On Wednesday, January 23, 2002, at 08:36 AM, Michael Motyka wrote: > Tim May <[EMAIL PROTECTED]> wrote : >> Escrow and online clearing. >> >> As you defined it, there is of course no way to resolve the problem. No >> more so than if Alice says she sent ordinary cash in the mail and Bob >> claims he never got it, or it was in the wrong amount, and so on. >> >> --Tim May >> > Forgetting for the moment the nature of the goods or services that are > exchanged couldn't the cash transfer be broken into many very small > transfers with acknowledgment on each? First there must be some sort of > agreement to pay. If the payment transactions halt when there is a > dispute there are three classes of money : paid, unpaid and disputed. At > least this leaves a smaller amount to be disputed. >
The general rule of thumb is: When the payoff for defection (fraud, failure to pay) is greater than the likely/expected future revenue stream from honorable behavior, defection is more likely. -- the local restaurant _could_ claim that a customer had ordered and eaten something he hadn't, and thereby pocket an extra $10 or whatever, but the loss of future business from this customer, those who hear about it, and even reports in the press or by rating agencies (who might be anonymous customers) will likely exceed the short-term payoffs for fraud. Importantly, neither the law nor being able to trace the true name of the restaurant owner is crucial to why this kind of fraud is rare. -- a bank can trivially claim "You withdrew your money yesterday." Scrawled signatures on withdrawal slips are meaningless as proof. And yet banks don't. The reasons are related to the above example. Granted, depositing real money in the First Nigerian Bank of Anonymous Accounts may be a problem, given their penchant for frauds. But the Bank of Orlin Grabbe probably will value reputation. And, as you and others (cf. Cyphernomicon, '94) note, breaking up a transaction into small pieces has many advantages. Not the least of which is "pinging" (testing) the account. One reason we trust banks is because of a Bayesian history of banks (in non-Nigerian-type places) always "honoring" their accounts. Inasmuch as transactions/accounts are untraceable, anyone can ping a bank, or moneychanger, or merchant, with lots of small transactions. Much has been written about this. Everyone has different language for this, and for the situation Mike describes, and so on. Without carefully knowing what it meant by "some sort of agreement to pay" and "dispute" and so on, I can't say whether Mike is on the right track (and, so that Mike doesn't waste his time writing definitions, I don't currently have any time to analyze English language story problems about payment failures....). Remember that international trade has gone on for centuries, even millennia, with various "transaction failures" being possible: a ship sails into a port carrying goods and the local satrap decides to simply seize the cargo rather than offer goods in return, a payment is made in gold that turns out to be painted lead bricks, and so on. Likewise, black market or ostensibly illegal transactions have been happening. Examples abound. Lots of transaction failures are possible. In these examples, there is little recourse to "courts." Of course, both examples are good examples of anarchies. (A point well-made by David Friedman over the years and by Bruce Benson in his exhaustive treatment of the Law Merchant, the anarcho-capitalistic system used by traders from various parts of the world to deal with each other in the absence of top-down authoritarian law.) The mistake so many critics/observers of digital cash (cryptographers especially) make is to expect digital cash to solve problems rigorously that are not rigorously solved when the forms of payment are chests of gold, IOUs, shipments of grain, or even Federal Reserve Notes. Or to solve the fraud problem that banks could easily pull off if they wished to (but don't). The protocol for digital cash must be seen as just one part of the larger ecology of economic transactions between actors with varying world models (beliefs in what other actors will likely do). Reputational/Bayesian enforcement is much more powerful than statist law. --Tim May, Citizen-unit of of the once free United States " The tree of liberty must be refreshed from time to time with the blood of patriots & tyrants. "--Thomas Jefferson, 1787
