Do as I do: leadership in the Wikipedia

Joseph M. Reagle Jr.

1 Introduction

"Do as I say, not as I do." Most of us have heard this expression, perhaps as children being protected from the bad habits of our elders. Of course, as adults, this saying is more often used as a comment on someone else's hypocritical leadership. We prefer to see others "lead by example." This preference is seemingly a requirement for successful leadership in voluntary communities; without coercive power or utilitarian rewards (Etzioni 1975), leaders must parlay merit resulting from "doing good" into a form of authority that can also be used in an autocratic fashion, though with a soft touch and humor, when—and only when—necessary.

In this paper I consider how notions of leadership operate in collaborative on-line cultures. In particular, I consider the seemingly paradoxical, or perhaps merely playful, juxtaposition of informal tyrant-like titles (e.g., "Benevolent Dictator") in otherwise seemingly egalitarian voluntary content production communities such as the Wikipedia. To accomplish this, I first introduce the Wikipedia as an open content community and review existing literature on the role of leadership in such communities. I then relate ethnographic and archival data on how leadership is understood, performed, and discussed in the Wikipedia community. I conclude by integrating concepts from existing literature and my own findings into a theory of leadership and note other communities and leaders against which this theory could be tested.

2 Wikipedia as an Open Content Community

What is often meant by the term "open" is a generalization from the Free Software, Open Source, and open standards movements. Communities marshaling themselves under these banners cooperatively produce, in public view, software, technical standards, or other content that is intended to be widely shared; these communities use computer mediated communications for much or all of their work. For example, the Linux kernel is one of the most famous free software projects and the many communities around it, such as the Debian distribution (which combines the kernel with other utilities and applications), are the most studied so far. Elsewhere, I define open content communities as those that commonly deliver products under a "free" or "open" copyright license and demonstrate transparency, integrity, non-discrimination beyond merit, and non-interference (Reagle 2004). Additionally, most participation is voluntary, meaning an "activity in which time is given freely to benefit another person, group, or organization" without precluding "volunteers from benefiting from their work" (Wilson 2000:215). The voluntary character and the principle of non-interference imply that if a constituency disagrees with the leadership's direction, the copyright license permits them to fork, or take the products and commence work elsewhere without interference. This notion of an open content community is central to this paper since it affects the type of leadership encountered in the Wikipedia community and hopefully contributes to a theoretical abstraction beyond any of the specific communities discussed.

The Wikipedia is an on-line "Wiki" based encyclopedia, the product of an open content community. "Wiki wiki" means "super fast" in the Hawaiian language, and Ward Cunningham chose the name for his Wiki project in 1995 to indicate the ease with which one could edit Web pages. In a sense, Wiki captures the original conception of the World Wide Web as a browsing and editing medium. The Wiki makes this possible by placing a simple editor within a Web page form and the functionality of formatting and linking on the Wiki server. Consequently, if a page on the Wikipedia (an encyclopedia on a Wiki server) can be read, it can be edited. With a Wiki, the user enters a simplified markup into a form on a Web page. To add a numbered list item with a link to the Wikipedia one simply types: "# this provides a link to [[Wikipedia]]". The server-side Wiki software translates this into the appropriate HTML and hypertext links. To create a new page, one simply creates a link to it! Furthermore, each page includes links through which one can sign in (if desired), view a log of recent changes to the page (including the author, change, and time), or participate in a discussion about how the page is being edited on its Talk Page—and this too is a Wiki page. The application of a general tool facilitates a surprisingly sophisticated creation!

Yet, the consequence of this quick and informal approach to content creation was not foreseen - or, rather, was fortuitous. Wikipedia is the populist offshoot of the Nupedia project started in March of 2000 by Jimmy "Jimbo" Wales and Larry Sanger. Nupedia's mission was to create a free encyclopedia via rigorous expert review under a free documentation license. Unfortunately, this process moved rather slowly and having recently been introduced to Wiki, Sanger proposed that a scratch-pad be set up for potential Nupedia content where anyone could contribute. However, "There was considerable resistance on the part of Nupedia's editors and reviewers ... to making Nupedia closely associated with a website in the Wiki format. Therefore, the new project was given the name 'Wikipedia' and launched on its own address, Wikipedia.com, on January 15 [2001]" (Wikipedia 2004hw).

Wikipedia proved to be so successful that when the server hosting Nupedia crashed in September of 2003 (with little more than 24 complete articles and 74 more in progress) it was never restored (Wikipedia 2004n). As of today, the original English version exceeds 900,000 articles, and there are scores of other active language Wikipedias (2004wa1). The Wikimedia Foundation, incorporated in 2003, is now the steward of Wikipedia as well as a new Wiki based dictionary, compendium of quotations, collaborative textbooks, and repository of free source texts.

The community that produces the Wikipedia is very much an open content community. Interaction is mediated by the Web, most all participation is voluntary, decision-making is fairly transparent, few restrictions are placed upon participation, and, most importantly, the disaffected can easily leave and resume their work elsewhere. For example, Wikinfo was forked from Wikipedia so as to operate under a different philosophical stance than neutrality required of Wikipedia articles (Wikinfo2004).

3 Leadership in open content communities

Existing literature on leadership, "influence exerted over other people to guide, structure and facilitate relationships in a group" (Yukl 1998:3), in open content communities identifies the salience of merit. In the Debian community, O'Mahony and Ferraro (2003, 2004) confirmed Fleming's and Waguespack's (2004, 2005) finding in the open standards community of the IETF that technical contributions, among other factors, are predictive of leadership. In a study of virtual teams, Yoo and Alavi (2003) found that emergent leaders send more and larger e-mail messages than other team members; the content of these messages also have a higher degree of task orientation. Additionally, in virtual contexts leaders provide frequent and predictable communication (the "heartbeat") as the basis for effective coordination. This interaction is often with a friendly and considerate orientation by the leader (Bosco 2004).

