symbol use, i.e., thinking, becomes possible. Self-perception begins in infancy with visuo-kinaesthetic discriminations of bodily actions giving the `sensori-motor self'. Internal motivational states add hedonic value, i.e., give feelings and a `feeling self'. Self-perception of one's own emotional displays forms the `emotional self'. Finally, the self-produced symbols `me' and `you' establish the `cognitive self', and `I' the self-concept. Neural feedback processes, from the level of recurrent collaterals of single neurones to that of cortico-cortical fibres of the neocortex, are not in themselves sufficient for self-awareness. Feedback from the actions of the individual to the neocortical sensory and perceptual systems is required if a discriminated self is to be represented in the association areas. These representations interact with a language system which associates `me' with the self-representation and `I' with the cortical and subcortical processes of perception, action and motivation to give full awareness of the self. Self-awareness in domesticated animals, as in other vertebrates, depends on the degree of external self-perception possible, on social organisation, on social communication through emotional signalling, and on acquisition of a symbolic language system. The phylogeny of self-awareness is briefly considered in terms of levels of awareness, from sensori-motor to cognitive self-awareness, and the evolution of symbolic vocal communication enabling a concept of self and a knowledge of this concept (i.e., cognitive self-awareness). The development of self-awareness in the human infant is similarly considered. Self-perception of emotion may be the basis of enjoyment and suffering but language may be necessary for a knowledge of these feelings and for continued suffering in the absence of the evoking stimulus or associated stimuli. The implications of this analysis for the welfare of domesticated animals are indicated.