[EM] Dave on approval, ranked ballots

[Abd ul-Rahman Lomax]

Warning, this post is long -- big surprise! :-) -- but it does actually 
stay on the topic of election methods, how to implement election methods, 
how to get from here to there, and evidence regarding these ideas.

At 03:51 PM 7/28/2005, Simmons, Forest wrote:
> From his remarks below I think that Lomax would agree with me that 
> candidate A in the following example would likely be better for the 
> electorate than candidate M:
>
>Sincere
>55 M>A>>B
>45 B>A>>M

Yes, if those were sincere votes. However, note that vote counts like this 
represent a very highly polarized electorate. It is quite possible that the 
best outcome for this election would be "None of The Above." Theoretically 
Asset Voting could result in such an election -- NTA --, if the candidates 
are allowed to give their votes to someone not on the ballot (as voters 
normally can with write-in votes.)

>On the other hand, M would be more likely to win the approval election, 
>unless a significant percentage of the M supporters were to vote 
>altruistically like Lomax, for A.

Yes. I'd quibble with the description of my voting behavior as 
"altruistic," though. I'd hope that it would be, instead, "enlightened 
self-interest." I think I and my family will be safer in a society which 
operates with a higher level of consensus. Think of Ruanda and the election 
of the Hutu government. The first choice of a majority of the people can be 
disastrous.

I don't think mass genocide is at all likely in the United States; Ruanda 
is raised simply to bring into relief the fact that a candidate acceptable 
to all major factions in a society may be a much better winner, for the 
society, than the first choice of a mere majority. This gets even more 
clear in an election environment where voter turnout is poor, and the 
winner is chosen by plurality, but there would still be an issue in 
environments that require a majority to win.

>In either case, we get a good winner.

By the low standard of plurality elections, yes. Note that in this case the 
result is probably the same as in a plurality election. Approval Voting is, 
quite certainly, not guaranteed to improve results. It merely allows for 
that possibility, at low cost in ballot complexity, in conversion costs, 
and, indeed, in the transitional political costs which may otherwise 
prevent significant reform.

I think that once Approval Voting is in place, people would then recognize 
that Approval Voting does not allow ranking. I think that Fractional 
Approval Asset Voting (my variation on Mr. Smith's proposal), which also 
does not allow ranking, nevertheless would be quite sufficient. Once you 
have the idea of proxy voting in the system, choosing one proxy, or 
dividing one's vote equally among proxies, is close enough to pure Asset 
Voting (which, as proposed, would be much more complex to use and to count) 
that the difference is moot. FAAV, though it would work just fine for 
single-winner elections, and as with Asset Voting, is designed for 
multi-winner PR elections. Fractional Approval as distinct from regular, 
full-vote Approval, is necessary in the proxy environment, else those who 
approve more than one candidate actually would gain voting power.

In FAAV, it is Asset Voting, but the only choice of the voters is yes/no on 
all candidates (just as in standard plurality elections, with overvoting 
allowed). If the voter casts n votes, each one of them counts for 1/n vote 
in the totals of those chosen.

So FAAV essentially gives the right to choose winners to an individual (one 
vote) or to a committee (more than one vote cast). It's been pointed out 
that voters could decide to trust an individual (and if a quota of them do 
this, then that individual is elected) or, for example, a *party* (perhaps 
the voter votes for all candidates from a party), or, for a caucus within a 
party (the voter votes for all candidates who belong to that caucus.) And 
in no case is the voter's vote wasted. In PR-FAAV, as in regular Asset 
Voting, almost every vote is eventually devoted to a candidate and brings 
that candidate to the quota. *Everyone* would be able to claim "my vote 
elected a candidate," though in some cases, they would not know exactly 
which candidate. (If they voted for A, who received more than the quota, 
then they would not know which votes, individually, were transferred to 
another winner.)

So FAAV ballots would look exactly like present plurality ballots, except 
that they might be longer, with more candidates. Possibly *many* more. I'd 
assume that rules might determine how candidates would get on the ballot, 
but presumably others could be written-in. And write-in votes, once again, 
would not be wasted, except in the case of intransigence of a candidate 
holding votes. "If I can't win that last seat, nobody can." Fine. The 
assembly would be a member short. Happens all the time, for other reasons. 
If you choose to vote for an intransigent candidate, you got what you 
chose. Intransigency is not a desirable trait in a democratic society, I'd 
suggest.

