Wikipedia talk:Five pillars

Proposal to remove the fifth pillar

 You are invited to join the discussion at Wikipedia talk:Ignore all rules § Should IAR be overturned?. WhatamIdoing (talk) 05:03, 23 October 2025 (UTC).

Misuse of the first pillar?

I have a vested interest in this, so I'll come clean. At a current deletion debate for List of dates for Easter, an editor has used WP:ALMANAC as justification for keeping the list. This links to WP:5P1, stating that Wikipedia combines features of various things, including almanacs. Clicking the link to almanac gives us the following definition

"It includes information like weather forecasts, farmers' planting dates, tide tables, and other tabular data often arranged according to the calendar."

Clearly Wikipedia is not here to provide weather forecasts. We don't provide tide tables. Is it genuinely intended that Wikipedia should contain such information? And if it isn't, how can 5P1 be reworded to describe the true scope of Wikipedia? Do we provide calendar things provided they can fit in a reasonable page-size (Easter is global, so it can be tabulated reasonably concisely; tides and weather are local, so they're clearly impossible)? What almanac-information actually fits in an online encyclopaedia? Or was the reference to almanacs a passing comment that is perhaps not so wise? Does this need clarification? Elemimele (talk) 15:28, 1 January 2026 (UTC)

• The Five Pillars doesn't stand alone, it is a guiding principle only. What is and isn't included is a function of consensus, not the contents of this page. See Wikipedia:What Wikipedia is not, which is an actual policy, decided by consensus. This is the core policy on what is allowed and not allowed. Dennis Brown - 2¢ 01:28, 2 January 2026 (UTC)
• WP:ALMANACWP:ALMANAC was originally a redirect to a stub essay, Wikipedia:Almanac. WP:5P1 talks about combining many features—it is not saying Wikipedia is an almanac. Dennis Brown has provided the correct advice above. Johnuniq (talk) 02:43, 2 January 2026 (UTC)

  Thanks both; that's the trouble with WP:shortcuts; they get thrown around in places like AfD as though they were the law. Elemimele (talk) 12:22, 5 January 2026 (UTC)

Neutrality of the word ‘mercilessly’ in guideline wording

I noticed the phrase in the guideline:

“…any contributions can and may be mercilessly edited and redistributed.”

I’m wondering if “mercilessly” is too strong, inhuman, or emotive for a neutral, encyclopedic tone. Would alternatives like “openly” or “without restriction,” or any other suitable synonym, better convey the intended meaning, or would simply omitting the word "mercilessly" be a viable option? Q3YZFU5RM4X (talk) 07:04, 27 March 2026 (UTC)

Wikipedia:Neutral point of view concerns text in articles. The use of "mercilessly" is intentionally strong to make a point. It could be argued that the term might provoke a reaction that makes readers disregard the idea being conveyed but the word has been in place for a long time and is likely to stay. Johnuniq (talk) 09:16, 27 March 2026 (UTC)

I’m in favor of strong wording where needed, but not wording that feels unnecessarily harsh or impersonal. We’re guiding, writing, and editing primarily for human readers, and the tone should reflect that. Does “mercilessly” really feel appropriate in a human-centered encyclopedic setting? It also seems slightly at odds with the spirit of civility. Could we consider other strong but more civil/kind alternatives? Q3YZFU5RM4X (talk) 10:14, 27 March 2026 (UTC)

The aim of WIkipedia is to produce a free encyclopedia. It is not a social support club where things should be left in so they don't offend the editor who put them in. It isn't human centered. However if you can think of another way of saying it that might work better with people then there's no harm in that. NadVolum (talk) 14:31, 27 March 2026 (UTC)

I agree that Wikipedia’s goal is a free encyclopedia and content isn’t owned. I’m just not sure “mercilessly” is needed, it can feel overly strong or loaded and might give the wrong impression to new human editors. Maybe a more balanced word would work. Does “mercilessly” fit Wikipedia’s usual tone and WP:NICE/POLITE/Civility? Q3YZFU5RM4X (talk) 17:08, 27 March 2026 (UTC)

"Mercilessly" has been there since 2009 and the only substantial discussion I found was in 2013 and ended without resolution (see Archive 8). Personally I find it to be useless hyperbole. As well as the word "merciless" having an unnecessary connotation of cruelty, the sentence also appears to say that there are no limits to editing, whereas in fact we impose multiple limits (RS, NOR, etc etc). I wonder what would be lost if we just deleted the word. Zero^talk 02:22, 28 March 2026 (UTC)

I support this proposal. - Butwhatdoiknow (talk) 04:43, 28 March 2026 (UTC)

Small addition regarding AI?

