The Pitfalls of Collaborative Democracy

For almost three years, a group of citizens have been pushing very hard for a new model of organising the decision processes in the CDS. They argue that the current model is too bureaucratic (they might be right!). They are unhappy about representative democracy, even the type we have, where anybody can submit bills and discuss them publicly in order to get them approved. Some, like Kaseido Quandry, even go further and consider our model of democracy hopelessly outdated and obsolete.

Back in 2006, when Rudy Ruml (Prof. Rudy Rummel iRL) did some seminars to discuss democracy in the CDS, we had some opportunities to talk to someone who actually has a background in political sciences. And he was also for a “lighter” model of government; he favoured a “tribal” style of government, without limiting who could participate in a “town council”. A “clan leader” would preside the audiences — which would be informal — and decisions would be taken by a majority vote, which would also decide who would become the next clan leader.

Since SL is a digital community, we’d immediately get chats which could be transcribed, so the whole community would be able to participate in the discussions and debates. But the final decision would be taken by the clan leader and their ad-hoc assembly.

The Simplicity Party used to defend a similar model. I guess they drew some encouragement from the words of Prof. Rummel. In any case, in my mind — and of some citizens at least — this “simple model” of the tribal council with a clan leader was, for a long time, what I thought that some people preferred.

I just got it all wrong.

There is nothing inherently “wrong” with this model. It’s just that from the perspective of the CDS, it’s a “too simple” model. We are a group of quite different people with so many different interests and ideas. We require a system of checks and balances to prevent abuse. That requires a bit more complexity…

Nevertheless, this doesn’t mean that we cannot try different models. For instance, the New Guild works pretty much as an adhocracy — every citizen can attend and vote, even vote for the NG Secretary. It works well. But of course the New Guild is checked by the RA and the Executive…

The Executive, for instance, also works a bit like that. The Chancellor is elected but can then invite a group of friends to join the Executive. The way the Executive meets and makes decisions is up to the Chancellor. Again, checks and balances are provided by the RA.

And the Regional Committees are also adhocracies. They also wield power: they can spend the budget. Meetings are informal and happen whenever there is a need for them. But they report back to the Chancellor, and, ultimately, to the RA as well.

In practice, however, we don’t need all those mechanisms of checks and balances: most of the time, after all, we’re all good friends (even if it doesn’t seem like that on the forums!) and we trust each other. But the system is designed to deal with problems when they occur. When everything is flowing smoothly, we don’t need to worry, but we always know that, when things go wrong, we have safeguards in place. At the very least, we have accountability and transparency, while still empowering citizens — any citizens, not just elected ones — to voluntarily participate in the decision processes of the CDS.

One fundamental safeguard is the ability to make a strong, public opposition when we disagree with the way things are going, and, on the next term, we might simply elect a new RA and a new Chancellor, and get a different style of management. And, of course, people interested in getting re-elected better watch out for those opinions! Having an opposition — people disagreeing with what the government is doing, and being allowed to be vocal about it, to the point of trying to push government into doing what they think is best for all, is one of the most important features of a democracy.

And the other fundamental safeguard is that we know who is responsible for the decisions. This responsibility mightg be purely political (do a bad job and get kicked out of office) or it might even be the cause of a complaint filed to the judiciary system (embodied in the SC) which might apply fees or even go further and decide on banning a citizen forever.

Now for the past two years or so, something started to change in the CDS. A group of citizens started to disagree that… people were allowed to disagree! This is a complex issue, because everybody agrees too much disagreement is disruptive (yes, we agree that disagreement is not good…); but if we have not enough disagreement, the government may think they’re doing a good job, while they might not be. The problem is always on how to find a balance. In my mind, the “balance” comes from politeness when stating one’s opinion, not by remaining silent or suppressing opinions.

