What is the best way to answer a question which has no answer?

Question

What is the preferred way to answer a question which asks “Provide me with X”, when it is overwhelmingly likely that X does not exist?

Assumptions

1. Assume the question is interesting on its own merits¹ and not something like Word for disrespecting eldest half-sister by referring to her husband as girly-girl-manly-boy though he's amused but the rest of the family isn't?.

2. Assume it is impossible to dispositively prove a negative.

3. Assume that it is relatively obvious to the experts on this site that it is very unlikely the question has a positive answer.

Example

Today a question was posted to which my immediate instinct was “there is no such word”:

Is there a word for punctuation marks favored by individuals?

Yet I found the question quite interesting on its own merits, enough to motivate me to pursue a micro-investigation.

That investigation only strengthened my suspicion that there is no such word. So if I were to post an Answer, it would likely take the same shape as the comments I made, something like this:

There is no single word for a person’s favored punctuation marks. You’ll have to use an adjective to modify “punctuation” or a synonym of that noun. Which adjective you choose will depend on the particular flavor you’d like to lend to the description.

Absence of Evidence (but not Evidence of Absence)
I say this on the strength of the lack of evidence that the word exists or that there is sufficient need or motivation for such a specific word to be coined.

In support of this conclusion, I'll first point out that the general discipline of identifying authorship from patterns, cues and clues in a text is known as stylometry. If anyone has sufficient reason to coin such a term, it is the scholars of that discipline.

Given that lead, I googled stylometry glossary and a series of variations on that, both including and excluding punctuation as a term, and found nothing of interest.

The top hit persisted in being the Wikipedia article on the topic. I read that article and found the related term forensic linguistics, and repeated my google searches with that term substituted for stylometry.

Still no glossaries, but I did come across this relatively comprehensive master's thesis with a full literature review:

INVESTIGATING THE USE OF FORENSIC STYLING AND THE USE OF STYLOMETRIC TECHNIQUES IN THE ANALYSIS OF AUTHORSHIP ON A PUBLICLY ACCESSIBLE SOCIAL NETWORKING SITE

It was authored by Colin Simon Michell in July 2017. Per its TOC, the section that deals with punctuation specifically is 2.7.3.3, starting on page 52. Despite the comprehensive nature of the study and the full literature review of the technical discipline that is most specifically pertinent to your search, that section offers no single-word term for “preferred or idiosyncratic punctuation”.

Consolation Prizes & Work-arounds
However, the thesis does offer us a couple of consolation prizes, multiple-word and/or insufficiently precise though they may be: style markers (not just about punctuation; general "clues") and punctuation devices (not a single word). My instinct is this is the best we will get.

The bottom line is you’ll have to use an adjective to modify “punctuation” or a synonym of that noun. Which adjective you choose will depend on the particular flavor you’d like to lend to the description. You might say “his particular use of ellipses” or “his idiosyncratic punctuation”, etc.

Does that Answer answer the Question? Does posting it improve the site for current and future visitors? Should it be posted²?

Methods

I can see three obvious potential methods to “answering”^†:

1. People post answers, doing their best to give evidence of absence (i.e. approximation of “proving no such thing exists”).
2. The question has no actual answer, so the posted Question receives no posted Answers.
3. The question is closed because there can never be an actual “answer”, so it is barred from receiving non-answers.

I include this last because it is certainly a possibility, but I would only recommend it when there’s an explicit or tacit premise in the question which is so obviously flawed that (a) everyone, including non-experts, can see it and (b) it’s impossible to repair (meaning repairing it would effectively “un-ask” the question).

Therefore, I’d reserve its use for the questions which are covert rants or otherwise have a hidden agenda which are already discouraged in the help center⁴.

That leaves us, so far as I can see it, with the first two options. I can see merits and demerits to both approaches.

Question

Leaving aside the general principle “take each question on its individual merits on a case-by-case basis”, which applies to all questions all the time, I'd like to get a sense of the community’s general preference, the default case, for such unanswerable questions³.

What is the community’s preferred approach?

In particular, please recommend one of the approaches above, or some other, and give a substantiated argument for it.

Recommendations which speak to the foundational goals of the site (“building a curated library of questions”, ”helping current and future visitors”, “developing a center of expertise on English”, etc) or cite precedent or policy on EL&U or somewhere else on StackExchange⁵ will be particularly helpful.

