If you saw the movie “The Last Picture Show” (one of my top five movies), you’ll remember the scene where an English teacher in a small Texas town, facing a class full of bored students, asks, “What are my chances of interesting you people in a little Keats today?” From their reaction, it was nil.
So I ask: what are my chances of interesting the readers in a little biology today, given the kerfuffle that is going on elsewhere? I’d suggest that we move onto chimps today, in particular, the question of whether, when they kill each other in the wild, such acts are “natural” behaviors versus behaviors tainted by human disturbance, which makes the chimps act “unnaturally.”
But first, if you haven’t looked at the post on birds of paradise yesterday, I urge you to do so. The videos are stunning, and I worry that people might have ignored them to go battle each other on the post after that.
So, chimps. . .. In a new paper in Nature with many, many authors (first author Michael L. Wilson, last author Richard Wrangham; reference and link to free download below), a group of researchers examined all the data on murder in chimpanzees: instances when one member of a species kills a conspecific (remember there are two species: chimpanzees, Pan troglodytes, and bonobos, Pan paniscus, both equally related to us). The researchers’ question was this: when chimps kill other chimps, is that a “natural” behavior? Or does it occur only when chimps are “disturbed” by human presence?
The rationale, of course, is not simply to look at how chimps behave. In a New York Times summary of the paper by James Gorman, one of the authors gives the motivation:
In studying chimpanzee violence, “we’re trying to make inferences about human evolution,” said Michael L. Wilson, an anthropologist at the University of Minnesota and a study organizer.
Well, that’s a bit dicey. After all, chimps are not our ancestors: they are primates that evolved from our common ancestor with the the two species, and they have had just as much time evolving on their own since the split as have humans. It’s not clear that our common ancestor would be warlike and murderous, even if chimps are. After all (as we’ll see in a second), bonobos are not very murderous at all, and yet they’re just as related to humans as are chimpanzees.
What did the researchers do? They combined a ton of data from various field studies—that’s why there are so many authors—and compiled every instance of either known or suspected chimpanzee murder (18 chimp and 4 bonobo communities were studied). They found 152 murders: 58 were observed directly, 41 were inferred from forensic data like bite marks) and 53 “suspected” killings were enumerated by observing disappearance of healthy chimps or suspected deaths after known attacks). All of these were in chimps; only one bonobo was murdered by another, and although fewer bonobos were watched than chimps, that probably reflects bonobo’s general and well-known peacefulness.
The authors then considered two hypothesis for these murders—both of the hypotheses could, of course, operate together.
1. Chimp murder is “adaptive.” That is, it is an evolved trait that increases the reproductive success of the killers in some way. To test that, they looked at three predictions: members of the “eastern” clade should kill members of the “western” clade of chimps more often (presumably because of xenophobia); number of adult males in a group should correlate with more murders (the more males the more murders, since presumably it is the males who murder for reproductive advantage); and higher density of a group, which is said to reflect competition for resources and number of interactions, should also be associated with more murders.
It’s a shame that these aren’t great surrogates for what we want to know: do the killers have more offspring than non-killers? But that would be impossible to know, since the killers are males and you’d have to do paternity analysis on every chimp to see who the father was. Still, I’m not happy with these surrogates for adaptation. I suppose they’re the best the researchers could do given the data, but it makes the paper’s conclusions weaker.
2. Chimp murder is an artifact of human impact. The authors used three measures of disturbance as well: size of protected area covered by a group (the assumption is that smaller protected areas experience greater human impact); whether or not a group was artificially fed by humans; and “disturbance” (an amalgam of five variables including habitat disturbance, human harassment of chimps, hunting, habituation to humans, and elimination of predators [a sign of human impact]).
The results favored hypothesis #1: murder is adaptive. There was little support for the notion that human impact played any role. The authors note this (my emphasis):
Of the 16 models we considered (Table 3), four of the five models in the resulting 95% confidence set included combinations of the adaptive variables; the fifth model included the three human impact variables. The best model included only males and density, and was supported 6.8 times more strongly than the human impact model (evidence ratio = wi/wj = 0.40/0.059 = 6.8). Considering model-averaged parameter estimates22, increases in males and density increased the number of killings; for all other parameter estimates, the 95% confidence intervals included zero (Table 3 and Fig. 2). Excluding one community (Ngogo) that had both an unusually high killing rate and unusually many males resulted in similar values for model-averaged parameters, but only the estimate for density excluded zero from the 95% confidence interval (Extended Data Table 5b; n = 17).