So I always thought… how could this mechanism of feedback from an opposition work with a tribal council? I supposed it would even be worse, as all opponents would simply join the meeting and depose the leader by the power of ad-hoc voting. What this model leads is to a different kind of “disruption” — a disruption of the normal flow of government, which needs some time to actually implement things. We set that time at 6 months (after testing with just 4 and finding out it was not enough). Now imagine a model where every week a new leader is elected, new policies are implemented, everything is started from zero, over and over again, as people disagreeing (and it’s impossible to please all!) would be the first to attend the next tribal meeting and kick the leader out… turning the existing council into the next opposition, and so forth. At some point, both sides would try to get as much people attending at possible, in order to push their ideas ahead. This won’t work, since we cannot get more than 100 avatars in a sim, and we have more than that number of citizens. So, clearly, we couldn’t get a tribal meeting going which would be representative of all citizens. Worse than that, can you imagine such a meeting with 100 people…? Nothing would ever be decided!

Tribal meetings are great if you have different ways to provide checks and balances and if they can ensure that the opposition is heard in a fair way. This is not easy to do with 100+ people, when all claim the right to join a meeting and discuss. While I can imagine that most people wouldn’t attend all the meetings, I’m nevertheless sure that the most vocal ones would always be there, always pushing their agenda. These days, after all, they do precisely that — with a subtle difference: they do it mostly on the forums, and only a subset is actually elected and votes at the RA.

So I really didn’t understand how this model would work in practice. In theory, it seemed simple and workable. In practice, it would utterly fail to set policy, make decisions, or achieve stability due to the wild swing of interests of the ones joining the ad-hoc meetings. And here is where things start to become dangerous. Democracy is the tyranny of the majority, but minorities are protected. Adhocracies are at the mercy of small, powerful groups that can join a meeting to disrupt it and just represent themselves (they’re not elected) and simply “swing” the style of government to their tastes. But since any group can do that any time… it becomes unmanageable. (That’s why we evolved out of tribal councils iRL 😉 )

But, as said, I’m afraid I got it all wrong. That wasn’t in the least what those people had in mind.

I have to profusely thank Kaseido Quandry for elucidating me. Finally, after almost 3 years, I understood what it’s all about, and what people are really after: collaborative democracy.

This is a concept which has been around for some time; I’ve found some references to it as early as 1989, but it’s probably even older than that. Basically — and you can look it up on Wikipedia, although the definition is biased there, for reasons that soon will become apparent — it’s a model of government somewhere between direct democracy and collectivism (aye, in the sense meant by the old Soviet republics, but without the encumbrance of the same kind of ideology). The basic idea is that “government” should happen at several levels, depending on the size of the community, and where citizens are empowered to discuss, debate, decide, and vote on all issues (real examples of collective democracy actually limit themselves to very specific items, though). Naturally enough, not all citizens wish to debate and vote on all issues. The model takes this into account by assuming that ad-hoc groups meet at each level, and that only interested parties will attend those discussions (and here, of course, is the difference between a Soviet-style model, where in theory each level is “fixed” with a number of representatives, which elect their own representatives among themselves to represent them at the level above — all the way up to the Supreme Soviet). Groups form and disband dynamically, depending on who is actually interested in discussing a specific item or idea. It is thought that this allows “experts” in an area to directly contribute — and vote! — on issues at a very low level, away from “central government”, and this encourages citizens participating at all levels, but specially at the lowest ones, where they feel they have the power to decide.

In RL, when we started to become a more and more digitally-integrated society, this model revealed some unexpected benefits in the online world. Now citizens don’t need to leave the comfort of their homes to attend ad-hoc meetings and discuss issues: they can do that from online forums (or similar tools). This also allows immediate publishing of all the meeting transcripts and agendas, and thus making them available for further groups if they require input to tackle their own issues. Databases can easily be assembled from all past decisions and future decisions can be made based on that data. So-called “experts”, instead of being part of powerful lobbies at the state/federal government level and thus become politicised, can be engaged at the local level, where their know-how is important, but their reach and influence does not go too far. But the most important aspect is that people, at the lowest levels, collaborate to make decisions, instead of having those decisions be mandated from “above”.

Now this utopian dream of collaborating towards a “common” government — where all citizens are part of government! — has actually been implemented in several cases, although not universally so. This model of collaborative democracy was made popular by Beth Noveck in her book Wikigovernment, where she explains her experience with collaborative democracy on a single issue in 2007, which she led, and then speculates (and argues) how this model could be employed at all levels of government. Noveck was in SL shortly after her book was launched; I managed to attend one or two of her lectures. She seems to be a nice, clever, calm, and engaging person. SL, of course, at least during the “hype years”, was seen as a good place to experiment with “democracy tools”, and things like Democracy Island were set up as “democracy labs” where models of ad-hoc decision-making were tested and recorded.

