Herding Cats

That which governs least governs best. – Thomas Paine
I.Nothing is perfect. Everything, this side of heaven, contains a flaw. The master rug makers of Persia go so far as to add a mistaken stitch into their carpets; perfection would be an insult to the greatness of God. For nearly everything else, and for nearly everyone else, we don’t have to worry about adding errors: we work from incomplete knowledge, we work from ignorance, and we work from prejudice. As Mark Twain noted, “It’s not what you don’t know that’ll hurt you, but what you know that ain’t so.” We believe we know so much; in truth we know nearly nothing at all. We have trouble discerning our own motivations – yet we constantly judge the motivations of others. Cognitive scientists have repeatedly demonstrated how we backfill our own memories to create a comfortable and pleasing narrative of our lives; this keeps us from drowning in despair, but it also allows us to be monsters who have no trouble sleeping soundly at night.

We constantly and impudently impugn the motives of others, carrying that attitude into the designs of systems which support community. We protect children from pedophiles; we protect ourselves from unsolicited emails; we protect communities from the excesses of emotion or behavior which – we believe – would rip them apart. Each of these filtering processes – many of them automated – serve to create a “safe space” for conversation and community. Yet community is at least as much about difference as it is about similarity. If every member of a community held to a unity of thought, no conversation would be possible; information – “the difference which makes a difference” – can only emerge from dissent. Any system which diminishes difference therefore necessarily diminishes the vitality of a community. Every act of communication within a community is both an promise of friendship and a cry to civil war. Every community sails between the Scylla and Charybdis of undifferentiated assent and complete fracture.

When people were bound by proximity – in their villages and towns – the pressure for the community to remain cohesive prevented most of the egregious separations from occurring, though periodically – and particularly since the Reformation – communities have split apart, divided on religious or ideological lines. In the post-Enlightenment era, with the opening of the Americas, divided communities could simply move away, and establishing their own particular Edens, though these too could fracture; schism follows schism in an echo of the Biblical story of the Confusion of Tongues. Rural communities could remain singular and united (at least, until they burst apart under the build up of pressures), but urban polities had to move in another direction: tolerance. Amsterdam and London flourished in the eighteenth century because of the dissenting voices they tolerated in their streets. It was either that, or, as both had learned – to their horror – endless civil wars. This essential idea of the Enlightenment – that men could keep their own counsel, so long as they respected the beliefs of others – fostered democracy, science, capitalism and elevated millions from misery and poverty. It is said that democratic nations never wage war against one another; while not entirely true, tolerance acts as a firewall against the most immediate passions of states. The alternative – repeated countless times throughout the 20th century – is a mountain of skulls.

Where people are connected electronically, freed both from the strictures of proximity and the organic and cultural bounds of propriety that accompany face-to-face interactions (it is much easier to be rude to someone that you’ve never met in the flesh) the natural tendency to schism is amplified. The checks against bad behavior lose their consequential quality. One can be rude, abrasive, even evil, because the mountain of skulls which pile up as the inevitable result of such psychopathology appear to lack the immediacy of a real, bleeding body. It has been argued that we need “to be excellent to each other,” or that we need to grow thicker skins. Both suggestions have some merit, but the truth lies somewhere in between.


While USENET, the thirty year old, Internet-wide bulletin board system remains the archetype for online community – the place where the terms “flame”, “flame war” and “troll” originated in their current, electronic usages – USENET has been long since been obsolesced by a million dedicated websites. We can learn a lot about the pathology of online communities by studying USENET, but the most important lesson we can draw involves the original online schism. In 1987, John Gilmore – one of the founding engineers of SUN Microsystems – wanted to start a USENET list to discuss topics related to illegal psychoactive drugs. USENET users must approve all requests for new lists, and this highly polarizing topic, when put to a vote, was repeatedly rejected. Gilmore spent a few hours modifying the USENET code so that it could handle a new top-level hierarchy, “alt.*” This was designed to be the alternative to USENET, where anyone could start a list for any reason, anywhere. While many USENET sites tried to ban the alt. hierarchy from their servers, within a year’s time alt. became ubiquitously available. Everyone on USENET had a passion for some list which couldn’t be satisfied within its strict guidelines. To this day, the tightly moderated USENET and free-wheeling, often obscene, and frequently illegal alt. hierarchy coexist side-by-side. Each has reinforced the existence of the other.

