It is demonstrated that the reduction of a physical theory S to another one, T, in the sense that S can be derived from T holds in general only for the mathematical framework. The interpretation of S and the associated central terms cannot all be derived from those of T because of the qualitative differences between the cognitive levels of S and T. Their cognitively autonomous status leads to an epistemic as well as an ontological pluralism. This pluralism is consistent (...) with the unity of nature in the sense of a substantive monism. (shrink)

The reduction from Einstein's to Newton's gravitation theories (and intermediate steps) is used to exemplify reduction in physical theories. Both dimensionless and dimensional reduction are presented, and the advantages and disadvantages of each are pointed out. It is concluded that neither a completely reductionist nor a completely antireductionist view can be maintained. Only the mathematical structure is strictly reducible. The interpretation (the model, the central concepts) of the superseded theory T′ can at best only partially be derived directly from the (...) superseding theory T; it is severely constrained by the mathematical structure, and it can involve qualitatively different central terms that cannot be logically related between T and T′. (shrink)