So far, so good. If you have never read anything detracting Noveck’s work or explaining what she and her followers are actually after, this model looks actually quite nice. It is a very compelling one. All governments want to empower their citizens somehow. The “old” models of representative democracy have been based on the assumption that it’s very hard to get people to meet. Electronic democracy shatters that concept and allows people effectively to debate, discuss, and vote from the comfort of their homes. So the argument that future democracies are based on models of electronic collaboration is actually quite compelling.

So compelling, in fact, that lawyers and political scientists have strangely observed techies and see how they tackle large projects without “hierarchical overseeing”. And they have found that the open source projects are actually already working that way. Nobody is “forced” to collaborate on them, and nobody “leads” the project: people pick just a bit of code which they want to work on, and engage others that are working on the same bit through electronic media (forums, bug-tracking software like JIRA, and so forth) and reach decisions at a low level. This all trickles up to create a robust piece of software which didn’t require planning in advance.

Wikipedia works the same way. Anyone can contribute to any page of Wikipedia, but, to ensure that information is unbiased and lists resources properly, pages are not randomly scattered — they’re assembled in themes, topics, areas of interest. At each level, discussion starts on the pages, and moderators deal with filtering out the data, so that only the most “neutral” information survives the strict criteria. “Good” pages tend to survive long, “bad” pages get quickly edited and re-edited until they become “good” and moderators let them stay around. Who defines what is “good” or “bad” depends mostly on the group interested in that particular aspect, idea, topic, or area of interest. The system works. Wikipedia is by far the largest encyclopedia ever created by Humankind, and it contains a staggering amount of relatively good information, and is an universal resource that is accepted for research, because it is peer-reviewed — even though we never know who our peers are.

We’re all familiar with those two examples, and so, a model of government that is based on the same mechanisms, is, as said, very very compelling — at least for techies, lawyers, and political scientists with at least passing knowledge of technology.

It is, sadly, not the whole picture. Just a convenient one.

Now let’s scroll back a few years, and read Archon Fung (an old article of him). Make no mistake, Fung is also pro-collaborative democracy. However, unlike Noveck, who basically just illustrates one example (and managed, through her popularity, to become an advisor to the Obama administration), Fung tackles a lot of examples, from several different areas, where principles of collaborative democracy have been applied, and tries to see how they worked and how they can be improved. He immediately found out one of the biggest drawbacks: there is no effective opposition.

This comes as a surprising result. One would assume that with low-level decision-making, opposing views would be represented at all levels throughout the system. But Fung found that this would hardly be the case. Opposition usually can gather enough support to be a force at the top level, or, eventually, at the bottom level. However, organisations that provide opposing views — think, for instance, about environmentalists — often cannot spread themselves so thin as to be present at all levels. What this means is that they might be unable to make valid contributions at some of those ad-hoc groups, and, because of that, valid opposing opinion is withheld. In fact, due to the way these ad-hoc groups are so dynamically created, quickly settle a vote and make a decision, and then disband again, organised opposition might never get a chance to emit an opinion — and when they do, it might be too late. Or, worse, in some cases they might not even be able to make that contribution at all, since the groups making the decision might refuse to meet when dissenting opinion-makers are present, and just meet at a different time. There are a lot of ways to subvert the system. Fung tries to analyse each and every one of them to see what can be done to make the groups mandatorily include the views of those who oppose the general consensus; he also argues why this model requires opposition and why it’s so fundamental for democracy to allow opposing views not only to be emitted, but be included as part of the decision process.

