Nine Theses

I

 submit these nine theses to Wikipedia's community and to the world. I do this, as Martin Luther said when he posted his famous 95 theses, "Out of love for the truth and the desire to elucidate it."

A quarter of a century ago, Jimmy Wales' company Bomis hired me to start a free encyclopedia. The first draft, from which we learned much, was Nupedia—it made slow progress. So, a year later, on January 2, 2001, when a friend told me about wikis, I immediately began imagining a wiki encyclopedia. I proposed it to Jimmy, then CEO of Bomis. He agreed and installed the software, and I went to work getting things ready. After I named it, we launched Wikipedia on January 15, 2001, and just nine days later, I was able to write, "Wikipedia has definitely taken on a life of its own; new people are arriving every day and the project seems to be getting only more popular. Long live Wikipedia!"

The title I claimed at the time was "chief instigator." My daily leadership for the 14 months after that was essential to transforming a completely empty, blank wiki into what would soon become the largest written resource in the history of the world. I was responsible for several policies that were and are fundamental to the project: the exclusive focus on an encyclopedia; neutrality; "no original research"; "be bold"; aspects of the verifiability policy; and other things. I even proposed the tongue-in-cheek "rule" to "ignore all rules." For more, see this page on my role in Wikipedia, my Slashdot memoir, and my book, Essays on Free Knowledge. I say these things not to brag but to show why my proposals deserve a careful hearing.

The Nine Theses

1. End decision-making by "consensus."

Wikipedia's policy of deciding editorial disputes by working toward a "consensus" position is absurd. Its notion of "consensus" is an institutional fiction, supported because it hides legitimate dissent under a false veneer of unanimity. Perhaps the goal of consensus was appropriate when the community was small. But before long, the participant pool grew so large that true consensus became impossible. In time, ideologues and paid lackeys began to declare themselves to be the voice of the consensus, using this convenient fiction to marginalize their opponents. This sham now serves to silence dissent and consolidate power, and it is wholly contrary to the founding ideal of a project devoted to bringing humanity together. Wikipedia must repudiate decision-making by consensus once and for all.

2. Enable competing articles.

Neutrality is impossible to practice, if editors refuse to compromise—-and Wikipedia is now led by such uncompromising editors. As a result, a favored perspective has emerged: the narrow perspective of the Western ruling class, one that is "globalist," academic, secular, and progressive (GASP). In fact, Wikipedia admits to a systemic bias, and other common views are marginalized, misrepresented, or excluded entirely. The problem is that genuine neutrality is impossible when one perspective enjoys such a monopoly on editorial legitimacy. I propose a natural solution: Wikipedia should permit multiple, competing articles written within explicitly declared frameworks, each aiming at neutrality within its own framework. That is how Wikipedia can become a genuinely open, global project.

3. Abolish source blacklists.

An anonymous "MrX" proposed a list of so-called perennial sources just seven years ago, which determine which media sources may, and may not, be used in Wikipedia articles. The page is ideologically one-sided and essentially blacklists disfavored media outlets. Wikipedians now treat this list as strict—but unofficial—policy. This approach must be reversed. Wikipedia should once again explicitly permit citations even from sources that the page currently blacklists. Rather than outright banning entire sources that can contain valid and important information, Wikipedia articles should use them when relevant, while acknowledging how different groups assess them. Neutrality requires openness to many sources; such openness better supports readers in making up their own minds.

4. Revive the original neutrality policy.

In short, Wikipedia must renew its commitment to true neutrality. The present policy on neutrality should be revised to clarify that articles may not take sides on contentious political, religious, and other divisive topics, even if one side is dominant in academia or mainstream media. Whole parties, faiths, and other "alternative" points of view must no longer be cast aside and declared incorrect, in favor of hegemonic Establishment views. Solid ideas may be found in some of the first policy statements, including the first fully elaborated Wikipedia policy and the Nupedia policy of 2000.

