Social Learning and Kin Recognition

A s Sherman (this issue) notes, I tried (Alexander 1990) to determine “the extent to which social learning, especially without self-reference, can account for the kin recognition findings of different investigators with different organisms . . . (p. 273). I had two motivations. First, I have had difftculty modeling the evolution of kin recognition via mechanisms that can function in the absence of social learning (Alexander and Borgia 1978; Alexander 1979, 1990). Second, it seemed to me that virtually all investigators of kin recognition were searching for unusual or cryptic mechanisms like use of self (without additional referents. Alexander 1979, 1990) or so-called “recognition alleles” or “genetic models,” and partly for that reason were in danger of biasing themselves away from thorough searches for opportunities for social learning. I think I provided evidence that the second argument is sometimes appropriate, by showing that several results previously regarded as unexplainable except as a result of self-referencing may indeed be outcomes of social learning opportunities that had been overlooked. Examples are Holmes’ (1986b) results with paternal half siblings in ground squirrels and Sherman and Holmes’ (1985) results with full and (maternal) half sister discrimination in ground squirrels. In Holmes’ experiment, juvenile ground squirrels were in fact exposed to the traits of fathers they shared with unfamiliar (nonlittermate) paternal half siblings through littermates that also shared that father. For the more complicated case of full and (maternal) half sisters in ground squirrels, I provided non-intuitive predictions (Alexander 1990, p. 277) that can be used to test my hypothesis based on imperfect social learning. I also tried to explain two effects that might be mistaken for either selfreferencing or “recognition alleles.” The first is that imperfections in social learning can give the impression of unusual mechanisms by causing organisms to behave as though they recognize relatives with which they have not associated when in fact they are simply mistaking them for associates. I