The Dolphin in the Mirror

by Diana Reiss

  Very soon the babies became even more adventurous, embarking on more extensive forays on their own, which led to a game of chase; the mothers would swim after the calves, corral them, and bring them back. Before long, the babies learned how to circle back on their own, and the chase developed into what I called a figure-eight swim, which is exactly what it sounds like. The big adult and the diminutive calf chased each other in a figure-eight path, occasionally porpoising, or arching clear of the water in repeated jumps. During the chase the pursuing mother would sometimes approach her calf from behind and nuzzle his genital area with her beak.

  Next, Delphi and Pan, in their initial and continued forays with each other during their first six months of life, did the same tango sequence with each other—the chase, the figure-eight swim, the dead man's float, the beak-genital nuzzling. Eventually, it culminated in what looked like sexual behavior. One of the boys would approach the other from behind and below and attempt to copulate with him. Dolphins copulate belly to belly. The boys would often switch roles, one pursuing, the other pursued.

  The light bulb went on in my head. From the standpoint of dolphin social behavior, this pattern matched the components of dolphins' courtship and reproductive behavior. The male chases the female; they swim in figure eights; she floats on her side; he nuzzles and then enters her. In short, young bottlenose dolphins appear to learn this socio-sexual behavior through interactions with their moms. And in the case of Delphi and Pan, it was natural that they would interact in the same way with each other.

  Growing up in close intimacy, Delphi and Pan were buddies. They had a strong alliance, interacting and playing together in many ways. During the summer months, when it got hot in the Bay Area, things got hot in the pool at Marine World too. Delphi and Pan engaged in these sexual behaviors quite frequently, and in every part of their pool. Bottlenose dolphins show among the highest incidence of same-sex sexual behavior of any animal. Bonobo chimpanzees are much more famous for it, but dolphins match them in this department. Same-sex sexual activity is very common in nature, much more than most people realize.2

  In both dolphins and bonobos, sexual activity of this sort provides behavioral glue in interpersonal relationships; in larger groups, it provides intergroup cohesion. Male dolphins form long-term alliances and work together cooperatively in many activities, including foraging and mating. This sociosexual behavior between male dolphins helps them form and maintain bonds; when performed with receptive female dolphins in estrus, it functions as reproductive behavior. Delphi and Pan's display of what looked like sexual behavior was actually social behavior—they may have been practicing, but they were also growing close.

  Long before they reached age seven, Delphi and Pan had engaged in this chasing, soliciting, and intromission attempts. What caused me to exclaim out loud that summer day in 1990 was something I had not seen before. Delphi and Pan had very deliberately positioned themselves in front of a large mirror we had recently installed in the pool. It seemed to me they were clearly and intently watching themselves in flagrante delicto.

  ***

  We installed the mirror in the pool as part of a very serious scientific experiment. While I hoped the dolphins would like the mirror and find it engaging, the real reason for the mirror was an experiment simple in concept yet profound in implications for our understanding of dolphin mind and human mind. We wanted to find out if Delphi and Pan would recognize that what they saw in the mirror was a reflection of themselves, not just other dolphins. If they could, it would tell us that bottlenose dolphins possessed a level of self-awareness that is rare in nature. Not so very long ago, it was assumed to be the sole province of Homo sapiens.

  Self-awareness—the capacity to have a concept of self and to know that one exists as an individual being—was long seen as a cognitive Rubicon: humans in supreme isolation on one side, specially endowed, and the rest of "base nature" on the other, possessed of brains of various sizes but with an absence of minds. It's a tempting thought for those who view Homo sapiens as the pinnacle of biological evolution, with cognitive competences far beyond and distinct from any other animal's. Conversely, for those who hold the view that humans are just one of many branches of evolution, with extraordinarily developed minds but not divinely ordained uniqueness, it is less shocking to find overlaps with other types of minds.