5. Repeal "Ignore all rules."

On February 6, 2001, I wrote this humorous rule—"Ignore all rules"—to encourage newcomers. Ironically, my joke now serves to shield insiders from accountability. It no longer supports openness; it protects power. Wikipedia should repeal it.

6. Reveal who Wikipedia's leaders are.

It is a basic principle of sound governance that we know who our leaders are. So why are the Wikipedia users with the most authority—"CheckUsers," "Bureaucrats," and Arbitration Committee members—mostly anonymous? Only 14.7% of such users reveal a full, real name. These high-ranking individuals obviously *should* be identified by their real and full names, so they can be held accountable in the real world. After all, Wikipedia is now one of the world's most powerful and well-funded media platforms. Wikipedia's influence far exceeds that of major newspapers, which follow basic standards of transparency and accountability. Such standards are not mere ideals but real requirements for any media organization of Wikipedia's stature. As of 2023, Wikipedia's endowment was $119 million, its annual income $185 million. Therefore, if safety is a concern, funds should be used to indemnify and otherwise protect publicly identified editorial leaders. Wikipedia, admit that your leaders are powerful, and bring them out into the open; great power requires accountability. If you continue to stymie accountability, government may have to act.

7. Let the public rate articles.

A system of public rating and feedback for Wikipedia articles is long overdue. Articles now boldly take controversial positions, yet the public is not given any suitable way to provide feedback. This is disrespectful to the public. There is an internal self-rating system, not visible to readers. The platform experimented with an external ratings system but scrapped it after a few years, and it didn't help readers. Wikipedia does not need a complex system to get started. An open source AI rating system would not take long to develop. The platform already collects relevant objective data such as number of edits and word count: make that public. As to human raters, they should be provably human, unique, and come from outside of the editor community. When articles are evaluated by a diverse audience, content quality and neutrality will be improved.

8. End indefinite blocking.

Wikipedia's draconian practice of indefinite blocking—typically, permanent bans—is unjust. This is no small problem. Nearly half of the blocks in a two-week period were indefinite. This drives away many good editors. Permanent blocks are too often used to enforce ideological conformity and protect petty fiefdoms rather than to serve any legitimate purpose. The problem is entrenched because Administrators largely lack accountability, and oversight is minimal. The current block appeals process is ineffective; it might as well not exist, because it is needlessly slow and humiliating. These systemic failures demand comprehensive reform. Indefinite blocks should be extremely rare and require the agreement of three or more Administrators, with guaranteed periodic review available. Blocks should nearly always be preceded by warnings, and durations should be much more lenient.

9. Adopt a legislative process.

Wikipedia's processes for adopting new policies, procedures, and projects are surprisingly weak. The Wikimedia Foundation (WMF) has launched initiatives, but these do not establish major editorial policy. Incremental policy tweaks cannot deliver the bold reforms Wikipedia needs. No clear precedents exist for adopting significant innovations. The project is governed by an unfair and anonymous oligarchy that likes things just as they are. This stagnation must end. Wikipedia needs an editorial legislature chosen by fair elections: one person, one vote. To establish legitimate and fair governance, the WMF should convene a constitutional convention to create an editorial charter and assembly. This assembly would be empowered to make the sorts of changes proposed in these "Nine Theses."

Further theses

When I began this project, I had more than nine ideas, of course. The following are some further theses, which I submit undeveloped. The fact that so many plausible proposals for improvement come so readily to mind underscores the platform's dysfunction.

Wikipedia should join the Encyclosphere.

The Encyclosphere is a project I started in 2019 to collect all the encyclopedia articles in the world in a single decentralized network, with each article shared according to the ZWI (Zipped WIki) file format. This is an enormous and very worthwhile project, and, by supporting both EncycloReader and EncycloSearch, the Knowledge Standards Foundation has made a credible start on this network. While the Encyclosphere has collected some 65 encyclopedias so far, Wikipedia could motivate the rest to contribute to the world's knowledge—by their own lights—by running an Encyclosphere node. If Wikipedia does not enable competing articles (i.e., Thesis 2), this would be an excellent fallback position.

