The thing about commodities is that they "plug into" a pre-existing worldwide commodity market. There are futures markets, places to store them, third-party methods to transport, insurance, and so forth. You can have a standard underlying type of thing to value. Things like promissory notes become much more difficult to value. Not impossible, though. The problem is that in order to understand and predict attacks on the currency, you don't want lots of things with different underlying physical realities.
There are two questions here. The first one, as you point out, is risk. The second one is the applicability of international ownership laws. If government X shuts down E-currency #48, everybody gets nothing. But in a decentralized system, if everybody can prove ownership of some tangible item, it might be possible to not lose your shirt. I'd much rather be making a case that these anonymous eight people owe me 10kg of gold -- and demand the government allow me to access it -- than asking for, say, something completely virtual and esoteric like Bitcoin. Asking for my physical property is well-covered by international law. Asking for some cryptographic stuff? You're back into uncharted waters.
Ripple (non-anonymous) and OpenTransactions (potentially anonymous using blinded tokens issued by each participant) kind of take that model.
What I'm unclear on is if Zerocoin can be efficient enough, maybe given an optimized, scrypt-based blockchain underlying, to work with either of those as a base, globally decentralized currency.
I might like Ripple a lot more if XRP got replaced with Zerocoin. Once Ripple's software is open sourced, it wouldn't be too hard to rip XRP out and replace it with Bitcoin or Zerocoin or anything else.