Marc Bekoff

Researchers are interested in animalawareness because they are curious todiscover what animals might knowabout themselves. There are, however, long-held and polarized views about the degree of self-awareness in animals. Some peoplebelieve that only great apes have rich notionsof self knowing who they are and/or having a theory of mind, which means beingable to infer the states of minds of others whereas others argue that it is methodologi-cally too difficult to address this questionbecause animal (like human) minds are sub-jective and private. Many in this latter categorydo not attribute any sense of self to animalsother than humans, and some, dismissingbehavioural and neurobiological research onanimal cognition, wonder whether animalsare conscious of anything at all.

What might animals know about them-selves? Most studies of animal self-awarenesshave been narrowly paradigm-driven. Thered spot technique was first used by GordonGallup to study animal self-awareness inchimpanzees; it and variations have beenused on great apes and monkeys, as well as ona few dolphins and elephants. For primates, aspot is placed on the forehead of an anaes-thetized individual and self-directed move-ments towards the spot are scored after he orshe awakens and catches sight of themselvesin a mirror, a high score indicating the pres-ence of some degree of self-awareness. But insome cases, the data are derived from tests onsmall numbers of individuals, many of whomfail it because they do not make self-directedmovements towards the spot. Those whopass the test might not be representative ofwild relatives because they have had extensive

human contact and previous experience withmirrors, factors that might influence theirtrainability and willingness to use a mirror.Those who fail the test might show somesense of self in other contexts, and otherindividual differences might also play a role.

Laboratory experiments can overlook thefact that many animals have evolved in com-plex webs characterized by variable social, sensory and other environmental features. Asingle technique based solely on visual cues is not the only valid test of self-awarenessbecause many animals regularly use sensorymodalities other than vision, either singly ortogether. Most wild animals do not know whatthey look like although birds can acquireinformation about their plumage colour fromreflections in drinking water but they seemto know what they sound or smell like. Manybirds distinguish their own from others songsand numerous mammals discriminate theirown from others scents. Some rodents useodour to distinguish self from other; theypreferentially associate or play with those who smell like them, or shun individuals whosmell like them to avoid inbreeding.

Discriminating between own andother is not necessarily an indication of self-awareness, but in most cases we dont know ifit is or isnt. Although no one has suggestedthat birds or rodents are self-aware to thedegree that chimpanzees might be, we do not know precisely what these animals know about themselves. We must be carefulneither to imbue animals with unknowncognitive capacities nor to rob them of skillsthat they might possess.

Researchers often dismiss non-primateswhen they investigate self-awareness. Socialcarnivores, such as grey wolves, who live inclosely knit, orderly and efficient packs, couldturn out to have well-developed self-aware-ness. Wolves engage in coordinated hunts andrear their young communally. Analyses of the behaviour patterns of individual wolves during these (and other) activities show thatthey have at least some awareness of what theyare doing, what others are doing, and whereeach is located spatially. Similar observationshave been compiled for other species.

Comparative data suggest that if we relysolely on mirror studies, we have been look-ing for self in the wrong places and on thewrong faces. The evidence also indicates thatno other animal has the same degree of self asthat possessed by most humans (an I-nessor I-self , perhaps mediated by language).On the basis of traditional measures, manyanimals appear to have a sense of body-nessor mine-ness, but not a sense of I-ness.

Answers to challenging questions aboutself-awareness have wide-ranging conse-

quences because they are often used byresearchers and lawyers as a litmus test fordefending the sorts of treatments to whichanimals can be ethically subjected. However,it is not clear that self-awareness or other cognitive capacities should be used for suchdecisions. Some argue that a sense of I-nessis morally relevant and necessary for experi-encing pain. But even if an animal does notknow who she is, this does not mean shecannot feel that something painful is hap-pening to this body. Even though the experi-ence of pain might not be the same acrossspecies, individuals of different species canstill suffer their own type or version of pain.Self-awareness is not a reliable test for anobjective assessment of well-being.

The concept of animal self-awarenessremains open to different interpretations,but we will probably learn more about themysteries of self and body-ness by usingnon-invasive neuroimaging techniques incombination with cognitive ethological stud-ies. If we look at self-awareness as body-awareness, we might also discover moreabout how animals think and the perceptualand neurobiological processes underlyingvarious cognitive capacities. Darwins ideasabout evolutionary continuity, together withempirical data (science sense) and commonsense, caution against the unyielding claimthat humans and perhaps other great apesand cetaceans are the only species in whichsome sense of self has evolved. nMarc Bekoff is in the Department of Environmental,Population, and Organismic Biology, University ofColorado, Boulder, Colorado 80309-0334, USA.

FURTHER READINGBekoff, M. Minding Animals: Awareness, Emotions, andHeart (Oxford Univ. Press, New York & London, 2002).Bekoff, M., Allen, C. & Burghardt, G. M. (eds) The Cognitive Animal: Empirical and TheoreticalPerspectives on Animal Cognition (MIT Press,Cambridge, Massachusetts, 2002); see especiallyessays on self-awareness by Gallup, G. G., Anderson,J. R. & Shillito, D. J.; Mitchell, R. W.; Shumaker, R. W. & Swartz, K. B. Mitchell, R. W. in Handbook of Self and Identity (edsLeary, M. R. & Tangney, J.) 567593 (Guilford, NewYork, 2002). Reiss, D. Nature 418, 369370 (2002). Rilling, J. K. et al. Neuron 35, 395405 (2002). Rilling, J. et al. Am. J. Phys. Anthropol. Suppl. 30,263 (2000).Rilling, K. et al. Biol. Psychiat. 49, 146157 (2001).

Animal reflectionsconcepts

NATURE | VOL 419 | 19 SEPTEMBER 2002 | www.nature.com/nature 255

AwarenessOn traditional measures, manyanimals appear to have a sense of body-ness or mine-ness, but no sense of I-ness.

Social work: is the pack behaviour of grey wolvesindicative of some degree of self-awareness?

L. K

EN

NE

DY

/CO

RB

IS

2002 Nature Publishing Group