Eprint note: this is the final submitted version; it hasn't been double-checked against the published version and there could be minor editorial discrepancies.

Moore, J. (1998). Commentary on The social behavior of chimpanzees and bonobos (by Craig Stanford). Curr. Anthropol. 39: 412-413.

Back to list of publications

Commentary on C. B. Stanford, "Social behavior of chimpanzees and bonobos: Empirical evidence and shifting assumptions. Current Anthropology 39(4) [1998].

Jim Moore
Anthropology Dept.
University of California, San Diego
La Jolla CA 92093-0532

After decades of research at multiple sites, it is perhaps surprising that we do not yet understand chimpanzees and bonobos well enough to tell them apart with confidence. Due to their utility as referential models of human ancestors and the ease with which such models become conflated with those ancestors (Moore, 1996), it is especially important for anthropologists to understand what we know and do not know about these apes--failure to do so can result in remarkably bad science (see e.g. Moore, 1992). Stanford's paper represents an important cautionary tale. For example, look at his Figure 1, illustrating mean numbers of males and females per party in chimpanzees and bonobos: the closest agreement occurs between points G and K, representing sites that have one primary investigator in common (Richard Wrangham); the next level of clustering appears to place K, G, and L (where most researchers have been European or American) together, with the predominantly Japanese sites M and W separate (L and W represent bonobos). Do methodological and/or definitional differences overwhelm taxonomic and ecological ones? Many factors are involved in our attempts to distinguish and define pongid reality, and as Stanford suggests our views of apes are subject to strong biases because, in part, they reflect biases about ourselves. I believe that in a small way the paper illustrates as well as discusses the problem, but I'll get to that later.

After that initial surprise, some readers may feel smug about progress in their own fields. That would represent a serious misunderstanding. The reason we do not "understand chimpanzees" (or bonobos) is the same as the reason that after about 2,000 years of formal inquiry, we still cannot say confidently that we "understand humans." Cultural ecologists and evolutionary psychologists can point to broad trends, but every such pattern is subject to the criticism, "But what about the Fugawi?" As long as we see instances (tribes, cultures, species) as categorically unique examples of different essences, our understanding is not likely to get much beyond naming the beasts of the field and birds of the air.

For example: there is a recent debate over whether Tai chimpanzees are "bisexually bonded" unlike the "male bonded" community described for Gombe (Boesch, 1996; Doran, 1997), and a similar though less marked difference has been noted between Gombe and Mahale for many years (Kawanaka, 1984). It also appears that the genetic structure of Gombe and Tai communities are different, with Tai lacking the greater male relatedness observed at Gombe (Gagneux et al., in press). As Stanford shows (and both Boesch & Doran note), bonobos can be seen in some sense as exhibiting an extension of the trend among chimpanzees toward larger, more stable parties in less seasonal habitats. This trend obscures the categorical meanings of "male bonded" and "bisexually bonded"; since average degree of relatedness and average party size/duration can in principle vary continuously across the full possible range, trying to determine the truth value of the assertion "chimpanzees are male bonded" is like answering "yes" or "no" the question, "Are humans monogamous?". Altmann & Altmann (1979) pointed out 20 years ago that such subtly different sociodemographic settings can have profoundly significant behavioral consequences.

Discovering the degree to which taxonomic categories (and by implication, genes) constrain sociodemographic continua is a central goal of primatology. Only by recognizing the underlying conceptual continuity can the meaning of local perturbations due to phylogeny be understood; only by recognizing the non-phylogenetic forces promoting categorical essentializing can we understand and compensate for them. This paper is a real contribution to this effort.

However, Stanford does not escape essentializing. A minor example: "eventually, female chimpanzees establish themselves as members of a new community" implies that this is universal; it is not (Moore, 1993). Sweeping the up to 50% of females who remain in their natal communities (Pusey et al., 1997) under the archetypal rug can only obscure the reasons why most of them emigrate. A more significant example: Stanford points out that our extravagant view of bonobo sexuality is based largely on captive studies, and implicitly dismisses these as "not reflect[ing] naturalistic mating patterns." Perhaps not, but if bonobos use sex to "cope" with social stress in ways that chimpanzees do not, then elevated sexuality in captivity for one but not the other is perfectly understandable and provides insight into how these apes work. "Captive artifacts" are only scientifically misleading if one assumes that there is one species essence, best revealed in one type of setting. (This is not to suggest that zoos should not strive to provide contexts conducive to naturalistic behavior!)

Finally, a case of a model threatening to overwhelm the data: do chimpanzees exchange sex for food? Stanford states that "males sometimes offer meat to females and receive matings in the process", citing Nishida et al. 1992 and Stanford et al. 1994; "in the process" is ambiguous but implies temporal proximity and a causal connection. These sources do not support this statement. Females in estrous ("swollen") are more successful at begging for meat from males (Teleki, 1973), and males are more likely to hunt in the presence of swollen females (Stanford et al., 1994). Males appear to use meat tactically amongst themselves in status politics (Moore, 1984; Nishida et al., 1992). Females are more likely to engage in restrictive (i.e., temporarily exclusive) mating with males who, on average, share meat more frequently with females (Tutin, 1979). Finally, females who receive meat more frequently tend to have greater reproductive success (McGrew, 1992).**

None of these studies directly address the "process" responsible for the patterns. It would be surprising if chimpanzee meat sharing were not involved in complex feedback processes contributing to long-term male- female relationships, sexual and otherwise (analogous to male-male relationships [Moore, 1984]). However, the only attempt to test this notion found no support for it (Hemelrijk et al., 1992). Stanford may turn out to be correct, but aside from Hemelrijk's study of captive chimpanzees, the problem simply has not yet formally been examined. As Stanford et al. (1994) state, "it is important to note that ... there are no data showing that he [sharing male] receives extra copulations from them [swollen female recipients]" as a result of his sharing.

McGrew failed to correct for female ages; when this is done, the effect is NS for survival to 1 year (U = 6, p = 0.222, 2-tailed) but becomes marginally significant for infants surviving to 5 years (U = 3, p = 0.056). That the effect increases with infant age is intriguing in light of nutritional arguments; given a relationship between female rank and meat access (Boesch, 1994) one wonders what might be the relationship among meat access, rank and female reproductive success (cf. Pusey et al., 1997).