[EM] Free Associations (was: Trees and single-winner methods)

[Abd ul-Rahman Lomax]

At 05:40 PM 3/25/2007, Juho wrote:
> > I'm considering an FA to be an organization that has formal rules
> > that prohibit the association from developing precisely those things.
>
>Ok, it seems that the FAs are in fact not without rules but have
>quite strict rules (to keep them "free").

No, that's not accurate. FAs have no rules that prevent them from 
becoming non-FAs. I've chosen to call the "FA rules," "Traditions," 
after AA usage. They don't bind anyone. Any meeting can violate these 
traditions, and nobody will prosecute them or expel them or the 
offending members.

(More accurately, it would be a violation of the Traditions to do so.)

The FA Traditions are mostly derived from the AA publication, Twelve 
Steps and the Twelve Traditions, in particular the latter part. There 
is a further publication, far less widely read, Twelve Concepts for 
World Service, which contains additional material.

Bill Wilson designed the Traditions and the Concepts specifically to 
avoid the organizational hazards that have been noted here. It's not 
an accident, it was quite deliberately done.

What keeps the FA going is *the members.* As long as there are some 
members who behave in accordance with these principles, at least 
generally, the FA still exists as an FA. The "official organization" 
might become something else, and what happens in the real world, 
because of the ubiquitous lack of understanding of these 
organizational principles, is that nobody is left, or not enough to 
maintain some coherent activity.

But that hasn't happened with AA, and the reason is probably the 
formalization of the Traditions by Bill W. Enough AA members 
understood this to maintain critical mass.

One of the Traditions bears mentioning in this regard: "AA as such 
ought never be organized, but we may create service boards or 
committees directly responsible to those they serve."

It's really a remarkable statement. It establishes "AA" as something 
different from the organization that exists. What is "organized" is 
not "AA." It is, variously, AA World Services, Inc., which is a 
"service board." Or it is, say, a regional Intergroup, which might 
operate an office. Or it is an individual meeting, which has a 
Secretary, Treasurer, coffee pot, and a box of books, and just about 
nothing else. The latter is closest to being "AA."

But don't these Traditions "organize" AA. Not really. Compliance with 
them is totally voluntary. An AA meeting that, say, accumulated a 
large treasury, contrary to traditions, wouldn't be delisted. What 
would happen, really, is that somebody would ask why it was being 
done. If the members want to do it, nobody is going to stop them. AA 
World Services, Inc., wouldn't sue them!

But consider what would happen. Suddenly there is an asset worth 
fighting over. And, Bill W. might have written, give alcoholics 
something to fight over, they will fight over it. He was really 
writing more about general human nature.... but it was especially 
true for alcoholics, at least to him.

>Power attracts power hungry people. The rules you mentioned (quite
>rigid and well tested ones) and "separation from power" may be needed
>to keep the FAs "free".

They are strong but not "rigid." Got a reason to disregard one. Go 
ahead. At your own peril.

Bill W.'s comment would have been that if it works, let us know. The 
Traditions recommend avoiding things that don't work, that are 
destructive of group unity, and maximizing unity was a major goal. He 
did not want to get unity by imposing some dogma, he wanted it to be 
natural, to represent true consensus.

A lot of what is supposed to be AA dogma is actually simply a general 
consensus. You can disagree with it at an AA meeting, and you 
wouldn't be expelled. The only danger might be from some newcomer who 
doesn't understand AA and who could feel threatened by it....

In my opinion, some of the common opinion in AA is simply not true. 
But it is true enough that it makes a good starting point. It isn't 
true that an alcoholic *can't* take only one drink. But it is far 
safer to assume that it is true than to test it....

>(A side observation. One option is to keep all the conclusions/
>recommendations/outputs of a meetings anonymous (not tied to any of
>the members) to keep the discussions neutral and to reduce the use of
>a meeting as a tool for personal career/image booster.)

