[EM] The poll is a good idea--suggestions for a better one

[Abd ul-Rahman Lomax]

At 04:40 AM 6/17/2005, James Green-Armytage wrote:

>         As I mentioned earlier, though, the point of the poll in my mind 
> wasn't
>to declare a winner, but rather to gain individually interesting bits of
>information from every voter's rating of every method. Hence, tally
>methods for the poll itself were not a primary concern for me.

Applying this to what I've been saying about voting methods in general, if 
there is a ratification step, or some kind of what we could call a 
consensus-analysis step, where the analysis is explicitly approved, then 
the voting method itself (i.e., the first step) is not so important. Simple 
Approval, for example, can often provide enough information to guide the 
subsequent steps.

As we know, if I am correct, every election method can display a defect in 
a particular election. Requiring ratification of the result -- and it might 
be desirable to require a supermajority for that in some organizations, but 
my view is that supermajority is highly desirable but fatal if required, 
makes the defects in the system used much less important.

The key, in my view, to making election systems fully rational is to 
include a deliberative ratification step. True run-offs are an approach to 
this, but inferior to full deliberation, which requires, in my view, either 
direct democracy or its close cousin, proxy democracy. But some forms of PR 
can get close (but since proxy democracy, in theory, would be so thoroughly 
simple, if we are designing an election system from scratch, my question, 
like my question about overvoting, is why not? rather than why? ...)

It may seem that direct democracy is sufficient in small organizations, but 
my analysis is that certain effects lead to problems even there that the 
voluntary designation of proxies can solve or at least ameliorate. The 
designation of a proxy should be (and indeed is, if I'm correct), a 
common-law right.

Some people think of "proxy" as a directed vote, and directed votes are 
poisonous to direct deliberation. It's essential that voters in a 
deliberative context be, in general, cognizant of the discussion and debate 
that takes place. But in the systems I'm proposing, there are no directed 
votes. The proxy is a free agent, chosen precisely because the proxy is 
trusted in that capacity; and where a voter does not trust any proxy, the 
voter remains free to directly participate to the maximum extent practical. 
And if you don't trust anybody, and you don't participate yourself, guess 
what? -- you are disenfranchised.

Ah, I do spin off, don't I/

To return to our subject, I fully support Mr. Armytage's intentions. No 
poll is going to be perfect, but that should not inhibit obtaining good 
results, it is only a problem if someone tries to force a conclusion from 
the poll, without allowing further polls. If anyone considers that there is 
a defect in the results, that person is totally free to create another poll 
designed as they think better. Or, in this case, fix the poll. If that is 
done, those who have previously voted should probably be specifically 
invited to vote again, i.e, to revise their vote if, on review, they think 
they could do better.