1. what is to prevent collusion btwn several bank gateways to falsify a ledger?
How do you falsify signed transactions? Any change to a ledger must be justified by a signed transaction whose deterministic operation generates the changes between ledgers.
2. why is this any different than the other iterations of digital cash that relied on digital signatures but failed due to centralization?
The design doesn't require any central authorities.
3. how often does the ledger have to be signed? after every tx, at the end of the day, or at the end of a month?
Every validator signs every new ledger. Typically, new ledgers are produced every 20 seconds.
4. with proof of work, any independent node all the way down to the individual, knows to pick the longest chain in case of a conflict in the blockchain ledger since it represents the most amount of work (Byzantine General's problem). How does Ripple resolve conflicts btwn ledgers?
Every validator signs every ledger. If somehow the overwhelming majority of validators you have chosen to trust are not signing the same ledger, you declare the network broken and do not consider any transactions to be confirmed.
5. how can gateways ever hope to be as secure than true p2p (down to the individual) like Bitcoin?
I think this question is based on the misconception that gateways and validators are the same. Gateways just enter into issue and redemption agreements with people. Validators establish consensus transaction sets and sign ledgers.