Implement term limits.

Administrators, as a class, tend to become too impressed with their own power on Wikipedia. If this really is a "janitorial" sort of duty (see Thesis 6), then a much larger body of people should be called upon to help. Therefore, I believe Administrators—and other positions of power and authority—should be subject to some system of term limits. I am not dogmatic about the length. One idea would be: two-year terms; may be elected to back-to-back terms; cannot be elected three times in a row; cannot be elected more than three times in a ten year period; otherwise, no limit to number of times one may serve as an Administrator. But there are many ways to implement such a system. Whichever is chosen, the election process would have to be made easier for experienced Wikipedians to get on board in this role.

Require yearly Administrator performance reviews.

Administrators, as a condition of their continuance in the role, should be subject to annual anonymous reviews of their Administrator work. Open source LLMs and other automated tools could be very useful in collecting data for such reviews.

Partner with an independent organization to handle appeals.

This is a much more ambitious way to solve the problems introduced in Thesis 8. Establish a fully and provably independent appeals body, which is nationally, politically, and religiously balanced. It must be answerable neither to the Wikipedia community nor to the Wikimedia Foundation. This body would oversee appeals against repeated blocking and on select editorial issues, ensuring decisions are balanced, just, and transparent—free from the internal politics of current administrative structures in which the foxes are guarding the henhouse.

End IP editing.

From the beginning, Wikipedia has allowed people to edit without logging in. This initially helped to attract contributors, but it is no longer needed and is now counterproductive. IP editing is now widely abused by insiders as a tool of gamesmanship, rather than making it easier for outsiders to contribute. It is long past time for this startup feature to be retired. Wikipedia has grown up. It is time for the community to act like it.

Replace or augment the edit counter with work assessments.

The edit counter has helped create an insider class that does not deserve the degree of power it wields in the system. Some of the most qualified people in the world have little time to edit Wikipedia, and so they will naturally not make many edits. But their opinion about their field of expertise ought to be worth more than that of a teenager with 50,000 edits. If not replaced, then maybe the edit counter could be augmented by independent work assessments (i.e., performance evaluations) by open source LLMs and other automated tools. It would be best to move away from the simplistic metric of edit counts and towards a more nuanced evaluation of contributions based on content quality and impact. This would reflect a true measure of a contributor's value to the project, if that is regarded as important. The use of automated tools for this task would help keep it free of corruption and cronyism.

End or loosen restrictions on "meat puppetry."

My understanding is that off-wiki collaboration is a thing that insiders do all the time anyway; the rule is selectively enforced, in a way that is extremely hypocritical. It should be possible to have meaningful discussions of how the Wikipedia article should look outside of Wikipedia. It is time for Wikipedia to become an open and explicit part of larger, off-wiki conversations. This is already happening. If this is not acknowledged, the conversations will take place sub rosa among secret confederates, which is much worse.

Label pages that are not appropriate for children under 13.

"Adult" content on Wikipedia should be labeled as such. By implementing age-appropriate labels to ensure the safety and appropriateness of content for younger audiences, Wikipedia would meet societal standards of protection for minors. The encyclopedia does not do so now. This is a problem I brought to Wikipedia's attention in 2012, when I proposed a solution. The proposal was never implemented.

Allow memorial articles about elders and deceased friends and family.

I claim that our elders are all noteworthy. Regardless of whether they were ever in the news, they have had a lifetime's impact on the rest of us. Therefore, the children, other relatives, and friends of persons over 65 years old should be permitted to memorialize their lives, but only if their next of kin agree. Existence could be confirmed through public records or reliable testimony. Such articles could be placed in a new namespace. Articles could be written based on oral histories. While the latter primary sources would not meet traditional reliability policies, they would be a valuable record of what family and friends said about our elders and dear departed, as permanent lore about a person. The result would be an amazing resource for future historians.

