Absence of Mirror Self-Referential Behavior in Two Asian Elephants
Moti Nissani, Donna Hoefler-Nissani
Interdisciplinary Studies Program, Wayne State University, Detroit, Michigan 48202, USA
Abstract: To date, one investigation failed to find mirror self-referential behavior in Asian elephants while two others reported positive results, a contradiction which could, among other things, be ascribed to the poor visual acuity of elephants. To throw additional light on these contradictory reports, the present study of mirror self-referential behavior in two captive Asian elephants bypasses the traditional mark test, relying instead on the elephants’ response to a far more visually conspicuous object, and on prior habituation to the presence of this object in their enclosure. In this study, neither elephant engaged in self-referential behavior in front of a mirror. Our simple experimental paradigm could serve as a more convenient alternative to the widely used traditional mark test, could meet some methodological objections which have been raised against the traditional mark test, and could profitably augment the traditional test in difficult or controversial cases.
Although elephants have been in close association with humankind for thousands of years, and although anecdotes about their wisdom and witlessness are many, they have been subject to a mere handful of psychological experiments (for recent reviews, see Nissani, 2004, 2006; Nissani et al., 2005).
This curious gap in the research literature provides a fertile ground for speculations and controversies. Indeed, views on the subject range from the assertion that the elephant is “exceedingly intelligent” (e.g., Sanderson, 1962, p. 292) to the assertion that it “is a stupid animal” (e.g., Sanderson, 1912, p. 78, see also Carrington, 1958, p. 73 et. seq.). Both sides cite anecdotes in support of their views. Likewise, behavioristically-inclined observers cite the elephant’s comparatively small brain, while observers with a more cognitive persuasion note the comparable rate of post-natal brain growth in humans (73%), elephants (65%), and chimpanzees (46%) (Eltringham, 1982). Some field studies likewise support the cognitive view; for instance, to date, the most extensive networks of vocal recognition in animals have been reported for African elephants (McComb et al., 2000). The few available laboratory studies, on the other hand, appear more consistent with the pessimistic view. Thus, one elephant took 330 trials to master a simple discrimination task (Rensch, 1957), while two others, according to a fragmentary report (Stevens, 1978), took more than a 1000 each.
In a 1989 study of two Asian elephants, Povinelli (1989) reported that they were capable of a sophisticated form of mirror-guided reaching behavior, but that they failed the mark test and responded to their mirror image as if confronted by another elephant. Several lines of evidence, however, put this conclusion in doubt. Thus, another Asian elephant passed the mark test (Simonet, 2000; Simonet et al., 2000). More recently, Plotnick et al. reported that one of three Asian females at the Bronx Zoo passed the mark test as well. African elephants in the wild likewise seem to possess “a sense of self” (Poole, 1996, p. 165). Not all chimpanzees show evidence of mirror self-recognition (Swartz & Evans, 1991; Anderson, 1996), and Povinelli’s negative results with elephants may thus be ascribable to the small number of subjects which is typically relied upon in elephant behavioral research. Elephants probably lack color vision (Rensch & Altevogt, 1953; Lair, 2001), and thus may be oblivious to color changes caused by the traditional mark test. Elephants routinely and habitually spray themselves with mud, leaves, and other materials. When wet, their skin noticeably darkens. Their size is enormous. Thus, elephants must be used to regular, and seemingly radical, changes in their companions’ appearance and they may be predisposed to disregard the comparatively slight changes in appearance which are typically introduced by the mark test (Anderson & Gallup, 1999). Recent studies of bottlenose dolphins (Reiss & Marino, 2001) suggest that the search for mirror self-recognition may be fruitfully extended beyond primates. Preliminary observations by our group (Nissani, 2004) suggest that elephants outperform chimpanzees in a variation of Povinelli & Eddy’s (1996) seeing experiments. These diverse lines of evidence make it worthwhile to subject the question of mirror self-referential behavior in elephants to additional scrutiny.
