Introduction

Predation pressure has long been considered a powerful selective force on primate behavior (Alexander 1974, Anderson 1986, Crook and Gartlan 1966, Dunbar 1988, Isbell 1994, van Schaik 1983, 1989). However, most studies that have investigated the importance of predation on primates have tended to focus on either patterns across different species (Anderson 1986, Cheney and Wrangham 1987, Goodman et al. 1993, Hill and Lee 1998, Isbell 1994) or patterns across individuals within a given population (e.g., Cowlishaw, 1997a, Isbell and Young 1993, Stanford 1995). In contrast, little attention has been paid to the consistency of antipredator strategies across different populations of the same primate species. Nevertheless, such patterns can potentially shed valuable light on intraspecific variation in tolerance to predation risk and the strategies used to reduce that risk, together with the differential costs and benefits of such strategies in different populations.

Baboons represent an ideal taxon for such an investigation, since predation risk in this species is relatively well understood (e.g., Cowlishaw 1994, 1997a) and the wealth of data available on their ecological and behavioral flexibility (Barton et al. 1996) permits detailed assessments on the impact of predation risk across a wide array of ecological conditions (e.g., Dunbar 1996). The evidence already available for baboons indicates that antipredator behavior is not fixed but that individuals invest differentially with response to the degree of predation risk. For example, baboons that live in small, high-risk, groups are more active in their use of refuges, such as tall trees and cliff faces, than are larger groups. Importantly, this pattern has been reported in more than one population: Amboseli, Kenya (Stacey 1986) and Tsaobis, Namibia (Cowlishaw 1997b).

Introduction

Competitive relationships result from competitive interactions over resources that can affect survival and reproduction. For male mammals, prospective mates are one such resource. For female mammals, food is more important than mates as a contestable resource (Trivers 1972). Its importance has been recognized by three ecological models that invoke either food distribution (van Schaik 1989, Wrangham 1980) or both food distribution and abundance (Isbell 1991) to explain variation in competitive relationships within and between groups of female primates (see also Isbell and Van Vuren 1996, Sterck et al. 1997, van Hooff and van Schaik 1992, Wrangham 1987). For both sexes, survival is also affected by other factors, such as disease and predation, and it is possible that individuals also compete for resources that in some way minimize exposure to disease or predators. In fact, one of the three models (here called the ‘predation hypothesis’) considers predation to be more important than food in its effects on the grouping behavior of females (van Schaik 1989).

The predation hypothesis assumes that predation ultimately forces females to live together and that variation in predation pressure causes variation in spatial cohesion within groups and among species (Sterck et al. 1997, van Hooff and van Schaik 1992, van Schaik 1989). Where predation pressure is high, females are predicted to decrease interindividual distances and thus live in groups that are spatially cohesive.

Introduction

Gibbons are small diurnal apes inhabiting the rain forests of South Asia (Preuschoft et al. 1984). The social system is predominantly monogamous and groups have typically only two to six members. Recent studies revealed the existence of extra-pair copulations, partner changes and at least occasional polyandrous or polygynous group composition (Brockelman et al. 1998, Palombit 1994a,b, Reichard 2001, Sommer and Reichard 2000).

Primate species that form single-female groups are relatively rare (Kleiman 1977). Selection pressures such as intergroup competition for food (van Schaik 1989, Wrangham 1980) and predation risk (van Schaik 1983, van Schaik and van Hooff 1983) are believed to favor the development of (polygynous or polygynandrous) multi-female groups. Two factors may thus have caused gibbon females to disperse (van Schaik and van Hooff 1983): (a) strong female–female competition for food; (b) a reduced predation risk due to the extremely arboreal lifestyle of gibbons, their capability of high-speed brachiatory locomotion, and the absence of large primate-hunting eagles (such as crowned hawk eagle, Stephanoaetus coronatus; harpy eagle, Harpia harpyja: great Philippine eagle, Pithecophaga jefferyi) in their habitat.

When scanning the literature on the extent of predation risk actually faced by wild gibbons, only scarce evidence emerges: a full grown siamang was found in the stomach of a python (Schneider 1906) and gibbon hair in two out of 237 samples of leopard faeces (Rabinowitz 1989). When wild gibbons were caught in canopy cage traps during a malaria study in Malaysia, pythons invariably surrounded them, obviously trying to prey on the gibbons (van Gulik 1967).

