This book is a relatively technical discussion of how we should think about altruism, as well as an argument for the reinstatement of group selection as a part of evolution. As such, it is more hypothetical than I had in mind, but hypotheses are the raw material of research. The book seems a bit repetitive, because the authors know that they must cover nuances to make a case to their fellow professionals. Wilson is a Professor of Biology and Sober is a Professor of Philosophy. The book is divided into two parts: Evolutionary Altruism and Psychological Altruism. I admit it got a touch above my head at times, especially when trying to represent things mathematically; I see what they are getting at, but I don't see where the numbers come from. The authors admit that they cannot prove all of their hypotheses, but argue that further study is worthwhile.

By "altruism," the authors mean that the altruist makes him/herself less fit in evolutionary terms, i.e., less likely to leave offspring, while making or attempting to make others more fit. It is only altruistic if the ultimate goal of the act is to help the other without hope of a reward. When a plane went down in the Potomac river, for example, one of the passengers repeatedly passed rescue device to one of the others trying to stay afloat, until he, himself, drowned. The authors think that our understanding of altruism is hindered by the assumption that all acts are basically selfish, if only to feel good about oneself for having helped another. Personally, I have always found that assumption absurd, since to be happy to have helped another is in itself proof of altruism. What appears to be altruistic can be selfish if the motive is to impress other people rather than actually help someone; I have known people to be publicly considerate to others that they are abusive to in private, but some people truly are unselfish.

The idea of group selection is that traits of individuals within a group may make the group stronger and better able to compete with other groups, even though they may not be related to all the group members. Thus, even if altruists are at a disadvantage compared to the selfish within a group, their contributions may allow the group, including other altruists, to flourish, thus increasing the number if not the percentage of altruists. In addition, societies often have ways of punishing the selfish and maintaining reciprocity. The authors give examples of groups of both related and unrelated individuals that prosper because the group works better. The idea goes back to Darwin, but it was considered to have been discredited in the 1960s. The authors blame this in part on the the averaging fallacy: averaging all the numbers together can mask the amount of variation. In his book How to Lie with Statistics, Darrell Huff points out that the average temperature in Death Valley is in the 60s, but that masks the extreme variation in temperatures. ( )