The experiments reported here were undertaken as part of a comprehensive research program on the mentality of the two Asian elephants of the Detroit Zoological Institute, USA. According to Zoo records, Winky, the older and friendlier of the two, was wild-born in Cambodia, probably in 1952, and acquired by the Sacramento Zoo in 1955. After an incident involving the death of another elephant, she was moved to Detroit in 1991. Wanda was wild-born in India in 1958 and moved to the USA around 1960. Before moving to Detroit in 1994, she had many owners. She is said to have suffered abuse at the hands of one owner, abuse which may have caused serious infection and partial amputation of her right ear. Wanda has reduced trunk mobility, and is unable to touch her forehead. Although both elephants were kept in protective contact, only Wanda is considered dangerous to humans. During their tenure at the Detroit Zoo, to the best of our knowledge the elephants have never been exposed to a mirror. Nothing is known about mirror exposure before coming to the Detroit Zoo.
Observations were either made in the elephants’ well-lit indoor enclosure or in the cable-enclosed courtyard inside the elephant one-acre outdoor compound. In either case, the elephants were prevented from destroying the mirrors by metal cables or bars. Written records were kept and critical phases of the experiment were filmed with a video camera. The video records were then entered independently into the computer, and the film and written records were compared for interobserver reliability.
In the outdoor enclosure, we used a 200 X 120 cm glass mirror. In some cases, the mirror was just outside trunk’s reach, while in others the elephant could actually touch its surface. The mirror was placed in a near-vertical position, leaning against a wooden structure. Care was taken to place the mirror so that an elephant could see herself and her forehead, and this was monitored throughout the experiments by looking at the mirror from the elephant’s position and height. If an elephant moved from this verified range, she was enticed back with a favorite morsel (bagels, marshmallows, or carbonated drinks). In the indoor enclosure, the same mirror was used, but was accompanied by a smaller (58 X 96 cm) mirror, placed some 5 m away from the edge of their enclosure and 1 m above the floor
Because we wished to monitor in this experiment the elephants’ response to their own changed appearance (that is, wearing a turkey feather on their head; see Fig. 1)—and not merely to strange objects which happened to be attached to their bodies—the period of mirror habituation consisted of habituation to white turkey feathers as well. Such feathers were scattered on the ground in the elephants’ indoor and outdoor enclosures and glued or taped to the cables, bars, and walls of their indoor and outdoor living quarters.
The critical tests consisted of three daily sessions. In turn, each daily session was comprised of a three-part sequence:
In Part 1 of each daily session, the elephant was exposed to the mirror for at least ten minutes and her reactions to it were closely monitored. This part was aimed at confirming earlier observations that by now the elephants were habituated to the mere presence of a mirror, and served as a baseline for subsequent responses to changes in their own appearance (changes caused by the later addition of a turkey feather to their forehead).
In Part 2 of each of the three daily sessions, the mirror was covered for a period of 10-30 minutes, to allow us to attach a feather with duct tape to each subject’s head and to monitor her reaction, before introducing her to the mirror itself. As well, during Part 2 of the second daily session (but not the first and third), an 8-inch white circle was drawn above the feather of one elephant and a cross of a similar size was drawn above the feather of the other, in an effort to supplement the feather test with the more usual mark test.
In the critical Part 3 of each daily session, a part which lasted 8-20 minutes, the mirror was uncovered and the elephant’s response observed and filmed in an effort to detect any self-directed actions. If the animal did not come to the mirror on its own, or started moving away, she was enticed to come near it and maintain a position where she could see herself, and her feather, in the mirror.