Opposite to predictions from the human impact hypothesis (Table 2), provisioned and disturbance both had negative effects [i.e. reduced killing]; the estimates for these parameters included zero in the 95% confidence intervals (Table 3 and Extended Data Fig. 2b). The highest rate of killing occurred at a relatively undisturbed and never-provisioned site (Ngogo); chimpanzees at the least disturbed site (Goualougo) were suspected of one killing and inferred to have suffered an intercommunity killing; and no killings occurred at the site most intensely modified by humans (Bossou).
Some other fun facts about chimpanzee murder:
- 92% of the killer chimps were male.
- The chimps most likely to be killed were other males and infants (killing someone else’s infant can be adaptive if you subsequently inseminate the mother of the victim
- Victims were usually members of other communities and hence unlikely to be related to the killers
- In killings observed, attackers outnumbered defenders by, on average, 8 to 1. Chimps attack in groups.
The conclusion:
We conclude that patterns of lethal aggression in Pan show little correlation with human impacts, but are instead better explained by the adaptive hypothesis that killing is a means to eliminate rivals when the costs of killing are low.
The implicit conclusion, I think, is that humans are innately warlike as well; that’s the reason why the paper got so much attention.
While I think it’s a good effort, there are problems with measuring disturbance and adaptiveness, though there’s no doubt that chimps kill each other in the wild, and they do it in relatively undisturbed areas. That tells us that chimps probably do murder in natural circumstances, but whether that’s adaptive or not is hard to gauge. (I’m not willing to say it’s adaptive just because it occurs.)
Further, bonobos don’t murder much. They are just as closely related to us as are chimps. Who can say whether the common ancestor was more bonobo-like or chimp-like? So while chimps murder in the wild, this doesn’t say whether our common ancestor, much less early hominins, were genetically warlike. Certainly our ancestors killed each other, and maybe that reflects a genetically conditioned xenophobia, but I don’t think the chimp studies say much about this.
At any rate, the New York Times piece reports some dissent by other scientists:
Robert Sussman, an anthropologist at Washington University who supports the idea that human actions put pressure on chimpanzee societies that results in killings, was dismissive of the paper. “The statistics don’t tell me anything,” he said. “They haven’t established lack of human interference.”
Brian Ferguson, an anthropologist at Rutgers University who has written extensively on human warfare and is working on a book about chimpanzee and human violence, also argued that the measures of human impact were questionable. The study considered whether chimpanzees were fed by people, the size of their range and the disturbance of their habitat. But, Dr. Ferguson said, impact “can’t be assessed by simple factors.”
“I’m arguing for the opposite of the method that’s being used here,” he said, adding that a detailed historical analysis was needed for each site.
If I were a reporter, I would have asked both Sussman and Ferguson to clarify their comments. What does the former mean by saying “they haven’t established lack of human interference.” Where is the interference supposed to be in the least disturbed site? (I’m not saying it wasn’t disturbed, but I’d like more specific criticisms.) And Ferguson’s objection does carry some weight given what I’ve described above, but saying disturbance “can’t be assessed by simple factors” overlooks the fact that five measures were combined into the index of disturbance. And what does he mean by “the opposite of the method used here” and “a detailed historical analysis”? It’s not clear; and that’s probably the reporter’s fault.
Finally, there’s dissent about what it all means for our own species:
Richard Wrangham of Harvard, the senior author of the new paper and Dr. Wilson’s onetime doctoral adviser, is the co-author of a 1996 book, “Demonic Males: Apes and the Origins of Human Violence.” Although the issue is not mentioned in this paper, he argues that chimpanzee behavior “is a reasonable start for thinking about primitive warfare in small-scale societies.” But, he added, “I certainly wouldn’t want to say that chimps have anything much to say directly about what’s going on in Syria.”
But what about the bonobos? Why is their behavior irrelevant? But he’s dead on about Syria, which is due largely to religion, something that chimps don’t have. (You could counter, though, that religion is just a surrogate for “othering,” which prompts an innate xenophobia.)
Dr. Sussman, who is skeptical of drawing connections between chimpanzee and human violence, said, “War has nothing to do with what chimpanzees do.”
Well, what if we’re not talking about organized state war, but simple person-on-person violence? Maybe that could be relevant to what chimpanzees do.
But the whole debate about humans from this paper alone seems futile, because we’re not looking at our ancestors, but species that diverged from our common ancstor with chimps 6 million years ago—and one of those species isn’t warlike at all!
***
If you want to see a group of chimps killing a member of their own band (not that common), there is a YouTube video here. But please be warned: it is fairly graphic, so don’t watch it unless you have an interest in how this occurs.
________
WIlson, M. L. et al. 2014. Lethal aggression in Pan is better explained by adaptive strategies than human impacts. Nature 513:414-417.
