Multiple conflicting "majority" rules

[Saari]

In a message dated 98-04-23 20:59:38 EDT, you write:

<< In like manner, each legislative proposal (i.e. bill) should only need
 majority acceptability.  If there are 2 or more conflicting proposals with
 such majority acceptability, then they should go head to head using Number
 Voting (1, 2, etc.).

I disagree with the pseudo-religious attitude that "majority" (i.e. 50.1% or
1:1 ratio) is the only sensible passing threshold for any and all democracies.
We should be able to be more generic and less restrictive.  

Assuming a voting method that allows for votes of "Support" or "Oppose" or
"Neither" or perhaps other combination such as "Mixed support and oppose".  We
can do well to simply say that the "score" for a proposal is the ratio of
support/opposition.

Please note that "support/oppose" ratio is not quite the same as "percentage
support".  Voting systems that only offer "Yes or No" make it quite impossible
to determine if a non-Yes voter was actively opposed or merely disinterested.
Under my proposed system, if out of 100 voters a proposal gets 20 support and
3 opposed then it would pass.  It is not so much the number of supporters that
matters, instead it is the proportion of supporters vs opponents that counts -
not quite the same thing.

ANY GIVEN GROUP may establish its own passing criteria, which among other
requirements would probably include a "passing score".  Some groups might
establish 1:1 for all decisions.  Another group might establish 3:1 as the
general passing score.  A more "consensus"-oriented group could simply
establish 10:1 as the passing score.

There is no provable "correct passing score" for all groups.  It should be for
each group to establish for itself.  Personally, I like something between 3:1
and 10:1.

One other note: it is common for some groups to have different "passing
scores" for different classes of proposal, e.g. "new taxes require 2/3
approval".  I believe this is a fundamental mistake, as it creates a need for
dictatorial/authoritarian/hierarchical oversight to "classify" each given
proposal.  This is contrary to my design goal for a true democracy.  On the
contrary, I believe it is quite feasible for a "one standard fits all" method,
meaning that a group would have a uniform passing criteria for ALL proposals
not a hodgepodge.
 
 DEMOREP: Any exaggerated supermajority requirement (5 to 1, or whatever) for
 legislation is by definition having a minority veto. I note the continuous
 minority veto in the U.S Senate due to its 60 percent requirement for
stopping
 debates.  I also note that most current laws have been passed by minority
rule
 gerrymander regimes.  Would such laws now need to get a 5 to 1 vote to repeal
 them ? >>

This would be up for the group to decide.  There also needs to be some method
for resolving conflicting passed decisions.  Requiring a ranked revote is
impractical, especially if the decision as to whether given proposals conflict
or not must necessarily be made hierarchically.

I could imagine a system where voting can continue, even after a proposal has
"passed".  We could then say "A proposal has passed when it reaches the "pass
threshold" support/oppose ratio (and other requirements involving sufficient
time to vote, etc.).  Once passed, the proposal is "in effect" but further
votes may still come in.  If the support/oppose ratio ever drops below the
"unpass threshold" then that proposal is nullified."  A group might then have
a 5:1 "pass threshold" and a 2:1 "unpass threshold", or any other desired
values.

If two proposals both are passed (whether at the same time or different times)
then you have a problem not dissimilar to problems we already have today
around conflicting laws.  Under this new system, it might resolve by the
weaker proposal attracting later Oppose votes such that it eventually un-
passes.

Some of these objections fall under the more generic category of "But what if
the group makes a bad decision?".  This is an inherent question with any
democracy that cannot be solved by a voting system.  The job of a voting
system is NOT to create "good" outcomes which are contrary to the desires of
the voters.  

The job of a voting system must be to reflect the desires of the voters and
not introduce artifacts, hierarchy and distorted outcomes.  In my view,
standard "majority" systems which require lots of support votes but do not
distinguish between apathy and opposition are distorted.  Also in my view, a
simple 1:1 passing criteria is inadequate - it says that it is acceptable to
make group decisions despite widespread opposition.

But if a group chooses 1:1 as a passing score then that is their right.  It is
NOT *our* right, as modelers and theoreticians, to make this decision for
them.  We should instead find voting methods that allow voters to express true
feelings without introducting incentives for distortion.  But as for a
"passing score" I feel we are better to not impose an arbitrary 1:1 passing
standard.

I believe a group can work very well under a 3:1 or 5:1 or maybe even 10:1
standard for all decisions.  And I believe the decisions reached will be of
superior quality than simple 1:1 decisions.  This belief has not yet been
rigorously tested, so it must remain tentative for now.  I have not yet heard
a convincing argument against it.  (Most arguments I have heard to date,
claiming impossibility or impracticality, would seem to equally imply that 5-
or 10-person "consensus" groups are also impossible or impractical.  But life
indicates otherwise.)

"Consensus" means that a single vote of Oppose is sufficient to cause a
proposal to fail.  It is fundamentally similar to a high passing ratio.  If a
10-member group is operating under "consensus" or under "20:1 passing ratio"
the result is the same - a single Oppose vote is sufficient to cause a
proposal to fail.  (But "consensus" doesn't distinguish whether the rest of
the group is actively supporting or merely going along, so I think the
"passing ratio" method is superior.)

If some groups get along under "consensus" (effectively 10:1 or 20:1 ratio)
and other groups get along under "majority" (effectively 1:1 ratio) then I see
no logical reason why an intermediate passing ratio such as 3:1 or 5:1
couldn't also be workable.

What I do see is many "majority" systems having exceptions - usually 2:1.  And
I see many "consensus" systems having exceptions where a sufficient number of
supporters can usually induce an opposer to back down - meaning the effective
ratio may only be 10:1 instead of 50:1.  Perhaps everybody is trying to move
toward a "happy medium" of 3:1 or 5:1 - where opposition is taken very
seriously but a sufficient number of supporters can still override.

Mike Saari