Set theory with and without urelements and categories of interpretations (Q2505133): Difference between revisions

The author proves the synonymity of two set theories, ZF and ZFU, in the sense that there are interpretations \(I:\text{ZF}\to \text{ZFU}\) and \(J: \text{ZFU}\to \text{ZF}\) with the property that \(I\circ J\) and \(J\circ I\) are ``identities''. (The theory ZFU carries an extra axiom dictating that the set of ur-elements is countable. The success of the present work crucially depends on this.) One can think of common-sense interpretations \(I':\text{ZF}\to \text{ZFU}\) and \(J': \text{ZFU}\to \text{ZF}\). (For instance, \(J'\) picks, as members of the domain, those elements whose transitive closures contain no ur-elements.) But neither has the inverse to produce identity. So the author resorts to ingeneous constructions to obtain the above \(I\) and \(J\). (\textit{A. Visser} embarked on classifying theories by manners of mutual interpretability in [Categories of theories and interpretations. Logic Group Preprint Series, Vol. 228, Univ. Utrecht (2004) (per bibl.); see also Lect. Notes Log. 26, 284--341 (2006; Zbl 1107.03066)]. The strictest condition is synonymity. \(I'\) and \(J'\) satisfy a weaker condition of homoty of ZF and ZFU. But synonymity was an open question.)