Qualities of both USENET and the alt. hierarchy have been embodied in the peer-produced encyclopedia-about-everything, Wikipedia. Like the alt. hierarchy, anyone can create an entry on any subject, and anyone can edit any entry on any subject (with a few exceptions, discussed below). However, like USENET, there are Wikipedia moderators, who can choose to delete entries, or roll back the edits on entry, and who act as “governors” – in the sense that they direct activity, rather than ruling over it (this from the original Greek kybernetes, from which we get “cybernetics,” and meaning “steersman”). By any objective standard the system has worked remarkably well; Wikipedia now has nearly 1.5 million English-language articles, and continues growing at a nearly exponential rate. The strength of the moderation in Wikipedia is that it is nearly invisible; although articles do get deleted because they do not meet Wikipedia’s evolving standards (e.g., the first version of a biographical page about myself) it remains a triumph of tolerance, carefully maintaining a laissez-faire approach to the creation of content, applying a moderating influence only when the broad guidelines of Wikipedia (summed up in the maxim “don’t be a dick”) have been obviously violated. The community feels that it has complete control over the creation of content within Wikipedia, and this sense of investment – that Wikipedia truly is the product of the community’s own work – has made Wikipedia’s contributors its most earnest evangelists.

There is a price to be paid for this open-door policy: noise. Because Wikipedia is open to all, it can be vandalized, or filled with spurious information. While the moderators do their best to correct instances of vandalism, Wikipedia relies on the community to do this nitpicking work. (I have deleted vandalism on Wikipedia pages several times.) For the most part, it works well, though there are specific instances – such as on 31 July 2006, when Steven Colbert urged viewers of his television program to modify Wikipedia entries to promote his own “political” views – when it falls down utterly. Wikipedia can withstand the random assaults of individuals, but, in its present form, it can not hope to stand against thousands of individuals intent on changing its content in specific areas. Thus, in certain circumstances, Wikipedia moderators will “lock” certain entries, allowing them to be modified only by carefully designated individuals. Although Colbert meant his assault as a stunt, with no malicious intent, he pointed to the serious flaw of all open-door systems – they rely on the good faith of the vast majority of their users. If any polity decides to take action against Wikipedia, the system will suffer damage.

With a growing consciousness of the danger of open-door systems – and a sense that perhaps more moderation is better – Wikipedia cofounder Larry Sanger has launched his own competitor to Wikipedia, Citizendium. Starting with a “fork” of Wikipedia (that is, a selection of the entries thought “suitable” for inclusion in the new work), Citizendium will restrict posting in its entries to trusted experts in their fields. The goal is to create a higher-quality version of Wikipedia, with greater involvement from professional researchers and academics.

While a certain argument can be made that Wikipedia entries contain too much noise –many are poorly written, have no references, or even project a certain point-of-view – it remains to be seen if any differentiation between “professional” and “amateur” communities of knowledge production can be maintained in an era of hyperdistribution. If a film producer is now threatened by the rise of the amateur – that is, an enthusiast working outside the established systems of media distribution – won’t an academic (and by extension, any professional) also be under threat? The academy has always existed for two reasons: to expand knowledge, and to restrict it. Academic communities function under the same rules of all communities, the balancing act between uniformity and schism. The “standard bearers” in any community reify the orthodox tenets of any field, blocking the research of any outsiders whose work might threaten the functioning assumptions of the community. Yet, since T.S. Kuhn published The Structure of Scientific Revolutions we know that science progresses (in Max Planck’s apt phrasing) “funeral by funeral.” Experts tend to block progress in a field; by extension any encyclopedia which uses these same experts as the gatekeepers to knowledge aquisition will effectively hamstring itself from first principles. In the age of hyperintelligence, expertise has become a broadly accessible quality; it is not located in any particular community, but rather in individuals who may not be associated with any official institution. Noise is not the enemy; it is a sign of vitality, and something that we must come to accept as part of the price we pay for our newly-expanded capabilities. As Kevin Kelly eloquently expressed in Out of Control, “The perfect is the enemy of the good.” The question is not whether Wikipedia is perfect, but rather, is it good enough? If it is – and that much must be clear by now – then Citizendium, as an attempt to make perfect what is already good enough, must be doomed to failure, out of tune with the times, fighting the trend toward the era of the amateur.

As Citizendium flowers and fails over the next year, it will be interesting to note how its community practices change in response to an ever-more-dire situation. The pressures of the community will force Citizendium to become more Wikipedia-like in its submissions and review policies. At the same time, additional instances of organized vandalism (we’ve only just started to see these) will drive Wikipedia toward a more restrictive submissions and editing policy. Citizendium overshot the mark from the starting line, and will need to crawl back toward the open-door policy, yet, as it does, it risks alienating the same experts it’s designed to defend. Wikipedia, starting from a position of radical openness, has only restricted access in response to some real threat to its community. Citizendium is proactive and presumes too much; Wikipedia is reactive (and for this reason will occasionally suffer malicious damage) but only modifies its access policies when a clear threat to the stability of the community has been demonstrated. Wikipedia is an anarchic horde, moving by consensus, unlike Citizendium, which is a recapitulation of the top-down hierarchy of the academy. While some will no doubt treasure the heavy moderation of Citizendium, the vast majority will prefer the noise and vitality of Wikipedia. A heavy hand versus an invisible one; this is the central paradox of community.


