At 07:11 PM 2/28/2007, Kevin Venzke wrote: >--- Abd ul-Rahman Lomax <[EMAIL PROTECTED]> a écrit : > > While it is already true that Range and Approval *do* > > satisfy that Criterion, in my opinion, on the argument that the > > majority has consented to a different outcome, > >"Majority criterion" has a specific meaning. It in effect says that a >majority can elect their first preference without concealing lower >preferences. You can hardly say that Range satisfies MF on the argument >that a majority could have concealed their lower preferences. That >would be like saying a product is free of charge to you because you can >choose not to buy it.
Yes. It has a clear meaning. I'm suggesting, however, that the Majority Criterion is based on an deeper criterion, which is Majority Rule. This phrase has been ripped from context, which actually weakened the apparent claim by acknowledging precisely what Mr. Venzke points out, that there is a degree of constraint imposed on the "consent" involved. And this phrase came only as an aside to the central point, which was that the Majority Criterion is fully satisfied by a Range election in which a top-two runoff is held between the Range winner and the Condorcet winner, if they differ. (I have not at all examined the question of how to handle Condorcet cycles, but the essence of the idea, with relation to the Majority Criterion, is that the alternative to the Range winner be one who satisfies the Majority Criterion.) (I will also note that the bare-minimum, simple election process where a candidate is presented for election, and the motion to elect may be amended, satisfies the Majority and Condorcet criteria, and, if preceded by a Range poll, also fully considers social utility as imputed by the poll. Quite properly, the majority may reject the Range results. The problem is when we try to stuff all this into a single deterministic poll, we lose necessary flexibility. The Range poll followed by the runoff reasonably satisfies this as a limited election process, full deliberation is always superior, except possibly from the point of view of efficiency. And the efficiency issue could be resolved with ... delegable proxy.) ---- election-methods mailing list - see http://electorama.com/em for list info
