Stack Exchange has three standard closure reasons, “details”, “clarity” and “opinion. We can add three site-specific reasons (we do so), which I refer to as “language-specific”, “text” and “trees” reasons. These reasons have been discussed over the years, and opinions on these reasons, their justification, explanation, and implementation have been variously praised and excoriated. The reason for this post is to identify and clarify those three most-important desiderata that should be our site-specific closure reasons, and their corresponding expression as rules as to what we wish to be clearly declared “off-topic”.
Here are reasons why there should be specific off-topic questions and clearly-stated descriptions of those subjects. The “null hypothesis” (the result of having no specific guidance) is that every person with the VTC privilege can rightly feel free to use any arbitrary reason to close. If we assume that all VTC-level users have “correct” views of the world and use “correct” reasoning in voting to close, there will not be a problem of people VTCing questions about Minimalist syntax on the grounds that they hate Minimalist syntax, since hatred is not a valid reason. However, I do not assume that we all agree on the scope of linguistics (is text-processing type computational linguistics really linguistics?), even if we are all fair-minded. I conclude that from the advanced-user’s perspective, it is desirable to have a clear declaration as to what topical standards a question should adhere to. I also accept without further discussion that LSE should have topical limits. I also accept that it is quite possible that a user will VTC for an ad hoc reason.
The primary utility of clearly-stated close reasons is to educate new users. It is inevitable that some new users will just ignore the help center. Warnings do not prevent the frequent OT question “Please do my intro Ling tree homeworks”. This is an unsoluble problem. What can be solved is the problem that new users cannot easily understand what is off topic. There are two manifestations of this problem. The first is that the text of the three site-specific VTC reasons is not available to new users. The text of don't ask, which is the obvious place to look for a description of what not to ask, contains nothing at all about site-specific negative desiderata. Since the content of that page is not rigidly fixed across SE sites, part of a solution would be to make the VTC reasons more-available in the obvious place. My first concrete proposal is that the site-specific text in the VTC pop-up should be copied into “don’t ask”. My second proposal relates to the description of “on topic”, in the “and not about” subsection.
There should be exactly three reasons stated there, which correspond exactly to the three pop-up reasons on the VTC box. Right now there are six reasons in the help-center list. First, we should eliminate the specific reasons “fictional/constructed/planned languages” and “advice or help on learning” and instead say separately (not as a reason to close: it’s a reason to move) that such questions will be migrated to Conlang or Language Learning SE. The reason “anything not related to/treated by Linguistics” is not useful information for the user, given the preceding description of what linguistics is. On the occasion that someone asks for a recipe on Ling SE, you can always opt for the “other reason” box, filling in “is not about linguistics”.
Next is a restatement of the core of the reasons “language-specific”, “text” and “trees”. The easiest is the “text” reason: LSE is not a translation service, and linguistics is not “the learning of lots of languages”. I see nothing that needs to be changed in the help center text or the VTC pop-up. However, the redundant line “Translation requests” should be deleted (look at the “identification” reason). Second-easiest is “please draw my tree”. The primary reason for this reason is that LSE is also not a homework-answering service, nor are we an essay-writing service. The main defect of the specific reason is that it only excludes a subclass of syntax homeworks. Here is the proposed revision of the “trees” reason:
Homework questions are off-topic here. If you have difficulties solving a homework question, ask your teacher who will know the specific approach that you are supposed to use. Questions should be asked to get information about linguistics that could be useful to anyone, and should not be request to assist a specific person in a class.
The first sentence goes in the VTC pop-up box.
There is an often-mentioned consideration that there is no single theory of syntax therefore there is no single correct tree, but that reasoning should also preclude asking a question about syntax that does not narrowly specify the framework, likewise questions about phonology, phonetics, semantics, morphology, historical linguistics and so on. Trees are not special. Saying "trees are off topic" does not communicate the right concept.
