Like other foundational concept in biology, as for instance genes and species etc., homology has a long history of frustrating the efforts of biologists and philosophers to define it. On its most general level, homology means correspondence of body parts across species, i.e. the continuity of character identity over evolutionary time. The difficulties arise, in part, from the fact that characters that are manifest the same in various animals can develop from different genes and along different developmental pathways. This fact is leading to the question whether it is at all meaningful to speak of the same character in different animals/species. In this talk I will defend my recent proposal, that character identity can be linked to our mechanistic understanding of development. The basis for my thesis is the realization that character identity and character states are caused by different developmental genetic mechanisms. My hypothesis states that character identity is tied to developmental mechanisms that enable differential expression of “realizer” genes. These character identity gene networks do not dictate whether a body part is in fact different from other parts or in what way it is different from the rest of the body. Homology is not only conceptually abstract, but also mechanistically abstract, in the sense that it stands for the potential to be different rather than for a particular way to be different. This conception of character identity, if validated, has consequences for the metaphysics of evolutionary biology that will be discussed in this talk as well.