On Jan 22, 2010, at 7:19 AM, Terry Bouricius wrote: > Arrow never uses the word "spoiler" in his theorem (original nor revised > version). You may be thinking about his independence of irrelevant > alternatives (IIA) criterion. While this could be expanded to have some > bearing on the concept of "spoilers," it is not the same thing. Firstly, > Arrow used IIA (as well as Pareto consistency and non-dictatorship) as > desirable characteristics of a social ranking of options, not finding a > single winner (or winning set).
It's a point worth keeping in mind. We toss Arrow's Possibility Theorem around pretty loosely, when strictly speaking we should be talking about Gibbard-Satterthwaite, or (better yet) Duggan-Schwartz. There are, of course, family resemblances. ---- Election-Methods mailing list - see http://electorama.com/em for list info
