03-29-2003, 11:33 AM
Hello.
Well, IIRC the problem here is to define the structure of the environment. If you impose too strong requirements on the environment everything is well provable but kind of trivial (e.g. all participants are totally incooperative then obviously any other strategy would loose). On the other hand if you try to treat the environment without any requirement you will end up with an chaotic highly non-linear system where you can not prove anything.
The gray area in between is quite interesting, though...
Additionally you can "play" with many parameters (e.g. reward for a successful cheat, etc.) which makes it challenging with respect to discrete mathematics.
For my taste that is a bit optimistic. If you would start up with an easily structured environment (e.g. 90% eye-for-an-eye, 5% totally cooperative, 5% totally incooperative) there could be some useful results from science (be it the mathematician's way (prove) or the computer scientist's way (simulate)).
But in most of the real-life cases I would rather assume that the environment is highly chaotic and thus proving fails (due to complexity) while simulating only gives short term "weather forecasts". Not a good thing to base an ethical system on this...
Bye,
ergates
Quote:Oh, and BTW, studies in game theory indicate that, under certain conditions, "cheating" (i.e., not cooperating) is a viable short term strategy. However, it tends to change the conditions in such a way that cheaters are no longer rewarded and co-operation is the best long range strategy. At present, AFAIK, those statements are conjectures based on observation, computer models, and the "gut feel" of mathematicians. They haven't been proven yet.
Well, IIRC the problem here is to define the structure of the environment. If you impose too strong requirements on the environment everything is well provable but kind of trivial (e.g. all participants are totally incooperative then obviously any other strategy would loose). On the other hand if you try to treat the environment without any requirement you will end up with an chaotic highly non-linear system where you can not prove anything.
The gray area in between is quite interesting, though...
Additionally you can "play" with many parameters (e.g. reward for a successful cheat, etc.) which makes it challenging with respect to discrete mathematics.
Quote:But at least the first few halting steps along the path of an ethical system based on reason have been taken, and it seems to agree, more or less in the few places where it can be compared at all, with the bulk of traditional morality.
For my taste that is a bit optimistic. If you would start up with an easily structured environment (e.g. 90% eye-for-an-eye, 5% totally cooperative, 5% totally incooperative) there could be some useful results from science (be it the mathematician's way (prove) or the computer scientist's way (simulate)).
But in most of the real-life cases I would rather assume that the environment is highly chaotic and thus proving fails (due to complexity) while simulating only gives short term "weather forecasts". Not a good thing to base an ethical system on this...
Bye,
ergates