We have a massive problem with close votes. There's no sense in placing blame — we've been on this path since before even I joined the site, and we all thought what we were doing was best for the site. However, that doesn't change the fact that there's real damage being done. And we need to fix it.
We closed 63.49% of the questions asked in 2022 according to the big list of close stats, which tracks closures across all SE sites. We've consistently been at the top of this chart year after year. Lately, when I look at close tools (10k only), I see that we are usually closing 80% or more of questions (but sometimes exceeding even 100%). When I look at questions that are being closed, I see so many questions that were not answered by a "general reference", some of which did include their "research" and got closed anyway (example: "Unrelentless" to mean "relentless"?). I see questions that were closed in spite of the expert answers that were posted to them, and questions that were closed before an expert answer could be posted. This isn't improving the internet. And I see what it's doing to the people here.
We've fostered a site where so many people who I respect greatly just don't like asking questions. This includes users who are native speakers, and even users who on top of that are also active on other Stack Exchange sites (and therefore know how this SE thing works), including moderators. Even users who have been around on this site for a really long time struggle to avoid close votes. That should be rare. It should also be rare for questions to be closed and reopened without significant edits. It shouldn't be the case where getting 1 or 2 close votes is an inevitability for almost every question. But all of these things are regular occurrences on this site and it causes many of our regulars (myself included) to spend way too much time on getting questions reopened instead of answering them. If you have an expert answer to an English question, your main obstacle to answering it should be the fact that you have to scroll all the way down to the bottom — you shouldn't have to beg and plead for it to be reopened or add "oh but these five dictionaries don't answer this question" or wait for a queue that might not even reopen the question at all without further intervention.
Now, I expect this may be hard for some of you to hear, and even harder to put into practice: We need to change.
The policy change
The policy is as follows: The "research" close reason should be used on questions which are answered in a reputable online dictionary. And no, you can't use the "other" close reason to get around this.
Dictionaries are great. There are so many comprehensive ones out there online from reputable sources. And they're simple to use. It makes sense to require our users to use one before asking. (For an idea of what dictionaries this covers, take a look at the main section under "General Dictionaries" and the section for "Learner's Dictionaries".)
How can I justify not allowing "Google" as a general reference? Relying on a search engine is the antithesis of our site. The best quality answer will not rise to the top of the search results, only the content with the best SEO keywords. Our experts cannot provide a better answer, or downvote the crap out there, or even leave comments on it. Nothing of value is guaranteed to stay on the internet either, as links go dead. And with one simple invention, suddenly all these problems can get worse quicker: Generative AI. Blogs will go from being one person's opinion to being no person's opinion and in record quantity. But answers here will be moderated for AI content; I'm making sure of it.
But how then do we differentiate ourselves from ELL? I think it's all in the answers. On ELL, most of the information in an answer will be stuff that a native speaker is likely to know already — it's just enough English to make it to the end of the day. But here, on English Language and Usage, an answer like that shouldn't be enough: We want enough depth to our answers that we can party all night. This is even possible with a question that looks easy on the surface, since an answer could draw from corpus evidence, for example.
A note on (up/down) voting
On other sites, bad questions that don't break any particular rule get downvoted instead of being closed. That works well, but before we become a site where everything is downvoted instead, please remember that short doesn't have to mean bad:
Questions that seed interesting answers should be rewarded. Questions that include their research should be rewarded (even if it wasn't perfect). Questions that cover something you don't think is covered anywhere else on the internet should be rewarded.
While moderators can't see your upvotes and downvotes, I hope that you'll keep an open mind and look out for what value could be under the surface.
The tech change
In order to reinforce the rules, there needs to be a change in the system. Years ago, close reasons got an upgrade. We never took advantage of this, but it's not too late now. We have 5 different "slots" to fill, according to Catija, most of which have a 500 character limit:
- Brief description (100 characters but should be just a few words) - this is the Bold part of the close reason that appears in the close vote UI when closers are voting to close the post.
- Usage guidance - this tells close voters when to use this close reason. It should clarify any edge cases and help voters feel certain they're choosing the correct reason
- Post notice close description - visible to all users. This is a general note about why the question was closed. It can include links to resources that explain the site's policy. It should always start and end with the same thing "This question was closed because it is... It is not currently accepting answers."
- Post owner guidance - this additional information appears in the post notice but only for the asker of the question. It should contain detailed information about how they can improve their post and may also include links to help here on meta or in the help center.
- Privileged user guidance - this additional information appears in the post notice but only for users with the close/reopen privilege. It is designed to help them know how to guide the asker in improving their question or inform them when the question should be reopened.
The "Usage guidance" will change
This is my vision:
Answered by a dictionary
This question has been answered by a free, publicly available dictionary. Make sure a link to that dictionary has been provided in the comments before closing.
This is a quick draft, so it could afford to be changed a little. I think this would be a sensible place to include specific guidance on when to migrate to ELL (if we had any; we don't).
The other parts of the close reason should also be updated, both in light of these changes and also because I think it would be better to provide askers different help than the rest of the community. I figure that's something we can work on drafting here.
Further discussions (TODO)
- Adding a tag warning to etymology and related tags pointing to Etymonline as a preliminary point of research.
- Adding guidance on the Ask Question page that actually explains how to ask a question here instead of telling people to just not ask questions here if they can avoid it.
- Switching over the other close reasons to the new format. We could be providing a lot more guidance to askers than we do. (We could mention thesauruses in our single-word-requests close reason.)
- How to write better answers. I can share how to do corpus research, and if you have another technique please consider posting your own discussion.