You might think that Fung is excessively worried about a minor issue that might, after all, get easily incorporated at some point. But I’ll just give you an example. If you start looking at all the entries for democracy under the Wikipedia, you’ll see how for each and every one, arguments pro and con are kept on the main page for the topic. This sounds like a good sign: opposing information is included. Then take a look at the entry for Collaborative Governance. It’s just a stub that points to the Metagovernment wiki, which is not a neutral and unbiased source: it’s an activist group promoting collaborative democracy as part of their ideology. So a reader will easily be confused with this — from the allegedly neutral Wikipedia, which has little to say about a system (which the Wikipedia itself advocates internally!), you just get dumped into the hands of militant activists, which have no qualms to explain that their purpose is to eradicate representative democracy and replace it with collaborative democracy instead. Other models similar to collaborative governance are, say, Open source governance, which, again, does not have a list of pro/con arguments or strengths/weaknesses. It is presented as a “perfect model” with little to be discussed about it (although at least the article says that there are many forms of open source governance). And this is typical of collaborative governance: little discussion, just do it implementation.

Now I’m not going to discuss in depth the merits and the disadvantages of Wikipedia or open source software projects. I’ll just point out that Wikipedia is by far not perfect. Hiding behind “wiki etiquette“, Wikipedia has in effect put all the power of moderation in the hand of a few. Well, those “few” are an estimated 150,000 moderators, but the point is, you will rarely, if ever, be able to publish your opinion on Wikipedia if you’re not friends of a moderator. I have experienced this a long time ago, with the three accounts I have had on Wikipedia. Getting accused of plagiarising myself was the most hilarious argument I ever got from a moderator who deleted my carefully prepared work of two weeks; most simply don’t bother to provide any argument whatsover, they just see a “newbie” posting some articles and get rid of them (150,000 moderators are plenty to edit the whole of Wikipedia, they don’t need more competition!). On another case, I had supplied a lot of supporting material on the discussion page and proposed to include that information on the main page. Other users did not disagree, so after several weeks, I implemented that article as suggested. A moderator saw the change and just deleted everything — he didn’t even bother to comment on the discussion page why, but it was quite obvious: he did it because he was a moderator and I was not, and if I insisted in reverting the page, I’d just be banned. Arguing, in those cases, usually leads to nowhere — ultimately, if you’re not a moderator, you have no power: everything you have written is a click away from a “Delete”.

You might argue that, in general, this happens seldom, since most of the content on the Wikipedia is relatively accurate. Indeed, it seems so. Most people don’t have the universal knowledge to recognise bias in all areas. But when we spot the bias and complain — when we as “experts”, so to speak — if we don’t support a view established by the moderators, we’re out of luck. Most of the information is unbiased, but the problem is that you’ll never know if it is or not. A further problem arises on the scientific areas, where we don’t know who did the peer-reviewing. We assume that the moderators are experts in the field. We assume their profiles actually link to their academic resumées. But we don’t know. Scientific knowledge presented on the Wikipedia does not follow the model of peer review of an academic journal — deliberately so, because Noveck & friends hate experts:

It is not necessary to pre-select authenticated and known professionals when structures can be put in place to ensure that informational inputs are discernable, specific, well-labeled, and easy to search, sort, and use. [Noveck, B. (2008). Wiki-Government. Democracy Journal, retrieved on June 7, 2010]

Open source software projects, as models of collaborative self-governance, are also a common fallacy. While I’m an enthusiastic supporter of free and open source software, and have both contributed to and launched a few of those projects, I’m also aware that there are basically three classes of OS projects, two of which lead to little results (except bickering), and one that is immensely successful. The typical example is the “lone programmer” who launches a new project, and perhaps attracts the attention of some friends. These projects tend to be short-lived, or, at least, little maintained — “lone programmers” lose interest quickly and move to different projects as soon as they’re bored; people still do it all the time, not because there is any revenue from them, but because of the fame associated to an open source project. Everybody knows the name of Linus Torvalds who pretty much created the Linux core single-handedly. Every one wants to become the next Linus Torvalds.

Some of those projects actually gather some attention (Linux is a good example) and start getting more and more programmers willing to contribute code. At the beginning, this model of collaborative governance seems to work fine, while there are few dissenting opinions. But at some point, opinions will diverge. More organised projects might have methods to deal with decision-making — like voting for features, for instance. What happens if someone is “voted out” on a feature? Well, the most frequent case is that they give up on helping out that project, and just move to a different project instead. Or they might “fork” the code and start a new project on their own with their friends. The Mambo split which created Joomla is one typical example. But the best example we SL residents have is how the Emerald viewer (itself based on code from Linden Lab) splits every week or so, when developers get angry and wish to implement ToS-breaking features and get kicked out by the team leaders. Ultimately, if the code repository has “leaders” or “moderators”, they will decide what code goes in, and what stays out — even if they allow contributors to do feature voting.