Related to an going talk at Wikipedia talk:Editing policy#AI information.

Considering the usage of AI (culminated in many policies and guidelines) is now basically discourage from article creation to editing, to generating images, sources and comments. Should we mention something here with a link.... something very simple. Like... Wikipedia is a human-driven endeavor, created, edited, and moderated without the assistance of artificial intelligence. Moxy🍁 03:54, 12 April 2026 (UTC)

I am opposed to AI editing (impossible to deal with slop) but I don't think anything should be added here. The five pillars don't address the fact that competence is required, or that we don't want spam or other bad stuff. Also, AI users won't read this page and they should be asked to study the new policy, not 5P. Johnuniq (talk) 04:19, 12 April 2026 (UTC)

I think you have a point okay. AI seems to be posing a very basic threat to Wikipedia. Wikipedia is one of the best sources for AI and if AI generated stuff gets fed back in a loop the whole business of having reliable data will go swirling down a plughole. It is important that Wikipedia be a firebreak to stop that sort of thing. Like to phrase a sentence which directs off to that policy, probably for the second pillar? NadVolum (talk) 14:22, 12 April 2026 (UTC)

While I personally agree with your sentiments (Wikipedia's strength in this nascent AI age is that we're written by humans and double checked by humans), there isn't enough anti-AI consensus yet for this to have a chance of becoming a pillar in an RFC. It was only recently that folks changed their minds enough to make WP:LLM a guideline. There's also the argument that anti-AI stuff does not have equal weight to the existing 5 pillars. –Novem Linguae (talk) 21:06, 12 April 2026 (UTC)

Another point is that it's not a pillar of Wikipedia that AI is not permitted. Some hypothetical future AI might edit constructively without hallucination. The legitimate concerns about AI are about the current crop of LLMs with their known defects. We don't know what the community would think about the coming-in-five-years good enough AI. Johnuniq (talk) 06:17, 13 April 2026 (UTC)

Any future AI will need to work from some approximation of ground truth if it is to work without hallucination. They'll get to the stage where they can fool most people but it will still lead to slop if they feed in a loop on what they produce - only we won't have any control anymore because too many editors will agree with the slop. What you are talking about is just hope and five years is enough to produce something which seems better but would still be doomed. There really is no way around the problem except having a break in important loops and the AI creators have already decided to use Wikipedia as a major source. We can use AI for checking purposes but the output will need to be severly restricted to not cause overload and people just copying over what they say. This is not a problem we can let loose on Wikipedia now and fix later, it is one that needs safety now and if your wonderful future turns out okay then it can be loosened later. It is bad enough we have people believing all sorts of stupid things now but we can cope with that, it is a known problem and part of our lives. In fact we have this already for people where they send things round in loops and what happens in the form of for example conspiracy theories. Believing AI is going to be a wondrous vision of clear and moral thought is wishful fantasy - with hallucinations it is going to be a faster more efficient and quite inhuman load of the same stuff humans are burdened with. NadVolum (talk) 11:54, 13 April 2026 (UTC)

By the way Wikipedia:Writing articles with large language models is a Wikipedia content guideline. Use of LLM's for generating content is prohibited. NadVolum (talk) 12:33, 13 April 2026 (UTC)

I totally agree—general intelligence won't be available in any of our life times. However, people will point out that LLMs are astonishingly good and AI might become good enough to edit. Whether that is true is not relevant for the proposed addition to 5P. Johnuniq (talk) 00:19, 14 April 2026 (UTC)

You're missing the point. They can already do edits that look reasonable. However they should not be allowed to. The total effect would stop Wikipedia being a reliable encyclopedia. And greater power brings greater problems, just think for instance what would happen if someone talks about removing inconsistencies in Wikipedia for instance - not too bad if an individual is concerned but what happens when every other AI then tries for more consistency and you loop around five time. Or ten times. Or a million? You're talking about systems that can analyse code that is supposed to be secure and find zero day bugs in it that have been there for years. NadVolum (talk) 09:17, 14 April 2026 (UTC)

For the record, there is some misunderstanding because I understand everything you have said (before coming here). I'm just saying that it's still not a pillar. You may have a different view but it's not because I'm missing the point. Johnuniq (talk) 10:52, 14 April 2026 (UTC)

I agree with Johnuniq. A pillar is something more basic than policy. The uses and abuses of AI is a matter for policy, but it is not a pillar. Zero^talk 12:03, 14 April 2026 (UTC)

