[EM] working paper on delegable proxy voting

Abd ul-Rahman Lomax abd at lomaxdesign.com
Wed Nov 16 12:45:20 PST 2005

At 03:07 PM 11/15/2005, Paul Kislanko wrote:
>  Abd ul-Rahman Lomax quoted:
>At 12:01 AM 11/14/2005, James Green-Armytage wrote:
> >. Issue independence: Even when there are multiple issues on the same
> >ballot, I should have the option of indicating separate proxies for
> >separate issues, while still voting directly on other issues, if I choose.
>This would appear to be nirvana, but I can't see any practical "method" for
>implementing it. How does one "vote if I chose" on an "issue?" It has been
>well-documented that that pollsters get different results on the same issue
>depending upon how they phrase the yes/no question.

As I wrote, I think that specific issue proxy assignment, within a 
single organization, adds far too much complexity, and, I suspect, 
breaks the delegability feature.

However, I don't think that Mr. Kislanko understood that it is direct 
democracy which is being considered. Because it is direct democracy, 
a member may vote on any issue directly. Under proxy voting, the 
proxy authority is only active when the delegator does not directly vote.

As to the framing of the question, I presume standard deliberative 
process. Robert's Rules, for starters. To proceed to a vote under RR, 
to close off the possibility of further debate and amendment, 
two-thirds must approve the Previous Question, i.e., must approve a 
motion to put the question on the floor to a vote; or the question is 
put to a vote by default when there is no more debate or amendment or 
other interference (such as a motion to table, etc.)

>I don't see how having multiple delagable proxies can do anything but make
>an already-intractable problem more complex. If I delegate to A for issues
>{a,c,e} and to B for {b,d} and I vote on issues {f,g} can we count anything
>except which issue gets the most support?

Two different things are being confused here. "Issues" are 
"questions," as in RR. Questions are not compared with each other. 
They either pass or they don't. The problem with the proposal is not 
that it is unclear, it is that it sets up a massively complex system 
of determining who votes for what, which must be assigned by 
question, i.e., every time that a vote comes up. This pretty much 
defeats the purpose of proxy voting....

However, if "issue" means a whole general topic, I suggest that there 
is a much simpler way to deal with it. Presently, in some state 
legislatures, members may vote by proxy in committee. They have 
assigned a proxy to another member of the committee. Committees are 
organized by general topic.

Basically, this is organizational division by topic. If you are a 
member of a body, organization, or committee, you may assign your 
vote to a proxy. Each organization, body, commmittee, whatever, has 
its own process, membership list, proxy list, etc.

One organization, one proxy. That's what I'm proposing. Simple.

>I talked offlist awhile back with Dave Gamble about Isaac Asimov's 1950s-era
>short story "Franchise", in which a a single "best fits the profile" voter
>is selected as a proxy for all voters, and his opinions on all the issues
>(with control questions on non-issues) uniquely determined the winners of
>all elections. Cool story.

Basic problem: who selects? How is the profile determined? If it is 
by polling the voters.... then why not just poll the voters? What 
does the single voter add?

Now, delegable proxy *could* end up with something similar; 
essentially, DP might create, without elections, a relatively small 
body which collectively represents almost all the citizens. I've 
called a "superproxy" a person who represents, directly or indirectly 
a majority of voters, or perhaps a supermajority. Ideally, of course, everyone.

>Subsequent to that, I suggested that it might actually be possible to hold
>an election where each voter anonymously provided input regarding every
>issue, and separately each candidate filled out the same ballot. A
>multivariate correlation of voters' responses to candidate's responses would
>select the "right" candidate.

An unbelievably bad idea, bad, that is, if human beings are involved. 
So a candidate, based on polls, determines how to fill out the ballot 
in order to best match what the voters will prefer. The candidate who 
is best at anticipating voter preference, perhaps the one with the 
most money to put into polling, focus groups, etc., wins. Kind of 
like we already have, only more complicated.

Of course, this being futuristic, the program that matches the 
ballots analyzes the candidate ballots and if they too closely match 
voter preferences, it determines that the candidate is lying, no 
human being would actually have that set of preferences. Of course, 
once the computers are that smart, there are plenty of science 
fiction stories where they simply run things directly. Sometimes this 
is cool, sometimes it's a nightmare.

Like, really, I want to spend my time figuring out whether we should 
issue sewer bonds or not, how much money to put into road repair, 
and, yes, the public employee retirement plan needs attention, and 
that is just for today. Tomorrow, more issues for me to fill out ballots on....

Ah! I know what to do. If I have the right to express my preferences, 
surely I can delegate that right to someone. I'll name a proxy! [If I 
can find anyone willing to serve. Ironically, some people thinking 
about delegable proxy worry about vote-selling. As if. It is much 
more likely that people will pay their proxies than that the money 
will flow in the other direction.

As Mikael Nordfors points out, naming proxies is what rich people do. 
What I've been calling the "proxy," he's been calling the "advisor." 
I like that, it emphasizes the other direction from "voting."

>Also science fiction/fantasy, I think, because of the problems involved in
>formulating the issues.

That's only the beginning of it....

More information about the Election-Methods mailing list