Finally we have “language-specific”, the most controversial and unclear reason. See this discussion that focuses on that reason. There is not even an attempt to explain this close-reason on the don’t-ask page, and no indication that it is a reason to close. This has become a overly-broad hodgepodge reason. It is not totally surprising, and somewhat distressing, to learn, as revealed in a data query by Lemontree that language-specific is by far the most common of all reasons for closing. My solution and proposal is to separate out some of the already-existing reasons, which would allow us to clarify the remaining core.
When one talks about “language-specific grammar and usage”, that is, as far as I can tell, about normative stylistic rules, like the rule that prepositions are things that you shouldn’t end sentences with. Questions asking “Is X proper English” are OT, and I support a rule to the effect that unscientific opinions about “proper” language have no place in LSE. There already is a pan-SE reason: opinion. It is a matter of personal opinion and not scientific fact whether English sentences “should” never end with prepositions.
A second class of justifications for “language-specific” as a close reason is to encourage closing some set of questions about a single language (as opposed to a language family, or all languages). The (small) portion of this reason that I agree with is when it is a language-learner’s question of the type “What is the rule (not the linguistic rule) for conjugating /kwn/ in Classical Arabic?”. There is a related linguistic question that should not be closed: “What is the rule deleting intervocalic glides in Classical Arabic?”. Language-learner question belong on a language-learner SE, not linguistics. Not all questions that are about a single language are language-learner questions.
A third class is the class of word-history questions, like “Where did we get the word ‘dog’?”, “Why are ‘language’ and ‘tongue’ cognate”. These fall under the “etymology” penumbra which has been recently discussed without clear resolution. I do not have a bottom-line rewrite proposal for “language specific” since I cannot discern what problematic fact this reason actually refers to. However, one other factor that has been cited in these discussion is that all questions should in principle be “general”, and not specific to the interests of a single person. We could, for example, have a rule to the effect that a question should reasonably be applicable to the interests of more than one person, but the patently obvious problem with that reason is that it calls for a personal opinion and is not a matter of objective fact (applying the goose and gander sauce rule).
The Ling tools 30 day synopsis gives enough information that one can extract (from actually-closed questions) an idea of what questions are closed because they are “language-specific”. At the moment, we have closed the following at least partially on the “language-specific” grounds. I divide them into “purely lg-spec” and “partially lg-spec”, and give a brief description of what the question is about. A-D are listed as purely “lg-specific”.
A An ordinary syntax question of the form “do my homework”. B Asks for a regional dialect survey of Egyptian Arabic w.r.t. a word. C Asks if there is an English translation of an ancient manuscript. D Asks “What is the ‘That-clause’ after nouns like ‘way’ or ‘fact’” (my retort would be “What do you mean by ‘what is?’”).
E-G combine lg-spec. with “trees” or “translation”:
E Split between tree vs. lg-spec: a tree request F Split on lg-spec. and translation. Asks all of the recorded meanings of a Turkish word. G Split lg.spec and translation. Asks the “meaning” (etymology) of Daniel in Hebrew
Also relevant is the Ashkinadze question which asks the etymology of a family name, closed as “translation”.
In reading these questions, I mostly agree that the questions should be closed, so the problem is that the apparent reason is not coherent, and coherence of close-reasons is the end that I am working towards. The “that-clause” question resembles a valid about linguistics, but it is badly formed, and since it presupposes that there is an agreed taxonomy of clause-types in linguistics and we could say e.g. “That’s a factive complement” or some such thing. Other evidence indicates that this is actually a language-learning question, and should simply go elsewhere.
Also relevant to the question of what is “on topic” is a list referred to in the help center “on topic” page, which gives a list of approved fields of linguistics. Please note that it includes etymology as well as language education. There are a number of problems with that list, which should be separately addressed, but “etymology” and “language education” specifically need to be eliminated now, unless we conclude that “etymology” is not off topic.
Summary: the help center needs specific revisions both in on-topic and don't-ask. The "tree" reason should be changed to "no homework". Finally, the "language specific" reason really needs discussion so that there can actually be some text in the help center that refers to the presently hidden close reason.