Classical cosmology predicts the age of the universe. Two degrees of freedom — the scale factor and a scalar field — evolve with respect to a time variable, and the model says the universe is 13.8 billion years old. This is a physical prediction: it tells us something about the world, it is confirmed by multiple independent observations, and it would be different in a different universe.
Note (2026-11-14; read against the paper's full text): the paper does not say that relational time defined from the scale factor and the scalar field yields conditional probabilities. The correlation a(phi) or phi(a) is the classical relation between those two variables, which the paper says does not contain the age of the universe. Conditional probabilities belong to the probabilistic interpretation (it cites Kuchar 1992 for it), which the paper treats separately from the internal clock interpretation and from the more radical relationalism that it also criticizes. The paper's conclusion is that the age is missing from quantum cosmology and that this should make us suspect something is wrong with this sort of approach; it says the worry is not that the same structures must reappear in the quantum theory, so the essay's 'must be recoverable' is stronger than the paper's wording. The paper also says the initial-singularity time of the simple two-variable model is only an approximation to the real age, which it takes from the best cosmological models and observations (Planck 2015: 13.8 billion years, accuracy of order 0.03 billion years); it does not describe the 13.8 billion figure as the output of the two-variable model itself.
Quantize the model and the time variable disappears. The Wheeler-DeWitt equation is a constraint equation, not an evolution equation. There is no external time parameter, no preferred clock, no before and after. The quantum state simply is. The problem of time in quantum gravity is well known, but this paper (arXiv:2603.09699) makes a precise version of the difficulty: the age of the universe, which is a genuine physical prediction of the classical theory, has no counterpart in the quantum theory.
The relational programs — attempts to recover time from correlations between subsystems — are examined and found wanting for this case. Defining time as correlations between the scale factor and the scalar field does not recover the age of the universe as a prediction. It recovers conditional probabilities: if the scale factor is this, the scalar field is probably that. But “the universe is 13.8 billion years old” is not a conditional probability. It is a statement about the absolute duration since a boundary condition.
The structural point is not that quantum cosmology has a philosophical problem. It is that it has a predictive gap. The classical theory makes a prediction; the quantum theory does not make that prediction; the prediction is confirmed by observation. Whatever quantum gravity turns out to be, this specific prediction must be recoverable. The age of the universe is not an artifact of classical approximation — it is data.