Copyright © 1999 Massachusetts Institute of Technology
Individuals in many animal species do not interact at random but have qualitatively different social relationships with different members of their group. The most stable, long-term bonds are typically found among matrilineal kin (paternity is generally unknown) and have their genesis in the close relationships formed between a mother and her offspring (Kummer 1971). For example, immature female African elephants interact at high rates with their mother and, through their mother, with their matrilineal aunts, siblings, and cousins. Matrilineal kin remain together throughout their lives, tolerating one another's proximity and cooperating to defend feeding areas against other matrilines (Moss 1988).
Old World monkeys, like the macaques of Asia or the baboons and vervet monkeys of Africa, are organized along similar lines but differ in the size of their social units. Groups of baboons may contain over one hundred individuals. Each group contains several matrilineal families. Immature animals acquire dominance ranks immediately below those of their mothers, and families are arranged in a linear rank order so that all members of family A rank above the members of family B, who rank above the members of family C, and so on. As in a Jane Austen novel, social interactions follow two general patterns. Most cooperative behavior (grooming, play, tolerance at feeding sites, and the formation of alliances) occurs among close kin, but cooperative behavior between the members of different matrilines can arise when a middle- or low-ranking individual attempts to groom, play, or form an alliance with the member of a higher-ranking family (for reviews, see Smuts et al. 1987).
The study of social cognition in animals attempts to identify the mechanisms that underlie these interactions. It is now clear that individuals in many species know a great deal about their own social relations. They recognize others as individuals, remember who has cooperated with them in the past, and adjust their future behavior accordingly. This is not surprising, given what is known about the learning and memory of animals in the laboratory, where they have been tested with objects. Social cognition in animals is most striking, however, when we consider what individuals know about the social relations of others in their group.
One hypothesis, derived from traditional learning theory, holds that animals develop their own social relationships and learn about the relationships of others through classical and operant CONDITIONING. Individuals recognize and respond differently to one another because of past experience, or because they have formed specific associations between particular individuals. A second view holds that social relationships in many species -- particularly long-lived animals like elephants, parrots, dolphins, and primates -- cannot be explained unless we assume that individuals differentiate among classes of relationships and make use of mental representations similar to our concepts of "kinship," "closely bonded," or "linear dominance hierarchy." These hypotheses have been tested most often on monkeys and apes.
In order to understand a dominance hierarchy or predict which individuals are likely to form alliances with each other, a monkey must step outside its own sphere of interactions and recognize the relationships that exist among others. Such knowledge can be obtained only by observing interactions in which one is not involved and making the appropriate inferences. There is, in fact, growing evidence that monkeys do possess knowledge of other animals' social relationships and that such knowledge affects their behavior. For example, in a test of individual recognition a juvenile vervet monkey's scream was played from a concealed loudspeaker to three adult females, one of whom was the juvenile's mother. As expected, the mother responded most strongly; however, even before this occurred the control females looked at the mother. Apparently, control females associated a particular scream with a particular juvenile and this juvenile with a particular adult female. They acted as if they recognized the kin relations that exist among others (Cheney and Seyfarth 1990).
Following aggression, a monkey may "reconcile" by engaging in friendly behavior with its former opponent. A fight may also cause a victim to "redirect" aggression and threaten a third, previously uninvolved individual. In both cases, individuals act as if they recognize their opponent's close kin. When animals reconcile they do so either with their former opponent or with one of their opponent's close matrilineal relatives; when they redirect aggression they do so most often against a close relative of their prior opponent (reviewed in Cheney and Seyfarth 1990; de Waal 1996).
Monkeys can also recognize the dominance ranks of others. When female vervet monkeys compete for grooming partners, competition is greatest for the highest-ranking female, next for the second-ranking female, and so on. When a high-ranking female approaches two lower-ranking individuals and interrupts their grooming, it is the lower- ranking of the two who is much more likely to move away (Cheney and Seyfarth 1990).
It is widely accepted, then, that nonhuman primates -- and very likely many other animals -- classify others according to their patterns of association and recognize the friendships and enmities that exist among individuals other than themselves. Humans, however, go further and classify different types of relationships into more abstract, superordinate categories like "friends," "family," and "enemies" that are independent of the particular individuals involved. Do social concepts of this sort exist in any animal species?
Dasser (1988) studied social knowledge in longtailed macaques that were members of a group of forty individuals living in a large outdoor enclosure. She trained three adult females to leave the group temporarily, enter a test room, and view slides of other group members. When shown two slides, one of a mother and her offspring and the other of two unrelated group members, subjects were rewarded for pressing a button below the mother-offspring slide. After training with five different slides of the same mother and her juvenile daughter, subjects were tested using fourteen novel slides of different mothers and offspring paired with fourteen novel unrelated alternatives. Mother-offspring pairs varied widely in their physical characteristics: they included mothers with infant daughters, mothers with juvenile sons, and mothers with adult daughters. Nonetheless, in all fourteen tests the subjects correctly selected the mother-offspring pair. They behaved as if they recognized a concept of "mother and offspring" that was independent of the particular individuals involved.
