The objective study of self‐recognition, with a mirror and a mark applied to the face, was conducted independently by Gallup (1970) for use with chimpanzees and monkeys, and by Amsterdam (1972) for use with infant humans. Comparative psychologists have followed the model (and assumptions) set by Gallup, whereas developmental psychologists have followed a different model (e.g., Lewis & Brooks‐Gunn, 1979). This article explores the assumptions in the definitions and methods of self‐recognition assessments in the 30 years since these initial studies, and charts the divergence between the developmental mark test and the comparative mark test. Two new studies, 1 with infant chimpanzees and 1 with infant humans, illustrate a reconciliation of the 2 approaches. Overt application of the mark, or other procedures related to how the mark is discovered, did not enhance mirror self‐recognition. In contrast, maternal scaffolding appears to enhance performance, perhaps by eliciting well‐rehearsed verbal responses (i.e., naming self). When comparable testing procedures and assessment criteria are used, chimpanzee and human infants perform comparably. A combined developmental comparative approach allows us to suggest that mirror self‐recognition may be based on a specific aspect of mental representation, the cognitive ability to symbolize.