Pretenses appear in six forms of activity in this volume:

• self-pretense, as in reproducing the appearance of, feigning, or inhibiting one's own acts and emotions (for play, teasing, deception, communication, or evocation);

• object substitution (including replica toy play);

• animation of objects (including doll play);

• pretending about imaginary objects (supported by real objects or not);

• pretending to be (or act like) someone else;

• pretending to have (imaginary) companions.

These pretend types are intended to be neither exhaustive nor mutually exclusive, but rather to direct attention to the pretenses observable in animals (or potentially so). All of these activities are much more common in children than in animals, and how to interpret them in either is (as this volume attests) subject to disagreement, providing a fertile field for reflection and investigation. (Imaginative nonpretend actions are discussed in the first chapter.)

In this chapter, I summarize research on pretense and direct attention to future research on animal pretense. A pattern should become evident – scientists interested in discerning pretense in animals need to observe individual animals ontogenetically in order to understand and interpret their behaviors, and also need to employ knowledge of the types of behaviors animals of the observed species use in potentially pretend actions, to discern the precursors of these behaviors (Mitchell, 1986, 1987, 1990). Put simply, extensive longitudinal study incorporating contextual analysis, focal animal sampling, and an (at least partial) ethogram are essential for discerning pretense in animals.

This book is a delightful collection of scientific writings about pretending and imagination in animals and children. The impetus for the present volume derives not only from observations of animals' activities similar (perhaps identical) to pretense (e.g., Groos, 1898; Mitchell & Thompson, 1986; Mitchell, 1987, 1990, 1991a, 1993c, 1994a; Byrne & Whiten, 1990; Miles, 1991; Russon, 1996), but also from developing ideas about children's understanding of pretense (Harris & Kavanaugh, 1993; Lillard, 1993a,b, 2001a), philosophy of art (Walton, 1990), and the evolution of image-making (Davis, 1986), all of which concern organisms' understanding or creating reproductions of various sorts (Mitchell, 1994a). The continuing influences of Piaget (1945/1962), Guillaume (1926/1971), Bateson (1955/1972, 1956), Vygotsky (1930–1966/1978), and Leslie (1987) are also apparent. The purpose of the book is primarily to present and examine evidence for the existence and nature of pretense in animals and children, and secondarily to examine various aspects of why or how. Evidence of pretense in animals may eventually allow us to provide a “psychologically and evolutionarily plausible account of ‘fictive acts of perceiving’” (Davis, 1986, p. 211; see Mitchell, 1994a; Reynolds, PIAC14). A full appreciation of what needs to be incorporated into such an account is provided by reading chapters in this volume, in which the topics range from relatively simple simulative actions to complex fantasizing about nonexistent objects and agents.

The emergence of the symbolic function at the end of the sensorimotor period plays an important role during the development of the human infant. It is the beginning of the attainment of signs and symbols – depicting an item (an object, a person, an event, etc.) by a differentiated signifier (language, mental image, symbolic gesture, etc.) specifically used for this particular representation.

Although the symbolic function cannot be tested directly, various behavioral manifestations, implying the use of differentiated signifiers, reflect it. Authors generally agree on the various behavioral manifestations appearing during the second semester of the second year of life:

• “generative” language (Piaget, 1936);

• deferred imitation (Piaget, 1945) or real imitation (Guillaume, 1925);

• the communicative function of immediate imitation (Baudonnière & Michel, 1988; Asendorpf & Baudonnière, 1993; Hart & Fegley, 1994);

• pretend play (Piaget, 1945; Kagan, 1981);

• stage 6 understanding of object permanence (Piaget, 1936);

• recognition of the specular self-image (Zazzo, 1975, 1977, 1985; Lewis & Brooks-Gunn, 1979).

These behavioral manifestations allow us to study the emergence of the symbolic function. The question is to determine whether these new competences emerge simultaneously, resulting from a common transformation, or successively. If the symbolic function appears as a singular modification influencing many behaviors, it should result in the simultaneous emergence of the various behavioral manifestations. On the contrary, if this emergence proceeds step by step, the different behaviors should emerge not synchronously but progressively, eventually in a hierarchical way.

The simulative modality

Play, imitation, and pretense can all be explained in terms of a theoretical construct called the simulative modality (Bruner, 1972; Reynolds, 1976; Mitchell, 1994a). All these phenomena imply a separation between the form of behavior, its normal motivation, and its typical consequences in the wider world. The concept of the simulative modality, first scientifically proposed in the work of Gregory Bateson (1955/1972), is now poised to become a major theoretical framework, encompassing data from field primatology, human evolution, and cognitive psychology.

