We elaborate on Israel Scheffler's claim that principles of rationality can be rationally evaluated, focusing on foundational development, by which we mean the evolution of principles which are constitutive of our conceptualization of a certain domain of rationality. How can claims that some such principles are better than prior ones, be justified? We argue that Scheffler's metacriterion of overall systematic credibility is insufficient here. Two very different types of rational development are jointly involved, namely, development of general principles that are (...) strictly constitutive of rationality as such, and development of specific principles determinative of our conceptualization of particular domains. For the first type a transcendental argument applies. As to the second, we show how foundational development is itself a condition of the possibility of its justification. In both cases only principles that are typical of the later stage yield the second order criterion in terms of which the evaluative comparison with former stages can be made and defended. In a discussion of problems involved we indicate to what extent Scheffler's idea of rationally justifiable rational development may be realized here, avoiding pitfalls of both foundationalism and relativism. (shrink)