I consider the mutual relationships between measures, and narratives, of corruption. Available measures of corruption mutually influence not just the merit, but also the nature, of its prevailing narratives. On the one hand, the latter define a demand for types of measures which they contemplate as pertinent and relevant, but not for others which could be pertinent within alternative narratives. On the other hand, the available measures feed into the narratives, which they constrain and bias. To account for such mutual relationships, I present a tentative conceptual framework. It assumes a simple taxonomy of types of corruption, where forms of “legalized corruption”, in representing the “non-existent narrative”, play an important role. It considers and compares two types of measures: perception-based ones, such as, Transparency International Corruption Perception Index, with the “objective” PACI and PACI and BPCI indices (Escresa and Picci, 2017; Picci, 2016). Last, it distinguishes explicitly between the problem of assessing cross-national levels of corruption vs. their within-country changes in time. The presentation is partly based on published results, and partly on what at this stage are mere speculations, reflecting the early stages of a research project which I tentatively title “the corruption of (anti)corruption”. Comments and criticisms are very welcome.