Class responses on relative importance of some of our criteria for evaluating voting systems:
A reminder of the a priori/a posteriori stuff (which is terminology I will not emphasize):
- a priori Smith fair plurality:
Any method can be "Smithified" by first consolidating to the Smith candidates and then running the method.
For instance, "Smithified plurality" means you first drop everyone outside the Smith set, then consolidate the
preferences and take the one with the most first-place votes after that as your winner. In the
book, the same thing would be called "the a priori Smith-fair plurality method" (see p33), but I think my name
is catchier.
- a posteriori Smith fair plurality: This one's a little different; it just means to take the Smith
candidate with the most first place votes. No consolidation required. You can do this for any method that does
not require consolidation, like plurality or Borda.
- What does it mean to ask whether a method is a priori Smith fair? That means consider
W, the winner set of the original election, and W', the winner set if you consolidate to S
first. If W=W', then we say the method is a priori Smith fair. (In the language of "Smithifying," this
just asks whether method M and Smithified M have to have the same winners!)