Consequently, leadership is the performance of consistent and substantive contribution within the community that affects its movement. Not surprisingly, leadership tends to be entangled in discussions about power (the extent of one's influence), authority (the legitimacy of one's influence based on accumulated merit), and governance (how to make group decisions). Coleman (2005), in her study of the Debian software community, found that: "Power, in other words, is said to closely follow the heels of personal initiative and its close cousins, quality technical production and personal dedication to the project" (p. 22). Also, in keeping with the character of voluntary community, Coleman, citing Donner and Mill, notes the power gained because of merit is a guiding force rather than coercive. Furthermore, the merit of exemplary behavior, particularly that of founding a community, might have a charismatic character. In an earlier study of the Debian community Garzelli and Galoppini (2003) found that:

In practice, this means that leadership (which translates directly into the formal and informal hierarchy and authority) is not just established in bureaucratic or rational fashion, but in charismatic fashion as well (Weber 1964 [1947], pp. 358ff.); here, charismatic authority mostly derives from earned respect often proven by leading a big, successful project. As a matter of fact, charismatic authority maybe, in some circumstances, more "efficient" than authority deriving its legitimacy from well-established rules (Coleman 1990; Langlois 1998). (Garzarelli and Galoppini 2003:18)

Perhaps one reason for this efficiency is that the reputation of such leaders has an additional benefit of being useful in circumstances where a community is otherwise deadlocked; charismatic authority can intervene in circumstances in which there are multiple simultaneous coordination costs that are too expensive to be addressed by "a complex system of rules such as a constitution" (p. 36).

While meritocratic leaders are granted much authority, this exists within—or besides, or in conflict with—other modes of governance. In the Debian project, Coleman (2005) found governance to be a blend of "democratic majoritarian rule, a guild-like meritocracy, and an ad-hoc process of rough consensus" (p. 7). Conflict between these models is central to crises within these communities. One interesting way in which this tension was exhibited is by references to "TINC" (There Is No Cabal) as a source of anxiety and joking about leadership (p. 23). Because of the informal character of many virtual communities the notion of a cabal and the "TINC" moniker is recurrent after its emergence on USENET, one of the earliest Internet discussion forums. Pfaffenberger (1996) details the historical and cultural development of governance in USENET including similar anxiety and joking; Dave Mack's 1991 tongue-in-cheek notice about the decision of the Usenet High Council is an exemplary parody of tyrannical authority (p. 379).

Consequently, one might think of founding leaders as the initial emergent leaders who fashioned a space in which the community comes to inhabit, and because of this are likely to garner merit and charismatic authority, which influences the community's culture. In this regard, Edgar Schein's model of organizational culture is compelling because of the salience of two factors: the seminal actions of community founders and the community's response to important incidents, including crises. Schein (1992) proposes founders and leaders embed and transmit culture via numerous mechanisms, the primary of which being:

  1. what the leaders pay attention to, measure, and control on a regular basis
  2. how leaders react to critical incidents and organizational crises
  3. observed criteria by which leaders allocate scarce resources
  4. deliberate role modeling, teaching, and coaching
  5. observed criteria by which leaders allocate rewards and status
  6. observed criteria by which leaders recruit, select, promote, retire, and excommunicate organizational members

It is worthwhile to note that while relevant, Schein's analysis, like most organizational research, is focused on traditional organizations, corporations even, rather than voluntary organizations or communities. Yet, leaders are still present in such settings. While the mechanisms might operate differently in voluntary contexts—e.g. symbolic rather than financial rewards—the mechanisms are present. (Indeed, as discussed below, a possible source of leadership/governance problems on the Wikipedia was that its founders did not appreciate that they were creating a community and culture at the outset.) For example, in Lois' (1999) study of a volunteer mountain rescue group, socialization, and therefore access to a leadership role, is dependent on the volunteer first "downplaying arrogance and egoism by displaying humility and respect" (p. 21). Only then will action of merit (paradoxically) be granted "heroic" status by established community members and leaders.

Finally, as already noted, an important attribute of open content communities is the possibility of a fork:

Fundamentally, the ability to create a fork forces project leaders to pay attention to their constituencies. Even if an OSS/FS project completely dominates its market niche, there is always a potential competitor to that project: a fork of the project. Often, the threat of a fork is enough to cause project leaders to pay attention to some issues they had ignored before, should those issues actually be important. In the end, forking is an escape valve that allows those who are dissatisfied with the project's current leadership to show whether or not their alternative is better. (Wheeler 2005)

Because of the voluntary and meritocratic character of open content communities it is not surprising that not only are leaders expected to lead by example, their very leadership is founded upon exemplary behavior—leadership emerges through action rather than appointment. And while a founding leadership role has some semblance of authoritarianism to it, at least in title as we shall see, it is eternally contingent: a dissatisfied community, or some constituency thereof, can always leave and start again under new leadership. Yet, regardless of whether a community does fork, discussion about such a possibility, the actions of leaders, and metaphors of governance are common topics of conversation.

4 Methodology

This analysis is informed by two decades of participation in computer mediated communities. While I began studying the Wikipedia specifically in January 2004, and made my first identifiable edit (rather than anonymous) in April of that year, I had made use of Wikis and Wikipedia before that. Sources of information that inform or are used in this paper include Wikipedia articles, discussion pages, blogs, and news articles encountered during those two years. Additionally, in September of 2004, I took the plunge from monitoring and searching HTML archives of two high traffic-mail lists (wikiEN-l and wiki-l) to subscribing to them, and I've received more than 22,000 messages since. These lists often include discussions of the administration and policy of the Wikipedia. The excerpted conversations that I use mostly span February 2005 through November 2005. I engaged all of these sources with the specific question of leadership and how it is understood—derived from my interest in politeness and leadership styles in open content communities—in the fall of 2005. In this time I also sought out and considered "small neighborhoods" of articles and collaborators, eventually settling upon the corpus of Harry Potter pages given the project's coherence, liveliness and my own, earlier, experience of advocating for a reform in how readers should be warned of the possibility of spoilers (i.e., having a plot of a new book spoiled when consulting a Harry Potter article).

On a suggestion, I developed a brief questionnaire to engage with editors of the Harry Potter Project pages but, as expected, received few responses. Open content communities are, presently, often studied (with similar questionnaires) and participants might have little interest in taking time away from their actual (volunteer) work to respond to yet another. (As a participant, I have never responded to such a questionnaire.) Contacting actual participants can be difficult as well, as Lorenzon (2005) noted: "Many editors have their own user page which give information about them but few give out their real names and contact information." I made my solicitation on the Talk page for the Project as well as the Talk pages of a handful of prominent editors, without much success. Additionally, because most all the discourse is public and the community is otherwise so reflective, there is an abundance of existing data situated in actual practice. This is not to say such research discussions are not useful; once I developed my questions I was interested in receiving answers and the single response was informative. Fortunately, while responses to questionnaires can be hard to obtain, I also do not think them necessary to understand this community. Instead, one must follow (or even engage) in the practice: "A culture is expressed (or constituted) only by the actions and words of its members and must be interpreted by, not given to, a field worker" (Van Maanen 1988).