Among the largest and most successful of the open source projects, these have evolved from the chaos of the example before, but suddenly “an organisation” emerged to direct the software. They’re usually foundations (e.g. Apache, Linux, Mozilla…), but they can be companies. Foundations allow chaotic open source projects to actually have a purpose, a direction, and, inevitably, planning and management. It also means that people can work on whatever code they like, so long as that code is aligned to the direction the management wishes the project to evolve.

More honest are companies like MySQL (bought by Sun; now Oracle) with the database manager of the same name; or Automattic with WordPress; or, well, Linden Lab with Snowglobe. These are open source projects that started inside a company and remained inside a company — but allowed a “community” to contribute code to the project in order to debug and improve it. As SL contributors point out, LL takes eons to implement any contributed code to Snowglobe, because most of it is not “aligned” will LL’s plans. I have no idea about the others.

There is also a mixed model, where originally the code was in the hands of a group of chaotic programmers, but became very popular, and companies start joining it, and, ultimately, they began shaping and promoting it for the future. Typical examples are some Linux distributions — companies like Red Hat, Novell, or even IBM are the big “shapers” of Linux, even though they’re not “exclusively” the only ones contributing and managing code. WebKit, starting as the “Konqueror” web browser for Linux, was “absorbed” by Apple and relaunched, and has as its biggest contributor Google — and Linden Lab as a very minor one. There are tons of similar examples throughout the industry. In the SL world, we have things like OpenSimulator and libopenmetaverse, originally “grassroot” development, which these days are mostly led by Intel and IBM developers, with a lot of contributors outside those companies.

Still, these managed open source projects are the most successful ones by far, and have produced consistently good, robust, industry-grade software. But they’re not the result of “collaborative governance” in the sense naively described by some innocent supporters of collaborative democracy. By carefully and selectively “ignoring” the way those projects are successful, and picking some key features of those projects as examples of how great collaborative democracy can be, the supporters of collaborative democracy are just throwing sand in our eyes. A typical example is how Wikipedia is touted as the biggest user-generated encyclopedia that anyone can edit — which is true, so far as the articles are approved by the moderators. The other example is explaining how immensely popular and successful open source software development is, when in reality the only projects that are “immensely popular and successul” are all lead by organisations or corporations, or at least by an elite cadre of administrators who are never “elected” into power.

Some of the proponents of collaborative democracy are naive but genuine believers. The appeal of an utopia that seems to be at our reach is immense, and there are really quite good points made by the enthusiasts of collaborative democracy. Some are serious proponents, in the sense that they’re aware of the pitfalls and suggest improvements to the model, but are usually ignored (Fung is cited, but not featured, on Wikipedia; Noveck’s page at least labels her as a “[…] biography of a person who has held a non-elected position in the government of the United States” [emphasis mine]). Most have an ideological agenda: they’re quite aware that collaborative democracy, if widely adopted, can totally subvert some of the pillars of democracy which so many find “annoying” — like the freedom of expression (dissenting opinions are easily ignored or suppressed on a wiki-based collaborative democracy “experiment”) or the protection of minorities (since all citizens are “empowered”, minorities don’t exist by definition — and that also means that their ideas will never be represented anywhere worthwhile, since, at each level, dissenting opinion will be discarded in favour of the majority who agreed with a decision, no matter how relevant that opinion was).

Collaborative democracy is a nice, politically correct ideology — and not merely a “system” or an “experiment” — which transmits the ideals of left-wing libertarianism (everybody is empowered at all levels to decide on their own) while in effect creating mechanisms to suppress dissenting opinion (unless one takes the trouble to implement the safeguards that people like Fung have suggested; one should observe, however, who is at the White House these days — it’s not Fung!) and to “confuse” issues throughout the process so that only the opinions of a few will trickle upwards to make policy. That’s not “evil” by itself — since at least there will be more people participating in the decision process — but it’s not representative. Nobody has “elected” the ones that contribute to the decision process. People just show up if they’re interested. They’re not held accountable for the decisions they make — they just participate in the process, but are not made responsible for their decisions, and they’re able to eliminate dissenting views from the public eye.

