In Aristotelian thought, the concept of episteme had a precise profile: it denoted demonstrative knowledge, anchored in first principles and in the necessary causes of phenomena, and capable of ensuring stability over time. The methodological criterion underpinning that framework required theories not merely to save the phenomena—as mathematical astronomy demanded—but to explain them through their causes, without introducing arbitrary postulates. Once validated, such knowledge constituted a structural heritage, meant to endure as a solid foundation for future generations.
In the transition to contemporary science, and particularly in theoretical physics and cosmology over recent decades, we are witnessing a striking asymmetry. On the one hand, the acceleration of research produces a continuous stream of provisional results, often superseded before they can consolidate into a shared framework. On the other hand, at the level of foundations, despite a proliferation of purely mathematical constructs—such as supersymmetry, (super)string theory, and the various models of quantum gravity—there has been no genuine, empirically verified advance for half a century. Fundamental physics is suspended between formally sophisticated but unfalsifiable speculation and the mere confirmation of old predictions.
Cosmology offers the clearest example of this drift. Faced with discrepancies between observational data—such as galaxy rotation curves or the redshift–distance relation on cosmological scales—and the predictions of classical gravitational equations or the standard cosmological model, the prevailing solution is not to revise the models’ assumptions, but to introduce non-baryonic, not directly detectable components. Dark matter and dark energy thus end up accounting for roughly ninety-five percent of the energy density of the universe, while remaining labels lacking not only any experimental confirmation in the laboratory, but even a unified theoretical framework to justify them. Instead of a solid theory, we see a continuous succession of hypothetical candidates and ad hoc models—from WIMPs to axions, from sterile neutrinos to the most exotic scalar fields—formulated after the fact to fill the same anomalies, and invariably disproved or scaled back with each new theoretical proposal or attempt at verification.
From a rigorous epistemological perspective, this approach inverts the classical principle: rather than adapting theory to phenomena, ad hoc corrective parameters are constructed to preserve the mathematical integrity of the model. The dynamic is not unlike the epicycles of the Ptolemaic tradition, in which additional variables served to keep the initial hypothesis standing despite observational anomalies.
The tension becomes even sharper in light of data from the James Webb Space Telescope, whose observations of the high-redshift universe reveal galactic structures too massive and evolved for the timescales predicted by the standard cosmological model. Given such critical issues, methodological rigor would call for extreme caution in treating as established the hypotheses outlined above, which remain, at present, purely conjectural.
The distance separating Aristotelian episteme from modern knowledge therefore lies not in the power of our measuring instruments, but in the status assigned to empirical verification. When a theory requires invisible hypotheses in order not to be falsified by the data, or takes refuge in mathematical architectures with no contact with experiment, the boundary between established knowledge and mere formal modeling becomes blurred. Recovering a clear criterion for distinguishing what has genuinely been demonstrated from what remains a computational convention would be the foremost task of an epistemology that is today evidently absent.
