Don: a few standards

[Mike Ossipoff]

I know we've been all over this subject, and I shouldn't repeat this
again for Don, but there's an aspect to it that I haven't really
emphasized sufficiently, and so it seems justified to mention it
one more time. Forgive me for talking about those democratic principles
again, but this is for Don, and, besides, I'm describing IRO's failure
of the 2nd one in a better way than I did before.

***

But 1st of all, the whole idea, Don, is to compare methods according
to standards, principles & criteria that are important for their
own sake, things that voters consider important (or whose 
consequences voters consider important, whether or not they've
actually heard of the standards, principles or criteria), or
that electoral reformers consider important. It goes without
saying that these must be things that we consider important
for reasons other than that they might be the same as the
rule that we propose.

Note that I never said that a standard couldn't be the
same, or almost the same, as my favorite method's rule. Actually
that's the case, in fact. However the LO2E standard & the
majority rule standard are imporant for their own sake, and
the fact that Condorcet's method's rule is transparently
pretty much a re-statement of those desired things doesn't
count against my use of that standard in an argument. On
the contrary, since the standards are justified in their
own right, the fact that Condorcet is practically synonymous
with them counts very much _for_ Condorcet.

***

Now, for the basic democratic principles:

#1: If a majority of the voters indicate that they'd rather
have A than B, then, if we choose A or B, it should be A.

This, Don, is something whose desirability is obvious for its
own sake. Majority rule. As I said, only Condorcet (& its
close relative, Simpson-Kramer) will never unnecessarily violate
that principle.

#2: If the number of voters indicating that they'd rather
have A than B is greater than the number indicating that they'd
rather have B than A, then, if we choose A or B, it should
be A.

There are lots of methods that will never unnecessarily violate
this 2nd basic democratic principle. Condorcet is one of them.
IRO isn't one of them.

But wait. It gets better:

IRO is the only rank-balloting method that can violate principle #2,
even if all of the voters rank, alone in 1st or 2nd place, the
alternative whose election wouldn't violate that principle.

***

But when I spoke of Condorcet transparently carrying out
what we want with the LO2E & majority rule standards, 
being practically a re-statement of what we want with
those standards, I was talking about the fact that it
counts "votes-against".

Tom mentioned that a good way to put the goal is that we'd
like our 2nd place ranking for a compromise to fully count
for it against things less-liked, but not against our
1st choice. Voters to whom LO2E is important insist on
casting a vote against, and our goal is so that they
can do that, while still voting their favorite over
everything. I point out that a method that counts
votes-against is transparently doing what LO2E is about,
is virtually nothing other than a re-statement of LO2E.

And, with majority rule, the votes-against count is the
very obvious way to respect basic democratic principle #1
(as numbered in this letter). So obvious that you could
accuse me of writing that rule for that purpose, & I
wouldn't even deny it. It's practically the same thing,
and not only do I admit that, but I assert it.

***

Mike



--