The Moonspeaker:
Where Some Ideas Are Stranger Than Others...
Many Times Wikipedia Should Not Be the First Search Result (2026-07-06)
Lithograph of the aurora borealis from William MacKenzie's 1891 National Encyclopedia, via rawpixel's collection of free high quality scans and openverse. The rawpixel's scan is licensed under CC Attribution (CC BY 4.0). This version has been slightly reduced in size.
The most famous wiki project, wikipedia, is one of those frustrating products of the attempt to create "web 2.0," a live-editable by anyone web. I appreciate the allure of "web 2.0" even though as a wholesale approach it is a bad idea, which is indeed disappointing. Security and content moderation are far from trivial challenges to address. Still, wikipedia has certainly shown itself to be a genuine use-case for the idea. Alas, in light of how wikipedia is actually owned and run, the various contributors cannot resolve the growing problem of articles written as fullbore propaganda, or at minimum presenting propaganda completely uncritically. (For clarity, I take "propaganda" to mean material created and presented with the deliberate intention to distract and deceive, not material a person may disagree with on grounds of politics or taste.) Nevertheless, it remains true that wikipedia is often a solid starting point for looking up STEM-related material. For non-STEM topics at least the reference list and footnotes can be fruitful, as they can point to useful articles and books or reveal effective terms for search strings. Those are solid achievements in themselves, and as the frequent appearance of wikimedia commons items as illustrations for thoughtpieces reflects, the media collection, much of public domain or creative commons licensed is too. All the same, it is truly unfortunate so many search engines insist on prioritizing wikipedia articles in search results, and this would be true even if there were not issues with propaganda and attendant editing wars, which are not always revealed by a helpful flag noting the content is contested.
The current intrusive ranking of wikipedia results is probably partly accidental. Since wikipedia is subtitled "The Free Encyclopedia" and has long had a permissive reuse license, including today's Creative Commons Attribution-ShareAlike 4.0, for at least several years reskinned clones of wikipedia were rife across the web. To keep the resulting mindless repetition out of search results, developers set about ranking wikipedia high and its clones low, a sensible approach. Then for the basic sorts of queries hard copy encyclopaedias traditionally covered, it did make sense to prioritize wikipedia results once it had won credibility. Wikipedia is almost unsurpassed for four W's type questions: who, what, where, when, so long as they are not part of a contentious recent or current event. Like its hard copy parent, wikipedia is quite weak on "why" questions, since it rejects original research, and rather shaky on "how" questions outside of STEM subjects. And fair enough, in my experience wikipedia does not tend to show up in search results for queries intended to explore the whys and hows of specific events, individual actions, or processes. But it is not this type of query, which is typically based around a phrase or even a naturally worded question that is the problem here.
UPDATE 2025-07-07 - This is one of those rougher thoughtpieces with loots of wrestling in it, in part because of missing information that writing it made me appreciate was missing on mulling it over after writing it. This realization was also helped along by Bob Leggitt at BackLit, who provided the missing piece and reminded me that in this decaying hypercapitalist world, it is critical to be cynical about these things. He wrote in a 10 july 2022 post,
The Truth About Creative Commons:
"We've already seen search and tech brands discriminating against robust copyright in favour of CC. For example, Google relentlessly prioritises the Creative Commons site Wikipedia over the more specialised source sites from which it cribbed its information. Indeed, Google has steadily buried those more specialised source sites deeper and deeper in its results, leaving Wikipedia with a virtual monopoly on most entry-point knowledge searches."
Leggitt explains a lot more about wikipedia's unsavoury and frankly evil habits when it comes to properly crediting source documents and authors, in Death by MMC: How Wikipedia Strangled The Information Highway. Learning more about how wikipedia actually works is part of what is leading me to reconsider where I find illustrations by others for use on The Moonspeaker.
The easiest way to experience intrusive wikipedia results taking up much of the first page of link results and often a summary box as well, is to try searching on a one to three words. This is not surprising in principle, it is reasonable to presume such a simple query is intended to pull up basic information. Probably there is no way to avoid this apart from adding more detail to the query and applying the search engine's options for excluding results from specific web domains. Much more annoying however, is when too-clever search engine algorithms extract key words from a query and end up identifying wikipedia pages as high-relevance results. If we are researching a question in part on the web, and no longer at the stage of needing basic definitions, this is an extremely frustrating outcome. How could this be, a reader might wonder, if we can go ahead and apply search engine options to exclude specific web domains? It can indeed be so alas, because alarmingly enough, search engines do not necessarily provide such options anymore, or even if they do their quality has collapsed due to competing factors in the business models of the corporations that run them. The flagship example is google with its hideously mutilated search to go with its relentlessly intrusive profiling and tracking of "browsers" which of course really means "person browsing the web." On the relatively more privacy respecting and still fairly well known side, duckduckgo does have means to exclude domains, but they don't seem to work consistently. And then there are not so well known, but promising "google-style" search engines like mojeek.
So, for deeper dives and controversial topics, wikipedia should not be the first search result. To help work around wikipedia being a bit too favoured in the search engine algorithm, it is critically important for the algorithm in question to include a means to exclude site domains. There is probably no way to adjust for controversial topics to push down wikipedia though, since it is not so easy to mechanically identify them or fairly adjust for them in real time via human participation. That does suggest an intriguing question though: do people often consult an encyclopaedia for information on disputed topics, meaning about matters more complex than "is a platypus a mammal?" or "which anthropologist wrote about the kula ring?"
Previous Thought Pieces
|
