Okay, let’s go through the scenarios. A big one is world war etc.
In this extreme scenario, the on-chain state wouldn’t be a concern any more. The banknote still has other security features it maintains. You can settle it once the war is over and just assume they are mostly valid. The country or DAO entity can also issue new banknotes and force the old ones to get settled. Using this as an opportunity to detect fraud, money laundering etc. The Cash-like small payments only require key A any way. The other keys are only used to resolve conflicts within a window. Arguably in that total collapse scenario, the whole Ethereum ecosystem would be dead any way.
What about partially offline scenarios?
Cash-like applies here of course too. So, if you hand a 5, 10, or 100 bill, it’s still to assume the key on there is enough to settle the transaction. You can just settle it once you got a connection again. A PoS, Smartphone, or some other device, still can keep the last state of the chain it receives. You can check the last on-chain state and verify if the banknote was included in that. Additionally, you may also carry proof in the data payload of the chip.
I wouldn’t see it as a need and more 2FA if you will. Another security feature. You still have to leave a trace behind in the physical world if you want to defraud it on-chain eventually. This may be at multiple intermediaries, but you leave a trace behind.
Mentioned it in this idea too, it’s not about making fraud, scam etc. impossible. It’s more of a game.
There will always be a certain amount of malicious actors. You won’t change human nature.
Instead of trying to make it impossible to defraud the system, just have good ways to detect those malicious actors within the system. The need to physically exchange to settle it on-chain adds this layer to detect them. The assumption here is, the Authority is involved in the money creation process that is locked in the smart account. So, the Authority is also able to insurance against fraud etc. Giving them even more incentives to detect fraud in this system. We also have to see it in the context of the reality of cash. It’s declining overall and is mostly used as an inconvenient backup solution in case infrastructure breaks down. Which typically, even in war scenarios, doesn’t break down completly, but sparsely. You could even argue a refugee with cash they can easily move into a “bank account” is quite the benefit overall.
The assumption is also that most actors are honest. The fact alone they have to settle it physically at some “official” place in a traceable way will be high enough barrier to not defraud it.
Happy to hear your input on how it can get improved.