[EM] Reply to Craig (Aug 1, '01)

[MIKE OSSIPOFF]

Craig wrote:

However, if an election method produces a winning opinion, it is impossible
to remove a pairwise comparison without killing a candidate.

I reply:

Now you're sounding like Demorep. I respectfully suggest that there's
a more nonviolent means for accomplishing that:

It's possible to remove a pairwise defeat from the
pairwise defeats list by taking an eraser in your hand, and pressing
the eraser down on the table entry representing that defeat, and
rubbing the eraser back and forth over that table entry until it
is removed from the table. Obviously if the entries are in ink, then
it would be more feasible to just cross the entry out.

In other words, Craig, we can remove the pairwise defeat from the
pairwise defeats table without killing anyone, and without claiming
that it's no longer true that more people ranked X over Y than
vice-versa. All I said in my previous message was that we drop a
pairwise defeat by removing it from the pairwise defeats table. Which
part of that don't you understand?


Craig continued in part:

If that opinion contradicts itself, you have to
change it so it doesn't contradict itself - which involves inverting the
pairwise comparisons with the smallest majority.

I reply:

Wrong. It can be done by removing that defeat from the pairwise
defeats table. It isn't necessary to invert it. Condorcet didn't say
to invert it. Condorcet said to eliminate it. I thought that error
had been corrected.


Craig continued:

pairwise comparison
is still there, because otherwise you don't get a complete opinion.  Markus
has come to the conclusion that this is what Condorcet means when he
discusses opinions and eliminating defeats.

I hope this might clarify your discussion a bit.

I reply:

Thank you Craig. You're pretty much repeating one side of a discussion
that I thought had already been resolved, or at least fully commented
on. I refer you to my earlier reply in the archives, when Markus
said what you're now saying. But I'll just briefly repeat that
your & Markus's interpretation requires drastically re-defining the
word "eliminate".

Mike Ossipoff




_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp