When a flagship scientific journal declares a four-thousand-year human consensus an"illusion," the burden of proof is substantial. This paper argues that Mastroianni and Gilbert's(2023) celebrated finding — that the universal perception of moral decline is unfounded —depends on an unstated and untested ontological assumption: that genuine moral decline, if itexisted, would manifest as a secular downward trend within a seventy-year observationwindow. We term this the Stationarity Premise and show that it constitutes an ImplicitlyVerified Domain — a boundary condition treated as self-evident that is in fact contestable.The study's data are equally consistent with an oscillatory model of moral and socialdynamics, supported by a twenty-five-century tradition from Plato and Ibn Khaldun throughSorokin's empirical synthesis and the comparative material-historical evidence of Sophia(2026a, 2026b). We then develop a positive theory: multigenerational collective judgmentfunctions as temporal integration — an extension of Hayek's dispersed knowledge argumentfrom its synchronic domain to the diachronic case — encoding information about long-cyclesocial phenomena that no single-era study is structurally equipped to produce. Finally, wederive five conditions under which collective knowledge is more likely to be epistemicallysuperior to scientific consensus and show that the moral decline case satisfies all five whilethe paradigmatic cases of legitimate scientific override satisfy none. The paper's conclusion isnot that moral decline is real; it is that science has not shown it to be an illusion, and that thecosts of misallocating the burden of proof in this direction are epistemically significant.