Embrace inclusionism.

The firm tendency to delete perfectly good articles because somebody thinks the topic is not "noteworthy" enough (called deletionism) is an innovation. Deletionist tendencies are toxic to a healthy, free, and open encyclopedia. Generally speaking, if someone can be found to write an article on a topic, and it otherwise meets Wikipedia's standards, it is best to include the article. Thus, Wikipedia's rules on what counts as "noteworthy" need to be revised, to be made more lenient and inclusive.

1. End decision-making by "consensus."

Wikipedia's policy of deciding editorial disputes by working toward a "consensus" position is absurd. Its notion of "consensus" is an institutional fiction, supported because it hides legitimate dissent under a false veneer of unanimity. Perhaps the goal of consensus was appropriate when the community was small. But before long, the participant pool grew so large that true consensus became impossible. In time, ideologues and paid lackeys began to declare themselves to be the voice of the consensus, using this convenient fiction to marginalize their opponents. This sham now serves to silence dissent and consolidate power, and it is wholly contrary to the founding ideal of a project devoted to bringing humanity together. Wikipedia must repudiate decision-making by consensus once and for all.

Note: The first four theses all concern different aspects of neutrality. This involves some repetition and expansion of analysis, because the issues involved are so central and important.

The Problem

When Wikipedia launched, we borrowed a principle from the original wikis of the 1990s: Wikipedia articles would represent a "consensus view."

A consensus is, of course, a position that everyone can agree to. Not on Wikipedia, though. On Wikipedia, an article that is completely one-sided and quite controversial is often declared—with furrowed-brow seriousness—to represent the community "consensus." If this sounds ridiculous, that's because it is. As someone who was there at the beginning, I can tell you that this is not Wikipedia's original notion of consensus.

But a consensus view was not a single view of a controversy. It was a frank admission that there were multiple, competing views; it was an exploration of the "lay of the land" that all could agree upon. Indeed, our original practice of representing multiple views fairly was why decision-making by consensus could be made policy in the first place. We, full of the foolish idealism of youth, imagined that motivated ideologues could be taught to write neutrally, all coming together to make the text express all relevant possibilities. The rule was simple: When we disagree, we should not fight over whose views should be stated by the article. Rather, we attribute our own views to their best representatives, and we allow others to do the same with theirs. In this way, we thought, we could avoid hashing out controversies and focus on recording facts. The practice of neutrality was a framework in which we could work toward a "consensus text." The consensus was not about the facts, but about how a neutral exploration of the debate should read. This was the original understanding of consensus—but now it is long forgotten. Of course we could not agree on the facts. What we could agree upon was a text that represented many different views of the facts side-by-side.

But that was, as I said, foolishly idealistic. We never made proper allowances for the harsh reality that there would be truly intractable disagreements, even among people who say they agree with the framework of neutrality—some people simply refuse to let others have their say at all, or not in any fair way. This became obvious even in the first year of the project, which cooled me on the very idea of "consensus" as a method of conflict-resolution.

Then, surely, the naïve idea of decision-making by consensus was dropped. Right?

Wrong. Instead, after I left, Wikipedia became increasingly strange and insular, and the notion of "consensus" was actually twisted into its opposite. Today, the new reality is admitted frankly:

Consensus on Wikipedia does not require unanimity (which is ideal but rarely achievable), nor is it the result of a vote.

...

When editors do not reach agreement by editing, discussion on the associated talk pages continues the process toward consensus.

A consensus decision takes into account all of the proper concerns raised. Ideally, it arrives with an absence of objections, but often, we must settle for as wide an agreement as can be reached. When there is no wide agreement, consensus-building involves adapting the proposal to bring in dissenters without losing those who accepted the initial proposal.

Long gone is any suggestion that neutrality is a framework that permits a true consensus to be achieved. We early Wikipedians find this sad. Let us analyze what has changed, in terms of the goal, the process, and the community.

