At 10:28 AM 6/7/2005, Chris Benham wrote:
So I don't agree that MMPO has a grave problem with indecisiveness.
Take this often-discussed example:
49: A
24: B
27: C>B
MMPO scores: A52, B49, C49.
The result is a tie between B and C. Which "one vote" would you change
(and how) to change this result into not a tie?
It is not clear that this is a "grave" problem. My answer to the question
is to change the "vote" that created an election system that tries to
automatically create a winner in such a badly divided electorate. My
generic solution to problems like this is to create a deliberative process;
automatic election methods cannot possibly handle the complex
considerations and decisions that might be necessary to avoid, for example,
civil war, if a real election turned out like that.
The most probable best resolution, under a deliberative process, for a
result like that, would be "None of the above." This is frequently true in
close elections. It is only if the result really doesn't matter much that
an automatic resolution through predefined method is truly acceptable. And
if the result doesn't matter much, resolving a tie with a coin toss should
be just fine, and is a lot easier to understand than some complex methods!
As to deliberative process, I think it should be kept in mind, in
discussing and refining election methods, that they are generically an
attempt to automatically choose a winner through a relatively simple
polling process, often with an attempt to minimize cost (i.e., avoiding a
true runoff election, which is one step toward a deliberative process,
albeit with an arbitrary restriction that the truly best compromise
candidate could get excluded in the runoff). Any process which incorporates
the runoff into the initial election is, by definition, non-deliberative.
It is merely a more sophisticated poll, as would be Condorcet or any of the
other methods.
Further, elections must be divided into two kinds, that is, elections for
officers and elections for representatives. Representatives exercise the
sovereign power of the voters on their behalf. Officers, generally, are
servants. (But many democracies confer a degree of sovereignty on officers,
both elected and appointed. Servants can be discharged from their duties at
any time, some elected or appointed officers serve for fixed terms or even
for life.)
Elections for representatives are inherently inequitable, they deprive
those who did not support the winning candidate of representation. Were
this a necessity, it would be justifiable. But it is not a necessity.
Rather, I strongly suspect, given that the alternative -- proxy
representation -- is practically a no-brainer and not at all a new
invention, elected representation were created because they transferred
power from individual citizens, not trusted by those in power, to mass
representatives, more easily controlled and manipulated. And, to be fair,
more generally sophisticated and educated. Supposedly not vulnerable to
"mob rule."
However, elections for officers are not so obviously inequitable, if it is
necessary to choose an officer. Yet a free association of peers, working ab
initio, would be, in my judgement, very, very unlikely to set up a single
officer election system using a fixed election method. Instead, the
traditional default in such associations (for-profit corporations and
non-profits) is to elect a board, a deliberative body, which then hires
officers. And which can fire them at will. For-profit corporations
generally elect boards through a proxy process, albeit a commonly defective
one. Nonprofits generally have either self-elected boards or, if there are
elections by qualified members, these elections, in my observation,
generally use an unsophisticated election method that acts to preserve
entrenched power even in the presence of substantial dissatisfaction; proxy
voting is only rarely allowed, and where it is allowed, in every case that
I've seen, it suffers from the same defect as afflicts for-profit
corporations, which is easy manipulation by existing management, thus
effectively creating an oligarchy even while maintaining the appearance of
democracy. It is no wonder that, given this default, proxy voting is rare
in nonprofits. Without additional protections, it simply doesn't add much
of value.
(Delegability, combined with traditions or rules that prevent existing
management from recommending proxies to members, and with the existence of
independent meetings of members, preferably decentralized such that each
meeting is not impossibly large, could be that protection. Such meetings
could be as simple as a mailing list, new members would be invited to join;
it's essential that traffic on such lists be less than overwhelming, so
fair means would exist to limit traffic if necessary. Such means have been
proposed; the bottom line is that there are two means: one is to keep such
lists relatively small in membership, probably through geographic
organization, and the other is to limit the unrestricted right to post to
lists to persons who hold a certain number of proxies; others can post by
the approval of any such person -- and there would be small lists which are
completely open.)
For a proposal of how to use delegable proxy as a deliberative election
method, see
http://wiki.electorama.com/wiki/Delegable_Proxy_Election
The "Electoral College" created in this system is a deliberative body, but
it could use any of the sophisticated election systems in its process, and
in this case, the voters would be generally more capable of understanding
even complex systems. However, I'd assume that its process would also
include a ratification of the election result, guaranteeing, at a minimum,
acceptance of the election by a majority. Wisely, however, such a college
would seek wider satisfaction. As should always be kept in mind, an
election result based on a mere majority, not to mention a mere plurality,
can lead to civil war, and, at best, it is likely to weaken the social
fabric, creating isolation between the citizens and government.
----
Election-methods mailing list - see http://electorama.com/em for list info