Possibly. But I don't mind that individuals gain publicity from what 
they say. I do mind if an individual is represented as being an FA 
spokesperson without being authorized, and I'd be very careful about 
authorizing anyone. There'd better be a good reason, and the scope of 
the authority should be clearly defined. For example, an officer can 
be charged with reporting the results of polls.

By the way, I'm more or less in violation of this myself, except that 
I will add that what I write is only my own opinion and is not an 
official pronouncement on behalf of BeyondPolitics.org. Where I or 
others have written wiki pages, they are drafts, proposals, not fixed 
doctrine. I take the liberty of writing as if I had authority, 
because it is simply much easier to write in that way, otherwise 
qualifying everything becomes far too cumbersome. I also get cut a 
little slack as a founding member. Bill W. was able to write the 
Traditions and put his own opinions in there.....

> > the owner of the FA domain decides to become a little tyrant
>
> > The proxies and other active members who don't agree with this can
> > simply recreate the FA with an altered name.
>
>This doesn't sound very good (if common). It'd be better to avoid
>this cycle and keep the rules such that (in most cases) the old
>structure can be kept "free".

Who enforces the rules? Essentially, structures put in place in an 
attempt to guarantee this freedom can quite possibly destroy it. 
Rather, FAs depend on the natural freedom of members. The key is that 
members can and do communicate directly; such direct communication is 
the core of the FA. If the central organization, such as it is, is 
corrupted, the members simply recreate what they need. This is 
actually how AA works, though there has never been a need to recreate 
the national organization. Meetings come in and out of existence as 
needed. Meetings tend to multiply. Instead of trying to create some 
enforcement structure that would prevent meetings from going astray, 
AA simply lets them do what they are going to do.

It is about as libertarian as can be imagined. It is *not* better to 
"avoid this cycle." What happens if there is disagreement over how to 
conduct a meeting is that members, generally, don't fight over the 
meeting. The majority in the original meeting, or the original 
leadership, depends, continue the original meeting, and whoever else 
wants to do so starts a new one. In a different location and/or at a 
different time. Everybody knows about this. And some people go to the 
new meeting and some go to the old and some, indeed, go to both. It 
simply isn't a big deal.

Now, if meetings accumulated property, suddenly the equation changes. 
There is suddenly something to fight over. Power. Control. And all 
the ensuing ugliness and distraction from the original purpose, which 
is mutual support in AA.

AA actually grows through this process. Instead of considering it 
chaos and loss, consider it cell multiplication. Growth. Meeting 
opportunities multiply, meetings become more convenient and thus more 
people can practically participate. Meetings also become diverse, 
with various special interests. While theoretically AA meetings 
should be open to all alcoholics -- self-defined -- in fact there 
come to be women's groups, men's groups, gay and lesbian groups, dual 
diagnosis groups, etc. These special groups to some extent violate 
the AA membership definition, and there is continuing controversy 
over this, but they continue to exist and probably they do more good 
than harm. The harm would come if such special meetings dominated an 
area, which doesn't seem to happen.

> > The FA traditions are a vast protection, even without DP. With DP,
> > I strongly expect, the structure becomes extremely robust and
> > extremely difficult to corrupt.
>
>What is the property of DP that gives this protection? How much do
>you refer to the chained voting mechanism? How much to the continuous
>election that gives immediate feedback? What other properties?

First of all, toss the idea that DP is about chained voting, even 
though it can have some effect like that. Thinking of DP as a voting 
process has been a severe limitation on the concept. It is a 
communications device that happens to be useful for voting, but 
voting is a small part of it.

The essence of it is the communication network. A proxy and client 
have, in my view, an agreement. It is an agreement of mutual 
accessibility for communication. By accepting a proxy, the proxy 
agrees to receive communication from the client, and, presumably, to 
respond to it. Likewise, by assigning a proxy, the client consents to 
communication from the proxy.