(1) The goal has changed; pluralistic expression of different viewpoints is not specifically preferred. Gone is any notion that consensus involves laying out a plurality of viewpoints in a coherent and balanced way. In fact, sometimes, when people attempt to explore various competing views in an article, this is rejected—wrong-headedly, I believe—as a "synthesis of published material," and thus original research. When there is conflict, positions often harden. Rather than allowing multiple views to emerge, the "community" winds up selecting one view, or a few leading views, and calling that "the consensus."

(2) The method of reaching "consensus" has also changed; real negotiation among equals has largely disappeared, regardless of what the guidelines say. Gone is the practice of friendly negotiation toward agreement or collaborating in flat, self-managing groups, usually without administrative interference. In its place is fiat judgments made by insiders, sometimes preceded by the adversarial process of pushing the issue through a complex dispute resolution bureaucracy. The end result of this often abusive process is cynically dubbed "the consensus."

The aim of the winning side, all too often, is to exclude ideological opponents. Thus, the consensus is engineered: one side's arguments are declared by an editorial bureaucracy to fit well with an alphabet soup of acronym-laden policies, guidelines, and "essays." This determination ultimately turns on which side boasts the most senior editors and administrators. Sometimes, the true heavies are called in, who rule peremptorily, as if they were high-ranking commissars settling matters between underlings. The Wikipedians themselves now rightly mock such displays of power, but without stopping the charade.

(3) And the community has changed—perhaps saddest of all, for those who remember the early days. A truly polite, collegial atmosphere has largely disappeared. I warned Wikipedians when I left to be "open and warmly welcoming, not insular." They did not take my advice. Long gone is the sincere, friendly collegiality of people who really are committed to synthesizing diverse viewpoints into a single cohesive document. In its place is the ill-will begotten of an adversarial game in which bureaucratic types face off, calling out every minor infraction and citing acronyms at each other. No wonder friendly, decent people are so often driven away by the sheer hostility of the Wikipedia "community."

The plain fact is that Wikipedian "consensus" is no consensus at all. That is the elephant in the room. I am pointing right at it. One is hard pressed to know what precisely to call the current decision-making process. Wikipedians deserve ridicule if they continue calling it "consensus." That is an institutional fiction, and a darkly cynical one.

The Reasonable Solution

To begin, stop calling your process "consensus." At least rename it. As to what description replaces it, this is important, but I leave that to the Wikipedians.

I will also abstain from proposing a different decision-making process. Mainly I am saying is that this institutional fiction must, for the sake of honesty, be dropped. I will say this, however. Anyone who has the honesty to admit that "consensus" was an impossible fiction all along should also be able to see that there is a need for some reform in how editorial disputes are resolved. The fiction itself plays a role in the Wikipedia game: it cynically papers over what is, in fact, the raw exercise of power. Yet, since the description of the existing process as "consensus" is official policy, it might be changed only through strong leadership within the community or imposition by the Wikimedia Board.

For those Wikipedians who are willing to try to think through the difficult issues involved in fair community decision-making, let me suggest just a few possible ideas:

1. Create an open editorial committee of persons known to be uniquely identified (if not known publicly), so that there is always one person, one vote. Controversies are settled by a vote of some randomly selected subset of the committee, who can escalate important issues upward.
2. As a variant on the foregoing, weight the votes in the same way that X.com does with its "Community Notes."
3. Those who submit a dispute to some deciding agency must precisely identify the issue on which the users disagree. They must spend at least 24 hours attempting to arrive at consensus on at least what the issue is that they disagree about.

If Wikipedia neither changes its decision-making practice nor changes the description of its practice as "consensus," it is clear that their editorial process has lost all credibility. The bickering baboons of bias will continue to fight among themselves until the most powerful emerges. Oblivious to the high comedy of it all, Wikipedia's self-appointed deciders congratulate themselves on being the voice of the "consensus"—of all who think exactly as they do.