In his 1993 book, Discovery and Explanation in Biology and Medicine, Schaffner compared the merits of two models of intertheoric reduction: the Causal Mechanical (CM) model and the General Reduction Replacement (GRR) model. The GRR model, a modified version of Nagel's model of intertheoric reduction, 1 was seen as the "appropriate framework in terms of which deeper logical questions", like the establishment of synthetic identities, "could be pursued" (1993, p 498); while the CM model, according to Schaffner, could "not be the whole story" by any means because it was occluding "important deep structural issues such as specific points at which identities need to be formulated between levels" (1994, p 293). In the Paris' conference on Reduction and Emergence, Schaffner made the following comment concerning his GGR model: "The GRR model, itself an extension of the Nagel model, is more suited to fully axiomatized sciences, of which there do not appear to be any" (emphasis added).It is not clear how to understand this statement. Should we conclude that Schaffner has significantly moved away from the GRR model, i.e. should we interpret him as now viewing it as a model of situations that never occur? In a way, "yes". But this has always been the case as Schaffner has presented GRR as forcing actual ongoing reduction into a "rational reconstructive" mode (1994, p 336):"The GRR is [. . .] useful in providing us with a kind of systematic summary and regulatory ideal, but it should not in general be confused with the process of establishing reductions in the ongoing elaboration of connections that typically unite the reducing and the reduced theory" (1993, p 497; my emphasis). 1 The modification is intended to avoid the derivability problem faced by Nagel's theory, that is, the fact that most cases of actual reduction would imply deriving a false theory from a true one as in the case where the lower-level theory corrects the higher-level theory