Functional irreversibility is the idea that lets the irreversibility register separate harms that can be repaired from the ones that cannot. Some conditions are literally irreversible; many of the most dangerous are not — a right can still be appealed, an authority can still in principle be withdrawn, an alternative can still in theory be built. What has changed is the cost. When reversal becomes prohibitively expensive, slow, or destabilizing, the formal possibility of it stops mattering, and for a living ecology that is the same as losing it.
It is why the register watches its pre-irreversibility thresholds most closely: an appeal process that exists but cannot change a decision, an override that is legal but never exercised, an infrastructure that could be forked but no longer realistically can — each is functionally irreversible while looking, on paper, perfectly open. Naming the category is what keeps a civilization from reassuring itself with rights whose effect has already drained away.