Aquatic Mammals 2006, 32(4), 413-422, DOI 10.1578/AM.32.4.2006.413 Dolphin Imitation: Who, What, When, and Why? Stan A. Kuczaj II and Deirdre B. Yeater University of Southern Mississippi, Department of Psychology, 118 College Drive #5025, Hattiesburg, MS 39406, USA E-mail: [email protected] Abstract Introduction The imitative ability of nonhuman animals has Early theorists tended to believe that imitation in intrigued a number of scholars and, in doing so, nonhumans reflected relatively primitive mental has generated a considerable amount of contro- abilities, but nonetheless recognized the signifi- versy. Although it is clear that many species can cance of imitation in the ontogeny of an individu- learn via observational learning, there is a lack of al’s behavioral repertoire (Wallace, 1870; Darwin, consensus concerning both what sorts of things 1871; Romanes, 1883; Baldwin, 1895; Thorndike, can be learned by watching others and what types 1898; Morgan, 1900; Washburn, 1936). Most of observational learning should count as imita- believed that the role of imitation in ontogeny was tion. These disputes have led to disagreements to facilitate an individual’s acquisition of behav- about the extent to which various nonhuman spe- iors already in the group’s repertoire rather than cies engage in imitation, based in large part on the creation of new behaviors. In this sense, imi- different definitions of imitation. An animal’s tation was thought to facilitate group conformity imitative success also depends on the context. and group traditions (the transmission of known For example, dolphins can be taught to imitate behaviors among individuals) more than individ- on demand, and studies using such elicited imi- ual innovation. tation tasks have yielded mixed results. Dolphins can imitate behaviors produced by other dolphins Imitation, Behavioral Traditions, and Innovation and other animals (including humans) and are The human capacity for imitation is well-known capable of deferred imitation. When dolphins (Piaget, 1962; Bandura, 1997; Meltzoff & Prinz, are asked to imitate, it seems easier for them to 2002), and the capacity for imitation has been reproduce familiar behaviors than novel ones. argued to be one of the hallmarks of human cog- Adult dolphins appear to be more successful than nition and culture (Meltzoff & Gopnik, 1993; juveniles at imitating on demand; however, young Tomasello, 1999). Imitation clearly facilitates the dolphins appear more likely than adults to spon- transmission of known behaviors from one indi- taneously imitate behaviors. Young dolphins fre- vidual to another (and so from one generation to quently spontaneously imitate the play behaviors the next), and may also be involved in the dis- of their peers, and sometimes acquire novel play covery and incorporation of innovative behaviors behaviors in the process. Following Kuczaj et al. into a group’s behavioral repertoire (see Galef, (2005), we suggest that the distinction between 2003; Kuczaj et al., in press). If imitation plays elicited and spontaneous imitation is important, some role in the acquisition of novel behaviors, and that understanding both types of imitation is then imitation may facilitate behavioral flexibil- essential. In addition to learning more about the ity—the ability of members of a group to adapt to factors that are influential when animals imitate, changing circumstances. it is also imperative to understand the types of The capacity for flexibility is another of the models and behaviors that are most likely to be hallmarks of humans, and it is reflected in our imitated, the types of animals that are most likely behavior, thought, and communication systems. to imitate others, and ontogenetic changes that The capacity for flexibility has resulted in many occur in imitation. diverse human cultures, for flexibility makes change possible (and inevitable). The human Key Words: imitation, observational learning, capacity for flexibility emerges relatively early in context, bottlenose dolphin, Tursiops truncatus, ontogeny (Piaget, 1952; Gopnik et al., 1999), but mental representation, mimicry the factors that influenced its phylogeny are a bit murky. Reynolds (1976) proposed that the evolu- tion of the human capacity for flexibility rested on 414 Kuczaj and Yeater four phenomena: (1) increasing delays in matura- piqued by observing other whales play with the tion, (2) increasing interest in manipulating objects, pup, and the observing whale’s subsequent inter- (3) increasing significance of play for social actions with the sea lion resulted in behaviors that development, and (4) increasing reliance on might benefit its future foraging efforts. observational learning. Although we agree with Goal Emulation—Individuals are enticed