---

¹ _{The sought item might be lexical gap. Of course, if it is a known lexical gap, then that is a a proper answer to the question and therefore it doesn't fall under this meta-policy's remit. The more interesting questions which do fall under this remit are words which give us some inner, thin hope that they might exist (because they "should"), which is is why they're interesting, but for most such questions, it will be "obvious" to English experts that they don't exist, which is the genesis of this policy question.}

² _{Based on early feedback to this Mets-question, I've gone ahead and posted it as an answer to the question proper. So now you can give direct feedback on the answer itself: if "no such word exists" answers should not be posted, on epistemological or other grounds, downvote it. If you agree with the philosophy outlined in, e.g., @ab2's answer, that such answers at least add information, upvote it. If voting won't help you express your position for one reason or another, don't feel any obligation.}

³ _{That is, assumption #2 guarantees that it is possible that someone, someday may find an actual answer. That is, such things are literally open questions, so the Questions should remain Open (same reason that questions aren't immediately closed when the OP "accepts" an answer; something even better may eventually come along).}

⁴ _{Consequently and ironically, then, I’d consider any posted Answer to this posted Question which recommends “taking such questions on a case-by-case basis” a non-answer to the actual question. Twists and loops and braids entangle us.}

⁵ _{If citing non-EL&U precedent or policy on SE, sites which are similar in aim or character to EL&U, or Meta.SE (which theoretically at least provides overarching policy) will lend the most strength to a given case as having more obvious applicability.}

† _{I've seen all such positions advanced. Consider this exchange in the comments of an old Meta-question:}

_{@JoeBlow: "There's no such word" is never a good answer to a word request. If that's true, it is adequately conveyed by lack of answers (or lack of positively voted answers). "I don't know of any such word" is actually defensible, but useless, so it's not a good answer either. – Ben Voigt Aug 13 '14 at 16:55}

_{I'm afraid I don't agree, Ben. If I was asking for a SWR, and "God" (as it were) could decisively tell me "There is no such word" - I'd want Him to tell me. – Fattie Aug 13 '14 at 18:42}

_{@Joe & Ben: whether or not to explicitly post a "there is no answer" answer on SWRs is a good question for meta - one of you guys want to post it? (It's worth considering whether finding in the affirmative will encourage multiple users to post identical "there's no answer" answers.) – Dan Bron Aug 13 '14 at 19:29}

Answer 1

I vote for #1, because, if done well, it provides information, and could supply the stimulus for someone else to delve further and perhaps discover an answer. Of course, some (many?) of the answers under Option #1 will be crap, and should be downvoted or converted to comments.

Options #2 and #3 preclude the possibility of providing any information. Where would science be if null results were discarded? See, e.g., Scientific American, Is Supersymmetry Dead?

Answer 2

Personally, not only do I think posting the answer you suggested in the question would absolutely be a good answer, I also think it's the kind of insightful, thought-provoking answer that is unfortunately difficult to find in single-word-request questions.

Don't misunderstand me here, I don't think there's anything wrong with SWRs, but I think we all know that they tend to elicit one line answers with a copy-pasted (I don't think this qualifies as copy-pasta) snippets from a dictionary. The kind of answer you offered as an example brings the process into the equation, which I think is part of what makes etymology yield such interesting answers.

Great etymology answers often tell a story to provide context for their claim. The story isn't just aesthetic in purpose; it gives the reader some idea of what process led to the conclusion found, whether examination of corpora, consultation of reputable dictionaries, or other kinds of research. This buttresses the conclusion and usually gives some idea of what the other possibilities might be and why the conclusion is possibly the "most likey" answer if there are alternative possibilities.

I find this Meta question especially interesting because I've often wondered where the boundary lies for etymology questions—that is, what makes a question so impossible to answer that it's ridiculous to ask. I've tested that boundary a few times myself. Some might have seen the +500 bounty I put on the etymology of slang. The bounty went to JEL's answer, which supported the proposition in the question itself, albeit while providing illuminating and informative research in the process.

My favorite example of a no-answer answer is Sven Yarg's answer to a question I posted about the word "growler." After an epic discussion of the history of the word, Sven concludes:

we may be dealing with a term whose source is unfindable.

This would be a disappointment if it appeared as a sentence in isolation, of course, but because the answer traverses the process and offers a thorough examination of what possibilities were considered and what history is known, it's a phenomenal answer.

To me, your non-answer to a single-word-request is very similar. You explain the process you went through, you show your research, and you explain why you've concluded that such a word very likely doesn't exist, all while offering close multi-word options as a consolation prize. I'd love to see analytical answers like that to SWRs.

Answer 3

I'm having trouble wading through your question so I'll answer what's in the title.

Just say something like 'There is no good answer' and give reasons, with content, suggesting why.

• The question may be too specific and you can give alternatives for less specific versions.

• There may be an archaic term which may be an answer for a thousand years ago, but no one would understand it now.