Before launching our experiment, more information was required about our elephants’ seeing ability. Elephants are known to be more dependent on olfaction than vision, to be often active at night in the wild, and there are even reports from the wild of a blind matriarch competently leading her troop (Groning & Saller, 1999, p. 78). Similarly, in an unrelated task involving insertion of a trunk into a 6-inch tube to obtain food, both elephants repeatedly found the hole by groping with their trunk (and through vision), taking 2-10 seconds in each trial. Moreover, the vision of elephants may deteriorate with age (Stevens, 1978). Fortunately, other lines of evidence suggested that our two elephants had adequate vision. Veterinary records indicated that Wanda had full vision, while Winky had limited vision in the right eye and normal vision in the left. Both elephants immediately retrieved distant food items attached to nearby ropes and chains ranging in diameter from 1 to 4 cm. When 1.4 cm3 (4g) white sugar cubes were tossed near them, they appeared to see the movement, looked in the direction of the rolling cubes, and retrieved them with ease. Finally, although Wanda and Winky daily respond to a variety of commands, each consisting of both arm movements and sounds, in ten trials with each (five in one session and five in another about one month later) they obeyed immediately 19 of 20 commands when only the gestural component of each was used, again suggesting that their vision was good enough to detect their reflection in the mirror (see also more recent observations on the visual acuity of Burmese elephants, Nissani et al., 2005).
We first introduced both elephants to their reflection by placing the 200 X 120 cm glass mirror on May 22, 2001 in the open-air courtyard. We continued with this preliminary baseline exposure for five additional sessions, for an approximate total exposure time of about 7 hours for Winky (the more dominant of the two) and 4 hours for Wanda. While observing the mirror, both elephants frequently extended their trunk towards it, wrapped their trunk around one of the cables which separated them from the mirror, manifested other unusual trunk movements, shook their head, and placed a leg on the lower cable of their enclosure. Throughout this first baseline period, their behavior suggested that they viewed their reflection as something like another elephant; for instance, Winky’s trunk movements were similar to those seen in her interactions with Wanda. Overall, they reacted intensely to the mirror, but they did not manifest unusual postures or an exploration of otherwise invisible body parts (e.g., relying on the mirror to inspect their mouth). In view of their strong emotional reactions to the mirror (involving an attempt by Wanda to charge one of the researchers), exposure to the mirror was discontinued on June 5, 2001.
On August 25, 2001, the same mirror was re-introduced and positioned just outside trunk’s reach on the floor of their sleeping quarters. The smaller (58 X 96 cm) mirror was placed some 5 m away from their cage and 1 m above the floor. The elephants had visual access to both mirrors in early mornings and late afternoons, for a total of some four hours a day. Their behavior in the indoor enclosure was monitored only occasionally and showed declining interest in the mirror. Both mirrors stayed in the same location until September 30.
Following the introduction of turkey feathers into their compound, the elephants first showed mild interest, picking up the feathers occasionally and tossing them away. But from the very beginning, they always tossed it to the ground within 10 seconds or less of picking it up. New turkey feathers were added every 2-3 days. By September 27, the predetermined habituation criterion had been reached: More than half of the feathers which were placed in the outdoors enclosure the day before were still in the exact same spot.
We also checked the elephants’ response to the taping of a bent feather to their forehead (see Fig. 1). From the very first trial, and throughout the critical mirror tests which started on Sept 30, Winky showed no interest whatsoever in the feather on her forehead, as long as it remained in its proper place. On the first few seconds after taping a feather to her forehead for the first time, Wanda tried to lift her trunk in an apparent attempt to remove the feather, but soon gave up (owing to somewhat reduced trunk mobility, Wanda cannot touch her forehead). Neither elephant ever tried to remove the feather by rubbing it against cables or other available objects.
Fig. 1. In the absence or presence of mirrors, Winky often removes a taped turkey feather from Wanda’s forehead—but never from her own.
Each elephant wore a feather on at least four different daily sessions, for a minimum of four hours each. Apart from Wanda’s initial reaction (described above), neither showed an interest in her own feather. On three occasions, at the end of each session, while the two elephants were in close proximity to each other, Winky removed, and tossed away, Wanda’s feather (Wanda never removed Winky’s feather, probably because Wanda is never seen to lift her trunk to Winky’s forehead—a gesture of dominance). Likewise, on two occasions, Winky’s tape loosened with the feather dangling down, which she then promptly removed. In two additional sessions, lasting at least one hour each, only duct tape was placed on their foreheads, and again neither elephant showed any reaction to it.
