[EM] Rejecting Universal Domain ...

[Kristofer Munsterhjelm]

On 05/02/2021 21.26, Forest Simmons wrote:
> How about this method for electing a candidate to a position ...
> 
> All candidates and other voters take an extensive multiple choice test.
> 
> Let p(i, j) be the number of questions that voter i and voter j answered
> the same.
> 
> For each candidate j let f(j) be the sum over i 
> of p(i, j).
> 
> Elect argmax f(j)
> 
> For an election where the alternatives are not voters use this method to
> elect a candidate voter (a voter willing to take the responsibility)
> whose responsibility is to pick the winning alternative.

This would elect a candidate who's most similar to the voters; it'd as
such be more analogous to sortition (Aristotelean democracy) than
ordinary elections (aristocracy).

The aristocratic analog would be to have each candidate take a number of
skills tests (IQ, physical performance, etc), and then the voters would
fill in their relative preferences for the different types of skill. The
candidate with the maximum weighted score wins.

In either the democratic or the aristocratic type of election, the tests
would have to be strictly supervised so that the candidates (or voters)
can't cheat. Otherwise, it would be tempting for the candidates to
maximize their scores in an aristocratic election and to try to appear
more like the median voter than they really are in a democratic election.