When the concept of homology is operationalized with synapomorphy and tested with character congruence, homology and homoplasy are treated as a complement relation, a and not-a, respectively. This leaves homoplasy to be defined nominally, something like operational 'error' in the inference of homology. In choosing the most severely tested and least disconfirmed cladogram, those errors are minimized, and the power of that cladogram to explain synapomorphies, as inherited from the same common ancestral condition, is correspondingly maximized. Tests of predictions of homoplasy can lead to the elimination of those kinds of error. The complementary relationship between homology and homoplasy is considered one of reciprocal clarification, not epistemological dependence.