The strong form of the theory presented here consists of four major theoretical propositions: (1) the simulative modality consists of neurohumoral states that evolve in conjunction with the cerebral cortex of the primate brain; (2) social interaction stimulates the simulative modality and induces a mental model of one's own body and the bodies of others called the cognitive body; (3) the major psychological differences among primate species are due to the types of subcortical systems that are integrated into the simulative modality and by differentiation of a species-specific cognitive body; and (4) each type of cognitive body is expressed through a characteristic form of social organization.

Although the dominant theories of primate evolution treat symbolic processing as a laboratory curiosity, the theory of the simulative modality indicates that it is a normal and essential aspect of simian social organization, required for the development of status hierarchies, enduring mother-child dyads, and normal sexual relationships among adults.

Examining symbolic, pretend play in a nonverbal individual, especially one who is not human, is inevitably fraught with difficulties in identification and interpretation. For example, a young wild chimpanzee maneuvered a small stick in a manner similar to the ant-fishing technique seen in older individuals, but there were no ants present. Goodall (1986) interpreted this action as possible evidence that the chimpanzee was imagining or pretending that there were ants present, although other interpretations are possible. Call & Tomasello (1996; Tomasello & Call, 1997), by contrast, suggest that such behavior may be more parsimoniously explained as simple manipulative play with sticks, and thus attribution of mental imagery and pretense is not necessary in the interpretation.

Still, there are accounts of captive members of all four great ape species producing pretend play. For example, they have been observed participating in pretend play with dolls, engaging in behaviors such as tickling and feeding (Hayes, 1951; Gardner & Gardner, 1978; Tomasello & Call, 1997). Controversy remains as to whether these behaviors truly constitute symbolic play. Arguments against the existence of symbolic object play in the great apes have been supported by a study conducted by Premack & Premack (1983) which suggested that young chimpanzees were unable to recognize the correspondence between a pretend room and its real-world referent (Call & Tomasello, 1996; Tomasello & Call, 1997). Chimpanzees watched as a small object was hidden within a model of a room.

Pretend play as a human characteristic

Fully developed pretend play, including role play and sociodramatic play, seems universal in all human societies and uniquely human, with only rudimentary pretence shown by some apes (Schwartzmann, 1978; Slaughter & Dombrowski, 1989; Smith, 1996). Cross-cultural variations exist in amount and type of pretence (Smilansky, 1968; Roopnarine, Johnson & Hooper, 1994), which appears dependent upon children's physical and cultural context. For example, among hunter-gatherers, children in mixed-age peer groups used sticks and pebbles to represent village huts and herding cows (Konner, 1976; Eibl-Eibesfeldt, 1989), and in the Marquesas Islands of Polynesia, children made mud bananas (Martini, 1994). Such symbolic object-use comprised half of the “fantasy play” episodes of the Marquesan children; the other half involved more complex scripts and roles, such as “‘ship’ play, ‘fishing,’ ‘hunting,’ and ‘preparing feasts’” (Martini, 1994, p. 84). These episodes “follow the same fantasy scripts from one performance to the next,” and often involve some imitation of adult actions.

The ubiquitousness of pretend play in human children suggests biological processes at work (Slaughter & Dombrowski, 1989; Harris, 1994), and raises the question of its biological function.

Pygmy chimpanzees or bonobos (Pan paniscus) have become best known for their varied and extensive repertoire of sexual behaviors and their use of sex in nonreproductive situations (Kano, 1986/1992; Furuichi, 1987; Hashimoto & Furuichi, 1994; de Waal, 1995; de Waal & Lanting, 1997). They are also quite adept at communicating their intentions and coordinating social activities, and many of these behaviors are closely linked to their highly cohesive social structure. In the wild, female genital–genital rubbing serves as a greeting among well-known individuals, as well as to form social bonds among unrelated individuals and ease tensions when entering a potentially competitive situation such as a fruit tree. Variations on a rocking gesture are used to request different kinds of close contact between individuals, ranging from close sitting to copulation (Kuroda, 1984; Ingmanson, 1992). Branch dragging is used to organize and direct group travel (Ingmanson, 1988, 1996), and play behavior is often signaled by carrying an object (Ingmanson, 1996). In captive studies, communication is also used to organize social activities – zoo bonobos use a clapping gesture to initiate grooming in a social group (Ingmanson, 1987, 1998) and have an extensive range of reconciliation behaviors (de Waal, 1989). Symbol-trained bonobos use lexicons on a keyboard to communicate both social and nonsocial desires, and respond appropriately to others' keyboard communications (Savage-Rumbaugh, 1984; Greenfield & Savage-Rumbaugh, 1993; Savage-Rumbaugh & Lewin, 1994). All of these behaviors coordinate social activities and reduce tension in the group by avoiding misunderstandings, thereby reducing aggression.

