[EM] using welfare functions in election methods
Abd ul-Rahman Lomax
abd at lomaxdesign.com
Mon May 15 19:13:29 PDT 2006
At 09:38 PM 5/15/2006, bql at bolson.org wrote:
>Though, the tricky thing I've always run into when trying to formulate a
>better social utility measure is that when trying to make sure no one is
>left too far behind, do we unfairly reward people who complain too much?
And if we don't, then we leave some people too far behind....
This is one of a general class of problems that I'd see Delegable
Proxy as solving. The key is Filtering. Delegable Proxy systems that
we are designing use the proxy as a filter. The filter judges
traffic, and passes some but not all. It does this in both directions.
It's fairly clear that someone or some process has to filter
information; indeed, this is the central problem of scale in
democracy. Delegable proxy allows the individual member (citizen,
voter, whatever) to choose the filter. Authoritarian models choose
the filter from the top, that is, a member is, at best, assigned
someone to go to to pass information up the structure, who may also
be the person who transmits information down from the center or top,
which may be, for example, orders.
The scale on which the proxies are chosen is crucial. If it is large,
proxies concentrate too much power at one level. The only harm from
"too small" is that the communication chain becomes long, and thus
not so efficient.
The ideal level is one such that no proxy is overwhelmed with
traffic, yet every member (in general) should be able to contact the
proxy and receive personal attention.
How does it address the problem of "rewarding complaint?" In a DP
system, a complaint does not ordinarily go directly to the top. It
goes to a proxy, who is, most often, a low-level proxy, perhaps
representing only, say 20 to 100 people. (The number depends on the
nature of the organization, it could be larger, if the traffic
permits, and it does work if it is small, just not as efficiently.)
The proxy decides whether or not to pass the information
("complaint") up. The proxy is in the middle, and may face pressure
from below to pass the information on ("or else I'll give my proxy to
someone else"), and will face pressure from above to not pass on
frivolous information and requests ("keep sending me garbage like
this and I'll withdraw my acceptance of your proxy.")
Crucial in the early implementations of DP is the Free Association
context. In FAs, there is no power other than the power to
communicate that is conferred by being at a high level, because power
in an FA remains in the hands of the members, generally. The FA
itself does not make controversial decisions, but individual proxies
can advise their members in any way they choose. Presumably after
communicating with others! The theory is that such an organization
will naturally seek consensus, because it is empowering. Otherwise
you end up with factions pulling against each other, with much less net result.
Yet there is no requirement that unanimity be reached. Essentially,
the organization, in theory, will value unity and will seek it, but
not obsessively. The DP structure actually allows individual
attention to be given to every minority, but only relatively large
minorities get direct consideration at the top.
If it is expected that proxies be available for communication to
their clients, this puts pressure on the proxies to not accept too
many clients. And especially not many clients who are complainers!
Many people, encountering this concept for the first time, miss this.
They assume that having more proxies gives you more power, therefore
people will crave and seek it. Probably they will, even without
direct power. There *will* be prestige coming with being a high-level
proxy. But if you get there by being available to thousands of
clients, directly, it won't be worth the work. So the way to get to a
high level, without working yourself to death, is by becoming the
proxy for others who themselves represent clients. And once you are
in that position, you want your clients to do a good job of
filtering. You don't want more than the optimum number of clients.
The structure will regulate itself.
This is theory. We aren't going to really know until we have
functioning DP systems. But the *theory* looks very good, from my perspective.
http://beyondpolitics.org/wiki
http://metaparty.beyondpolitics.org
More information about the Election-Methods
mailing list