PHY.10:6 - Bias-Annotation
The cases use controlled laboratory comparisons and idealized response laws. Astronomical, geological and other field questions can require naturally available contrasts whose backgrounds cannot be independently fixed. The relevant physical premises must then support the observational comparison.
A finite candidate set can exclude a useful account before the test begins. A result inconsistent with every candidate, or a new physical dependence exposed by the experiment, can justify developing the set. It need not become a competition among the original labels.