In Part 1 of each of the three experimental daily sessions, the elephants showed little interest in the mirror, spending less than 5% of the time responding to it in any way (e.g., looking at it, extending their trunk towards it), owing probably to the preceding long period of mirror habituation (see Methods). In Part 2 of each experimental session, neither elephant evinced the slightest desire to remove the feather and duct tape, nor showed signs of being bothered by the them.
During the critical Part 3 of the first experimental daily session, Winky showed considerable interest in the mirror, remained within sight of the mirror even though she was not enticed to do so, and spent 16% of the time extending her trunk towards it. This was in sharp contrast to her reaction to the mirror in Part 1 of the same session, when she had to be enticed to stay near it, and when not even a single trunk extension was observed. In Part 3 of the second and third experimental daily sessions, she spent less than 3% of the time extending her trunk to the mirror. Throughout the 3 experimental daily sessions, Winky never directed her trunk towards herself nor engaged in other self-directed behaviors, e.g., trying to use the enclosure cable to rub the feather off her forehead while looking at the mirror.
In Part 3, second daily session, the feather had to be re-taped to Winky’s forehead within sight of the mirror, but this too failed to produce self-directed behavior. At the end of Part 3, third daily session, Winky’s trunk was guided towards the feather; upon touching the feather, she immediately removed it and tossed it away.
Wanda, in contrast, never extended her trunk to the mirror in all three parts of each daily session. Particularly, during the critical Part 3 of all 3 daily sessions, Wanda, like Winky, never directed her trunk towards herself nor engaged in other self-directed behaviors.
The marking procedure described here, a procedure which is particularly appropriate for animals with poor vision and in situations where anesthesia is inapplicable, may perhaps help to confirm or disprove mirror self-recognition in difficult or controversial cases (Anderson & Gallup, 1997; Hauser et al., 1995), and it may provide an alternative methodology for the study of self-referential behavior. Our results likewise raise the possibility that animals undergoing the mark test may at times be reacting to a novel, interesting, mark—and not to a change in their own appearance.
Our observations—which utilized a more conspicuous object than the traditional paint mark (Gallup, 1970), habituated the subjects to that object, and failed to observe self-directed behavior—are consistent with the view that most captive Asian elephants are incapable of self-referential behavior when confronted with a mirror. To resolve the contradictory results reported so far (Povinelli, 1989; Simonet, 2000; Simonet et al., 2000; Plotnik et al., 2006), additional experiments will have to be carried out. Ideally, such experiments would (i) involve a group of 10 or more healthy animals with excellent trunk mobility, (ii) make use of some younger subjects between the ages of 12 and 20 (such subjects may have better visual acuity than the older subjects used so far, and may enjoy greater behavioral flexibility, cf. Nissani et al., 2005), (iii) employ subjects which have spent their formative years in the wild or in semi-wild conditions, (iv) habituate the animals to the mark before placing that mark on them, (v) rely on more visually conspicuous objects than those that have so far been employed, (vi) carefully test the visual acuity of the subjects, and (vii) follow Plotnik et al’s (2006) elegant procedure of presenting elephants with a large mirror which they can tactilely explore.
We warmly thank Scott Carter, Ron Kagan, Erin McEntee, Bettie McIntire, Mary Mutty, Patti Rowe, Michelle Seldon, Kim Van Spronsen, and Rick Wendt for friendship and support, and the two anonymous reviewers for helpful comments.
Anderson, J. R. (1996). Chimpanzees and capuchin monkeys: Comparative Cognition. In: Russon, A. E., Bard, K. A. & Parker, S. T. Reaching into Thought. Cambridge: Cambridge University Press.
Anderson J. R. & Gallup G. G., Jr. (1997). Self-recognition in Saguinus? A critical essay. Anim. Behav. 54:1563-67
Anderson, J. R. and Gallup, G. G., Jr. (1999). Self-Recognition in Nonhuman Primates: Past and Future Challenges. In: M. Haug & R. E. Whalen. Animal Models of Human Emotion and Cognition, pp. 175-194.