• A translation may not be word for word or hit exactly all and only the semantic features, but nearby alternatives may be sufficient.

Since negatives are hard to prove succinctly (there's no word to point at because that's the problem it doesn't exist!), there is some reliance on authority. But we're somewhat anonymous here. Just saying 'no' is not enough. Explicitly say 'no' but give some content reasons as to why you think so. Weasel words, like 'for the most part' or 'it is expected' actually increase accuracy.

Note that existence questions are usually of the questionable SWR variety, where a simple thesaurus look up usually can give you a straight yes or no. So it is likely not a good question to begin with, but maybe your non-answer answer can at least shed some useful light on it.

Answer 4

First I know you put plenty of effort into this, but let us make no unproven assumptions. I think part of your problem is that you have over-complicated the problem.

What Do We Desire?

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.

This is the closure reason we usually apply when we suspect an answer can not be adequately corroborated. Sometimes, it seems to me as if Primarily Opinion Based is misread, as if it was just "Opinion Based" but I do not believe that is the case. As you can see, from the closure reason itself, there is some amount of leeway, for "some degree of opinion".

If you can base your answer in "facts, references or specific expertise" for us to serve as a basis for our votes, then we welcome it to be audited by our system of peer review. What we generally try to avoid is answers that are merely "based almost entirely on opinion." The definitive post for what constitutes a Primarily Opinion Based question on the Stack Exchange network is Good Subjective, Bad Subjective.. In my opinion, the standards are not especially high, and I think we can even go all the way down to anecdotal evidence based on the moms4moms example, provided that the anecdote is sufficiently explanatory and relevant.

What we dislike is for people to just claim "There is no such word!" and without any further reason to believe them. That is just an opinion that does nothing to settle anybody's doubt, since there is a high likelihood that it may very well be wrong. According to Improve Your Speed Reading by Phil Heartman, a joint Google/Harvard study revealed approximately one million words in the language, and the adult only vocabulary ranges between 35,000–75,000, which only represents about 5% of the typical vocabulary in Phil's approximation. Just to stress how good of a vocabulary 75,000 words is, Noah Webster's magnum opus, An American Dictionary of the English Language contained approximately 70,000 words upon its first publication in 1828. It took him 26 years to complete, and was the largest dictionary available of its time. I doubt that anybody else is as renowned for mastery of the English Language vocabulary as him, as his surname has become something of an eponym for the very concept of a dictionary in the U.S.A.

This is one of my principle interests in single-word-requests, because the Stack Exchange system can be exceptionally good at sussing out especially rare words by crowd-sourcing information.

This is not to say that "no such word exists" is never a good answer. The community has already established that "no such word exists" is a perfectly valid answer to single word requests in these questions:

• How should we handle single-word-requests that don't have a matching word?

• For [single-word-requests], should we encourage "there is no such word" as an answer?

I am not against people doing that if they want to do so. It seems like the best solution to the problem, if somebody is willing to do it properly. However, I do not think we should be suspending our normal expectations of what constitutes a good answer.

So How Can We Answer Questions Regarding Nonexistence Well?

Although I do not want to foreclose upon any good methods of doing it, I believe a principle concern in this question is how to corraborate the answer. The way I see it, there are at least a few ways:

Specific Expertise

Some people need certain kinds of words more than others, and are hence more likely to know if a word within a certain category exists than other people, so one way to answer is for such a person to claim no such word exists. This seems to be the approach Professor John Lawler took to answering Is There One Term for Grammar, Punctuation and Spelling? Now, I somewhat disagree with the answer and hope to eventually write my own answer arguing that grammar is the best word based on historical usage, and my own interest in the subject matter of what grammar means is why I know of this question in the first place.

Nevertheless I still think it serves as a good example of the sort of value that type of answer may have especially since Prof. Lawler is a very well respected expert within our community, and it probably still goes to show that the existence of another such word within the scholarly community is quite unlikely, since it would not be so troublesome for him to suggest it if he had ever heard of one, and I think most of us can agree it would probably be him if anybody had ever heard of such a word.

References

Another way to show that something does not exist is to show that it is not where anybody would expect it to be if it did exist. This is similar to the tactic of looking through the closet and shining a light under the bed to assure to frightened children who do not know any better that the boogeyman is not going to eat them as soon as they are tucked under the covers and close their eyes.