This chapter examines four possible cases of pretending in rehabilitant orangutans. Only about a dozen other cases are known, so these cases are worth dissecting because they are so rare. Seven incidents have been reported for Chantek, an enculturated language-trained orangutan, spanning pretending to feed a toy animal at 1 year 8 months, deceptive self-pretense at 1 year 10 months (acting as if needing to urinate in order to stay in the bathroom) and pretend play at 4 years 10 months (clay man symbolic play) (Miles, Mitchell & Harper, 1996). Two incidents reported as deceptions may also qualify as pretending: a wild adult male lulled a competitor into dropping his vigilance by pretending to lose interest in combat; and an adolescent female rehabilitant fooled a human by pretending to lose interest in forbidden goods (Byrne & Whiten, 1990). Some imitative behavior in free-ranging rehabilitants resembles other-pretend, or pretending at behaviors observed in others, at representational levels (Blake, 2000; McCune & Agayoff, PIAC3). Examples are: nonfunctional tooth-brushing, and siphoning fuel from an empty fuel drum (Russon & Galdikas, 1993, 1995; Russon, 1996).

These new cases offer a chance to explore whether characteristics linked with human pretending, especially complex ones like metarepresentation, feature in great ape pretending. All four cases show complex social maneuvering, so they offer a rare window on complex social cognition in a species best known for its minimalist approach to sociality.

As the only comparative psychology graduate student at Clark University, Massachusetts, I was surrounded by graduate students studying human development assured in their belief that humans were distinctly different psychologically from other animals. When imaginative pretense was brought up as one of many “uniquely human” capacities (including language, intentional deception, imitation, and self-recognition), I mentioned the chimpanzee Viki's imaginary pulltoy (Hayes, 1951), to the general response that it was only one example compared to the myriad instances exhibited by children. The intellectually stimulating debates that followed my frequent disagreements led to this book, and fueled my desire to publish three other books, which I initiated prior to this one – Deception: perspectives on human and nonhuman deceit (Mitchell & Thompson, 1986), Self-awareness in animals and humans (Parker, Mitchell & Boccia, 1994), and Anthropomorphism, anecdotes, and animals (Mitchell, Thompson & Miles, 1997). I hope the skills I learned from my talented co-editors – all ardent educators – are evident in my first run as solo editor.

While I was at Clark, Ina Užgiris introduced me to Paul Guillaume's (1926/1971) Imitation in children in her course on “Imitation, internalization, and identification,” and to Piaget's (1947/1972) Psychology of intelligence in her course on “Piaget's theory.” Reading Guillaume and Piaget was a gift, but reading them with Ina's guidance was an extraordinary gift, for which I am grateful.

This chapter discusses cases of spontaneous behaviors resembling human pretend play in hand-reared lowland gorillas studied longitudinally over a period of several years. In their early years these gorillas had regular contact with humans and human objects. We discuss instances of behaviors that apparently meet some criteria of pretense (such as object substitution or role enactment). Our argument is that there is little evidence that these apparently pretend behaviors are of the same kind as the pretense shown by human children. However, they might reflect the operation of some basic cognitive mechanisms that played a role in the evolution of pretense in the human lineage. The spontaneous presence of these precursors of pretense may also explain why more complex cases of possible pretense can be induced in apes subject to formal procedures of symbolic training.

The notion of pretend play

Pretend play, also known as symbolic or make-believe play, is a characteristic behavior pattern of the human species that appears in children during the second year of life (Piaget, 1945; Leslie, 1987; other chapters in this volume). It consists of the ability to “act as if” something is the case when it actually is not. For example, acting as if one were a monkey (if one actually is a human, that is) or as if a wooden block were a camera and one were taking pictures (perhaps of an imaginary monkey!). Pretend actions usually display only a fraction or an otherwise modified version of their serious counterparts.