Truth exists, but we have limited knowledge of it. Scientific theories last for barely 200 years before they are replaced by more accurate new theories. The corpus of science is the state-of-the-art latest version, and scientists have a positive attitude to falsification. In comparison, religions are anchored to some central historical event that remains irreformabilis. Believers expect the original "type" to be preserved and might easily feel betrayed when told that something they were asked to believe turns out to be (partially) false. This poses different constraints for dogmatics compared to the epistemology of science. Looking closely at what exactly is supposed to be irreformable, they have much in common: there is reformability and discontinuity of signifiers and at the same time a continuity of the signified (ceterum paribus), which however is partially hidden [Potschka 2024]. Doctrine tries to capture this signified with ever immature and historically situated formulae that develop over time. "The substance … is one thing, and the way in which it is presented is another" [John XXIII 1962]. A dogma might be improved by limiting the range of applicability, or eliminating contextual values. It may not apply to a changed historical context, and may even become obsolete because it has fulfilled its historical role [Kasper 1965]. Ratzinger [1965] considered it even acceptable that a dogma turns out to be wrong. I will illustrate my thesis with the growth and maturation of the concept of truth based on papal promulgations on the relationship of faith and reason. My epistemic proposal allows for the development of doctrine without denying the concept of irreformability. It is at variance with John Henry Newman's assumption that principles are permanent [Newman 1845], and with Lakatos' concept of research programs where the core principles are irreformable. It remains to be seen if all disturbing epistemes of theology can be explained in this conservative way.