The longest-running hoax

Op-Ed

The longest-running hoax

Enwebb is the organizer of WikiProject Bats and founder of the Tree of Life Newsletter.

On August 7, WikiProject Palaeontology member Rextron discovered a suspicious taxon article, Mustelodon, which was created in November 2005. The article lacked references and the subsequent discussion on WikiProject Palaeontology found that the alleged type locality (where the fossil was first discovered) of Lago Nandarajo "near the northern border of Panama" was nonexistent. In fact, Panama does not even really have a northern border, as it is bounded along the north by the Caribbean Sea. No other publications or databases mentioned Mustelodon, save a fleeting mention in a 2019 book that presumably followed Wikipedia, Felines of the World.

The article also appeared in four other languages, Catalan, Spanish, Dutch, and Serbian. In Serbian Wikipedia, a note at the bottom of the page warned: "It is important to note here that there is no data on this genus in the official scientific literature, and all attached data on the genus Mustelodon on this page are taken from the English Wikipedia and are the only known data on this genus of mammals, so the validity of this genus is questionable."

This is not a Mustelodon.

Editors took action to alert our counterparts on other projects, and these versions were removed also. As the editor who reached out to Spanish and Catalan Wikipedia, it was somewhat challenging to navigate these mostly foreign languages (I have a limited grasp of Spanish). I doubted that the article had very many watchers, so I knew I had to find some WikiProjects where I could post a machine translation advising of the hoax, and asking that users follow local protocols to remove the article. I was surprised to find, however, that Catalan Wikipedia does not tag articles for WikiProjects on talk pages, meaning I had to fumble around to find what I needed (turns out that WikiProjects are Viquiprojectes in Catalan!) Mustelodon remains on Wikidata, where its "instance of" property was swapped from "taxon" to "fictional taxon".

How did this article have such a long lifespan? Early intervention is critical for removing hoaxes. A 2016 report found that a hoax article that survives its first day has an 18% chance of lasting a year.^[1] Additionally, hoax articles tend to have longer lifespans if they are in inconspicuous parts of Wikipedia, where they do not receive many views. Mustelodon was only viewed a couple times a day, on average.

Mustelodon survived a brush with death three years into its lifespan. The article was proposed for deletion in September 2008, with a deletion rationale of "No references given; cannot find any evidence in peer-reviewed journals that this alleged genus actually exists". Unfortunately, the proposed deletion was contested and the template removed, though the declining editor did not give a rationale. Upon its rediscovery in August 2020, Mustelodon was tagged for speedy deletion under CSD G3 as a "blatant hoax". This was challenged, and an Articles for Deletion discussion followed. On 12 August, the AfD was closed as a SNOW delete. WikiProject Palaeontology members ensured that any trace of it was scrubbed from legitimate articles. The fictional mammal was finally, truly extinct.

At the ripe old age of 14 years, 9 months, this is the longest-lived documented hoax on Wikipedia, topping the previous documented record of 14 years, 5 months, set by The Gates of Saturn, a fictitious television show, which was incidentally also discovered in August 2020. Based on the edit history of List of hoaxes on Wikipedia, new hoaxes are identified regularly at English Wikipedia. Dealing with this hoax and its fallout left me ruminating over some questions: How can we better identify hoaxes to keep them from reaching their tenth (or even fifteenth) birthdays? How can Wikipedia co-ordinate more readily across its different language versions once a hoax is discovered in one language? Does English Wikipedia harbor hoaxes that have been deleted elsewhere? Happy to hear your ideas.

References

^ Kumar, Srijan; West, Robert; Leskovec, Jure (April 2016). "Disinformation on the Web: Impact, Characteristics, and Detection of Wikipedia Hoaxes" (PDF). Proceedings of the 25th International World Wide Web Conference: 591–602. doi:10.1145/2872427.2883085.

