[EM] proxies and confidentiality

[Jobst Heitzig]

Dear folks!

A question concerning proxy voting: Does anyone know of a mechanism for
delegable proxy which assures that nobody knows what any other voted? In
particular, it seems to me that no person X must know whether or not
s/he is a proxy for some other person Y, and Y must not have a
possibility of proving to X that X is Y's proxy. Otherwise Y could force
X to name him as proxy! Is this possible without the use of advanced
technology like, say, public key cryptography?

Yours, Jobst


Abd ul-Rahman Lomax wrote:
> At 05:04 PM 2/27/2006, Simmons, Forest wrote:
> 
> 
>>It does my heart good to see this kind of discussion.
> 
> 
> It takes a good heart to be warmed by this....
> 
> 
>>>Yes, that is the solution I came up with. A meeting determines what
>>>is the optimum size, N, then the N proxies with the most votes are
>>>full members. It would actually not be difficult to do.
>>
>>
>>My response:
>>
>>It seems to me that each connected component of the DP network 
>>should have a seat in the assembly, unless that makes the assembly too large.
> 
> 
> Consider the context of my proposal. Every member of the network has 
> a *vote* in the assembly, so every member participates in determining 
> if there are to be any restrictions on full participation. Generally, 
> those restrictions aren't necessary until the assembly actually gets 
> too large, though even in small organizations, it can be found 
> necessary to restrict an individual. This is simply the right of the 
> assembly to protect itself. (in FAs, no member can ever be expelled, 
> but a corollary is that nobody is forced to meet with someone 
> personally, with whom they do not wish to meet, except and unless 
> both parties are high-level proxies and they both desire to seek some 
> kind of organizational unity, which I presume high-level proxies will 
> generally value.
> 
> The only reason a loop (your "connected component") would not have 
> *full* representation would be that the loop closed or terminated 
> without having reached a sufficiently-trusted proxy. The continuation 
> of that situation would be voluntary, unless the loop were so 
> "deviant" from the large majority of members that they couldn't find 
> rapport with any of them. Remember, if they can find anyone to trust 
> outside their group, they have enlarged the loop. Outsiders may 
> simply negotiate with each other to agree on a common proxy, bringing 
> them over the threshold.
> 
> But it is not necessary to have full participation rights to 
> participate. Participation really only involves finding a proxy with 
> full rights willing to enter a motion or pass an idea on. Of course, 
> if a proxy does that in a way that harms the group, that proxy might 
> face censure. But if the idea has any merit, I'd think it would not 
> be difficult to find a proxy who would think it worthy of 
> consideration, if for no other reason than to bring the "deviant" 
> group into the fold by giving its ideas consideration, to the point 
> that these ideas would find proper inclusion in the consensus.
> 
> 
>>In that eventuality one possibility is to require each of the 
>>smaller components to attach itself to one of the N larger 
>>components by delegation.
> 
> 
>>But this solution bothers me because the "require" is a constraint 
>>that violates the otherwise voluntary delegation principle.
> 
> 
> It ought to bother you. Great minds think alike. :-) I think the 
> value of fully voluntary delegation, aside from natural consequences, 
> is so great that the harm from lack of full privileges is outweighed, 
> by far. Essentially, if you can't find agreement outside your small 
> group, you *deserve* to be somewhat isolated. But FA/DP organizations 
> would present so many opportunities for connection, and so many of 
> the fully-qualified proxies would want to be as inclusive as 
> possible, that I doubt that serious isolation would take place.
> 
> Restriction of full participation is actually essential; however, 
> suppose I'm wrong about that, that there is a way to allow everyone 
> to participate fully with incurring the well-known harms from it. 
> Meetings, remember, establish their own rules (that is actually part 
> of Robert's Rules, as it should be); all it takes is one full member 
> of the meeting to move that restrictions be lifted, and all those 
> "excluded" members can vote on it.
> 
> I value consensus, but I do not want the seeking of it to become, as 
> it often does in consensus organizations, minority rule where the 
> status quo favors a minority. The majority retains the right of 
> decision and the natural power that it has merely by being the 
> majority. (If it is a true majority.)
> 
> 
>>Another thing that bothers me is that even the largest of the 
>>natural components of the DP network might be so small that that the 
>>N=500 largest components together might fail to represent even one 
>>percent of the population, total.
> 
> 
> I think that as proxy level rises, the proxies will be better and 
> better able to find good proxies for themselves. We won't really 
> know, however, until we create such organizations.
> 
> But my opinion is that there is so much benefit from collecting 
> proxies in single individuals on a large scale, in terms of 
> efficiency and cost -- remember, I consider that the time that 
> members put into an organization its "cost," and in an FA, it is 
> almost the only cost, or at least the overhead, there will be project 
> costs, paid directly to projects by the members who support the projects.
> 
> People will be motivated to seek those they trust, since it will 
> relieve them of the burden of redundancy. There are *not* 500 
> factions in any organization, and trust in a proxy does not have to 
> be absolute.
> 
> 
>>These two objections might not obtain in typical cases, but it would 
>>be nice to have a method that finessed these potential problems, 
>>because when there are problems at the extremes, manipulators have 
>>incentive to press towards the extremes that favor them.
> 
> 
> FA/DP organizations will be, I predict, practically invulnerable to 
> manipulation, if they have a certain set of founding principles that 
> are well-understood by the members. FA/DP organizations are free to 
> become whatever they want, they could abandon the freedoms and 
> restrictions that preserve them, but they won't if they understand 
> why they are there, that is, if they understand the organizational 
> hazards that the rules are designed to avoid.
> 
> The power of FA/DP is in its fractal structure; in particular, if the 
> assignment of direct proxies takes place on a small scale, where, 
> say, only 20 people directly choose a single individual as their 
> proxy, corruption would be extremely difficult and ineffective. I've 
> thought, sometimes, that this number should be limited, but, same 
> principle as you stated above, it would be an infringement on the 
> members, who must remain free to make what we think are mistakes.
> 
> I just know that *I'm* not going to choose a proxy who represents so 
> many people that I cannot reliably communicate with him or her. And, 
> on the reverse side, I would not want to *accept* more proxies than I 
> could comfortably call in a few hours. That's *direct* proxies. There 
> is no natural limit to the number of indirect proxies.
> 
> (And this is why delegable proxy is an inevitability. A single layer 
> of proxies runs into the problem of scale, just at a bit higher level.)
> 
> We aren't going to truly understand the problems of delegable proxy 
> until there are real organizations using it. Hence my project to 
> start FA/DP organizations, or to encourage others to do so. Try it. 
> You'll like it.
> 
> And I really want to encourage anyone who cares to join with me. 
> BeyondPolitics.org, in its conception, is itself an FA/DP 
> organization, I am merely a trustee at the moment, and I would never 
> abuse that trust contrary to the FA principles. (I.e., suppose that a 
> majority of members want to do something that I just cannot abide 
> with. I might, as one possibility, turn the web site into a simple 
> page referencing two new sites: one being controlled by me and 
> whoever agreed with me, and the other being controlled directly by 
> the original majority, however they wished, together with a consensus 
> description of what happened (a consensus description can include 
> unresolved controversy by stating multiple points of view). In other 
> words, I would not use my position as trustee to bias the future of 
> the organization in whatever direction I might happen to personally 
> want. There are ways to ensure fairness and equity without controlling others.
> 
> FA/DP organizations can quickly and easily split, but they can just 
> as quickly and easily merge. And there is usually much more power in 
> merging, if ways can be found to cooperate.
> 
> ----
> election-methods mailing list - see http://electorama.com/em for list info
>