by the Reynolds that each of these factors undoubtedly end result of an observed behavior. Consequently, influenced the evolutionary path of our capacity they attempt to achieve the same result but not for flexibility, observational learning is the focus necessarily with the same behaviors that the model of this paper. used. For example, Haggerty (1909) devised a task Observational learning, behavioral flexibility, in which a monkey had to climb up the side of a and culture are closely intertwined. Culture is cage, stick its arm into a wooden chute, and pull acquired through various forms of social learn- a rope in the chute to release food. After watching ing (Boyd & Richerson, 1996, 2000; Rendell & a monkey go through this process on four sepa- Whitehead, 2001), one of which is observational rate occasions, another monkey was provided an learning, which may help maintain behavioral tra- opportunity to obtain the food. The monkey did ditions by facilitating the transmission of behav- not reproduce the behaviors it had observed, but iors from one animal to another. Innovation is instead, attempted to obtain the food in various necessary to add new behaviors to a group’s reper- other ways, finally succeeding via trial and error. toire, however, and Poirier & Fitton (2001) noted More recently, Call et al. (2005) reported that that innovators, the agents of cultural change, chimpanzees (Pan troglodytes) were more likely have been relatively little studied. Morgan (1900) to reproduce the outcomes of observed actions suggested that more “interesting” models would than the actual actions themselves, while human be imitated more than mundane models, and one children (Homo sapiens) were more likely to pro- might expect innovative behaviors to be more duce the actions. Thus, there appear to be spe- interesting than familiar ones. Consistent with this cies differences in terms of what is learned from view, bottlenose dolphin calves are more likely observing others. than adults to produce innovative play behaviors. Herman (2002) suggested that bottlenose dol- Dolphin calves are also most likely to imitate phins (Tursiops truncatus) may sometimes focus the novel play behaviors produced by their peers on the functional aspect of a modeled behavior (Kuczaj et al., 2005, in press). rather than the actual form of the behavior and thus produce goal-emulated behaviors rather Types of Observational Learning than imitative ones. For example, a dolphin that Observational learning and imitation are not syn- watches a model place a ball in a basket might also onyms. Tomasello (1999) described various ways, place a ball in the basket when asked to mimic the such as the following, in which observational behavior, but it may do so in a different way than learning could occur yet not involve imitation. the model. The extent to which dolphins engage Exposure—Individuals learn about their envi- in goal emulation rather than actual imitation is ronment by virtue of maintaining close physical not clear, but we suspect that they are capable of proximity to individuals that have more experi- various forms of observational learning, including ence. For example, a young dolphin might learn imitation. the location of an underwater mount teeming with As the above examples illustrate, there are fish simply by staying in close proximity to its ways in which animals can learn via observa- mother. tion that do not involve the imitation of observed Stimulus Enhancement—Individuals become behaviors. Heyes (1993) distinguished imitation interested in an object as the result of observing and non-imitative social learning in the following others interacting with the object (first described way: imitation occurs when animals learn about by Spence, 1937). Increased interest in an object behavior from observing conspecifics, whereas may result in object manipulation, thereby facili- non-imitative social learning occurs when ani- tating the acquisition of new object-related behav- mals learn about the environment from observing iors via trial and error learning. For example, a others. Both types of observational learning are young killer whale (Orcinus orca) might become important, but imitation has proven much more intrigued about playing with a sea lion pup after controversial. watching other whales toss the pup about. Playing with the pup might then result in the killer whale’s What Should Count as Imitation? acquisition of foraging behaviors appropriate to The extent to which a model behavior must be such prey. In such a case, the killer whale did preserved in a reproduction for the reproduction not learn to prey on sea lions by observing other to count as imitation is a matter of some debate. whales do so. Instead, interest in the sea lion was Thorpe (1963)