Online Norms and Collective Choice

by Henry Farrell on January 16, 2009

Melissa Schwartzberg at Columbia and I have an essay in the new issue of “Ethics and International Affairs”: on the ways in which norms structure majority-minority relations on Wikipedia and the Daily Kos. The journal has made it “freely available online here”:

Building on case studies of Wikipedia and the Daily Kos, we make three basic claims. First, we argue that different kinds of rules shape relations between members of the majority and of the minority in these communities in important and consequential ways. Second, we argue that the normative implications of these consequences differ between online communities that seek to generate knowledge, and which should be tolerant of diversity in points of view, and online communities that seek to generate political action, which need less diversity in order to be politically efficacious. Third, we note that an analysis of the normative desirability of this or that degree of tolerance needs to be tempered with an awareness that the actual rules through which minority relations are structured are likely the consequence of power relations rather than normative considerations.

The other lead essay is “Michael Walzer on democracy promotion”:, which should keep Daniel entertained …



tom s. 01.16.09 at 4:38 pm

Very interesting. I very much like the focus on particular communities rather than the familiar generalizations: it is a big help in keeping feet on the ground.

I don’t go to Kos and I only rarely edit Wikipedia, so the inner workings of both are fascinating to me. But I come away from the article with a sense of familiarity: these are the same old dynamics that are at work in offline communities, writ digital. The willingness of some groups to win by attrition, the use of procedures to establish one’s own position, the conflicts of private agendas and so on are all familiar to anyone who has been in political parties or in other groups with a purpose, from community orchestras to protest movements.

If the Internet has a unique character in how these conflicts are resolved, surely it comes from code. Although you say that online projects are vulnerable to disruption by outsiders, both the examples you choose still have the founder at the top of the hierarchy. Such stability, over periods of 8 and 6 years, may speak to the importance of control over code in determining the tolerance of dissent.

Within a few years, we will see how Internet organizations handle transitions of power to new generations. Will Wikipedia still flourish when those who feel a sense of ownership from their connection with the founding of their creation move on to other things, or will a new generation be able to feel ownership? Will there be ritual “handing over of the code” ceremonies to anoint a successor to each owner?


Henry 01.16.09 at 7:19 pm

Tom – that is very interesting and helpful. We do need to think more about code – what the piece really does is to transplant some arguments from standard institutional theory and see how they do in this new context, but code does give a degree of veto power that is still hard to capture. Another comparison which might be interesting would be to look at Kos vs Kuro5hin – both began from the same basic code base (Scoop), but both have evolved (or degenerated in the latter case) in very different directions.


c.l. ball 01.16.09 at 8:07 pm

A thoughtful piece, and helpful for tracking down data on this, including the “Self-Segregation or Deliberation?” piece.

In the Kos case, what counts as a “collective choice” is unclear. In some ways, Kos is more tolerant than Wikipedia. It would appear that the Kos code changes enabled minority groups to proceed in their dissent rather than getting out trolled by the majority. If HRC supporters were under the old system, they would have been marked as trolls, whereas with the code reforms they mostly received negative comments. The HRC debate went on until her supporters chose exit rather than continuing to voice dissent.

At Wikipedia, the choice is the text of the article. Once the minority concedes, the debate ends, although they might continue to participate in Wikipedia on other articles.


John Quiggin 01.16.09 at 11:14 pm

A point you might have mentioned about Wikipedia is the development of recourse to external authority in resolving content disputes, embodied in policies like Weight, Fringe and Reliable Sources. Particularly on scientific issues, this policy (and the willingness of administrators to enforce it even on pages where they have no particular interest) makes it very difficult even for determined groups of editors to get a fringe viewpoint established.

Things are a bit trickier in areas like social science of course, where there’s much less agreement about what is fringe and what is not. Fans of the Austrian school of economics got into Wikipedia early, for example, and for a long while established ownership over a lot of their pet topics.