Carrington, R. (1958) Elephants. London: Chatto & Windus.
Eltringham, K. (1982). Elephants. Dorset: Blandford.
Gallup, G. G., Jr. (1970). Chimpanzees: self-recognition. Science 167:86-87
Groning, K. & Saller, M. (1999). Elephants: A Cultural and Natural History. Cologne: Konemann.
Hauser, M. D., Kralik, J., Bott-Mahan, C., Garrett, M, Oser, J. (1995). Self-recognition in primates: phylogeny and the salience of species-typical traits. Proc. Natl. Acad. Sci. USA 92:10811-14.
Lair, R. 2001. An interview on May 30, All Things Considered, National Public Radio, USA.
McComb, K., Moss, C., Sayialel, S. & Baker, L. (2000). Unusually extensive networks of vocal recognition in African elephants. Animal Behaviour, 59, 1103-1109.
Poole, J. (1996). Coming of Age with Elephants. New York : Hyperion.
Nissani, M. 2004 () Theory of Mind and Insight in Chimpanzees, Elephants, and Other Animals? Chapter 7 of: Comparative Vertebrate Cognition. Rogers, Lesley J & Kaplan, Gisela (eds.), pp. 227-261. [No. 4 of Series: Developments in Primatology: Progress and Prospects. Russell H. Tuttle, Series Editor.] Kluwer Academic / Plenum Publishers: New York.
Nissani, M. (2006). Do Asian elephants apply causal reasoning to tool use tasks? Journal of Experimental Psychology: Animal Behavior Processes, 31: 91-96.
Nissani, M., Hoefler-Nissani, D. Lay, U T., & Htun, U W. (2005). Simultaneous Visual Discrimination in Asian Elephants. Journal of the Experimental Analysis of Behavior, 83: 15-29.
Plotnik, J. M., de Waal, F. B. M., & Reiss, D. Self-recognition in an Asian elephant. Proceedings of the National Academy of Sciences, USA, vol. 103, November 7, 2006 (www.pnas.org_cgi_doi_10.1073_pnas.0608062103).
Povinelli, D. J. (1989). Failure to find self-recognition in Asian elephants (Elephas maximus) in contrast to their use of mirror cues to discover hidden food. Journal of Comparative Psychology, 103 (2) 122-131.
Povinelli, D. J. & Eddy, T. J. (1996). What young chimpanzees know about seeing. Monographs of the Society for Research in Child Development (Vol. 61, No. 2, Serial No. 247).
Rensch, B. (1957). The Intelligence of Elephants. Scientific American, 196, 44-49.
Rensch, B. & Altevogt, R. (1953). Visuelles Lernvermoegen eines indischen Elefanten (Capacity for visual learning in an Indian elephant). Zeitschrift-fuer-Tierpsychologie, 10: 119-134.
Reiss, D. & Marino, L. (2001). Mirror self-recognition in the bottlenose dolphin: A case of cognitive convergence. PNAS 98: 5937-5942.
Sanderson G. P. (1912). Thirteen Years among the Wild Beasts of India. Edinburgh: J. Grant.
Sanderson, I. T. (1962) The Dynasty of Abu. New York: Knopf.
Simonet, P. (2000). Self-recognition in Asian elephants, preliminary findings. Elephant, 2 (4), 103.
Simonet, P, Myers, M., Lamoureux, L., Quashnick, J., Jacobs, J., & Willison, K. (2000). Social and Cognitive Factors in Asian Elephant Mirror Behavior & Self-Recognition. Unpublished manuscript.
Stevens, V.J. (1978). Basic operant research in the zoo. In: Markowitz. H. & Stevens, V. J. (eds) Behavior of Captive Wild Animals, Chicago: Nelson-Hall, pp. 209-246.
Swartz, K. B., & Evans, S. (1991). Not all chimpanzees (Pan troglodytes) show self-recognition. Primates, 32, 483-496.