In the course of this work I have developed a number of methods and software for capturing, organizing, and citing documents and discourse within this community. (For example, with the help of the MARC e-mail list archivists I was able to ensure the unique message-ID of a message cited has a similarly unique and easily dereferenceable URL. Or, I developed a small script to collect and analyze the distribution of contributions (edits) to a Wikipedia page.) This work is documented on my research blog (Reagle 2003).

Finally, because of my long-standing participation in open content communities it is difficult to place myself in a particular methodological school. First, I often come to these communities as a user so I cannot take for granted the distinction between researcher and subject. I tend to consider myself a reflective practitioner that sometimes, also, becomes a researcher. This also complicates the relationship between theory and data. Since I often engage in practice first, my approach is clearly not deductive (i.e., posing a theory prior to exposure to the community). However, it is not purely inductive either (i.e., allowing my own concepts and theories to develop solely on the basis of the ethnography) as I have already encountered many theories as part of my reflective practice. Yet, there is a goal that I aspire to: "Ethnography should be empirical enough to be credible and analytical enough to be interesting" (van Maanen1988:29); I hope to make a convincing contribution (Golden-Biddle and Locke 1993) by providing an account that has authenticity, "the ability of the text to convey the vitality of everyday life encountered by the researcher in the field setting" (p. 599), plausibility, "the ability of the text to connect two worlds [of the writer and reader] that are put in play in the reading of the written account" (p. 600), and criticality, "the ability of the text to actively probe readers to reconsider there taken-for-granted ideas and beliefs" (p. 600).

5 Leadership and the Wikipedia

The conception of leadership does not play a prominent formal role within the most of the Wikipedia community. In a place as reflexive as the Wikipedia, in which there are dozens of documented norms—such as assume good faith, be polite, be prepared to apologize, and forgive and forget (Wikipedia 2004ww)—it is surprising to find no such page on the topic. I encountered no recommendations on how to be a good leader or leadership mentoring. In fact, the only page on leadership I found was the actual, extensive, encyclopedic article on the topic.

This may be, in part, due to the egalitarian, if not anarchistic, character of the community. There are no designated leadership roles for editing encyclopedic articles. While co-founder Larry Sanger was "editor in chief" of Nupedia, he was informally known as the "chief organizer" of the Wikipedia and he was careful not to make any formal titular claim. Neither role was ever claimed again after he resigned from the project. Instead, the documentation, such as the Administrators page discussed more fully below, stresses that everyone is an equal editor. Those that demonstrate themselves to be good editors may request extra responsibilities but "are not imbued with special authority." Yet, while the culture stresses editorial egalitarianism over administrative responsibilities, this does not mean there are no leaders in Yukl's (1998) sense of "influence exerted over other people to guide, structure, and facilitate relationships in a group" (p. 13). Consequently, before turning to how the community speaks about leadership, I first present an account of the three "places" I looked for it in operation: with the Wikipedia founders, its administrators, and more "locally" in a collection of Harry Potter related pages.

5.1 Founders

The two people in the history of the Wikipedia stand out as having the ability to "guide, structure, and facilitate" are its co-founders: Larry Sanger and Jimmy Wales. In keeping with Schein's analysis of leadership and culture the actions of these two, particularly "how leaders react to critical incidents and organizational crises," very much affected the Wikipedia culture. While a history of Wikipedia's development is beyond the scope of this paper—which fails to account for others at Bomis, the Nupedia board, and early Nupedia and Wikipedia contributors—the following brief account of the crisis of Nupedia's demise, Wikipedia's rise, and Sanger's departure provides an excellent introduction to leadership in the Wikipedia context.

Wales, a co-owner of Bomis, an Internet content and search company, hired Sanger in February 2004 to launch and act as the editor in chief of the Nupedia project. Until he resigned, Sanger was the most prominent leader of Nupedia (the original peer review project) and Wikipedia (its Wiki complement and eventual successor). As Sanger writes in his April 2005 memoir:

The idea of adapting wiki technology to the task of building an encyclopedia was mine, and my main job in 2001 was managing and developing the community and the rules according to which Wikipedia was run. Jimmy's role, at first, was one of broad vision and oversight; this was the management style he preferred, at least as long as I was involved. But, again, credit goes to Jimmy alone for getting Bomis to invest in the project, and for providing broad oversight of the fantastic and world-changing project of an open content, collaboratively-built encyclopedia. Credit also of course goes to him for overseeing its development after I left, and guiding it to the success that it is today. (Sanger 2005)

What precipitated Sanger's resignation? Stretched between continuing frustration with Nupedia's progress, problems with unruly Wikipedians, and a widening gap between the two, Sanger failed to save the Nupedia and alienated some Wikipedians who saw his actions as unjustifiably autocratic. (I use the term autocratic to describe leadership actions which do not derive their authority solely from legitimate group decision making processes such as unanimity, supermajority, majority, etc.)

Additionally, with the burst of the Internet bubble, Sanger, among others, was laid off from Bomis and resigned his Wikipedia role shortly thereafter. Subsequent commentary from the sidelines, particularly advocacy for Wikipedia to respect expert authority, has prompted additional criticism from those Wikipedians with a strong commitment to radical openness. In any case, Sanger's account recognizes the uneasy tension between title, authority, and cultural momentum at the founding of this community:

My early rejection of any enforcement authority, my attempt to portray myself and behave as just another user who happened to have some special moral authority in the project, and my rejection of rules—these were all clearly mistakes on my part. They did, I think, help the project get off the ground; but I really needed a more subtle and forward-looking understanding of how an extremely open, decentralized project might work. (Sanger 2005)

Such an understanding might've been that of Teddy Roosevelt's recommended leadership style: speaking softly and carrying a big stick. While Sanger did have special authority in Nupedia, such was not the case in Wikipedia and Sanger's corresponding "loudness" was a later cause of regret:

As it turns out, it was Jimmy who spoke softly and carried the big stick; he first exercised "enforcement authority." Since he was relatively silent throughout these controversies, he was the "good cop," and I was the "bad cop": that, in fact, is precisely how he (privately) described our relationship. Eventually, I became sick of this arrangement. Because Jimmy had remained relatively toward the background in the early days of the project, and showed that he was willing to exercise enforcement authority upon occasion, he was never so ripe for attack as I was. (Sanger 2005)

