Here's the intro:
Chimpanzees share many traits with humans but altruism, it seems, is not one of them, scientists said on Wednesday.
Although chimps live in social groups and co-operate and hunt together, when it comes to helping non-related group members, they don't put up with any monkey business.
When given the opportunity to help themselves and other chimps they often choose the selfish option.
The experimental setup gave the subject an option between two alternatives:
If the subject (hereafter referred to as the actor) chose option 1, the actor obtained a food reward and another chimpanzee simultaneously received an identical reward (hereafter referred to as the '1/1 option'). If the actor chose option 2, the actor obtained the same size and type of food reward, but no food reward was delivered to the other chimpanzee (the '1/0 option'). As a control, actors were presented with exactly the same reward options when there was no other chimpanzee present (Silk et al. 2005:1357).
So it's not a benefit/cost comparison, but a benefit/benefit. Sort of like if you won a house party from VH1, and you could either decide to invite other people or have the party all to yourself.
The chimpanzees didn't choose the "1/1 option" any more often when another chimpanzee was there (and got the reward) than when there was no other chimpanzee there. Those unfeeling primates!
I'm of two minds about the study. On the one hand, I'm not entirely sure how untutored humans would perform on this one. I think my two-year-old twins would pass -- when we are giving out treats, one will insist on an extra treat to bring to her sister. That situation is pretty analogous to the experiment, I think -- it's not like there's any cost to asking for an extra, since we know they are going to take it to the other twin. Nor is it really analogous to "sharing", which they do inconsistently. But we've had to work pretty hard to teach them to give out treats in that way, and they get direct feedback from us and the grateful sibling.
Considering how complex even this simple case is, I'm not too surprised that the chimpanzees would fail to give out the treats to their groupmates. Nobody has taught them how to do it, and there is relatively little direct feedback (although at one study site, the potential recipients sometimes made begging gestures). And I don't think that untutored humans would do it without explanation and feedback -- it's just that humans have a pretty sophisticated verbal and nonverbal ability to give that kind of feedback. So there is a genuine cognitive difference between humans and chimpanzees that may be involved in the result, although it is not perfectly clear that it is "indifference" in the chimpanzees.
On the other hand, look at the claim at the end of the paper:
These results complement observational and experimental studies that indicate that chimpanzees cooperate mainly with kin and reciprocating partners and show no aversion to inequitable exchanges that benefit themselves (Silk et al. 2005:1358).
This raises a question: would chimpanzees give out the treat to their kin? If not, then we're not seeing a failure to be empathetic toward the "unrelated other", we're seeing a failure to be empathetic at all. But we know that chimpanzees do behave preferentially toward kin in many contexts. So if this test failed to show empathy toward kin, it would be a failure of the test, and not a real indication of chimpanzee behavioral capacities.
So I think there are some missing steps here. Coming up with clear psychological demonstrations of the concept of empathy, or altruism, or welfare of other individuals is tough.
Silk JB et al. 2005. Chimpanzees are indifferent to the welfare of unrelated group members. Nature 437:1357-1359. Full text (subscription)