The intention was to introduce a new point within one of the existing pillars, not to introduce a new pillar. Multiple pillars would be very badly affected by using LLM content, I see no problem with warning a new editor that use of LLMs is seen as a dangerous act which would pervert WIkipedias aims and therefore is prohibited. I was considering the neutral point of view pillar but now think the one about treating with respect and civility may be the most appropriate. LLMs do not actually treat people with respect and civility though they may be programmed to have a veneer like that. NadVolum (talk) 12:30, 14 April 2026 (UTC)

Or the first pillar 'Wikipedia is an encyclopedia' might be best since usng LLMs would directly act against that. NadVolum (talk) 18:01, 14 April 2026 (UTC)

Too many people think that WP:5P1 means that WP is an almanac/gazetteer or inclusion in an almanac/gazetteer should mean inclusion on WP

I'm not sure how to fix this given that my last effort to fix it basically didn't work, but for reference this is what very typical, representative, actual gazetteers and almanacs actually look like:

As you can see, this is exactly the kind of content that WP is resolutely WP:NOT.

Typically the response is "well, WP:5P1 doesn't actually say this", which I agree, it does not. Instead it only means we can have features of gazetteers/almanacs. But lots of people say it does say this, and their misunderstanding is driven by the fact that gazetteers and almanacs are ranked equally with encyclopaedias in WP:5P1, and Wikipedia clearly is an encyclopaedia so why would it not be these other things?

The other response is typically along the lines of "WP:5P isn't as important as they are making out". And, again, I agree with this. WP:5P is not some "constitution" for Wikipedia (the idea that we need such a document is anyway very American). But if it isn't then we shouldn't be pinging it to every new editor as some kind of essential code that they should follow.FOARP (talk) 08:59, 11 June 2026 (UTC)

Are you talking about this November 2021 edit? Everything is misunderstood—the trick is to make a proposal for an improvement. Johnuniq (talk) 09:53, 11 June 2026 (UTC)

As far as I'm concerned amending "gazetteers and almanacs" to "other reference works" was an improvement for the reasons given then (i.e., there's other reference-works that could be called out such as compendiums/chronicles/genealogies, so better to use a more encompassing term). Most of the discussion ended up being about process though. FOARP (talk) 10:19, 11 June 2026 (UTC)

I agree, almanacs and gazetters gives altogether the wrong impression. Other reference works is much better or as you suggest some other specific ones that aren't so likely to go out of date so quickly. NadVolum (talk) 15:06, 11 June 2026 (UTC)

What if we linked that statement to the actual P&Gs like all of the other pillars? We don't leave "Wikipedia is neutral" open-ended because that would obviously be misused to treat all opinions equally, so why would "Wikipedia has elements of a gazetteer" be any different?

It's not a guideline, but WP:GAZETTEER is actually a pretty well balanced essay that could fill that role with a bit of tweaking. I think most editors would be comfortable with its explanation of what our "gazetteer function" actually means. –dlthewave ☎ 16:58, 11 June 2026 (UTC)

IMO, as evidenced by it's prominence, degree of consensus, the very high bar for making changes to it and other factors, 5P is a document that is above quidelines and policies, albeit vague (like a Constitution), not an essay. We just fail to have a suitable name/status/category for such things. North8000 (talk) 18:01, 11 June 2026 (UTC)

I have a different point of view, which is that WP:5P was never intended to be anything more than what it says it is: a non-binding summary of some principles. In the case of the addition of gazetteers, this was an undiscussed 2008 edit by Unitedstatesian, so not even summarising of a principle. FOARP (talk) 09:26, 12 June 2026 (UTC)

The first encyclopedia that I read as a child was The Children's Encyclopædia and that would look equally quaint and unfamiliar compared to Wikipedia. The essential feature of a general encyclopedia is that it is a comprehensive compendium which covers everything worth knowing. Alamanacs and gazetteers are a bit more specialised but their facts would certainly be expected in a comprehensive reference work like Wikipedia. The current wording is fine. Andrew🐉(talk) 21:24, 11 June 2026 (UTC)

You would expect us to identify all the bends in a river? You would expect us to give all the dates of markets, details of tides and so forth? FOARP (talk) 05:26, 12 June 2026 (UTC)

I was looking at the Tideway article recently when working on the Boat Race 2026. This details the tides and topography of the tidal Thames and I'd expect some detail. The main difference is that we should have a historical perspective whereas an almanac tends to be forward-looking and focused on a single year. Andrew🐉(talk) 12:02, 12 June 2026 (UTC)

For better or for worse, I think the effect is for "almanac" to give a tilt in favor of list articles and "gazetteer" to give a tilt in favor of geo articles compared to being strictly "enclyclopedias", the third member of that phrase. Sincerely, North8000 (talk) 16:50, 11 June 2026 (UTC)