Rather than proposing yet another mechanism, it’s worth stepping back and asking a simpler question:
What properties should a consensus system have if it is meant to secure history, not just coordinate blocks?
One principle stands out as surprisingly absent from most designs:
Permanence of consequence.
A system passes the permanence test if the following is true:
Actions that meaningfully affect consensus must leave behind consequences that cannot be cleanly undone.
This is not about punishment for its own sake. It’s about asymmetry. Writing history is asymmetric with reading it. Rewriting history should be more expensive than participating honestly. Not just in theory, but in ways that cannot be arbitraged away.
Proof-of-Work achieves this accidentally. Burn energy once, and it’s gone. Proof-of-Burn does it explicitly. Proof-of-Stake, by contrast, tries to simulate permanence using time-boxed penalties, which are better than nothing but still reversible.
The permanence test doesn’t demand a specific mechanism. It doesn’t require burning tokens, locking collateral forever, or inventing exotic cryptography. It simply demands that influence accumulates inertia.
If a participant gains more ability to shape consensus, they should also gain more difficulty in escaping the consequences of misbehavior. Influence without inertia is leverage without responsibility.
This principle also clarifies why certain designs feel “off” even when they look sound on paper. Systems that allow validators to scale influence rapidly but shed risk linearly tend to create hidden cliffs: points where rational behavior flips from honest participation to strategic exit.
A permanence-aware system resists this by design. It makes exit boring, slow, and costly in proportion to past influence. Not punitive, just honest.
If this principle sounds uncomfortable, that’s a good sign. Security that feels comfortable is often security that’s been deferred to someone else.