Proxy assignment lists based on this agreement *can* be used to 
expand votes, on an assumption of a relationship of trust, that the 
client trusts the proxy to serve when the client is unable to act 
directly. Because I'm working with the FA context, I don't have to 
nail all this down, since votes aren't going to bind the clients, as 
FAs have no authority at all over their members -- beyond, possibly, 
denying that the member represents the FA.

The protection comes from the fact that proxies will have lists of 
their clients, including contact information. So a proxy does not 
depend on a central organization for communication with the clients, 
and vice-versa. The central organization *cannot* hijack this link, 
unless the proxy and client have relied upon it, which I don't 
recommend. So suppose somebody manages to gain a majority at the top 
level, either legitimately or through fraud, and the trustee holding 
the domain keys is either complicit or goes along with the vote, and 
the vote is to change the FA in some crucial way. The proxies who 
agree with this simply go along with it. Those who don't can 
*immediately* create their own organization to continue what they 
consider to be the "true vision." And, *at the same time*, they can 
remain within the original organization, assuming that it will 
tolerate them. Because of DP, they don't have to duplicate a lot of 
effort, since they can particpate -- in either organization -- 
through proxy. In a major organizational split, what is created is 
two or more organizations that have links through proxies. Those 
links are only broken if members choose to break them, or if one of 
the organizations has gone so far from the FA traditions that it 
expels members.

(While it can occasionally be necessary to limit participation in 
some ways, AA would never "expel" a member for bad behavior. In the 
FA context, a member might be sanctioned (by majority vote at a 
meeting), but would not lose voting rights for any reason I can think 
of. However, the vote might have to be exercised by proxy.... There 
are details involving fraudulent registrations that I won't go into. 
Suffice it to say that such aren't the problem that they might be 
considered to be.)

As the trustee for BeyondPolitics.org, what would I do if the 
registered membership voted to change what I considered to be a 
crucial FA characteristic? I'd make the judgement at the time, but 
what I might do, if I considered the matter truly important, would be 
to freeze the domain. On the home page would appear referrals to any 
domains set up by the various factions, and the home page would be, 
as far as possible, NPOV about it. Essentially, I would protect the 
overall FA concept, while allowing the factions to pursue their own 
interpretations and decisions. If I kept the original domain as an 
active site, I'd have the problem of property, of an asset.

The point is to grow through disagreement and diversity, without 
becoming biased as an organization.

> > Some kind of body to create and enforce international law makes
> > sense, though. Present structures are pretty inadequate.
>
>Some level of enforcement is needed but I'd be careful not to
>establish a one centrally controlled unit to do that.

Right. It is better if there are *many* units capable of some level 
of enforcement, that voluntarily cooperate. And, in a more 
enlightened world, that don't normally fight with each other in equal 
matches. The idea is to find consensus, which can then be properly 
enforced against what dissident elements remain *when necessary*.

Much of the harm of force comes when it is inadequate to prevail immediately.

> > I'm not proposing FAs as the "main working method," i.e., the main
> > method of carrying out the business of government. FAs are
> > thoroughly libertarian, an FA "government" would pretty much be an
> > oxymoron. But large FAs would essentially be able to keep
> > governments in check. It could be pretty interesting.
>
>Ok, a method for keeping check of the decision making process, not
>part of that process. You should state this clearly when promoting
>the method and when justifying the details of it. Rules and
>optimisation criteria for the legislative and other decision making
>structures may be often different.

My discovery has been that if the supervisory intelligence is 
created, the power structures that already exist are generally 
adequate. It is only when they are not supervised that they are 
dangerous. (More accurately, when the power structures are also the 
only systems for developing group intelligence, the ensuing blindness 
can be fatal.)


> > Okay, suppose the employees of the FBI form an FA/DP organization.
>
>This example points out that FA style structures can be used also
>inside otherwise closed organisations. Companies and "bureaus" differ
>from democratic decision making in that they are centrally and
>hierarchically led. FA style approach probably will have somewhat
>different role here.

