Reply to Gilbert: On the relationship between association and altruism

Gilbert (1) contends that the observed heritability and fitness consequence of helping behavior in ground tits (2) should be explained by mechanisms involved in individual association, instead of by those underlying altruism per se. This contention rests largely on the association theory proposed by Gilbert himself, which was believed to have potential to replace the current dominant paradigm of social theory (3). We show in our critical response to his questions that this theory might not be the case.

Individuals are able to interact when they are coming together. The simple logic leads to the core idea of Gilbert’s theory (3): association is a necessary prerequisite to expressing social actions like altruism. However, this idea obviously ignores a common fact that altruistic actions can spontaneously lead to an association, or the association is a by-product of altruism. In such case, the association theory’s core idea will become a meaningless statement: “if two individuals do not live in the same region, it is impossible for them to interact with each other.” In other words, social contact or association must be an integral part of an altruistic behavior, according to the definition of altruism: a social behavior that causes individuals to benefit others at the expense of their own fitness (4). Therefore, altruism cannot be broken down into association and action. In addition, Gilbert (1) assumes that all birds in our study population could be altruistic, differing only in their propensity to associate. This is still a misinterpretation of the definition of altruism, because it makes no sense to define individuals as altruists who never take an altruistic action (5).

Even though in some scenarios association (or local aggregation) arises before altruism, here Gilbert (1) treating association as the prerequisite of altruism actually conflicts with his theory, which emphasizes that association and altruism are distinct events driven by different factors. Association may be promoted by some “automatic” benefits, such as sharing information and avoiding predators (3), and altruism typically by indirect fitness benefits (4). For example, while a comparative analysis suggested family living as an essential stage during the evolution of avian cooperative breeding, it also showed that the lifestyle and helping should result from different selective pressures (6). Many bird species live in family groups in which retained offspring do not help at all (7). Our focus was on altruism (2), which should be independent of association.

Finally, we would like to note a few misunderstandings by Gilbert (1). He cites that in a social amoeba there was little evidence for stable coexistence of altruistic and nonaltruistic phenotypes (8). This incorrectly equates nonaltruists in ground tits with cheaters in microorganisms, because the latter exploit public goods and thus are analogous to the recipients of altruism (9). It is difficult to understand the statement that our study on a single population represents one data point for comparative study. Virtually, our study is characterized by testing inclusive fitness theory with an empirical system based on the rigorous program of population genetic theory followed by Hamilton (4).