Ontology is taken by Moulines as supervenient on science: what kinds of things there are is determined by our well-confirmed theories. But the fact is that today, science provides us with a multiplicity of well-confirmed theories, each having its own ontological commitments. The modest, ontological form of reduction advocated by Moulines (this volume) restores hope of putting some ontological order in the “huge chaotic supermarket of science”. In this paper I show that any claim on the amount of order obtained by reduction does not only always remain “temporally qualified” but, worse for the reductionist with a taste for ontological order, that the very notion of orderliness must be relativized to the capacities and interests of knowers.