This is part of the approach I took with Are The “Paraphrase Brackets” (Not Square Brackets) Used in English? With all of the evidence I provided that such a thing is unlikely to exist in our language, my answer received seven votes for it, as opposed to the three Hotlicks received for posting the same conclusion without any evidence at all. Hotlicks has more reputation points than me, and posted his answer earlier too, so it is out of appreciation for my evidence that my answer outshone his. You and Guildenstem even honored me by providing some direct approbation for my research:

The bit about the Manding languages was really intriguing. I wondered if these languages needed paraphrase brackets because paraphrasing is an important or fundamental part of their discourse (relatively more than other languages, I mean). The question occurred to me because I know there are some languages, like Pirahã, which inflect for the "source of knowledge" (1st hand, 2nd hand, etc), which would be more important in oral-only communities. Sadly I didn't find anything interesting on a quick search, though Wikipedia does mention these communities have "a strong oral tradition.—Dan Bron

What an exceptionally thorough and well-researched answer.—Guildenstern

Now this is much more difficult with words than a proper dictionary, since there are so many more of them. However, the place you would expect to usually find those are dictionaries.

Nevertheless, despite however difficult proving an absence may be using this method, it is still hypothetically possible. As incredible as it may seem, there is at least one person out there who have read the entire 20 volume Oxford English Dictionary second edition. Ammon Shea undertook the task by reading eight hours a day for a year, documented the experience in a book entitled Reading the OED: One Man, One Year, 21,730 Pages. He was interviewed by reputable news outlets like N.P.R. and the British Broadcasting Corporation.

Additionally, this does not have to be a full time job: Despite however many volumes it has, the latest print edition of the Oxford English Dictionary is not an especially word entry dense publication with its 218,632 total words (its many pages mostly contain examples of historical use, with many extra definitions). I think it might be realistic for us to read at least a single volume of an unabridged dictionary. Those advertise word totals ranging from 315,000 like the Random House Webster's Unabridged Dictionary, Second Edition, to 700,000 words like Collins English Dictionary—Complete and Unabridged. The Oxford Dictionaries Blog claims that the task of reading the Oxford Dictionary of English only takes 41 hours of continuous reading.

I admit, that is plenty of work, but some people have done it with no real end goal in sight, and if somebody wanted to undertake the task, the result could be widely applicable to the range of word request questions on our site, and constitute a form of expertise that would be unwise to turn away. I have also noticed that people have undertaken the task of reading the whole Oxford Canadian Dictionary, such as Nikki Love (A, B, C, D-H, I-O, P-S, T-Z). Additionally, the use of specialty dictionaries can help to lessen the task or find especially arcane words.

Now the issue of trust has been raised, and I do agree that this is a somewhat legitimate source of concern, since anybody can claim they did something when they did not, but usually, when somebody discloses their credentials, such as being a native speaker of a certain dialect or a professional linguist, we give them the benefit of our trust. I don't see why "I read an entire dictionary" should be much different, especially since a skeptic can audit the claim by another reading of the same dictionary. At any rate, if we should verify what a person's credentials are and how we should go about it, is a problem that should be considered separately, in my opinion.

Explaining the Facts

Sometimes a word is more likely to exist than not simply due to the facts surrounding it, and providing a good explanation may just be able to prove that. It may require the use of some creative logic.

In order to exemplify some of what I mean, I submit some answers which I wrote that which take novel approaches with etymology to trying and disprove the existence of a word:

You should recognize my etymologically based attempt to disprove the existence of the word lexophillia. Granted, this is not a single-word-request but perhaps similar guidance regarding how to disprove the existence of the word can be taken from other is-it-a-word questions.

More directly applicable is in Is *There a Word Analogous to “Dual” for Three or More Options?, I try to select and dismiss the closest possible match to the question * I claim that if there was to be a word within that series, that it must be trial, and then I show how trial is not the word.

Propose Alternative Solutions Instead

Sometimes even a good question does not have a good direct answer, but if a question such as the one described is asked well enough, we can explain what the next best alternative is. Although I believe I have adequately demonstrated that there is no perfectly analogous word to the word dual, the other members have tried to suggest would-be synonyms which would match the given context. The accepted answer is Triple, and the top voted answer is Threefold. Stack Exchange encourages this in How Do I Write A Good Answer.

Indeed, if the questioner describes what it is they want well enough an alternative answer may even invoke serendipity:

The Happy faculty, or luck, of finding, by "accidental sagacity," interesting items of information or unexpected proofs of one's theories; discovery of things unsought: A factitious word humorously invented by Horace Walpole.[—The Century Dictionary and Cyclopedia Supplement]

Combining Solutions

Although implementing just any one of these solutions should be good enough if done well, there is no need to restrict yourself to implementing just one. Some of these work especially well in combination. For instance, reading an entire dictionary may just give you a sense of how people would communicate the desired concept, if no match for the precise concept exists, and so you can comprehensively document which alternatives you have encountered that seem best to better prove you have undertaken the task.