And most interestingly, Sanger attributes a root of the problem in a way that Schein might appreciate: failing to recognize the seed of a community and culture around the project:

For months I denied that Wikipedia was a community, claiming that it was, instead, only an encyclopedia project, and that there should not be any serious governance problems if people would simply stick to the task of making an encyclopedia. This was strictly wishful thinking. In fact, Wikipedia was from the beginning and is both a community and an encyclopedia project. (Sanger 2005)

Upon publication of Sanger's memoirs a controversy arose over whether Sanger even deserved credit as a cofounder of Wikipedia. However, other responses engaged more directly on the importance of leadership authority:

Now, I must say... I think a project of such a type can only work *without* a strong authority. It is important to let people built their own organisation. Jimbo has this very powerful strength, in this that he lets most of the organisation be a self-organisation. For those who know a bit about leadership, it is a rather rare occurence. For the sake of wikipedia, and to let all the international projects grow up (without a strong hand to lead them), it was important that the role of the editor in chief disappear. (Anthere 2005rss)

Sanger actually concedes as much in the development of editorial policies but is still concerned about controlling abusive editors and attacks, particularly when they alienate high quality expert contributors. These questions of authority and leadership are common, as will be seen in the discussion of dictatorship.

5.2 Administrators and the Board

It can be difficult to speak of leadership without a basic understanding of a community's structure. Yet, a novel characteristic of Wikipedia is that anyone that browses the Wikipedia may edit it—in rare circumstances pages, such as the main page, are locked if they do not need to be updated often, do not benefit from popular edits, and otherwise would be constantly experimented upon or vandalized. Contributors who signed up for an account and log in—no longer "anonymous"—receive no additional authority in editing a page, instead they have access to useful features such as a user page and the ability to track the pages one cares about. Additional features are made accessible to experienced users in the role of an administrator (sysop). These features permit an administrator to enact Wikipedia policy and group consensus, particularly with respect to the management of protected pages, the deletion of pages, or temporarily blocking computers from which vandalism attacks are being launched. Yet, the Wikipedia's Administrators page quotes Jimmy Wales as saying, "This should not be a big deal." Indeed, an association with editorial authority is purposely disavowed:

Administrators are not imbued with any special authority, and are equal to everybody else in terms of editorial responsibility. Some Wikipedians consider the terms "Sysop" and "Administrator" to be misnomers, as they just indicate Wikipedia users who have had performance- and security-based restrictions on several features lifted because they seemed like trustworthy folks and asked nicely. However, administrators do not have any special power over other users other than applying decisions made by all users.

In the early days of Wikipedia all users acted as administrators and in principle they still should. Any user can behave as if they are an administrator, provided that they do not falsely claim to be one, even if they have not been given the extra administrative functions. Users doing so are more likely to be nominated as full administrators by members of the community and more likely to be chosen when they are finally nominated. (Wikipedia 2004)

Essentially, administrators are able to quickly prevent and intervene in any destructive edit. (However, textual vandalism isn't truly destructive as the previous versions are available; one administrative feature is the rollback which permits the quick reversion of such edits.)

In the time since its founding, additional levels of authority have appeared as Wikipedia evolved from a small community to a massive project that is now formally constituted as a nonprofit foundation. In addition to the 700+ administrators (Wikipedia 2005), bureaucrats appoint those administrators and other bureaucrats within their specific project, and stewards can, respectively, change any such role. Orthogonal to administrative and governance roles there are also developers, those who actually write the software and administer the servers.

At the time of incorporation, Wales delegated some of his authority to five directors of the Board of Trustees, in which he serves as the chair. The Board "has the power to direct the activities of the foundation. It also has the authority to set membership dues, discipline and suspend members (article III), and to amend the corporate bylaws (article VI)" (Foundation 2005). In the realm of editorial disputes between users (including administrators) dispute resolution can be facilitated by mediation or arbitration, the latter can issue a binding decision. However, disputes are encouraged to be worked out civilly between the participants as the mediation and arbitration processes can be slow and tedious. The Arbitration Committee, the Board, and Jimmy Wales himself, ultimately, have the authority to penalize or remove abusive users.

Finally, while consensus is preferred for most decisions, voting does occur in some elections and on pages like VfD (Votes for Deletion, on pages that are so poor or on topics of little note). There is a mixed governance model, as Wales himself notes:

Wikipedia is not an anarchy, though it has anarchistic features. Wikipedia is not a democracy, though it has democratic features. Wikipedia is not an aristocracy, though it has aristocratic features. Wikipedia is not a monarchy, though it has monarchical features. (Wales, in Wikipedia 2005mtb)

5.3 The Harry Potter Project

To focus solely on the bureaucratic apparatus that has evolved to address disputes during Wikipedia's growth would miss much of the day-to-day experience of most Wikipedians. This is what led me to follow the Harry Potter WikiProject: "a collection of pages devoted to the management of a specific family of information within Wikipedia."

Yet, even here, moving from the hundreds of thousands of articles and thousands of editors of the larger Wikipedia to the more manageable site of Harry Potter fans, it can still be difficult to discern the rich data of place, performance, relationship, and motivation one might find in a traditional ethnographic study. Indeed, a strength of the "peer-to-peer" production model (Benkler 2002) of open content communities is the accretion of asynchronous, sometimes anonymous, micro-contributions over time. In the case of the Harry Potter Project, even while there is a self-nominated list of members, the work is not evenly distributed. In fact, adding oneself to this list might be the last Harry Potter edit one does. Relative to the small number of contributions made by many, I was only able to discern a few consistent contributors—this is consistent with the literature on emergent leaders in these communities: a few doing much and many doing a little. For example, a brief analysis shows that on the Harry Potter page, of the 295 editors who made the last 500 edits, the top 10% of editors made 29% of the edits. A similar pattern is found on other Harry Potter Project pages. Even so, these "leaders" do not act in an overtly leading manner; in addition to persuasive arguments, influence is exerted, simply, through the massive amount of work done by these highly active contributors and their ability to otherwise ignore or exhaust those not willing to commit a similar amount of activity.