Martin Bento 01.17.09 at 2:43 am

I’m not sure that wikipedia is more open. Look at the format: both have an essay and comments on it. But at Kos, as most blogs, the comments are held to be valuable content in their own right, at least potentially, whereas at wiki the comments are pretty much just meta – their purpose is to direct the changing of the article. With a group authored article as the main content, consensus is very important: though differing points of view can be presented in a schematic fashion, still the article must strive for coherence. On kos, the article has a single author, so it is coherent if he is. However, various views are presented in the comments with no one being able to edit another’s. The Hillary supporters weren’t censored; they just felt shouted down. In an edit war, though, the majority will often end up actually censoring the minority.


Seth Finkelstein 01.17.09 at 7:14 am

[Abstain, on a risk/reward basis :-(]


The Raven 01.17.09 at 11:36 am

Wikipedia and Kos both have owners which control the code. Code is a mechanism of policy enforcement; as Lessig says, “Code is law.” I like Bento’s point very much. There is what I call a “grinding consensus” in lightly moderated on-line discussion groups. The ability to get a majority to mass-flame a minority (or even a dominant minority to mass-flame individual representatives of a majority) can completely stifle debate. A single troll, even, or a handful of trolls can stir up so many attacks as to make going against consensus in a lightly moderated group extremely unpleasant and that, even, without outside incentives like pay or organization affiliations.

I am very cynical of troll dominance of any on-line social space. It is intellectual brutality and cannot fail to produce poor intellectual results, as well as much pain. In areas where site owners simply allow trolls to operate without check, one gets the equivalent of street-gang domination of an urban space, and this is much of Wikipedia. In areas where site owners ally themselves with trolls, one gets the equivalent of bad cops, and that was Kos at its worst, and a major part of why I got sick of that community. (To forestall attacks I will say that I was a very minor participant on Kos and supported neither Clinton or Obama in the primaries. Happily, Kos was never very important to my intellectual life–I just stopped liking the neighborhood.)

Some of you may remember the NYT magazine article on with serious aggressive trolls–criminals and near-criminals. The little trolls are criminal wannabes and will turn an unmoderated space toxic, and if allied with site ownership or moderators, create intellectual tyrannies. I reject that–I don’t think it’s appropriate anywhere. In a site like Wikipedia it makes the site unreliable and contribution unpleasant, and a place not to contribute my knowledge. In a place like Kos it is party discipline and, as Tom S. points out, very much like non-internet political organization behavior.

Cynical, much? Well, what do you expect from a raven? Krawk!


Meh 01.17.09 at 1:43 pm

Dkos has lots of issues, but there’s a tendency amongst those who disagree with the politics of the majority there to consider it a “nasty internet neighbourhood.” I’m no fan of Dkos, but as someone who watched the decline of kuro5hin, I have to flag up that Dkos is nowhere near as bad as it gets.

The theory of K5 was largely “community moderation” and in my view it broke down at least in part because the community just wasn’t big enough. A determined troll who would create a number of identities on the site could easily disrupt the “community moderation” system. Ironically to some degree it was the lack of sites like kos and redstate that started the slide at K5. The first major breakdown in community spirit occured around the 2000 election arguments…

9 01.17.09 at 1:50 pm

I saw Henery Farrell on CSPAN this morning and I have a new model of political action, namely taking the political fight to the contributors of the Republiklan party. If taking care of the poor and the middle class and insuring social justice and product safety and worker safety and keeping unions strong require partisanship then I highly believe in partisanship when it means that the Republiklan party fails to stop the advance of progress.

I don’t go to Kos and facebook, and myspace much anymore etc. because they purposely discourage the wide transmission of a plan that I have to offer. Even some liberal sites they limit the number of friends you can add each day and messaging each day, use captcha to stop their own liberals on their own site from broadcasting a new message far and wide. I have gotten banned from Facebook more times than I care to mention. I wonder what would have happened to Paul Revere back then going from house to house, saying the British were coming would the people on the street call him a spammer?

Here’s my plan to accomplish the following:

“Making the Republiklan party do our bidding for us or their contributors lose money. A progressive agenda strategy.”

