[EM] IRV and Brown vs. Smallwood

[Abd ul-Rahman Lomax]

At 10:17 PM 1/18/2009, Terry Bouricius wrote:
>Markus,
>
>The source you cite is a Minnesota League of Women Voters report that
>refers to an article with legal analysis in the Minnesota Bar
>Association's journal, dealing with the Minnesota statutes and Minnesota
>constitution.
>
>"Municipal Voting System Reform: Overcoming the Legal Obstacles." Bench
>and Bar of Minnesota. Vol. 59, No. 9, October 2002
>
>FairVote is not responsible for reports by the League of Women Voters or
>lawyers writing scholarly articles.

Perhaps that's technically true. However, that article has been cited 
many times by FairVote, and one of the coauthors was with FairVote:

"TONY ANDERSON SOLGARD is chair of the board of FairVote Minnesota, a 
nonpartisan, nonprofit organization educating the public about voting 
systems and their effect on the quality of democracy."

The interpretation of Brown v. Smallwood, many times argued by 
FairVote, including as an intervenor in the present case, has been 
one which would indeed render any Condorcet method unconstitutional 
in Minnesota. Not anywhere else, because the Brown decision was 
idiosyncratic and the arguments in it not accepted elsewhere. 
However, the Later-No-Harm argument isn't what the Brown court 
actually decided upon, and they repetitively expressed 
dissatisfaction with any kind of alternative vote, where a candidate 
preferred by one voter faces many other candidates preferred by 
another voter. In this argument, that the other candidates are 
presented together or one at a time doesn't seem to be relevant.

As you know, it's a bizarre decision, because on the one hand, they 
quoted with obvious approval that what mattered was the number of 
voters supporting a candidate, not the number of marks on the ballot, 
but then they complained about the marks on the ballot, with Bucklin, 
exceeding the number of voters. On an IRV ballot, generally, the 
number of marks on the ballot exceeds the number of voters!

In the end, with Bucklin, the winner is the candidate who was found 
to have been voted for by the most voters. How the method gets there 
is something else. IRV sequentially assigns votes, Bucklin considers 
them in ranks and ultimately all of them at the same time if a 
majority hasn't been found before that, a Condorcet method considers 
them all simultaneously as well (though you can think of it as a 
series of round-robin elections). The problem with IRV is that some 
votes are counted and some are not. I agree that there is a problem 
with this, but disagree that it makes IRV unconstitutional.