One additional note is merited about the Harry Potter Project and other "local neighborhoods." First, the larger Wikipedia cultural norms such as neutrality, no original research, and verifiability are present. And, no one would be permitted to claim any exception to these norms, to claim ownership or formal leadership title, even if this privilege was claimed on merit gained in the local context. In keeping with Lois' (1999) finding of a mountain rescue group, contributors would first have to orient themselves towards the larger Wikipedia with worthwhile contributions, and become widely known among other leaders, before making a substantive policy proposal.

6 Discussing leadership

In the previous sections I introduced the notion of leadership in open content communities by way of existing literature, particularly that of the Debian community, and, more specifically, the structure, operation, and culture of leadership in Wikipedia by way of histories and Wikipedia's own self-documentation. In this section, I focus on how the concept of leadership is understood and discussed in the community. Some important themes in these conversations are the frustrations resulting from the voluntary and consensus character of the community, and the use of metaphors (e.g., dictator or constitutional monarchy), politeness, and humor in understanding and resolving those frustrations.

6.1 Dictatorships and Jimbo's Role

Open content communities with a single prominent, or paramount, leader sometimes characterize that leader as a type of benevolent dictator (Wikipedia 2006)—like the cabal oligarchy of USENET (Pfaffenberger 1996) or Debian (Coleman 2005). Linus Torvalds, the original author of Linux, is known as a benevolent dictator. Guido von Rossum, author of the Python programming language has the additional honorific of being benevolent dictator "for life," or BDFL. Jimmy Wales is also often characterized as a benevolent dictator though it is not a designation he accepts, as we will see.

The need for "dictatorship" arises from the difficulty inherent to decision making in large, voluntary, and consensus oriented communities. NSK (2005) writes, "Wikipedia suffers from many voices, often contradictory. I think you need an influential leader to take final decisions (after community input of course)." In addition to differing opinions among those of good faith, an informal and consensus based approach does not seemingly deal well with those who act in bad faith:

What is needed in obvious cases like this is a "benevolent dictator", whether it's Jimbo Wales or the arbcom [Arbitration Committee], to examine the editors' contributions then ban them, because these are not bona fide Wikipedians who happen to have a strong POV [point of view]. They are fanatics acting to promote the views of a poltical cult, and they're here for no other reason. Yet here they remain, making a mockery of everything Wikipedia stands for. (slimvirgin 2005)

Where possible, Wales has delegated authority, particularly to the Board of Trustees and Arbitration Committee, but much authority remains with Wales:

Wikipedia is "at the mercy of" Jimbo. Jimbo has delegated his "mercy", to use your term, to the Arbitration Committee that he convened over 15 months ago, and which he periodically refreshes the membership thereof as guided by the wishes of the community. Significant disciplinary matters in Wikipedia are thus guided by a number of editors who are held in high esteem by the community at large (or, at least, so one hopes). (Forrester 2005)

Anthere, a member of the Board of Trustees, described this balance of reserved authority and delegation as one of facilitating or hindering a direction:

I think that what is especially empowering is the leadership type of Jimbo. Jimbo is not coaching at all, and rather little directing (though hints are sometimes quite clear), as well as rather little delegating (I think the foundation would sometimes benefit from more delegation from Jimbo). His type is essentially supportive. Very low direction but very high support. This leaves basically as much opportunity to work in certain directions as one would dream of. However, one moves in a direction supported by Jimbo much more quickly than in a direction not supported by Jimbo. I[t] can take a long time to find a satisfactory decision, but prevents from travelling in an unsafe direction. (Anthere 2005)

However, this balance can lead to ambiguities that prompt discussion, such as the question of editorial authority. In February of 2005 an enormous debate erupted over the illustration included in the encyclopedic article on autofellatio. (Images tend to prompt many debates and raise questions of censorship, free speech, cultural differences, and on the age appropriateness and quality of Wikipedia. A similar debate arose for the image in the clitoris article, as well as a cinematic still of Kate Winslet, in the Titanic article, wearing nothing but a diamond necklace.) When Wales deleted the photographic image of autofaellatio, which had replaced the less contentious illustration, Eric Moeller challenged this action:

Perhaps you could clarify that this was not done in your role as trustee. I don't believe it was, as you did not consult with Angela and Anthere, so I consider it just like an edit by any other Wikipedia editor, only that, of course, you hope that people will take it more seriously because of the reputation that comes with your role in the project, past and present. That's completely reasonable, if done rarely and in cases you consider important.

The page is currently being edit warred over, and one editor uses the comment "rv [revert] to Jimbo's approved version". It would be helpful if you could state here that you are not in the business of approving articles. I believe your edit summary "This image is completely unacceptable for Wikipedia" could be misconstrued to be an official statement, when it is your personal opinon. Some people still see Wikimedia as being governed by a benevolent dictator, and any explanation would help to eliminate that misconception.

I still remember how the Spanish Wikipedia forked over some discussion on advertising. I'm somewhat worried that people might misunderstand your comments, and assume that you are acting as "Chief Editor". On the other side, those who do support the removal of the image might deliberately seek to create that impression in order to further their agenda. (Moeller 2005)

Wales did not respond to this particular e-mail message, but continued discussion with respect to the role such an image would serve for educational purposes. However, Wales' role was further discussed when it was feared that Wikipedia would be the target of a concerted neo-Nazi "attack." This led Wales to clarify that he would prevent such an attack though he also recognizes the dangers inherent to such action:

The danger of course is that the benign dictator may turn out to be biased or wrong himself. So I hestitate to do this except in cases where speed is of essence, or where it's just very clearcut and easy. What I prefer is that I can act as a temporary bridge and "person to blame" while we work on community solutions.

If 300 NeoNazis show up and start doing serious damage to a bunch of articles, we don't need to have 300 separate ArbCom cases and a nightmare that drags on for weeks. I'll just do something to lock those articles down somehow, ban a bunch of people, and protect our reputation and integrity. And then we can also work in parallel to think about the best way to really take care of such problems in the long run.

But if a handful of LaRouche fans want to come in and do pseudo-NPOV on a handful of relatively obscure articles, I'm not in favor of me just cracking heads over it. We can't just ignore it and hope it goes away, either, of course. We just start thinking about it and working on it until we come up with something useful. (Wales 2005)

Seven months later, on the same thread, Wales further defined his role as a "constitutional monarch"

I do not believe in the "benevolent dictator" model for Wikipedia. Our project is of major historical significance, and it is not appropriate for any one person to be the benevolent dictator of all human knowledge. Obviously.