Right. The FA isn't "inside" the closed organization, it is, rather, 
in parallel to it, connecting the *participants.* In the FBI example, 
the top-down structure is not challenged, it is, rather, enhanced by 
a rapid feedback, independent structure.

The biological analogy would be that the traditional organization is 
like the organization of cells by diffusion of chemical messengers, 
these messengers aren't specifically targeted, they do not just 
connect two cells, but they affect all the cells in the environment. 
Then the DP structure is like adding a nervous system to this, 
linkages that are specifically between individual cells and that link 
all cells, or at least key cells in various locations, in a rapid 
communications network.

The important thing to take away from this is that parallel FA/DP 
organizations can be formed, at any time, and they do not depend upon 
the consent of the traditional organization, in general, as long as 
the members of the latter are free to independently associate.

One of the suggestions is that the shareholders of a large 
corporation could form an FA/DP shareholder organization. This 
organization would serve, among other things, to make suggestions to 
the members as to whom to name as official corporate proxies. This 
could give small individual shareholders power commensurate with what 
is already done by large institutional shareholders (who often hire 
professional proxy organizations).

Shareholders, by this means, could recover control over corporations 
that are presently far too self-directed, with management and the 
board in bed together, too often at the expense of shareholders (and 
the public in general).

The DP structure, it must be realized, is maximally efficient. It can 
concentrate deliberative power down to a very few people, it does not 
require that every member, or even that many members, put in a lot of time.

My vision is that people will eventually belong to *many* FA/DP 
organizations, maybe hundreds of them, something that would be 
unthinkable with traditional organization. The key is the proxy, who 
manages communication with the organization for the member. And it is 
important that this also work for the proxy. Proxies may be paid, but 
if they are to be paid, it must be by their clients.

(It is perfectly acceptable for a proxy to charge for the service; 
the proxy-client relationship should be as free as possible, assuming 
mutual consent. No mutual consent, no relationship. This is why the 
proxy lists we have created so far have an acceptance field. We want 
the proxy-client relationship to be mutual. *This* is the protection 
against the bugaboo of some famous person collecting millions of 
proxies. If that famous person has to consent to communication from 
all those millions of people.... well, he won't do it. Rather, he 
will pass off the requests to someone else, building a network, and 
if the client in the end isn't satisfied, *personally*, with the 
person he ends up stuck with, he'll go somewhere else.)

>One (just one) possible reason is that in the open process people
>tend to say what they think others expect them to say.

It's an obvious possibility, but if you actually went to Town 
Meeting, I don't think you'd offer this possibility.... People pretty 
much say what they please.

>  One reason
>behind the well established idea of keeping ballots secret is to
>allow voters to make their decisions free of any external pressure
>(of the community, of the husband, of the election officials, of the
>media, of friends, of the FA, and the most vocal members of it).

Sure. It's a protection, and a sound one.


> > The town would know in advance if it was going to pass or not,
> > because of the proxy structure.
>
>It could be best to keep the discussion forum just a discussion
>forum.

The FA is exactly that. It doesn't control anything, it merely 
informs and makes recommendations. Technically, the FA makes *no* 
recommendations, but the proxies who form the structure can and will. 
The FA may report poll results, that is about it. Poll results are 
facts, not opinions.

>  It could give some indication on what the outcome of the
>ballot might be. But there may be another FA round the corner with
>different discussions, and the people in the FA are not bound to the
>opinion of the most vocal persons at the FA meeting.

There might not be any FA meetings, routinely, other than scattered 
meetings between proxies and their clients.

It is important to realize that I'm not trying to set up a complex 
independent structure to take up people's time. Indeed, this was the 
concern of one town official. It is already hard to get people to 
volunteer for town boards and committees, and to come to Town 
Meeting. The last thing we needed, was the fear, was another new 
activity to take up people's time.

But that is a misunderstanding of what is being proposed. All that is 
really being proposed is that a proxy list be set up and used. The 
rest is what I expect will fall out from that, with little extra 
effort. Active proxies will largely be those who would go to Town 
Meeting anyway, and what they do in their position as proxy is to 
communicate with their clients. I.e., talk with their friends about 
town affairs.