I do suggest DP for the resulting electoral college, as have others. That 
allows minor "winners" to choose to pass their votes on and not participate 
further, unless they want to. (You are a "winner" in this sense if you get 
even one vote. Which might be your own. In what is on the wiki, however, 
you must be registered as a proxy to participate further, and you exercise 
your own vote in that way, in the post-election process, which is why 
registered proxies would not vote in the general election by secret ballot. 
The secret ballot layer is necessary for conditions where coercion or 
extortion are reasonable possibilities.) Whether or not proxy *voting* 
would be allowed in the preliminary votes of the electoral college, I'd 
suggest that it might be advantageous under dangerous conditions to require 
that all votes be cast directly by the "electors," secretly. This would 
deal with coercion directed against low-level "winners." There are also 
technical possibilities that I won't go into; generally, my preference is 
to avoid trying to define in detail what final operational structures would 
look like, but only to sometimes suggest possibilities simply to allow the 
impression to form that this might actually work, that this might actually 
be practical and realizable.

>The case I worry about is
>
>sincere
>60 A>>B>C
>20 B>C>>A
>20 C>B>>A
>
>where the corporate polls are reporting:
>
>disinformation
>40 A>B>C
>30 B>C>A
>30 C>B>A

Disinformation is, of course, a crucial issue. I have my own solution for 
it, which is, naturally, part of the FA/DP concept. If enough of the public 
self-organize using the FA/DP principles, they can easily afford their own 
media and their own surveys. *They already spend, directly or indirectly, 
quite a bit of money on media.* Directing some of that money into stock in 
media corporations, and assigning corporate proxies in concert through an 
FA/DP structure, could result in media where honest and depth coincide with 
profit motive.

Disinformation, if not countered by reliable and not only trustworthy but 
trusted information, can distort any election in much more profound ways 
than by causing incorrect strategic voting.

Note that the disinformation posited here was vastly different from the 
truth, and even a fairly primitive survey that was not distorted would 
uncover that. In any case, it is crucial that we begin to have reliable 
media. I say that the best way to do that is to either buy existing media 
or create new media.

*How* such media would be controlled is crucial. If all we do is to set up 
yet another non-profit corporation, which will inevitably develop its own 
self-interest, and which will act on that self-interest unless there is a 
restraining force, we will merely be going down the same old road again. I 
think NPR is great, I listen to it every day, but neutral and fully 
reliable, it is not.

-- Though it may be *more* neutral than other alternatives out there. 
Neither is Radio Free America reliable, it is a progressive-side reflection 
of the trash on the other side. In other words, at least some of it is 
trash. Believe it or not, this is not a criticism! (or, perhaps more 
accurately, it is certainly not a condemnation.) It is merely a reflection 
of the shortcomings inherent in creating polarized media.

>In this case all of the pundits are going to encourage A supporters to 
>approve B as well.  I'm afraid that most of the A supporters would fall for it.

B is not a bad choice, though, in that election. B is the first or second 
choice of *all* the voters. Whether or not B is a better choice, I'd say, 
depends on the depth behind the ranking of A over B. To know that, we'd 
need something like range voting, assuming that voters vote sincerely and 
do not merely give all their vote to their favorite. But Smith's Asset 
Voting is much better.

>That's why I think that there should be some standard or target that comes 
>out of the primary or preliminary vote that sets the stage for 
>Approval.  "Approve or Dare" or "Lottery Informed Approval" or some other 
>source of binding (hence reliable) information.
>
>Small groups could use repeated approval balloting where the scores 
>accumulate so that manipulation is counter-productive.

I agree that there are means to make Approval more reliable. I also point 
out that such means potentially exist outside of the election system, that 
could function without changes in election law. Plurality works fine if the 
effective field has been winnowed by pre-election negotiation among the 
parties. And the voters.

*Within* what would be election law changes is Mr. Smith's Asset Voting, 
which essentially solves certain election problems by including a 
deliberative phase in which every vote has power.

>Large groups could form a small subset of voters by random sample to 
>choose a "target candidate."

It's called "delegable proxy," only instead of random sample (imagine the 
trust issues!), there is a *choice* process. A small subset of voters end 
up negotiating the election results (in political DP) or what the 
recommendation to members (in FA DP) will be. (For the most part, in pure 
delegable proxy, any voter may still participate, but few will actually do 
so, and those few will not gum up the works, because of how the rules, I 
expect, will develop.)

>   Then this target candidate is the winner unless some other candidate 
> gets more than fifty percent approval in the general approval election, 
> in which case the candidate with the most approval wins.  Since the 
> proceedings of the small group are public, the general voters get an idea 
> of the relative strengths of the candidates, etc. before they vote.