← Previous "Op-Ed"

Next "Op-Ed" →

In this issue

30 August 2020 (all comments)

News and notes

In the media

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

It turns out that our article on work-life balance has 13 year old copyvios in it. I'm not surprised of this hoax lasting 14.5 years. MER-C 17:39, 30 August 2020 (UTC)[reply]
- Coppeboy continues on their merry way despite being a serial copy writer. Wikipeedia claims it isn't the place for "exhaustive logs of software updates," yet Windows 10 version history continues to be exactly that. And lets keep linking to World War Two in every single article in which it gets a passing mention (it's even better if you can link to the article multiple times!). But good luck finding those hoaxes. Oh, DrPizza! (talk) 12:00, 3 September 2020 (UTC)[reply]
"How can we better identify hoaxes?" - mercilessly delete unreferenced articles over 5 years old. Staszek Lem (talk) 18:51, 30 August 2020 (UTC)[reply]
- That's an idea, especially to emphasize verifiability, but we probably need a bit more than that to snuff out more elegant hoax articles. Some hoax articles can have skillfully crafted fake references that won't be caught unless someone thoroughly examines them. ComplexRational (talk) 19:26, 31 August 2020 (UTC)[reply]
  - Yes, I found one of those fairly skillful hoaxes last year. The hoax had already passed through AfC using fabricated references that looked like offline books. The only reason the hoax got caught was because the editor pressed their luck a little too far by making this edit, also with a fabricated reference (though I wish someone would publish Sexual Relations of the Mammals of Bolivia and Peru!). Enwebb (talk) 19:55, 31 August 2020 (UTC)[reply]
    - Some hoaxes have been caught when an erroneous ISBN was supplied for a fake reference. Hawkeye7 (discuss) 00:59, 1 September 2020 (UTC)[reply]
- Perhaps what we need is a WikiProject that systematically goes through all old articles and marks them off, similar to NPP, as meeting 2020 standards, or deals with them if they don't. It'd have to be done cautiously, though, since many of these pages won't have anyone around to defend them, so if a reviewer PRODs a notable topic because they were too lazy to check for sources, it could end up being deleted. {{u|Sdkb}} ^talk 23:06, 31 August 2020 (UTC)[reply]
Actually quick google search shows "real" images of a Mustelodon. Can we replace it in the article? Staszek Lem (talk) 19:03, 30 August 2020 (UTC)[reply]
Please be joking. That other-wiki entry was created in 2015 as a near identical copy of the WP hoax article (translated into Spanish, or maybe just lifted from one of the Spanish-language Wikipedia versions), and never showed any meaningful differences other than the image used. But even there, similarities abound in that both are completely devoid of any documentation regarding authenticity / provenance that would lend them credibility. Two sites perpetrating the same hoax, even independently, does not make it any less a hoax. -- FeRDNYC (talk) 22:22, 30 August 2020 (UTC
Of course I was joking, didin't you see the quotes around the word real? Not to say that every joke contains a grain of joke. In fact, the image from the copycat site with the caption "This is not..." would be helpful. Staszek Lem (talk) 16:22, 31 August 2020 (UTC)[reply]
I was trying to figure out where that picture came from and if we may already have it on Commons. It might be a miacid, probably Paroodectes, someone better at (extinct?) animal ID may be able to tell. ☆ Bri (talk) 16:34, 31 August 2020 (UTC)[reply]
Its Paroodectes feisti, see [1]. ----Jules (Mrjulesd) 22:33, 31 August 2020 (UTC)[reply]
At the ripe old age of 14 years, 9 months, this is longest-lived documented hoax on Wikipedia The "documented" qualifier is critical there, because it would be naive to imagine there aren't even more subtle/obscure hoaxes out there right now that still haven't been uncovered. -- FeRDNYC (talk) 22:29, 30 August 2020 (UTC)[reply]
- Breaking the record for oldest documented hoax twice in one month certainly indicates there there are plenty more where that came from. Enwebb (talk) 02:31, 31 August 2020 (UTC)[reply]
  - May I offer Molesworth Institute (now converted into an article about the hoax), which also lasted over 14 years and 5 months. In general though, we should also be aware that sourcing standards for new articles and edits to existing articles have become much stricter since 2005. It's more about actually applying them, i.e. editors should feel encouraged to liberally question and if necessary remove older unsourced (or poorly sourced) information. Regards, HaeB (talk) 02:47, 31 August 2020 (UTC)[reply]
    - Last year I found one about an alleged university secret society, supposedly formed in the nineteenth century, that had no evidence of existance beyond legend and heresay. It only takes one editor to write an article but it takes about a dozen to get one deleted. Blue Riband► 14:55, 31 August 2020 (UTC)[reply]
A long-running hoax with (say) 1 pageview per month isn't something that bothers me particularly (though, of course, the article should be removed). A badly written article with significant errors that gets (say) 100 pageviews per day - something like that would concern me a great deal more. -- John Broughton (♫♫) 23:59, 30 August 2020 (UTC)[reply]
I think this is absolutely right.—Brigade Piron (talk) 13:54, 7 September 2020 (UTC)[reply]
"Mustelodon remains on Wikidata, where its 'instance of' property was swapped from 'taxon' to 'fictional taxon.'"

