Sunday, January 22, 2012

Anonymity and governance

Commenter 'Belette' has questioned my attempt to connect anonymity with the failure of democracy.  I thought this would be obvious, but fair enough: this is a logic blog and the argument I gave in the previous post was far from demonstrable, being full of all sorts of missing assumptions, handwaving, enthymeme and absent premisses. Let's try and tidy it up and take it from the top with the following pseudo-syllogism:

Anonymity leads to conflict of interest
Conflict of interest leads to bad governance
Ergo, anonymity leads to bad governance

Even this is not deductively valid, but it will do for now.  I will leave the minor (COI leads to bad governance) unless anyone objects, for while not self-evident, it can easily be made so by supplying some missing assumptions. Note also that I have substituted the idea of governance for that of 'democracy'.  Democracy does not always result in good governance, and it is governance we are interested in.  I also use 'governance' rather than 'government'.  The latter almost always implies the action of a state, whereas it is governance in the widest sense that I am concerned with.

So let's focus on the major premiss: does anonymity lead to conflict of interest? Certainly.  A conflict of interest arises when an individual under one description F1 has an interest which is opposed to the the interest of that individual under the description F2.  For example, let F1 be 'director of charity X' and let F2 be 'commercial supplier of provisions and services to charity X'.  Clearly the individual under the first description has an interest and a duty to procure the same service at the lowest possible price, whereas, under the second description, he has an interest in supplying them at the highest possible price. Allowing the directory of a charity to  be the same person as its chief commercial supplier is therefore a conflict of interest.  But if we cannot identify the individual satisfying each description, it will be practically impossible to prevent such a conflict.  Let F1 be satisified by John Smith, a publicly identifiable individual. Let F2 be satisfied by some individual under a fake user name.  Unless we can connect the public name with the made up one, we will never know if there is a conflict of interest.

One recent example of such a conflict on Wikipedia was when the PR agency Bell Pottinger was found to be editing Wikipedia under the fake user name 'Biggleswiki'.  Another less recent example was when London Labour councillor David Boothroyd was not just an editor of Wikipedia under the pseudonym 'Sam Blacketer', but was a member of its  important 'Arbitration Committee', probably the most senior decision-making body on the project. Paul Williams, then director at Wikimedia UK, said: "Sock-puppeting is a very serious offence for anybody. But for someone on the Arbitration Committee it is even more so. It can result in a lifetime ban. The problem with Wikipedia is that you can hide behind user names, but there is an expectation that you don't write for self-interest. In this case there is a conflict of interest."  There you go.

There are other instances of politicians anonymously acquiring important positions in Wikipedia, and who don't want to be publicly identified, but I will leave those for now, as they are for use in the book I am writing on Wikipedia.  I will leave you with an example from last weeks 'Blackout vote'.  The full vote for the blackout was on this page.  Over 700 members of the Wikipedia 'community' voted in support of an action that affected hundreds of millions of people across the world.  Closer inspection reveals that many members of this community may have voted twice, or even more.  About 150 of the support votes were from accounts opened specifically for voting, or re-used. There were a number of obvious 'sleeper' accounts, opened some time ago, almost certainly by existing accounts who wanted to vote twice or more. Typical of the first kind is Korney654*, who only edits twice, the first time to the article on Oboe, displaying knowledge of Wikipedia that is untypical of a first time user, the second time to vote. Typical of the second kind is Phauxcamus* who opened his or her account in 2006, made a handful of edits since then, and now votes for the blackout.  It is highly probable that all these votes were second or third votes from more established Wikipedia account names.  There is no more evident and absurd form of conflict of interest than pretending not to be yourself under a different, anonymous description.

Belette will probably argue that such conflict of interest only arises in the real world, because it has dishonest motives, whereas Wikipedians only act in the best motives. There is no reply to that except to laugh or shrug, with a wink at the gallery.

*With apologies to any editor who was really acting in good faith.  The point is, it is difficult to know under a system that almost guarantees anonymity in this way.


William M. Connolley said...

I'll look at your examples in a bit; thanks for posting them.

As a holding action: your syllogism is inadequate (i.e., just demonstrating, even if you could, that anon -> bad governance isn't good enough). Bad governance is everywhere; most countries in the world are mis-governed, and whilst you could attribute some of that to a sort-of-not-really-anon civil service, rather a lot of it is attributable to known politicians being influenced to the public ill (only slightly o/t: have you ever read ?)

Edward Ockham said...

