Computationalialism, functionalism, and instrumentalism-behaviorism all appear to some philosophers to facilitate representation without utilizing resemblance. I show that in all these systems, some sort of isomorphic correlation, covariance, pattern, set of relations, etc. is maintained, an expression of a structure or order that is shared by or is the same in both the representation and the represented. Such resemblance is not sufficient, but it seems to be necessary for representation. Philosophers such as Goodman and Cummins base their denial that resemblance is needed on their restriction of the sense of “resemblance” so that many forms of likeness and similarity, such as that maintained by covariance, are not called resemblances. But that means that there is really only a vocabulary difference between, say, Cummins and me. From my broad-based viewpoint, Cummins in fact elaborates arguments that make my point. It is certainly important to distinguish the various ways in which things can be like one another — by sharing properties, by covarying, by being exemplifications of the same set of abstract relations, and so on. But the genus of all these species is that venerable sameness, which, coupled with difference, produces the variety of things in our world.