Get 3 people to make these phone calls and have those 3 people to get 3 other people to make these phone calls and so on. Accordingly pass this message onto 3 people and so on.

In 2008 Brown-Forman, the maker of Jack Daniels Whiskey and Southern Comfort gave Mitch McConnell money for his campaign.


Call GOP contributor and war contractor General Electric Corporation at 800 386 1215 or 203 373 2211 and tell the person who answers, to tell the GE CEO Jeffrey Immelt that you want him to get the president to end the war in Iraq and until that happens you will not buy any GE products and that you will tell your friends about this.

Call GOP contributor Rite Aid at 800 325 3737 and tell the person to get the CEO to get congress to enact HR 676 Single payer health care and repeal Medicare Part D and place the drug benefit in Medicare Part B covering 80% of the cost of drugs with no extra premiums, no extra deductibles, no means tests, no coverage gaps, and remove the means test for Medicare Part B and until that happens, you won’t buy ANYTHING from Rite Aid.

Call GOP contributor Wendy’s restaurants at 800 443 7266 and tell the person in public relations that you want their CEO to get congress to enact a $10/HR MIN. WAGE and the employee free choice act into law and until this happens you will not go to a Wendy’s Restaurant.

10 01.17.09 at 1:58 pm

Another aspect of blogs that peeves me comes from moderation. I don’t come to a site to curse or offer mlm or work at home schemes. I come to places like this to offer something new. Moderation occurs at MSNBC sites like Countdown with Keith Olbermann who I like to watch and respect but when even liberal sites moderate fellow liberals I wonder what’s going on. How about taking a chance that liberals like myself can offer something of value and not have to wait for some moderator to allow it? How about taking the chance and if something does appear extremely offensive then remove it after the fact? Oh and I hope someone finds the inventor of captcha and throws a pie in his or her face.


The Raven 01.17.09 at 11:47 pm

Meh writes, “I have to flag up that Dkos is nowhere near as bad as it gets.”

Sure; I used to hang out on the Usenet social groups, which have become places, as Teresa Nielsen Hayden memorably put it, much like Sunnydale, where the mouth of Hell opens and sooner or later you see one of every kind of monster. I am–except for the support of some particular candidates–fairly sympathetic with the politics of Dkos. Dkos became unpleasant, nonetheless.

Trollishness apparently is fractal–the bigger the community, the more the trolls. The trolls always outsmart the automated moderation systems. At the largest scale, this ultimately is an argument for governance.


Martin Bento 01.19.09 at 4:49 am

If a troll is someone who intends to disrupt a discussion by diverting it into bitter arguments, I don’t know that it makes sense to speak of an alliance of site owners and trolls. Tolerance for the sake or low or non moderation is one thing, but if the site owner is actively encouraging or reinforcing the trolls, wouldn’t that make the trolls part of the moderation system? I don’t think we’re worked out our intellectual categories very well yet.


The Raven 01.19.09 at 6:47 am

A good point, Bento. The key, however, is to aim the bitterness at particular targets, and to get otherwise good-willed netizens to participate. Trolls excel at this. It is not, I think, so different from the psychological dynamics of a lynching, though the immediate harm is usually less. I am aware of several people who attempted suicide after being targeted, however, so do not dismiss its potential for harm.


Gregory Kohs 01.19.09 at 2:19 pm

When I come across a thoughtful, “outside” assessment of Wikipedia, I get very excited and settle in for a good read. But, I have to confess, when I arrived on this article’s description of Jimmy Wales as “founder” of Wikipedia, I had to stop reading. Is it that this is just the “easy” way to describe him, and the authors simply don’t care how that might distort reality and subtly discredit Dr. Larry Sanger’s work? Or, even worse, is it that the authors actually have “bought in” to the revisionist fiction that Jimmy Wales has perpetuated, fashioning himself as the “sole founder” of Wikipedia? In either case, if the authors need more background material, I highly suggest:

Larry Sanger’s role in Wikipedia

Comments on this entry are closed.