But we have retained a 'constitutional monarchy' in our system and the main reason for it is too _support_ and _make possible_ a very open system in which policy is set organically by the community and democratic processes and institutions emerge over a long period of experimentation and consensus-building....

It is not possible for 10,000 NeoNazis (if such numbers exist) to storm into Wikipedia and take it over by subverting our organic democratic processes because I will not allow it. Period. So we don't have to overdesign those processes out of a paranoia of a hostile takeover.

But this also means that we don't need to over-react right now. We can wait and see. They'll talk a big game but just review those message boards and then look around here. A battle of wits between Wikipedians and Nazis? I know who I'm betting on. (Wales 2005rrn)

Wales' concern with not over designing the "organic democratic processes" echoes Garzarelli's and Galoppini's (2003) notion that the judicious use of charismatic authority can be preferable to a "complex system of rules" (p. 36). And even though Wales is seemingly conscientious about the use of his authority, others note that the "charismatic" character of his leadership can become unsavory. If others appropriate what Wales has said or done as the justification for their own position, some will object:

This kind of hero-worship begins with Christians who find it more chic to parrot Christ's words than to live them. In our context this translates into using "Jimbo said ..." as an argument that would stop all debate. (Saintonge 2005)

Interestingly, concern about this role and title led to a consideration of alternatives for "benevolent dictator" including constitutional monarch, the most trusted party (TMTP, Linus Torvalds' preferred moniker), éminence grise, and Deus ex Machina (Wikipedia 2006mbd). And while the notion of constitutional monarch has achieved some stabilization and acceptance within the community, "dictator" will never disappear from the conversation given its long history within virtual communities. Indeed, the notion not only serves as a measure of the leader's actions, but that of other participants. In one of the many threads about sexual content on the Wikipedia a participant wrote to another: "So your opinion is now law? Wonderful. We don't need all of those nasty little polls or votes.... All we have to do is have you make the decision for us. I thought Jimbo was the benevolent dictator. You seem just to want to be dictator, period" (Rick 2005).

6.2 Politeness, apologies, and joking

In addition, or in response, to the failsafe solution of a leader acting as a tie-breaker and defender of last resort, two cultural practices attempt to lessen the frustration and conflicts common to Wikipedia interaction: politeness and humor.

Among the many politeness norms of Wikipedia, the "Please don't bite the newcomers" is exemplary:

New contributors are prospective "members" and are therefore our most valuable resource. We must treat newcomers with kindness and patience—nothing scares potentially valuable contributors away faster than hostility or elitism. While many newcomers hit the ground running, some lack knowledge about the way we do things. (Wikipedian 2005pdb)

Wales himself frequently writes with:

  1. patience: on a thread regarding Serbo-Croatian dialects: "For those who find Mark irritating, and who may not tend to listen to him on those grounds, I would like to say, listen to him on this point" (Wales 2005ras).
  2. politeness: in response to someone who spoke of a threatened fork over a Frulian dialect and challenged "ARE YOU CRAZY!?!!!!?!!?!?!" Wales responded, "Good luck with that. 'Not yelling at people' is a critical trait of leadership in an all volunteer project" (Wales 2004).
  3. humility: in response to someone concerned about perennial problems, including language policies, Wales wrote, "I'm very sympathetic to all these points. I don't have an easy answer what to do" (Wales 2005rsw).
  4. and apologies: when Wales recommended some text be added to a page when it was already present he wrote, "Ok, my mistake, I'm very very sorry. I didn't see that. I apologize for any confusion" (Wales 2005rrs).

Indeed, the occurrences of frank apologies are quite striking: Skyring wrote to the list, "I've been a real jerk recently, and I'd like to take this opportunity to apologise to everyone who has suffered unnecessary stress as a result" (Mackay 2005). (Skyring was permitted back on the Wikipedia-l list but his one-year ban by the arbitration committee on editing the Wikipedia remains in effect until October 26, 2006.) People are cognizant of the fact that in stressful interactions their tempers might flare:

To Angela, Anthere, Jimbo, Developers and anyone else I might have scolded a bit to much on the Sorbs issue. I was way out of line especially with regards to my remarks towards Angela, Anthere and Jimbo in mail and the developers on IRC, all of whom I know would help at all times. I am sorry and my apologies.

My problem is not that I do not understand why the measures are taken (at least my brain does), they are unfortunately necassary at times. It is just that the heart doesn't understand it and flares up. It is a normal problem with me. The heart takes over from the brain and the damage is done. (vanKalken 2005)

Even Willy on Wheels, an infamous high-speed page move vandal who has attacked many Wikis, might have had a change of heart:

Specifically, I am one of six people who have chronically trolled numerous wikis, notably the English version of Wikipedia. Recently, Wikipedia saved me from failing a research project horribly, and thus, I have decided to repent for my past colourful actions. (on Wheels 2005)

Additionally, joking serves to create general camaraderie, as well as address anxiety about leadership. In response to a message about an April fool's joke about Wales as dictator someone responded:

These jokes don't have a "point". If you scour the list for all messages, you will find that I am not the only one who has a sense of humour and knows how to make jokes. In fact, this extends to Ant, Mav, Jimbo, etc. who can occasionally be found to be making a joke on this list.

I don't know how it is with you, but as far as I know the point of humour is to lighten up a situation, and only occasionally to make a point. (Williamson 2005)

Ironically, April fool's jokes are a source of difficulty themselves since this tomfoolery isn't present and understood in all cultures, some use it as an excuse for outright vandalism, and many object to any change of encyclopedic articles for humorous purposes.

Interestingly, much like the ancient USENET parody on a cabal (Pfaffenberger 1996), Shannon wrote (tongue in cheek) a message entitled "how to join a cabal" to the list:

I have reliable information that an over-zealous Australian is about to launch a coup to gain control of the wikimedia cabal (and hence all international commerce, and politics). I am told that he goes by the code name of Ta bu shi da yu (which may well contain demonic anagrams, several super-computers from wikimedia's secret service are currently working on the problem). I attach a letter i recently intercepted where he goes so far as to claim that control is all ready his; this suggests that he believes a sizable number of editors will join him in his rebellion. (Shannon 2005)

Politeness and humor permit the community to discuss a source of anxiety without actually accusing or attacking one another.

