gsmith@brahms.BERKELEY.EDU (Gene Ward Smith) (10/10/86)
In article <117400125@inmet> janw@inmet.UUCP writes: >Utilitarianism does not pass its own test. Case by case optimiza- >tion is not optimal - for at least two reasons: the overhead is >too high, and human weakness distorts judgement, exposing one to >the temptations of the moment - cumulative temptations of >many moments. This leads to what I call the Principle Principle - >or the Rule Rule - which is that one needs inflexible rules. In >legal sphere, this justifies written law - and as laws are >changeable too, constitutions. This seems to be an argument, not against Utilitarianism per se, but in favor of what is sometimes called Rule Utilitarianism. In any case, to say flatly that Utilitarianism doesn't pass its own test is manifestly a fallacy. If I decide my goal is to optimize X, then it is nonsense to claim optimizing X is not the best way to optimize X. It might be true that the best way to optimize X (in this case human happiness) is not via a situational method. Then we move the utility considerations to the "meta level": we use this as a basis for choosing our rules. ucbvax!brahms!gsmith Gene Ward Smith/UCB Math Dept/Berkeley CA 94720 "What is algebra exactly? Is it those three-cornered things?"J.M. Barrie
tedrick@ernie.Berkeley.EDU (Tom Tedrick) (10/11/86)
>>Case by case optimiza- >>tion is not optimal - for at least two reasons: the overhead is >>too high, and human weakness distorts judgement, exposing one to >>the temptations of the moment - cumulative temptations of >>many moments. This leads to what I call the Principle Principle - >>or the Rule Rule - which is that one needs inflexible rules. In >>legal sphere, this justifies written law - and as laws are >>changeable too, constitutions. Well said. In any case, we are at a point in time when it is becoming possible to formalize "rule-based protocols" as mathematical objects and study this problem scientifically (as I was saying from my soapbox last year ... :-)