In: Martin R. Jones and Nancy Cartwright (eds.), Idealization XII: Correcting the Model. Idealization and Abstraction in the Sciences (Pozna! Studies in the Philosophy of the Sciences and the Humanities, vol. 86), pp. 59-115. Amsterdam/New York, NY: Rodopi, 2005. James R. Griesemer THE INFORMATIONAL GENE AND THE SUBSTANTIAL BODY: ON THE GENERALIZATION OF EVOLUTIONARY THEORY BY ABSTRACTION 1. Introduction This essay is about the nature of Darwinian evolutionary theory and strategies for generalizing it. In this section I sketch the main argument. In this essay I describe some of the limitations of classical Darwinism and criticize two strategies to get around them that have been popular in the philosophical literature. Both strategies involve abstraction from the entities that Darwin talked about: organisms in populations. One strategy draws on the notion of a fixed biological hierarchy running from the level of biological molecules (like DNA and proteins) through cells, organisms, and groups on up to species or ecological communities. As a consequence, evolution is thought to operate at any level of the hierarchy whenever certain properties hold of the “units” at that level. The other strategy rejects this hierarchy, in part because it seems strange to imagine the evolutionary process “wandering” from level to level as it alters the properties of organisms and populations, creating and destroying the conditions for its operation. The second strategy constructs a different, two-level hierarchy of “interactors” and “replicators” by abstracting functional roles of organisms and their genes. The second strategy simplifies analysis of the sorts of properties deemed essential by the first, but also describes the key process of natural selection as operating at a single ontological level (when it does operate). Along with these virtuous ontological simplifica- tions, however, the second strategy raises metaphysical problems for both strategies, suggesting that neither is on the right track. I trace these further problems to two features the strategies share. To do this, I offer a twist to Cartwright and Mendell’s (1984) view of abstraction in science which relativizes the degree of abstractness of an object to an Aristotelian theory of explanatory kinds. I focus on science as a process rather than on the