[EM] To ws, re: Condorcet vs Approval

[Abd ul-Rahman Lomax]

At 09:44 AM 7/19/2007, Howard Swerdfeger wrote:
> > It is also important to understand that simple Plurality on a Yes/No
> > question, where the question is formed through deliberative process,
> > will settle on the Condorcet winner, when there is one.
>[...]
>
>I can see no reason why a deliberative process should end in the
>Condorcet winner.

Perhaps Mr. Swerdfeger does not understand deliberative process.

In standard deliberative process, a member introduces a motion, a 
resolution to make some statement or take some action, as described 
directly or indirectly in the motion. Whatever the motion is, it must 
properly become a Question that is answerable as Yes or No.

If a motion is introduced, it must be seconded or it fails immediately.

If it is seconded, debate is open.

At any time during debate, an amendment to the motion may be 
proposed. If the amendment is seconded, debate on the main motion is 
suspended and debate on the amendment is opened.

At some time, any member who thinks that debate has proceeded 
adequately may move the Previous Question, which is a call that the 
question on the floor proceed to vote. If seconded, this is 
undebatable and must proceed immediately to a vote, which vote is an 
action to close debate. The rules often require a two-thirds majority 
for Previous Question to close. The U.S. Senate has reduced that to 
sixty percent.

So let's say that a member moves that John Smith be elected president 
of the society, and this is seconded.

Suppose that there is a member who prefers Jane Doe. This member then 
moves to amend the motion to elect, instead, Jane Doe. If seconded, 
and when debate is closed, the vote determines who remains in 
nomination, it will be John Smith or Jane Doe.

Every amendment like this is a pairwise election, and, as part of the 
process, there is debate and open communication among the members; 
it's quite possible that members could poll themselves -- Range would 
be an excellent method -- to attempt to discover if there is a better 
proposal than the current nominee. But such external polling is not 
part of the method.

If there is a Condorcet winner, it would be unusual that a supporter 
of that winner would not attempt to amend the motion in favor of that 
candidate. If the Condorcet winner is nominated, that candidate will 
prevail, that's actually an initial condition, isn't it?

And there is nothing to prevent the Concorcet winner from being 
nominated (through a motion to amend). And with *any* election 
method, a candidate not nominated is unlikely to win!

While this process might seem tedious, in actual practice it would 
not be. If people wanted to debate each motion, yes, it could become 
quite tedious, but, remember, two-thirds vote can cut of debate at 
any time. I've seen lots of meetings where it was practically this: 
Motion [to do such and such]! Second! Previous Question! Vote! -- and 
voting is by voice vote, and is only actually counted if the winner is unclear.

So the assembly could drill down through dozens of candidates in a 
matter of minutes, if that is what it wanted to do. But the process 
is also completely open. That is, the candidates are not restricted 
to some list -- unless there are special rules that do so.

What would prevent such process from electing the Condorcet winner? 
It would be something that prevents the person from being nominated, 
and every Condorcet method is vulnerable to that kind of "failure." 
The most common cause would be that the preference on which the 
Condorcet designation was based was weak, and it simply wasn't worth 
the trouble to make the nomination. In other words, full deliberative 
process can deviate, a little, toward Range Voting, by action of 
preference strength.

If it's a strong preference as a basis, then, presumably, at least 
two members will introduce the motion and second it, so that 
candidate will prevail.

If there is a Condorcet cycle, all bets are off. However, 
deliberative process can deal with that directly, without being fixed 
to some particular method. I've seen Approval Voting used to move 
beyond the Majority Criterion, where the group strongly valued unity 
and consensus as being crucial to its mission.

Deliberative procedure *must* be superior to any fixed voting method, 
because it can use any fixed voting method if it so chooses. (The 
argument against this position would be that the assembly might 
choose something inferior, and, while this is true, it is also true 
that it can make any inferior decision at any time, a hazard of 
democracy, and it is not improved by establishing inflexible rules -- 
except for a few that slow down the "tyranny of the majority" and 
cause it to realize the dangers of barging ahead.)