Hull's project has two aspects -one empirical, the other conceptual. I take them in turn.
THE EMPIRICAL ASPECT
The empirical one involves finding out exactly how scientific theories and conceptual systems originate and are transmitted. One needs to investigate the psychology of the creative process; the social structure of scientific communities; the economics and politics involved in the funding of research programmes; the dynamics of the publication process; the politics and logistics of experimental testing; and like matters. In this connection Hull has provided interesting data and insights, which I would not wish to challenge. He does not, however, go so far as to draw on the work of the Edinburgh school of sociology of science (Bloor 1976; Barnes 1977; Compare Knorr-Cetina 1981; also Pickering 1984); but that may be because he would not wish to endorse their anti-rationalist, relativist conclusions (for more on which, see below). I think Hull's notions of conceptual inclusive fitness and of the demic structure of science offer the prospect of fertile application and extensions in this area. His analysis of the balance between cooperation and competition among scientists puts many aspects of the scientific enterprise into illuminating focusespecially the practices of mutual citation, priority disputes, and the ethical norms within the scientific community governing fraudulent, plagiarized or shoddy research. I think one of the most interesting implications of Hull's analysis, which he could have emphasized more, is that if science is best understood (descriptively or normatively) as a matter of conjecture and refutation, then the labour tends to be divided: one research group's conjecture tends to be the subject matter of a rival research group's attempts at refutation.
THE CONCEPTUAL ASPECT
The other aspect of Hull's project -the conceptual one -is more problematic. Hull puts forward a very general evolutionary model, designed to accomodate both cultural evolution (or what he repeatedly calls "conceptual change") as well as different kinds of selection processes at the biological level. He wishes to avoid the routine sort of analogising that involves projecting from relatively well-understood biological models to something roughly similar at the cultural level. Instead. he abstracts to a more general level from which one can later descend to recover the particulars of all these different processes at both the cultural and the biological levels. Now while I have no objection to this procedure in principle, I do have reservations about the way Hull has carried it out. Firstly, I have some quibbles about his generalized notions of replicator, interactor, and lineage, as they apply to the better understood process