Addressing the Methods of Dealing with the Problem

Preferably speaking, I would prefer for solution 1 to be implemented, if somebody is willing to do it properly. By using the aforementioned techniques, we can actually learn something about the language. If somebody wants to try, I see no really good reason to stop them.

However, in implementing the first solution, we have to acknowledge that solution no. 2 may be the one that is implemented most often in practice, because this is a volunteer run website. Sometimes people just are not going to want to answer a question, especially if it is an odd and relatively insignificant request. As long as a question remains open, it is up to the people who consider answering to decide whether or not they want to try.

A Solution I Think We Should Avoid

In a case such as the one you describe, this option is not one that I would like us to implement this possible solution to the problem:

3. The question is closed because there can never be an actual “answer”, so it is barred from receiving non-answers.

I do not see much merit in this approach to the problem over just ignoring the question. The only benefit I see in it over ignoring the question is that it forecloses upon wrong answers, but some wrong answers, and other qualitatively bad answers, are supposed to be on the website for the sake of giving meaning to our system of peer review. That is why Stack Exchange allows voting against answers, unlike hackernews. If adequate evidence to the contrary is provided, then we can just let votes sort out any incorrect claims to the contrary.

We do not want to be answering meaningless questions like questions like What's Your Favorite Word that other Q & A. websites without our policies like Yahoo Answers or Quora gets, where all answers are equally correct, without any need for explanation. The relevant closure reason has been quite good at keeping our website neat and tidy. It may just be that English Language & Usage never got such a question because this type of problem was spotted early on Stack Overflow, or perhaps all such questions have been deleted, but we do not receive that type of question here.

Another more commonly applicable purpose for this closure reason is naming questions, on the grounds that a proper noun could hypothetically borrow from any word, or be created from scratch to address this issue. Whatever these, and other Primarily Opinion Based questions may have, they do not exist for the sake of the pursuit of truth, and may even obstruct it, so Stack Exchange as a whole has decided to eschew them. I have seen these are expressly singled out in our help center.

However, when I cite those standards, it needs to be remembered that too much closure can be an obstruction to the pursuit of truth too, since the main purpose of it is to preclude all answers to a problematic question. This includes any potential good answers.

The over-application of closure has been a problem for the Stack Exchange network in the past. Stack Exchange noted that also rejected a more specific "too narrow" closure reason to replace it, on the grounds that it is very hard to objectively determine just what people need. Similarly, the closure system as a whole was reworked in 2013 in part to clarify the reasons for closure, and in part to restrain them. As an example of such restraint, The Too Localized closure reason was long, confusing and over-applied, and the Stack Exchange network has recanted upon it, and even rejected a network-wide 'too narrow' closure reason as still being too much. Similarly, we used to have a Not Constructive closure reason:

As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or specific expertise, but this question will likely solicit debate, arguments, polling, or extended discussion.

This has been reworked into Primarily Opinion Based, which is somewhat more permissive:

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.

If we took the closure approach of trying to answer questions like this, I might not have had the chance to post my answer to What Is A Noun For Something That Edifies Us during the few days while I was waiting for the questioner to fulfill my request divulge research. In the meantime, somebody suggested no such word exists, and proposed using an "adjective" instead of a noun. With a little bit of luck, and a somewhat novel approach, I did very easily find a word that almost no current dictionary contains, with the exceptions being Merriam-Webster's New International Dictionary, Third Edition, and its online counterpart. I also almost certainly would not have had the chance to answer the aforementioned paraphrase bracket question which went unanswered for a period from November of 2016 to march of 2017.

Also worthy of note is that question closure is "supposed to be a (mostly) spontaneous action", so if the community achieves a consensus on a certain point of policy, then it is possible that other positive contributions I have mentioned could be swiftly blocked by a few people who overestimate the worth of their independent vocabulary.

Sometimes good things come to those who wait, and I do not think we should be so fearful of a question that is not being answered anyway that we should foreclose upon it being answered entirely. Denying other people the opportunity to report if they have discovered such a thing does little to reassure people that no such thing exists. It is fine for hypothetically answerable questions to wait for an answer. That is why the unanswered questions tab exists.. In a case like this, I see no benefit over solution number 3 (closure) over solution number 2 (just ignore it). I think it is needless to prevent answers to a question if nobody is answering it anyway.

Finally, I think it is also worth noting that we already close most of the questions we receive. I have seen numbers reported as high as 60%. Can we please us not grow even more trigger happy with the closure button than we already are, and lose decent contributions in the process: Please? Pretty please? I'll put sugar on top if I must!