抄録
Conspecific individuation is crucial in primate societies. Face recognition is the most prominent information to achieve this in the visual domain of perceptual processing. However, the contribution of various factors such as innate components and lifetime experience to the ability of individuation remains elusive. In most studies, there is no way to distinguish among these factors. Here, we test the own- versus other-species effect with chimpanzee participants of two age groups (means: “young”: 10 years vs. “old”: 30.3 years), allowing us to disentangle effects of innate face tuning and lifetime experience. Chimpanzees performed a delayed matching-to-sample (DMS) task and were required to select the one face out of two simultaneously presented faces that showed the same identity as a previously presented face. Face stimuli consisted of chimpanzee (own-species) and human (other-species) faces. Under the assumption that the innate representation biases perception toward a conspecific face morphology, we would predict a facilitation of face identification for own-species as opposed to other-species faces for the young chimpanzee group. This effect, however, is hypothesized to be reduced or equalized for the old chimpanzee group due to lifelong exposure to humans and exposure to a limited number of members of their own species. Our results show the predicted tendencies in the following way: Young chimpanzees have an unstable bias toward chimpanzee faces, i.e. in many sessions (but not all) performance rates for chimpanzee faces exceed the performance rates for human faces. Old chimpanzees have a clear bias toward the human faces, i.e. performance for human faces is drastically increased as opposed to chimpanzee faces in literally all of the sessions. To break down this effect in more detail, we apply exploratory data analysis techniques, such as similarity estimation using an eigenface approach or support vector machines (SVM). The two stimulus classes do not show differences in their overall similarity, nor does (dis-)similarity among the individuals influence the performance. However, SVM segregates sessions testing chimpanzee faces from those testing human faces with relative high accuracy. Thus, the most-likely explanation is, there indeed is an innate representation of the own-species morphology and there is a very strong effect due to lifelong exposure to the other-species' face, which not only facilitates the ability to successfully identify out-group members, but ironically overwrites the initial high-level performance of identifying in-group members. Together, perceptual expertise comes at a high cost.