  In 1970, Gordon Gallup Jr., then at Tulane University in New Orleans, upset the comforting, self-serving picture of human uniqueness when he demonstrated that the common chimpanzee, Pan troglodytes, also possesses a degree of self-awareness.3 Gallup had shown that chimpanzees can recognize their images in a mirror as themselves. Gallup, who had long been a skeptic of "animal mind," was inspired to perform the mirror experiment by the account of a visit by Charles Darwin to the London Zoo in late March 1838. Jenny, a newly arrived orangutan (the zoo's first), was attracting a lot of public attention. Darwin wanted to see for himself. A few days later, he wrote the following in a letter to his sister, Susan Elizabeth Darwin:

  "The keeper showed [Jenny] an apple, but would not give it her, whereupon she threw herself on her back, kicked & cried, precisely like a naughty child.—She then looked very sulky & after two or three fits of pashion [sic], the keeper said, ‘Jenny if you will stop bawling & be a good girl, I will give you the apple.'—She certainly understood every word of his, &, though like a child, she had great work to stop whining, she at last succeeded, & then got the apple, with which she jumped into an arm chair & began eating it, with the most contented countenance imaginable."4

  Darwin was impressed by Jenny's humanlike behavior, writing in his notebook of the day, "Man in his arrogance thinks himself a great work, worthy the interposition of a deity. More humble and I believe true to consider him created from animals." Darwin visited Jenny at the zoo two more times and wrote that she appeared to be "astonished beyond measure" when she saw her reflection in a mirror. Gallup recognized the significance of this little anecdote, and he devised what has since become known as the mirror test. Whether Jenny had indeed recognized her own reflection in the mirror that day, Gallup determined to find out if chimpanzees possessed the intellectual capacity to do so.

  When an animal first sees itself in a mirror, it usually explores the mirror itself, touching it, looking behind it and over it. Then it typically behaves as if the reflection is another individual and it tries to engage it in social behavior of one kind or another, sometimes being aggressive, sometimes trying to elicit play. Gallup put a mirror in front of each of four chimps, two preadolescent females and two males, who all did exactly this at first. Very soon, though, the chimps started to engage in another type of interaction with the image, what is called contingency testing. Each chimp appeared to realize that what it did had some influence on what the chimp in the mirror did. It bobbed up and down, turned its head from side to side, that kind of thing, all the time intently monitoring what the "other chimp" was doing.

  Relatively quickly, within three days, each of Gallup's chimps seemed to figure out that the chimp in the mirror was in fact itself, and they started to use the mirror as a tool. They would "[groom] parts of the body which would otherwise be visually inaccessible without the mirror, picking bits of food from between the teeth while watching the mirror image ... making faces at the mirror, blowing bubbles, and manipulating food wads with the lips by watching the reflection."5 From this point on, each chimp stopped all attempts at social interaction with the image in the mirror.

  To any reasonable person this self-directed behavior, using the mirror as a tool to inspect themselves, is compelling evidence that the chimps recognized the reflections as themselves and thus were self-aware. But as a scientist, rather than just a "reasonable person," Gallup felt he needed to go one step further to prove beyond doubt that the chimps' behavior was what it seemed. An ultraskeptic could always argue, for instance, that the chimps' apparently self-directed behaviors were simply sophisticated forms of interaction with the chimp in the mirror. So, as Gallup
wrote in his paper in Science, "In an attempt to add direct experimental support to the idea of self-recognition of the reflected image," he carried out what came to be called the mark test.

  Gallup anesthetized the chimps and used a simple, odorless red dye to mark two spots on each: the uppermost portion of an eyebrow ridge and the top half of the opposite ear. After each chimp regained consciousness, Gallup observed its behavior for thirty minutes and noted how often, if at all, it touched the marked areas. Then he positioned the mirror close to the chimp's cage and watched its behavior. While the chimps had paid no attention to the marked areas before the mirror was reintroduced, they were now deeply curious, repeatedly touching the marked spots, sometimes inspecting their fingers, sometimes smelling their fingers after they had touched the marks. Clearly, they knew who was who. Gallup claimed that this was the "first experimental demonstration of a self-concept in a subhuman form."

