Responding to this again, from a somewhat different perspective.
At 05:03 PM 5/11/2008, Fred Gohlke wrote:
Good Afternoon, Abd ul-Rahman Lomax
re: "Mr. Gohlke, do you care to look at this?"
OK. Absent a specific definition of the group of voters to which
you've assigned a ratio of 'p', 'p' can be taken to represent any
group of people who have an identifiable political orientation, and
'x' is the balance of the electorate.
This isn't what was proposed, actually. p represents the proportion
of representation of some characteristic of the voters in those whom
the voters vote to represent them.
I.e., if the choice process amplifies the representation of some
trait, unless some countervailing process intervenes, that
multiplication repeated over the selection process stages continues
to increase the representation. Equating this with "political
orientation," and assuming that it must be "indentifiable," restricts
the application such that the point is more easily missed, even
though that is one example.
A single stage election ordinarily amplifies like this. Single-winner
elections inherently, as ordinarily accomplished (contested
elections, where some voters win and some lose), have this problem.
That a particular political party is disproportionally represented is
only an example. Proportional representation systems counter the
effect. *Some* of these are party-based, but not all. STV, for
example, in common use, is not intrinsically party-based and party
choices aren't a critical part of the process, but, typically, voters
have been allowed to vote for a party slate (which is ordered in a
way that the party has decided) rather than for candidates
individually), but it is still the voter's choice, the voter can vote
entirely without regard for party. And then the degree of warping of
representation varies inversely with the number of seats elected from
a particular district. A single parliament elected for one single
district does not warp much.
Therefore, as you say, "With many layers, as is necessary for this
system to represent a large population the proportion of p rapidly
approaches zero ...", which shows that ideologues ... of any stripe
... will be eliminated, leaving the non-ideological majority of the
people to select the best among themselves as their representatives.
Two problems with this comment: first of all, the assumption made was
that there were at least two exclusive traits or sets of traits, and
that one of them was in the majority. However, that's not the core of
the problem. The problem is selection bias. If some characteristic of
the voters leads them to preferentially select for some trait, then
that trait will be amplified over its natural frequency among the
voters, in those whom the voters select. Political affiliation
*could* function this way. However, the problem is much more general.
Suppose people tend to choose taller people, other things being
equal. We could expect that average height would increase with stage.
What I want to happen is that voters select preferentially for
trustworthiness. They will probably tend to do this when they are in
relatively homogenous groups. People who think alike are more likely
to trust each other. For *representation*, trust is crucial. I can't
really be represented by someone I trust. The system proposed allows
people to be misrepresented: either they got stuck with a group where
there were two others who more easily agree with each other, or in a
group where there were three who could not agree, and neither of
these is necessarily due to any fault of the voter. And the process,
inherently, does not allow the necessary time for getting to know
each other; and whatever time is spent doing this is wasted when the
next election round occurs, because, presumably, the groups will be different.
The complex rules which Mr. Gohlke made up are an ingenious solution
to *certain* problems. Unfortunately, he neglected to solve the
fundamental problem, which is representation in deliberation; and his
structure will leave a substantial portion of the population,
essentially, out in the cold. As I mentioned, for making a single
decision where people's minds are already made up, it would work
quite well, but that's a lot of complicated process to apply when
there are much simpler ways of doing that!
----
Election-Methods mailing list - see http://electorama.com/em for list info