When our robot overlords start slaughtering us, that's going to be one of the many similar reasons why they're furious at us. EllenCT (talk) 08:38, 31 August 2020 (UTC)[reply]

@EllenCT: #SRSLY. "Do you want to get a feral race of human stragglers eking out a meagre existence beneath the ruins of humanity's once thriving cities? Because that's how you get a feral race of human stragglers eking out a meagre existence beneath the ruins of humanity's once thriving cities!" -- FeRDNYC (talk) 13:13, 31 August 2020 (UTC)[reply]

Depending on how you view this, the human element of verifying and copyediting is always essential. I'm not surprised as the traffic of creating Wikipedia articles peaked sometime around 2005-2007 before quieting down. There have been some articles I made back in 2005-2006 that I was surprised weren't really touched and actually made me go back and improve on–or altogether PROD/CSD them realizing they don't meet guidelines. Doesn't hurt to review your oldest articles. – The Grid (talk) 14:46, 31 August 2020 (UTC)[reply]
It's so funny that we have an article that can now count as the "longest-running hoax" on Wikipedia! LOL! xD -ias postb□x+ 21:02, 31 August 2020 (UTC)[reply]
¿Qué tan bueno es tu español, OP? Interesting article nonetheless, wikihoaxes are a treat to read about -Gouleg🛋️ (Talk • Contribs) 21:05, 31 August 2020 (UTC)[reply]
It's a great shame that the PROD mechanism is so fragile, and that when the big red "CANCEL" button is pressed - the work of a moment - there is nothing to flag up the action for review. Of course PRODs can be mistaken or even malicious, but a review by a patroller would not be troubled by that. Chiswick Chap (talk) 08:21, 2 September 2020 (UTC)[reply]
Worse are probably hoaxes inserted into, often multiple, genuine articles, such as the exploits of Emile Campbell-Browne that made it’s way into several West Country and zoological articles, complete with references to obscure or fictitious off-line sources. Backed up by a series of sock-puppets that contested deletions of the hoax material as vandalism! Murgatroyd49 (talk) 09:07, 2 September 2020 (UTC)[reply]
The more often a hoax gets read the, the more likely it will be that it is discovered as a hoax. So presumably the main reason for long lasting hoaxes is because hardly anyone reads them. In such a case, it hardly matters. Paul August ☎ 10:23, 2 September 2020 (UTC)[reply]
There are many times more hoaxalicious factoids placed in articles than there are hoax articles. Is it fair to just remove statements tagged as needing references a year ago, or are we expected to spend our time searching for a reference to back up someone else's assertion? Edison (talk)
Love the word hoaxalicious! The answer is, it depends, if I think something looks reasonable but unreferenced I might research it. If it is fairly obviously untrue I would delete it as unsourced. Murgatroyd49 (talk) 15:32, 2 September 2020 (UTC)[reply]
I think inclusionism, WP:NEXISTS, and WP:AGF are to blame. Abolish those and many problems will go away. Chris Troutman (talk) 01:12, 8 September 2020 (UTC)[reply]
@Chris troutman: Well, aren't you a ray of sunshine? My view, personally, is that if those are "to blame", then the occasional hoax article is a small price to pay. Preventing hoaxes is not more important than preserving Wikipedia's core ideals. Your suggestion is the sort of thing that's always heard from draconian voices in society, the type who want to establish a police state to ensure there are no more jaywalkers. Overreaction much? Things like WP:AGF have gotten us this far, and it seems like it's mostly worked out alright. (I'm also going to go out on a limb and posit that there have likely been multiple occasions where you, too, have benefited from assumptions of good faith.) -- FeRDNYC (talk) 00:59, 14 September 2020 (UTC)[reply]

Want the latest Signpost delivered to your talk page each month?

Home

About