Basically, this is my idea. I've just detailed the mechanism by which such 
a group could be created, without having elections and thus without 
excluding *anyone* from the process. Because the group size would become 
large, there would need to be certain restrictions on how individuals 
participate at a high level; my suggestion is that large "meetings" (real 
or virtual) define a subset of members, based on level of proxy assignment, 
who may directly speak (or post); others would have to convince a member of 
this subset to present their ideas or to enter their motions. Any qualified 
member. Including, most specially, those whom they chose, directly or 
indirectly, and who presumably have given them permission to communicate 
personally with them. (Would you choose someone who did not give you that 
permission, or who did not demonstrate, whenever you tried, that the person 
actually read and considered what you wrote?)

Sometimes the word Free Association conjures up an image of a free-for-all. 
But anyone who has actually participated in the close analogies to Free 
Associations that already exist would know that this is not the case. 
People, even quite damaged people, are much better behaved than that. And 
the Free in Free Association also implies the freedom *not* to associate.

In Free Associations, decision-making is doubly isolated from dominance by 
a clique or by the equivalent, a set of members who habitually take the 
time of everyone with long speeches; as in the words of a local preacher in 
our Town Meeting town, people who have nothing to say and who are willing 
to take a long time to say it. The fact is, though, that I haven't seen 
this actually happen at Town Meeting. People have been brief and to the 
point. But this is a small town. What could be called the "crank effect" 
becomes more and more of a problem as the population sample becomes larger, 
and this is one of the two reasons why Town Meeting has always been 
abandoned as towns grow. A local town, Amherst, still calls itself a Town 
Meeting town, but their "Town Meeting" is actually a representative 
assembly -- though a really large one. You have to be elected to speak.

Indeed, the Amherst Town Meeting is a step toward delegable proxy, but has 
its problems because the proxy concept was missing and instead there is a 
large elected body, I think it is elected by neighborhood. There have been 
two attempts to discard the Town Meeting in Amherst and both failed by only 
a very few votes.

I do wander. Were this a face-to-face meeting, I'd think this quite 
offensive. I argue that in written material, it may be less than fully 
effective, but it is not offensive, because anyone can simply skip it. It 
would be quite different at a Town Meeting which is already running late 
into the night.... (And, by the way, I've received multiple congratulations 
for my actual participation in Town Meeting, where I was often able to 
state the consensus, leading to immediate vote and decision, almost always 
unanimous. I don't like long speeches by others and I don't want to inflict 
them on anyone.)

With Delegable Proxy, if there is a meeting where a bore has the right to 
speak, the other participants can indeed simply go home, or reconvene as a 
caucus. If for some reason the meeting running long is actually important, 
it is not merely a hearing, but a vote is going to be taken, the 
participants who leave can leave behind a proxy or a few, to ensure that 
nothing untoward happens from what would otherwise leave the cranks in 
charge (or a radical clique, as in alleged previous Communist union 
take-over techniques).

So that's one protection. The other protection is that in a Free 
Association, the Association itself does not take a position on any issue 
of controversy. It exists solely to facilitate communication, and one 
aspect of that is the measurement of consensus. A vote in an FA never 
*decides* anything, unless it is a purely internal matter, what in Robert's 
Rules is called a Question of Privilege (i.e., should we turn down the 
thermostat or open the window?). Rather, the Association reports the 
results of polls.

Only if a poll reveals unanimous consent to a position (voting directly or 
by proxy, and I'd include procedures for a member who is offended by their 
proxy's vote, and who remains unconvinced by the proxy to change their 
vote) would the Association announce the result as the position of the 
organization. In a large political FA, I'd think this would almost never 
happen. Indeed, I'd advise against the organization taking a position if 
that position might inhibit others, as yet unconvinced, from joining. But 
it could be reported that the poll was unanimous. Anyone outside who 
doesn't like the prospect of that could simply join and vote in the poll!

But consider the power of an open FA, one with no particular bias in 
membership (other than a willingness to join an organization specifically 
designed to make joining and participating as easy as possible). Pick *any* 
other member and name that person as your proxy. If the proxy refuses to 
accept, most proxies would, I think, suggest a replacement. And you can 
name your husband or wife, mother, partner, father, sister, brother, 
neighbor or any other member. Perhaps a movie star.

(About that, my views are quite different from what is often said. For me, 
it is only important that the proxies, on average, be smarter, more 
reliable, of higher character, or whatever characteristic that would make 
the proxy a better participant at a higher level, than those who choose 
them. If people are so careless as to choose a media figure whom they do 
not personally know, or at least know well enough to warrant the choice, 
then I'd suggest that they would not be good proxies themselves, and they 
probably have raised the level of the average proxy by choosing that movie 
star rather than one of their peers. But I also think that most movie stars 
would refuse to accept many proxies, because of the way that proxies would 
work in a Free Association)

You would *not* be forced to choose *any* proxy, and you could still 
participate. In a DP organization, your vote is never wasted, it is present 
and active at the highest levels, because any "meeting" of the whole 
organization would provide means for people to vote absentee. In 
governmental DP, this would be a bad idea -- under anything resembling 
present conditions -- but in an FA, the purpose of a poll is to measure the 
level of consensus among the members. *All* the members. And if a member 
does not feel represented by his or her proxy on an issue, the member is 
free to directly vote.