Not a huge additional burden.

But the difference is that there is this formal structure, and a 
means of representing what is going on. We can generally presume that 
the votes of proxies at Town Meeting -- or at hearings or other 
special meetings, including on-line forums -- will approximate the 
position of the town, to the extent that the collection of proxies 
represents the whole town.

Present Town Meeting can't do this because the membership is skewed 
due to participation bias. Proxies de-skew this, I expect, at least 
to some degree. It doesn't have to be perfect.

No proposals are being made that laws be changed. What is proposed is 
totally legal, now, and it is simple and cheap. And astonishingly 
difficult to advance as a proposal. I understand why, which is why 
I'm happy to get one or two more people who grasp the concepts and 
consider this great success. For now.

The problem is more like inertia than like an actual obstacle. I'm 
not running into true opposition, at least not yet. And with inertia, 
the solution, if the force you can apply is limited, is sustained 
effort, which accumulates.

As I've said, I've got a lever, and I've found the fulcrum, and I'm 
attempting to move the earth.

And it is moving. A little. As is exactly what I'd expect at this point.


>The strongest driver might be the fact that if there is some power or
>other benefits available there will be people trying to reach that.
>My medicine for this would be to isolate the FAs from decision
>making, career building etc., just like you did (in most places).

As I mentioned, I'm not averse to "career building." But when the FA 
starts making, itself, "decisions," there has come to be a power 
focus, and thus an attractor for corruption. Theoretically, clients 
can assign personal funding power to their proxies, but this is still 
highly decentralized and thus far more difficult to corrupt than a 
central treasury subject to disposition by a central committee or 
even by vote of an assembly or the entire membership.

To move that funding, you have to convince the members or their 
proxies. Without that delegated spending authority, you've got to 
convince the members. The trick is that the proxies will do it if 
*they* are convinced, so if you can convince the proxies, you have a 
good shot at convincing the members, since the members chose the 
proxies, presumably, for general trustworthiness.

You may try to corrupt a few high-level proxies, but the likely 
result is that proxy relationships will be disrupted. It is hard to 
maintain lies in a large number of person-to-person relationships.

(If proxies are given spending authority, generally I would assume 
that it would be of limited scope, perhaps enough funding for a 
month.... and this kind of thing is going to happen in caucuses, it 
must be noted. Caucuses can take on projects, hold controversial 
opinions, and generally do things that the FA as a whole can't do 
without violating traditions.)

> > Remember, the FA/DP organization never "gets some power." There is
> > no center to grab and corrupt.
>
>Yes, "freeness" and "separation from power" go together.

I'm glad that has been accepted. It is quite difficult to get across 
to some people.

> > You try to corrupt it, you break the links that you grabbed, and
> > the elements that were connected to them reattach elsewhere.
>
>This is supposed to happen in the regular formal democratic decision
>making process as well. Corruption may hit both (unless well protected).

Well, DP is nothing but a collection of links. This is true of 
regular democratic process, but only very informally and irregularly.


> > The puppet master
>
>One important part of the "mathematics of democracy" is that groups
>that believe in majority decisions within the group and strong
>discipline within the group have more power in the decision making
>process than (more fragmented) groups whose members always sincerely
>vote as they personally feel. The members of these groups may thus be
>controlled by "puppet masters" or maybe rather "elected masters" and
>group majority decisions voluntarily (and there may be nothing
>suspicious about that (within these particular groups)).

What would happen in an FA/DP organization is that groups of people 
like this would exist as caucuses. The economics of consensus still apply.

I think. We really won't know until it is tried. What I do expect, 
quite strongly, is that FA/DP won't do harm. It may not solve every 
problem -- though it has a better shot at it than anything I can 
imagine -- but it is highly unlikely to make things worse. It doesn't 
exhaust the members, or at least it shouldn't.