[EM] Re: majority rule criteria--alternative nomenclature

[James Green-Armytage]

This is James Green-Armytage replying to Ralph Suter.

>I wish you would read and think more carefully before
>you respond.

	Maybe I've said some catty things in this exchange myself, and if so, I
apologize. Anyway, if possible, I'd like to try to skip the insults and
stick to the subject matter from now on.
>
>Are you really incapable of imagining what I might mean
>by an "informal group needing to make quick decisions"?
>Are you really incapable of imagining that it's possible
>that even a "high stakes" decision may sometimes need to
>be made so quickly (e.g., during an emergency in which
>lives are at stake) that time consuming ranked ballots
>are out of the question and that in such situations
>approval voting is by far the best option? If there is
>a better option for such situations, then please tell
>me what it is.
>
	Do you mean a situation where there is less than a few minutes to come to
a decision? I could see approval being helpful there. But if there is at
least that much time, it's quick enough to fill out a ranked ballot and
then count the key pairwise comparisons by show of hands.

>someone else (you) who thinks approval voting
>is close to worthless

	I don't think that it's close to worthless. I just think it should be
understood that minimal dominant set methods are much, much better. I've
said it several times: I would rather have approval than plurality for
single-winner public elections. I suspect that approval would be a very
substantial improvement over plurality, although it's hard to say for sure
before it's put into practice. It's just that I prefer to think of
approval as a short-term, provisional goal rather than a superior
organizing principle that should be used in a wide range of situations.
>
>An even more interesting debate would be a 3-way
>one that included Rob Richie of CVD or another IRV
>advocate who dislikes approval as much as you but thinks
>IRV is the best method.
>
	Sure, that would be interesting; I'd be happy to participate in something
like it. I just objected to what you said because it seemed that you were
using the idea of the debate with Brams to put me down.
>
>You know as well as I do that no method will "guarantee"
>majority rule under all situations and voting conditions
>unless you define "majority rule" in a arbitrary way that
>favors the voting methods you prefer.

	Well, you can't necessarily guarantee majority rule with respect to
sincere preferences, but you can have a method that enables voters to vote
all of their transitive pairwise preferences, and guarantees specific
types of majority rule with respect to expressed preferences. The minimal
dominant set methods guarantee what I call strong majority rule with
regard to expressed preferences. The mutual majority methods guarantee
what I call weak majority rule with regard to expressed preferences.
Approval, range, and Borda guarantee neither.
	The only difference between guaranteeing majority rule with respect to
expressed versus sincere pairwise preferences is voter strategy. Strategy
can distort virtually any voting method, Condorcet methods, approval, IRV,
plurality, and so on. So, the remaining challenge to strong majority rule
methods is to prevent voter strategy from getting out of hand.
>
>
>You failed to address my point. If, in actual voting
>situations with all their uncertainties resulting from
>potential human error, approval voting resulted in a
>Condorcet winner as often or nearly as often as your
>favorite ranked ballot method, how could you insist that
>approval is not a "majority rule method" and your method
>is? If all you care about is hypothetical "perfect world"
>scenarios, fine, but the only problems I'm interested
>in right now are real world ones, especially ones with
>serious human consequences that urgently need to be
>solved or at least made less serious than they are now.

	Okay, let's say that you're provisionally accepting the Condorcet
criterion as the definition of majority rule. A method like ranked pairs
allows all transitive pairwise preferences to be expressed, and guarantees
majority rule with regard to expressed preferences. Approval doesn't make
this same guarantee, because it doesn't allow all preferences to be
expressed (you are forced to vote a lot of candidates as being tied with
one another). Range voting can potentially allow people to vote all of
their transitive pairwise preferences, but it does not make this guarantee
of majority rule, which can cause very real problems. Yes, range and
approval may often elect the Condorcet winner, and if so, that's good for
them, but it's much better to automatically elect the Condorcet winner
whenever one exists with regard to expressed preferences.
>
>
>Again you have replied without bothering to carefully read
>what I said. You entirely skipped over the word "major"
>before "ethnic and other interest groups." Sure, anyone
>can form their own interest group (and people often do),
>but to form a "major" one is a lot easier said than
>done. 

	Well, if the problem isn't interest groups being fabricated out of
nowhere, the problem will be 'minor' interest groups jockeying to
represent themselves as 'major' interest groups, and trying to raise hell
if they don't get a seat. In general, it just seems like a process which
is liable to become somewhat phoney and divisive.
	I mean, in the USA, how many 'major' interest groups do you think there
are? Can you list them? What are the chances of reaching a genuine
consensus on such a list? Would all of the interest group representatives
need to consent to every new law that was made?

>As to how best to deal with the problem of
>representation in "many different situations," there are
>many possibilities that you clearly have not given even a
>moment's thought to. Until you do, your dismissive comments
>are nearly worthless. I'd like to recommend a couple
>of books you should read if you are really interested
>informing yourself about this subject. One is "Is Democracy
>Possible?" by John Burnheim (U. of Cal. Press, 1985).
>The other is "Random Selection in Politics" by Lyn Carson
>and Brian Martin (Praeger, 1999). Neither these nor any
>other books I could recommend have all the answers, but
>they will force you confront problems involved with
>collective decisionmaking that you clearly have never
>confronted until now. As for the general problem of
>representation, it is much more complicated and debatable
>than you are apparently aware. This is a subject I would
>like to study more but (as with many others) haven't had
>nearly enough time. One recent academic article that
>looks very worthwhile is "Rethinking Representation"
>by Jane Manstridge in American Political Science Review
>97.4 (December 2003).
>
	It might be a while before I have the chance to read those books. Feel
free to summarize the points in them that you think are relevant. If you
can't do that, you're not really helping our current discussion very much,
although I do appreciate the recommendations. Anyway, I know that the
tyranny of the majority can be an extremely serious problem for the rights
of minorities, and perhaps supermajority requirements do not solve the
problem entirely. I'm just trying to point out that the alternative you
suggest (representation by interest groups) is problematic as well.

Sincerely,
James Green-Armytage
http://fc.antioch.edu/~james_green-armytage/voting.htm