Let me explain a specific example: If a person A cheats person B out of his money, it affects me negatively. It poses a risk in the future that person A, being a cheater, will repeat this many times and eventually will have enough money to cheat me. (Both B and me will agree to be cheated by A at the time when we will be cheated, yet it won't be our preference. Kind of like getting a forced move by check in chess.)
Now I am not sure your claim, that this problem can be expressed as property rights, is correct. At very least, these would have to be collective property rights. In fact, I think what we are getting here, if we are to express this correctly as property rights, to is state control over transactions, in one way or another.
> Being absurd doesn't necessarily make it less philosophically valid.
I would say it's a good way not to be taken seriously, but YMMV.
> On a more concrete level, the rejection of quantifiable utility is based on rejection of the extra assumptions it requires.
Which is a bullshit excuse. We have mathematical capabilities to deal with these kinds of problems - probability, fuzzy logic, whatever. It's just laziness of economists to show that the free market assumptions hold under these new assumptions. (To be fair, it's probably not laziness, I doubt they would hold.)
In particular, we can model things like this in computers. I don't know of much cases where this was actually done, but I heard that free market in silico very quickly collapsed into an oligopolic mess.
Anyway, your assumptions should be close to the real world as much as possible. That should be the criteria for choosing them. It's a nice curiosity to have free markets work under some weird, unrealistic assumptions, but it's rather useless in the real world.