[EM] the "meaning" of a vote (or lack thereof)

[Jonathan Lundell]

On Aug 24, 2011, at 7:33 AM, Warren Smith wrote:

>> Lundell:
>> Arrow would not, I think, quarrel with the claim that a cardinal ballot has a pragmatic/operational "meaning" as a function of its use in determining a winner.
>> 
>> But but it's an unwarranted leap from that claim to use the ballot scores as a measure of utility. Arrows objection to cardinal scores, or one of them, is that they are not and cannot be commensurable across voters.
> 
> --(1) using, not range voting, but DOUBLE RANGE VOTING,
> described here:
>   http://rangevoting.org/PuzzRevealU2.html
> the ballot scores ARE utilities for a strategic-honest voter.  Any
> voter who foolishly
> uses non-utilities as her scores on her ballot, will get a worse
> election result in expectation.  This was not an "unwarranted leap,"
> this was a "new advance"
> because the Simmons/Smith double-range-voting system is the first
> voting system which (a) is good and which (b) incentivizes honest
> utility-revelation (and only honest) by voters.

It still seems to me that you're arguing in a circle. A utility score needs to have meaning logically prior to a voting system in order for a voter to vote in the first place. What is utility, from the point of view of a voter? 

Let me put the question another way. Suppose I'd rank three candidates A > B > C.

On what grounds do I decide that (say) A=1.0 B=0.5 C=0.0 is honest, but A=1.0 B=0.7 C=0.0 is dishonest?


> --(2) I agree that it is difficult to measure utilities commensurably
> across voters.
> However, range voting and double range voting do not do so, and do not
> claim to do so.  What IS commensurable across voters, are the scores
> voters give to candidates (since those by the rules of the voting system
> lie within fixed bounds).   Double range voting will extract honest
> utilities from
> each voter, but not commensurably, i.e. with different and not-known scaling
> factors for each voter.
> 
> As a result, neither range voting, nor double range voting, are
> "perfect" regret-free voting systems, and they were never claimed to
> be.  What I am claiming,
> is that a double range voting ballot (honest part) has a MEANING.  It
> has a very definite, very unique, very clear, meaning, which due to
> the Gibbard-Satterthwaite theorem is clearer than the meaning of
> ballots in any (rank-order, deterministic) voting system Arrow ever
> considered in his life.  NO such rank-order system exists or ever can
> exist, in which meaning is as clear as in double range voting.
> 
> Therefore, Arrow's "meaning"-based argument against score-type and in
> favor of rank-order-type ballots, is busted and has no validity.