Opportunities and constraints when studying social learning: Developmental approaches and social factors

KRISTIN E. BONNIE 0 0 Beloit College , Beloit, Wisconsin and Lester E. Fisher Center for the Study and Conservation of Apes , Chicago, Illinois Identifying social learning in wild populations is complicated by the relative lack of ability to conduct controlled experiments in natural habitats. Even in more controlled captive settings, tracking the innovation and spread of behavior among known individuals can be challenging, and these studies often suffer from a lack of ecological validity. In recent years, a host of new approaches have been undertaken to attempt to provide more quantitative control and empirical demonstration of social learning, both in the wild and in captive settings that more closely mimic natural contexts. Developmental approaches are being undertaken more regularly that allow us to study the ontogenetic trajectory of complex skills in a variety of taxa. Likewise, a spirited focus on the social context of social learning has emerged, and researchers have begun to meticulously analyze the influences of social systems and the characteristics of demonstrators and observers. Here, we provide a review of these studies and summarize the opportunities and constraints that exist when one attempts to study learning in social species. We suggest that although the study of social learning in nonhuman animals is becoming much more complex, addressing this complexity provides a fruitful model for understanding the evolution of human cultural behavior. - Monkey see, monkey do is a phrase commonly used to describe behavioral mimicry among young children. Data gathered from around the animal kingdom demonstrate that if individuals see, they very often do. However, recent research has increasingly shown that this monkey see, monkey do process, known as social learning, is not always straightforward and is commonly affected by the social dynamics that exist among individuals and the social setting in which they find themselves. Anthropologists, psychologists, and behaviorists have long been intrigued by the task of untangling the transmission processes, developmental trajectories, and cognitive requirements of behaviors that are purported to be learned socially. But social learning is exceedingly difficult to study in wild animals, due to a combination of the lengthy observation times necessary to capture behaviors that may be rare, the long developmental period of some species (in which many such behaviors are learned), and the lack of an ability to conduct the requisite experimental controls. As a result, few field studies provide strong evidence for social learning in wild populations (Laland & Janik, 2006), although more examples are beginning to emerge within the literature (see below). In the simplest of terms, in order to conclude that social learning has taken place, researchers must observe a novel behavior and be able to test that its presence is related to interaction with an experienced individual (or its products) and is not a result of ecological or genetic causes (West, King, & White, 2003). In field studies of wild animals, this is largely impossible, which has resulted in a variety of approaches taken to try to quantify social learning in other ways. Although statistical and mathematical modeling techniques are increasingly prevalent within the literature (see, e.g., Franz & Nunn, 2010; Hoppitt, Kandler, Kendal, & Laland, 2010; Kendal et al., 2010; Lycett, 2010), the two most predominant approaches remain ethnographic and experimental. The ethnographic method pools observational data from intensive and long-term field studies and infers social learning as the causative agent for differences between social groups when genetic or ecological explanations seem implausible (bonobos, Hohmann & Fruth, 2003; capuchin monkeys, Perry et al., 2003; cetaceans, Rendell & Whitehead, 2001; orangutans, van Schaik et al., 2003; chimpanzees, Whiten et al., 1999). This approach contributes valuable quantitative measures of scope of behavioral variation but still seeds much debate on controlling for genetics and ecology (Laland & Hoppitt, 2003; Laland & Janik, 2006; Lycett, Collard, & McGrew, 2007, 2010; McGrew, Ham, White, Tutin, & Fernandez, 1997). The aim of experimental methods with captive individuals (groups or, more often, pairs of individuals) is to control for and delineate the cognitive processes thought to support social learning (Whiten, Horner, Litchfield, & Marshall-Pescini, 2004). Therein lies the conundrum, or gap (Whiten & Mesoudi, 2008); that is, in most reports of social learning among wild populations, social learning is inferred from existing behavior patterns, rather than studied over the developmental trajectory of behavior. The underlying learning processes are difficult to determine in the field, and the diffusion is tricky to track, since the initial innovation is rarely observed and the patterns of association between individuals are rarely well known. In contrast, studying social learning in experimental settings allows for control of behaviors and relationships between individuals, but then suffers from a lack of ecological validity and may not reflect transmission patterns in the wild. A happy medium may be natural experiments, where different wild animal groups are seeded with alternative behaviors and transmission is carefully documented (Kendal et al., 2010; Whiten & Mesoudi, 2008), or studying free-ranging populations in protected areas that, although provisioned regularly (e.g., the capuchin monkeys in Brazil; Ottoni & Izar, 2008), also show natural wildlike foraging and social behavior. We refer the reader to Reader and Biro (2010) for a detailed review of these studies. Adding to the complexity of studying social learning is the breadth of behaviors that animals may acquire sociallyfrom behaviors with important fitness consequences, such as foraging and predator detection, to arbitrary behaviors such as idiosyncratic gestures or vocalizations (e.g., stone handling in Japanese macaques, Huffman, 1996; raspberry vocalizations in orangutans, van Schaik et al., 2003; arbitrary conventions in chimpanzees, Bonnie, Horner, Whiten, & de Waal, 2007). The picture is further clouded by the increasing assortment of taxa, including fish, birds, and mammals, that are thought to show social learning and the diverse range of developmental and social systems that are represented in these species. Transmission of behaviors can occur vertically (from adult to offspring) or horizontally (between conspecifics outside of the parentoffspring relationship), and a myriad of proposed mechanisms have been defined, studied, and debated (see Hoppitt & Laland, 2008). Since social learning is proposed to be a key driver of cultural differences among populations, an understanding of the dynamics and factors affecting social learning gives us insight into human and nonhuman cultural evolution. However, as we detail below, untangling and identifying these i (...truncated)