7 A theory of leadership

The theoretical approach of this paper has not been to contest, or even reconstitute (Burawoy 1991), a single existing theory. Simply, there is, as of yet, little (to no) literature on leadership in the Wikipedia. Instead, I've identified features of leadership in related communities (e.g. emergent leaders, discussion of cabals, humor, etc.,), confirmed their existence in the Wikipedia community, and subsequently subsume, extend, and marshal them under the theoretical notion of an open content community.

  1. Leaders tend to be emergent; the few who do much lead the many who do a little.
  2. Leaders operate within a mix of governance models: meritocratic (setting the direction by leading the way), autocratic (acting as an arbiter or defender of last resort), anarchic (consensus) and occasionally democratic (voting).
  3. Leadership is not granted formal status, (meritocratic action and egalitarian discourse reign), except for paramount leaders, those at the very top, such as a founder.
  4. Early leadership (i.e. founders) lends direction and momentum to the development of a community's culture.
  5. Paramount leaders convince by persuasion and example though they also retain charismatic authority accumulated from their merit in order to act, as a last resort, as an (autocratic) arbiter between those of good faith or defender against those of bad faith.
  6. Paramount leaders whose autocratic actions exceed their accumulated merit/charisma risk being hounded to death within their community, or the community forking.
  7. Humor and politeness facilitate camaraderie between all participants and eases the exercise of authority and the related anxiety about it.
  8. Only those leaders that tread carefully and continue to make important contributions (including, now, the judicious exercise of autocratic authority) are granted the "dictator" title. While this term might not be the most appropriate in capturing the genuine character of this role, it serves as a warning: a good-natured joke balanced on the edge of becoming a feared reality. It serves as a caution to such leaders, as well as a metaphoric yardstick for discussing any participant's action.

This theory could be tested explicitly against other paramount leaders of open content communities; for example, those who are known to be of a good humor and referred to as "dictators," such as Linus Torvalds and Guido von Rossum, and those who are not.

8 Conclusion

If one were to draw lessons from the case of Wikipedia for aspiring leaders in similar communities, the first truth to be recognized is that it takes a lot of work. In fact, the passion needed to dedicate oneself to the often voluminous, mostly voluntary, and possibly thankless work undercuts my supposition; people don't set out to be leaders, they end up as such like the farmer generals who left their fields for the love of Rome. They were dedicated to some small project (e.g., software or an encyclopedia) around which a community developed and must then be guided and protected. However, when a person comes to be responsible for more than he or she can do by dint of will alone, new responsibilities and authority pull taut a tightrope that must be carefully walked before the eyes of one's peers. Sanger's insightful reflections about his exit from the community and Wales' moniker of benevolent dictator are testaments to this.

Bibliography

Anthere (2005). Re: NYTimes.com: Google may host encyclopedia project. wikipedia-l. Retrieved on February 14, 2005 from < http://marc.theaimsgroup.com/?i=4211339B.6030803@yahoo.com >.

Anthere (2005). Re: Sanger's memoirs. wikipedia-l. Retrieved on April 23, 2005 from < http://marc.theaimsgroup.com/?i=4269F0E4.8080501@yahoo.com >.

Benkler, Y. (2002). Coase's penguin, or, Linux and the nature of the firm. The Yale Law Journal, 112(3):369-446.

Bosco, G. (2004). Implicit theories of good leadership in the open-source community. Master's thesis, Technical University of Denmark. Retrieved on January 13, 2006 from < http://opensource.mit.edu/papers/bosco.pdf >.

Burawoy, M. (1991). Reconstructing social theories. In Ethnography Unbound: Power and Resistance in Modern Metropolis, chapter 2, pages 8-27. University Of California Press, Berkeley.

Coleman, E. G. (2005). Three ethical moments in Debian: the making of an (ethical) hacker, part III. In the social construction of freedom in free and open source software: actors, ethics, and the liberal tradition. Retrieved on September 09, 2005 from < http://healthhacker.org/biella/coleman-abstract.pdf >.

Etzioni, A. (1975). Comparative analysis of complex organizations. Free Press of Glencoe, NY, revised edition.

Forrester, J. D. (2005). Re: Your golorous [sic] leader. wikien-l. Retrieved on April 08, 2005 from < http://marc.theaimsgroup.com/?i=200504081652.j38GqMtd009896@mail-relay-2.csv.warwick.ac.uk >.

Foundation, W. (2005). Board of Trustees. Retrieved on January 13, 2006 from < http://wikimediafoundation.org/wiki/Board_of_Trustees >.

Garzarelli, G. and Galoppini, R. (2003). Capability coordination in modular organization: Voluntary FS/OSS production and the case of Debian GNU/Linux. In Economics Working Paper Archive at WUST, number 0312005. Industrial Organization. Retrieved on September 03, 2004 from < http://opensource.mit.edu/papers/garzarelligaloppini.pdf >.

Lois, J. (1999). Socialization to heroism: individualism and collectivism in a voluntary search and rescue group. Social Psychology Quarterly, 62(2):117-135.

Mackay, P. (2005). RE: Skyring has been allowed back on. wikien-l. Retrieved on November 01, 2005 from < http://marc.theaimsgroup.com/?i=20051101102530.IIDR16720.omta01sl.mx.bigpond.com@skyringstudy >.

Moeller, E. (2005). Re: [[talk:autofellatio]] - Poll. wikien-l. Retrieved on February 13, 2005 from < http://marc.theaimsgroup.com/?i=420F062F.50805@gmx.de >.

NSK (2005). Re: Flags. wikipedia-l. Retrieved on January 10, 2005 from < http://marc.theaimsgroup.com/?i=200501101748.06696.nsk2@wikinerds.org >.

O'Mahony, S. and Ferraro, F. (2004). Hacking alone? The effects of all online and offline participation or an open-source community leadership. Retrieved on January 13, 2006 from < http://opensource.mit.edu/papers/omahonyferraro2.pdf >.

O'Mahony, S. and Ferraro, F. (2003). Managing the boundary of an 'Open' project. Retrieved on December 05, 2003 from < http://opensource.mit.edu/papers/omahonyferraro.pdf >.

on Wheels, W. (2005). Clearing my Conscience. wikien-l. Retrieved on October 30, 2005 from < http://marc.theaimsgroup.com/?i=BAY18-F19D77AAACD6C687D36E7D1A96D0@phx.gbl >.