  By way of contrast, Gallup put three species of monkey through these same experimental hoops. None of them showed any indication that it thought the mark on the animal in the mirror had anything to do with itself. Moreover, unlike the chimps, the monkeys never stopped trying to interact socially with the image in the mirror, continuing to think it was another individual. "Our data suggest we may have found a qualitative psychological difference among primates," Gallup wrote in his now classic paper, "and that the capacity for self-recognition may not extend below man and the great apes."6

  We had introduced the mirror into the dolphin pool at Marine World in Vallejo in July 1990 to test whether mirror self-recognition (MSR) was indeed the sole province of large-brained primates. In the two decades since Gallup had published his original study, orangutans, bonobos, and, to a less dramatic degree, gorillas had also demonstrated mirror self-recognition. Old- and New-World monkeys continued to fail the mirror test. In the evolutionary lines closest to mankind, a threshold had been drawn.

  Why not bottlenose dolphins? After all, we knew by then that dolphins showed comparable abilities with the apes on other cognitive tests. And you don't have to be in close proximity to dolphins for very long to get the strong impression that they are indeed self-aware. Despite my scientific intuition on that matter, the notion that dolphins could exhibit MSR ran counter to the accepted idea of a single mountain of "higher intelligence," with humans at the summit. But the origin of human intelligence—including the capacity for spoken language and higher levels of consciousness—was typically cast as an enhancement of capacities that existed only in the common ancestor of humans and the great apes. From this "continuity" perspective, the intellectual capacities of Homo sapiens were viewed as superior to those of our closest primate cousins in quantity but not in kind. Such thinking was Darwinian—none of these scientists argued for divine intervention—but it was self-centered. Evolution was seen as a scaffold for humans, building on the kind of mental machinery that also exists in chimpanzees, gorillas, and orangutans and extending it to the levels we humans enjoy today. The continuity perspective is a primate-centered lens to explain why humans have the cognitive skills we do. It posits a cognitive gap between humans and great apes, and then between great apes and the rest of the animal world.

  My intuition that Delphi and Pan were indeed self-aware was, of course, not acceptable as scientific evidence. The boys had to pass the mirror test. And to do that, they had to pass the mark test. We faced one blindingly obvious problem with the test: unlike primates, dolphins have no hands with which to touch foreign marks on their bodies. We used a highly technical term for this: the no-hands problem. As simple as it is, the mirror test is a much greater challenge for a species whose members lack hands (or trunks, or some such appendage) with which to touch their bodies.

  A study of mirror self-recognition hadn't been on my original research agenda at Marine World. A few years earlier, Danny Povinelli, a colleague and later director of the Cognitive Evolution Center at the University of Louisiana, had called me and suggested the idea, and although I thought the dolphins would be excellent candidates for a test of MSR, at that point it just wasn't the right time, for me or the dolphins. We were newly into the keyboard work, and it was a very busy and sensitive period at the lab, so I opted not to pursue the collaboration then.

  A year or so later I got a phone call from Lori Marino. She explained that she was a doctoral student in Gordon Gallup's lab, now in Albany, New York, and that they knew of my research and wanted to collaborate on a study of mirror self-recognition in dolphins. I didn't personally know Gallup yet, but of course I was quite aware of his groundbreaking work and greatly admired it. I was also intrigued by several of his ideas. Foremost among them was that MSR and empathy might be linked. For example, children begin to recognize themselves in the mirror at about the same time that they start showing concern or empathy for others, sometime between eighteen and twenty-four months old. Apes also show both empathy and MSR abilities, whereas monkeys show neither. This led Gallup to speculate that other species that showed social complexity and empathy, like dolphins and elephants, would likely have the capacity for MSR.7 This time the timing was right, and I decided to do the study with them. Gordon, Lori, and I agreed to team up and test these minds in the water.8 Lori came out to the lab shortly afterward, and we spent time pondering the no-hands problem and thought about how to conduct the study. We were in uncharted and very challenging territory.