This can be dangerous, and we can only see that on extreme examples: suppose that a local group starts running a discussion to make a decision about increasing the speed limits to, say, 250 km/h inside towns. This starts to gather support from a huge number of reckless youth, who all vote for it. The car industry, seeing an opportunity, mobilises their employees and family to vote for it too. Of course, a lot of dissenting opinions will pop up, and say how crazy that measure is. But their unpopular “grumbling and complaining” makes their opinions be voted down, until they’re merely noise that disappears from the discussion. The measure is passed. After a few weeks, a drunken teenager drives at 250 km/h against a wall and dies. Who is to blame? There is no “authority”, no “group”, not even a “politician” to step up and claim responsibility for this.

Of course, in the CDS, we’re not likely to cause serious accidents by reckless decisions. We might still ignore minorities; make decisions that are completely insane (but popular!), like, for instance, getting rid of tier payments as a requirement for citizenship (nobody likes to pay taxes!); we might not have a “tyranny of the majority” but just a “dictatorship of the minority”, where small groups are present at all ad-hoc groups at all levels, making sure their ideas are carried away, even if the majority disagrees with them — but the majority is not organised enough to be in all groups with enough votes. Collaborative democracy is also very good to “hide” political affiliation, since, by definition, all groups are informal and all meetings are ad-hoc; there is no real “group” behind everything pulling all the strings (there is no conspiracy theory with a hidden cabal, if, by definition, nobody is formally grouped to implement policy). Nevertheless, this doesn’t mean that, by tactfully and selectively placing key elements on all possible levels of discussion/decision, a group might effectively take control of the CDS under this system, effectively forever, and not be even accused of doing so, since there is no group, there are only individuals empowered to make decisions. We can argue that in real countries this “taking over” is next to impossible to accomplish because of the sheer size and number of groups spontaneously meeting to deliberate issues (Fung even argues that this is one of the shortcomings of collaborative democracy: organisations cannot be represented all the time at all the places), but not in tiny communities like ours, which is much more limited in scope.

So, again, I have to thank Kas for having opened my eyes. I was completely off the track; although, as said, I was quite aware of Beth Noveck and her theories (specially because at some point some members of the CDS even collaborated with Noveck’s people on democracy-related tools and projects), I didn’t understand that this is what is supposed to be “implemented” in the CDS as an alternative to representative democracy.

I can only finish with what I said before. Representative democracy might be outdated, but we have actually gone beyond it and implement a lot of what Noveck has in mind — but avoided the pitfalls. Thus, any citizen can propose and discuss legislation, in public — but voting is limited to elected representatives, who bear the burden of being responsible (politically at least) for what they pass. Any citizen can collaborate on the planning and building of our sims — the New Guild is not part of Government and it is an adhocracy, which nevertheless has an elected leader (who is accountable). Regional committees are adhocracies too, and they are allowed to use public funds in any matter they like, and discuss the spending in public with whomever attend meetings. And, finally, discussion is free on the forums, which, in spite of everything, have collaboration from half of the citizens, which is not that terrible.

No, the difference is that minorities, who are felt to be excluded from the decision process, can appeal to the RA, or, if they’re not heard by the RA, they can appeal to the SC. Representatives are known, and their ideas are public, and they’re responsible for what they do or don’t. At the levels below RA and Executive, everything can be made as informal as possible, but we know who is made accountable. And opinions cannot be suppressed, diminished, deleted, rewritten, or buried — at most, they can be ignored. This is pretty much a model of semi-representative, semi-collaborative democracy that we have already implemented for the CDS, that neatly evades the pitfalls of a full collaborative democracy, according to Fung and others. It is not the left-wing libertarian utopia envisioned by Noveck and her friends.

For further reading: an old article on my blog where I tended to discuss pretty much the same problems and issues… in 2008.