Pfaffenberger, B. (1996). if {I} want it, it's okay: Usenet and the (outer) limits of free speech. The Information Society, pages 365-386. Retrieved on November 05, 2005 from < http://pfaff.sts.virginia.edu/bphome/docs/pdf/usenet.pdf >.

Reagle, J. (2003). Open communities, media, source, and standards. Retrieved on January 12, 2006 from < http://reagle.org/joseph/blog/ >.

Reagle, J. (2004). Open content communities. M/C: A Journal of Media and Culture, 7. Retrieved on July 01, 2004 from < http://journal.media-culture.org.au/0406/06_Reagle.rft.php >.

Rick (2005). Re: Re: Writing about sexual topics responsibly is not. wikien-l. Retrieved on February 14, 2005 from < http://marc.theaimsgroup.com/?i=20050214062541.66792.qmail@web60609.mail.yahoo.com >.

Saintonge, R. (2005). Re: Wikimedia Foundation Internal Radio. wikipedia-l. Retrieved on April 12, 2005 from < http://marc.theaimsgroup.com/?i=425C3FB4.2070401@telus.net >.

Sanger, L. (2005). The early history of Nupedia and Wikipedia: a memoir. Slashdot. Retrieved on April 19, 2005 from < http://features.slashdot.org/article.pl?sid=05/04/18/164213&from=rss >.

Schein, E. H. (1992). Organizational culture and leadership. Jossey-Bass Publishers, San Francisco, 2nd edition.

Shannon, R. (2005). Re: Re: How to join the cabal. wikien-l. Retrieved on January 03, 2005 from < http://marc.theaimsgroup.com/?i=623d733805010301457eace82c@mail.gmail.com >.

SlimVirgin (2005). Re: Original Research versus Point of View,. wikien-l. Retrieved on January 18, 2005 from < http://marc.theaimsgroup.com/?i=4cc603b0501181558569cb84f@mail.gmail.com >.

van Kalken, W. (2005). my apologies. wikipedia-l. Retrieved on July 06, 2005 from < http://marc.theaimsgroup.com/?i=42CC23D9.7080207@vankalken.net >.

van Maanen, J. (1988). Tales of the field: on writing ethnography. University Of Chicago Press, Chicago.

Waguespack, D. and Fleming, L. (2004). Penguins, camels, and other birds of the feather: the emergence of leaders in the open innovation communities. Retrieved on January 28, 2004 from < http://faculty.haas.berkeley.edu/elf/ECON222/Waguespack Fleming.pdf >.

Waguespack, D. and Fleming, L. (2005). Penguins, camels, and other birds of a feather: brokerage, boundary spanning, and leadership in Open Innovation Communities. Retrieved on January 13, 2006 from < http://opensource.mit.edu/papers/flemingwaguespack.pdf >.

Wales, J. (2005). Re: Ant: Serbo-Croatian Wikipedia - a policy question? wikipedia-l. Retrieved on October 16, 2005 from < http://marc.theaimsgroup.com/?i=43524BFD.9000501@wikia.com >.

Wales, J. (2005). Re: Neo-nazis to attack wikipedia. wikien-l. Retrieved on February 07, 2005 from < http://marc.theaimsgroup.com/?i=20050207162911.GO29080@wikia.com >.

Wales, J. (2005). Re: Re: A neo-nazi wikipedia. wikien-l. Retrieved on August 27, 2005 from < http://marc.theaimsgroup.com/?i=4310D2CB.1060401@wikia.com >.

Wales, J. (2005). Re: Re: Static HTML dumps. wikipedia-l. Retrieved on October 18, 2005 from < http://marc.theaimsgroup.com/?l=wikipedia-l&m=112961483513485&w=2 >.

Wales, J. (2005). Re: Status of Wikimedia. wikipedia-l. Retrieved on November 07, 2005 from < http://marc.theaimsgroup.com/?i=436FDD60.1060301@wikia.com >.

Wheeler, D. A. (2005). Why Open Source Software/Free Software (OSS/FS)? Look at the numbers! Retrieved on October 01, 2004 from < http://www.dwheeler.com/oss_fs_why.html >.

Wikinfo (2004). How we differ from Wikipedia and why. Retrieved on April 29, 2004 from < http://internet-encyclopedia.org/wiki.php?title=Wikinfo:How_we_differ_from_Wikipedia_and_why >.

Wikipedia (2004). History of Wikipedia. Retrieved on April 29, 2004 from < http://en.wikipedia.org/wiki/History_of_Wikipedia >.

Wikipedia (2004). Nupedia. Retrieved on December 22, 2005 from < http://en.wikipedia.org/wiki/Nupedia >.

Wikipedia (2004). Wikipedia:About. Retrieved on December 17, 2005 from < http://en.wikipedia.org/wiki/Wikipedia:About >.

Wikipedia (2004). Wikipedia:Administrators. Retrieved on April 25, 2004 from < http://en.wikipedia.org/wiki/Wikipedia:Administrators >.

Wikipedia (2005). Please do not bite the newcomers. Wikipedia. Retrieved on October 12, 2005 from < http://en.wikipedia.org/wiki/Don't_bite_the_newbies >.

Wikipedia (2005). Wikipedia:List of administrators. Retrieved on January 13, 2006 from < http://en.wikipedia.org/wiki/Wikipedia:List_of_administrators >.

Wikipedia (2006). Benevolent Dictator for Life. Wikipedia. Retrieved on January 04, 2006 from < http://en.wikipedia.org/wiki/Benevolent_Dictator_for_Life >.

Wikipedia (2006). Meta:benevolent dictator. Wikipedia. Retrieved on January 04, 2006 from < http://meta.wikimedia.org/wiki/Benevolent_dictator >.

Wikipedia (2006). Meta:talk:benevolent dictator. Wikipedia. Retrieved on January 04, 2006 from < http://meta.wikimedia.org/wiki/Talk:Benevolent_dictator >.

Williamson, M. (2005). Re: Wikimedia Foundation Internal Radio System. wikipedia-l. Retrieved on April 12, 2005 from < http://marc.theaimsgroup.com/?i=849f98ed050412011473cf4b91@mail.gmail.com >.

Wilson, J. (2000). Volunteering. Annual Review of Sociology, 26:215-240.

Yoo, Y. and Alavi, M. (2003). Leadership in virtual teams: what do emergent leaders do? unpublished.

Yukl, G. A. (1998). Leadership in organizations. Prentice-Hall, Englewood Cliffs, NJ, 4 edition.