>>As a holding action: your syllogism is inadequate

I'm not sure 'inadequate' is in the lexicon of logic or not. I've already said it is not formally valid. What I think you mean, however, is that proving that anonymity leads to bad governance is not what I want to prove.

I reply, proving that anonymity leads to bad governance is exactly what I want to prove.

Edward Ockham said...

It is a long time since I read any science fiction. I liked Jack Vance a lot when I was 16. I believe the eponymous Wikipedia editor took his handle from that story, and that editor is also a strong proponent of anonymity.

The question is whether this kind of anonymity leads to conflict of interest. I would say the interest of a ruler includes the possibility of being deposed by his or her people. An anonymous ruler cannot be thus deposed, thus, conflict of interest.

William M. Connolley said...

> Bell Pottinger

Yes, but that ended pretty quickly.

> David Boothroyd... Sam Blacketer

Not familiar with that. Since I endure endless allegations of COI, I'm not terribly favourable to the idea. What evil did he commit?

> Typical of the first kind is Korney654

Not a good example. He isn't on the final tally page. Any number of SPA's were tagged in the voting (,, and their votes are ignored. Everyone knows this happens, and it doesn't matter.

> Phauxcamus

Can't see him in the final list, either.

> Belette will probably argue...

Nope, or at least not yet, cos I don't need to.

William M. Connolley said...

> proving that anonymity leads to bad governance is exactly what I want to prove

What I was trying to say was that bad governance is everywhere. You could just as easily prove that people lead to bad governance. What you need to prove (or rather, demonstrate, since I can see no way to prove it) is that anon leads to worse governance than known-identity.

For example:

1. known identity of leaders leads to concentrated lobbying
2. lobbying leads to decisions that favour the lobbyer and not the public at large
3. therefore, known-identity leads to bad governance.

Edward Ockham said...

>>Any number of SPA's were tagged in the voting

Thanks for pointing that out, I'll look at the tally page.

Edward Ockham said...

>>anon leads to worse governance than known-identity.

OK let's go for that case next.

Edward Ockham said...

>>Yes, but that [Bell Pottinger] ended pretty quickly.

So? The point was to demonstrate how anonymity implies conflict of interest, which it does. As soon as the anonymity ended, the coi was spotted. And did it end pretty quickly? I thought they were editing for a long time? In any case, it wasn't Wikipedia systems and controls which ended it, but an external investigation. BP admitted to the undercover investigators that they were editing Wikipedia anonymously.

Edward Ockham said...

>>Can't see [Phauxcamus] in the final list, either.

He is Phaux. As for the other one, a number of the accounts gave different signatures from the actual account name, which is yet another feature of the system which leads to bad governance.

William M. Connolley said...

> So? The point was to demonstrate how anonymity implies conflict of interest, which it does

No, the point was to demonstrate that anon -> bad governance. BP isn't a convincing example of that.

> As soon as the anonymity ended, the coi was spotted

I'm not convinced. Do yo have a good timeline for it? The WP article is ambiguous.

> He is Phaux

Good point. And Korney is Josh. But anyone with a redlink talkpage isn't going to get any notice.

Edward Ockham said...

>>But anyone with a redlink talkpage isn't going to get any notice.

Are you saying that there was an ‘official’ count of the votes and that the redlink votes weren’t counted? Where is this official count, please? What about redlink user pages? What about the bluelink socks who bothered to write a few words on their talk and user pages? What about redlinks who have had a long editing history (yes, they exist).

>>I'm not convinced. Do yo have a good timeline for it? The WP article is ambiguous.

Biggleswiki began editing in November 2010. So, editing for a year before he was discovered by accident. The socking was crude and unsophisticated, too. For my research on the book I have been talking to much more polished operators, who have been working for years without being uncovered.

The discovery of BP was the result of an undercover investigation by the Bureau of Investigative Journalism. Reporters from the Bureau posed as agents for various clients, including the government of Uzbekistan, which has a poor human rights record, and its the cotton industry, which uses child labour, to investigate the ethical standards of the agency. Ten firms were contacted, of which five were prepared to work for the 'Uzbekistan' agents. The discussions were secretly recorded. The Wikipedia sockfarm was mentioned, and the Bureau published its report, which was then covered by the Independent. The rest is history.

>>No, the point was to demonstrate that anon -> bad governance. BP isn't a convincing example of that.

It’s a perfect example of governance failure, if we mean by that a failure of systems and controls.