  So, when we introduced the mirror into the dolphin pool at Marine World in Vallejo in July 1990 to investigate Delphi's and Pan's responses to their reflected images, we challenged the primate-centered explanation of the origin of higher levels of self-awareness. After all, humans and dolphins have been on divergent evolutionary paths for ninety-five million years. If we were able to demonstrate that Delphi and Pan had the self-awareness to recognize images in the mirror as themselves, it would challenge the nature of human self-awareness. It could no longer be viewed as the culmination of smarter and smarter primate brains; self-awareness could arise in brains that were very different from ours in architecture and circumstances of development. Perhaps consciousness could arise in nature via many potential paths.

  There were two circular pools at Marine World, each fifty feet in diameter and sixteen feet deep. They were joined by a channel twenty feet long, ten feet wide, and eight feet deep. We rigged up a PVC frame at the exit to one of the pools into which we could drop a three-by-five-foot two-way mirror, which we could cover with a black rubber tarp. We stationed an underwater video camera behind the mirror, in the interconnecting channel. During my early experience as a set designer, I had learned how to construct scenery from whatever was available and on a very limited budget. This experience as a bricoleur has served me well.

  Underwater windows in aquariums usually have some degree of reflectivity, like poor mirrors. But the pools at Vallejo had no such windows, and so for the first seven years of their lives, Delphi and Pan had had no experience with their own reflections. We were dealing with completely mirror-naive test subjects. Our immediate task was to see how they would behave when they first saw their reflections.

  Dolphins are curious creatures, always ready to investigate novel objects in their environment. So it wasn't surprising that when we lifted the tarp from the mirror, Delphi and Pan were eager to check it out. They touched its surface with their beaks and aimed a lot of echolocation signals at it. It must have been very odd for them to see what appeared to be three-dimensional objects in front of them, their reflections in the mirror, while getting a sonar signal of an essentially flat surface. But if they were surprised, it wasn't obvious to me. They quickly began to interact with the reflections in the mirror as if they were other dolphins. Very soon, though, they moved into what had every appearance of contingency testing. While staring at themselves, they circled and cocked their heads, rocked their bodies back and forth, and opened and closed their mouths. Delphi and Pan definitely appeared to be testing the consequences of their actions, just as Gallup's chimps had done. So far
, so good.

  On the eighth day of one-hour-per-day exposure to the mirror, both Delphi and Pan moved to the next level. They appeared to engage in self-directed behavior. For instance, they each got very close to the mirror, putting an eye just a few inches away, staring intently. They opened their mouths very wide, as if examining the inside, and then wiggled their tongues, something chimps did with mirrors. At different times, Delphi and Pan each positioned himself so he could see his own ventral surface and then protruded his penis, an action that is under voluntary control in dolphins. And of course there was the courtship dance and reciprocal intromission attempts with each other that I described earlier.

  It was all suggestive evidence of self-recognition. But a skeptic could argue that the penis display, and even the intromission attempts, might be a sexual demonstration for the "other" dolphin in the mirror. The skeptic could also argue that the open-mouth behaviors were really open-mouth threats, something dolphins do during aggressive encounters. But this open-mouth behavior was very different; their mouths were relaxed and then held open in a prolonged manner. Also, the penis display had occurred when no other social behavior was directed toward the mirror. As for the intromission attempts, we noticed that when the mirror was covered and unavailable, Delphi and Pan did it in many different locations in the pool, but when the mirror was in plain view, they did it only in front of the mirror, always intently gazing at the reflection. And tongue wiggling, as far as we knew, was not part of any aggressive behavior.