Actually, zwenger was right. Without a proof of work, they do.
What's a blockchain? It's a data structure with a cryptographically strong CRC mechanism. But there's nothing stopping me from writing a new set of data into a new chain and passing it off as the original.
Sure, you may notice, but you can't prove my new chain is forged. I'll say the backup chain you're using as proof is the forgery...
It takes a proof of work to make the chain effectively uneditable. (Uneditable without paying for the work, which has been selected to be too expensive to be practical.)
If the items in the chain have digital signatures, you can prove that the people they're signed by contradicted themselves (or lost control of their signing keys). In Bitcoin people have felt that this wasn't good enough by itself, perhaps because there's not necessarily recourse against a counterparty who double-spends money, but you can still show that the double-spend attempt happened and who (in the sense of a signing key/address) was responsible for it.
It looks like the underlying mechanism that they use for this was described in
The parties don't have to be trusted. It's the opposite. A good blockchain needs mutually mistrustful parties who are not anonymous. If somebody tries to tamper with the ledger, all the other parties can easily tell who did it. It takes collusion of all parties to change history invisibly. Thus, you want parties who are not entirely cooperative.