Such an organization would advise its members in two ways: directly from 
the top, by publishing the results of polls on recommendations, and, in 
addition, down through the proxy network, where a trusted person is 
advising the one who trusted. FA-DP networks are networks of trust, and 
this is why I expect that FA-DP recommendations would carry real weight. 
Ideally, each proxy only accepts

>Alternatively, the small group could designate candidates or other proxies 
>to pick the candidate to be "targeted" in the general approval election.

Once again, I'm glad to see the proxy concept being mentioned. Mr. Smith's 
Asset Voting, once I realized what it was, I recognized as being a form of 
delegable proxy, and specifically governmental DP. See 
http://wiki.electorama.com/wiki/Delegable_Proxy_Election

Substitute "Candidate" for "Registered Proxy", and perhaps twiddle a few 
rules, and Asset Voting is DPE.

It's too bad that Mr. Smith does not recognize that we don't have to wait 
for Godot to come in order to begin building organizations that will 
effectively *implement* Asset Voting, not merely advocate it or another 
method. (If you can control the nomination process, that is, the process of 
naming candidates to be on the ballot as representing large groups, then 
you can control the outcome of elections, and plurality works just fine as 
a ratification method. Essentially, the final vote would be an independent, 
secret-ballot confirmation of the decisions made openly in a large FA.)

Mr. Smith, in a post to the ApprovalVoting list, wrote:

>If you cannot even get the 3rd parties to support this, who the hell
>are you going to get?  (And frankly, I tire of this hogwash about
>somehow
>ignoring all parties and the government and just organically creating
>magic FAs
>everywhere that come from the spontaneous joy of free love that
>conventional
>thinkers like me are too closebrained to see.
>Go back to the Land of Oz as rapidly as possible please.)

I responded to this on the Approval Voting list, and it appears that I was 
put on moderation without notice, prior to the post; my response was 
deleted without notice. This inspired me to create an alternative list, for 
posts rejected or disappeared by the moderator of the ApprovalVoting list. 
This is an FA-like response. It does not attempt to take over existing 
organizations, just to create alternatives. If nobody wants an alternative, 
fine. If nobody wants to read my posts, I can't say that I like that, but 
it is absolutely the right of individuals. As I said, in FAs -- and society 
as a whole is an unconscious FA with various caucuses which sometimes 
pretend to speak for the whole -- the freedom includes the freedom to *not* 
associate. If the moderator of the Approval Voting does not want my posts 
to appear, it's his privilege to reject them. But in a Free Association, 
such an action by a person with a position of control will *typically* 
inspire the formation of an independent group. This is actually one of the 
ways that Alcoholics Anonymous grew so rapidly. Part of the genius of the 
founders of AA was to institutionalize this freedom. If a group of people 
say that they are an AA group, they can and will be listed in the meeting 
lists. And members can and will check them out. It doesn't matter if those 
people, for example, believe that controlled drinking is just fine. The AA 
position would be that any member is free to advocate that position, and if 
the member manages to stay sober, they won't be offended! But those who 
tried it and ended up in the gutter once again will drag themselves back to 
meetings and report their experience. This is why the "controlled-drinking" 
position is hardly ever heard in AA. Too many have tried it already.

So a defacto organizational position appears without any need to repress 
minority views. In AA, the position that controlled drinking is impossible 
for a true alcoholic is so much a matter of consensus that it is expressed 
in organizational publications, but organizational publications in AA, 
quite clearly, only express the views of the writers. AA itself "takes no 
position on issues of controversy." So when you see newspaper articles 
excoriating AA for being fascist or dictatorial or a bunch of religious 
fanatics (I've seen all these), the organization itself remains silent.

As to my response to Mr. Smith, such as it is, it can be found at
http://groups.yahoo.com/group/AVFA/message/3

And another rejected post, which was clearly relevant to Approval Voting 
(as the moderator admitted), but which allegedly included irrelevant 
material, can be found at
http://groups.yahoo.com/group/AVFA/message/4

And I just discovered that, apparently, a whole series of messages were 
deleted without notice or comment. That's not necessarily relevant here, 
though if relevance appears to me, I might mention it again.