Editing Openai/6897769e-4ee4-800f-aba5-69cca34f701c (section)

=== A scientific programme must give testable predictions. Examples: ===
# If decay amplitudes depend on boundary charge distribution, changing the local EM environment (extreme fields, plasmas) could shift decay rates slightly. Existing measurements limit such shifts; compute expected magnitude and compare with experimental bounds.
# If CP-violating phase is geometric and linked to a Wilson holonomy, then experiments that alter boundary holonomies (if you can do that in a controlled system — think engineered mesoscopic analogues) should change interference phases.
# Compare relative sizes: QAT may predict correlations between the sizes of geometric constants (e.g. α, golden-angle geometry constants) and the magnitudes of certain CP asymmetries. Look for correlations across systems.

If predicted geometric dependences are absent at the sensitivity level you compute, that would rule against the simplest versions of the model.