We have, as yet, no idea how monkeys might represent social relationships in their mind. They may use physical resemblance as a cue, because members of the same matriline often (but not always) look alike, or they may use rates of interaction. The latter seems unlikely, however, because kin do not always interact at higher rates than nonkin. Mothers and infants, for example, interact at high rates whereas mothers and older sons do not. Both were classified by Dasser's subjects as falling within the same category. Similarly, although bonds within matrilineal kin groups can be extremely variable (depending, for example, on the ages and sex of family members), monkeys nevertheless treat competitive interactions as pitting one family against another (Cheney and Seyfarth 1990).
In sum, monkeys seem to use a metric to classify social relationships that cannot be explained simply in terms of physical features or the number and type of interactions. To do this, they must compare relationships according to an underlying relation that has been abstracted from a series of interactions over time. Monkeys take note of the elements that make up a relationship (grooming, alliances, etc.), then make judgments of similarity or difference not by comparing specific elements but by comparing the different relationships that these elements instantiate. By this view, conditioning and the formation of mental representations are not alternative mechanisms underlying social cognition; instead, they interact, with the former helping to give rise to the latter (see also PRIMATE COGNITION).
The hypothesis that monkeys classify relationships into relatively abstract categories receives additional support from experiments indicating that vervet and diana monkeys classify vocalizations into categories according to their meaning. When asked to compare two vocalizations, vervets judge them to be similar or different not according to their acoustic properties, which are measurable and concrete, but according to their referents, which are more abstract (Seyfarth and Cheney 1992). Female diana monkeys give their characteristic leopard alarm when they hear either a leopard's growl, a male diana monkey's leopard alarm, which is acoustically different from their own, or the shriek of a duiker, a small antelope that is often hunted by leopards (Zuberbuhler, Noe, and Seyfarth 1997). These results suggest that, in the mind of a female diana monkey, some sort of MENTAL REPRESENTATION of a leopard serves as an intervening variable between hearing one type of call and producing another (see also ANIMAL COMMUNICATION).
It has been suggested that natural selection has acted with particular force in the domain of social interaction (Jolly 1966; Humphrey 1976), and that, as a result, the cognitive mechanisms that underlie social interactions are different from those that underlie, for example, foraging or ANIMAL NAVIGATION. The processing of knowledge about social relationships, it is argued, constitutes a mental module, much like the processing of language or music (Fodor 1983; Jackendoff 1987; Cheney and Seyfarth 1990; see also EVOLUTIONARY PSYCHOLOGY). Despite great theoretical interest, however (e.g., Byrne and Whiten 1988), this view is largely untested.
Byrne, R., and A. Whiten, Eds. (1988). Machiavellian Intelligence. Oxford: Oxford University Press.
Cheney, D. L., and R. M. Seyfarth. (1990). How Monkeys See the World. Chicago: University of Chicago Press.
Dasser, V. (1988). A social concept in Java monkeys. Animal Behaviour 36:225-230.
de Waal, F. (1996). Peacemaking Among Primates. Cambridge, MA: Harvard University Press.
Fodor, J. A. (1983). The Modularity of Mind. Cambridge, MA: MIT Press.
Humphrey, N. K. (1976). The social function of intellect. In P. P. G. Bateson and R. A. Hinde, Eds., Growing Points in Ethology. Cambridge: Cambridge University Press, pp. 303-318.
Jackendoff, R. (1987). Consciousness and the Computational Mind. Cambridge, MA: MIT Press.
Jolly, A. (1966). Lemur social behavior and primate intelligence. Science 153:501-506.
Kummer, H. (1971). Primate Societies. Chicago: Aldine.
Moss, C. J. (1988). Elephant Memories. Boston: Houghton Mifflin.
Seyfarth, R. M., and D. L. Cheney. (1992). Meaning and mind in monkeys. Scientific American 267:122-129.
Smuts, B. B., D. L. Cheney, R. M. Seyfarth, R. W. Wrangham, and T. T. Struhsaker, Eds. (1987). Primate Societies. Chicago: University of Chicago Press.
Zuberbuhler, K., R. Noe, and R. M. Seyfarth. (1997). Diana monkey long distance calls: Messages for conspecifics and predators. Animal Behaviour 53:589-604.
Cheney, D. L., and R. M. Seyfarth. (1990). The representation of social relations by monkeys. Cognition 37:167-196.
Essock-Vitale, S., and R. M. Seyfarth. (1987). Intelligence and social cognition. In B. Smuts, D. Cheney, R. Seyfarth, R. W. Wrangham, and T. Struhsaker, Eds., Primate Societies. Chicago: University of Chicago Press, pp. 452-461.
Seyfarth, R. M., and D. L. Cheney. (1994). The evolution of social cognition in primates. In L. Real, Ed., Behavioral Mechanisms in Evolution and Ecology. Chicago: University of Chicago Press, pp. 371-389.
Copyright © 1999 Massachusetts Institute of Technology