A well-run online community walks a narrow line between anarchy and authoritarianism. To encourage discussion and debate, a community must be encouraged to sit on a hand grenade that always threatens to explode, but never quite manages to go off. In general, it’s quite enough to put people into the same conversational space, and watch the sparks fly; stirring the pot is rarely necessary. Conversely, when the pot begins to boil over, someone has to be on hand to turn the heat down. Communities frequently manage this process on their own, with cool minds ready to reframe conversation in less inflammatory terms. This wisdom of communities is not innate; it is knowledge embodied within a community’s practices, something that each community must learn for itself. USENET lists, over the course of thirty years, have learned how to avoid the most obvious hot-button topics, and regular contributors to these lists have learned to filter out the outrageous flame-baiting of list trolls. But none of this community intelligence resides in a newly-founded community, so, in an absolute sense, the long-term health of any community depends strongly on the character and capabilities of its earliest members.

The founding members of a new community should not be arbitrarily selected; that would be gambling on the good behavior of individuals who, insofar as the community is concerned, have no track record. Instead, these founders need to be carefully vetted across two axes of significance: their ability to be provocative, and their capability to act like adults. These qualities usually don’t come as a neat package; any individual who has a surfeit of one is more than likely to be lacking in the other. However, once such “balanced” individuals have been identified and recruited, the community can begin its work.

After a time, the best of these individuals – whose qualities will become clear to the rest of the community – should be promoted to moderator status, assuming the Solomonic mantle as protectors and guardians of the community. This role is vital; a community should always know that they are functioning in a moderated environment, but this moderation should be so light-handed as to be nearly invisible. The presumption of observation encourages individuals to behave appropriately; the rare examples when a moderator is forced to act as a benevolent and trustworthy force for good should encourage imitation.

Hand-in-hand with the sense of confidence which comes from careful and gentle moderation, a community must feel empowered to create something that represents both their individual and collective abilities. The idea of “ownership,” when multiplied by a community-recognized sense of expertise, produces a strongly reinforcing behavior. Individuals who are able to share their expertise with a community – and help the community build its own expertise – will develop a very strong sense of loyalty to the community. Expertise can be demonstrated in the context of a bulletin board system, but these systems do not easily adapt themselves to the total history of interactions experts have within the community. A posting made today is lost in six months’ time; a Wiki is forever. Thus, in addition to conversation – and growing naturally from it – the community should have the tools at its disposal to translate its conversation into something more permanent. Community members will quickly recognize those within its ranks who have the authority of expertise on any given subject, and they should be gently guided into making a record of that expertise. As that record builds, it develops a value of its own, beyond its immense value as a repository of expertise; it becomes the living embodiment of an individual’s dedication to the community. Over time, community members will come to see themselves as the true “content” of the community, both through their participation in the endless conversation of the community, and as the co-creators of the community’s collective intelligence.

This model has worked successfully for over a decade in some of the more notable electronic communities – particularly in the open-source software movement. The various communities around GNU/Linux, PHP and Python have all demonstrated that any community with room enough to pool the expertise of large numbers of dedicated individuals will build something of lasting value, and bring broad renown to its key contributors, moderators, and enthusiasts.

However, even in the most effective communities, schism remains the inevitable human tendency, and some conflicts can not always be resolved, drawn from deep-seated philosophical or temperamental differences. Schism should not be embraced arbitrarily, but neither should it be avoided at all costs. Instead – as in the case of the alt. hierarchy – room should be made to accommodate the natural tendency to differentiate. Wikipedia will eventually fork into a hundred major variants, of which Citizendium is but the first. The LINUX world has been divided into different distributions since its earliest years. Schism is a sign of life, indicating that there is something important enough to fight over. Schisms either resolve in an ecumenical unity, or persist and continue to divide; neither outcome is inherently preferable.

Every living thing struggles between static order and chaotic dissolution; it isn’t perfect, but then, nothing ever is. Even as we feel ourselves drawn to one extreme or another, wisdom wrought from experience (often painfully gained) checks our progress, and guides us forward, delicately, into something that is, in the best of worlds, utterly unexpected. The potential for novelty in any community is enormous; releasing that potential requires flexibility, balance, and presence. There are no promises of success. Like a newborn child, a new community is all potential – unbounded, unbridled, standing at the cusp of a unique wonder. We can set its feet on the path of wisdom; what comes after is unknowable, and, for this reason, impossibly potent.

One thought on “Herding Cats

  1. This is a most beautiful piece. Congratulations.

    Expertise is something that is hard to prove and even harder to quantify. The way you’ve explained the internal maturation of a well-behaved community is elegant and succinct. A collective understanding of both the forward progress of the group and the invisible implied moderation from above are key ingredients, indeed.

    Been doing this long? 🙂 I think you really get it.


Leave a Reply

Your email address will not be published. Required fields are marked *