Before this blows up in edit and comment wars (it is already happening), I want to get meta consensus on what to do.

One particular user prefers to use an uncommon variant spelling for certain words: reëntry instead of reentry or re-entry, reüse instead of reuse or re-use, etc.

While there can be some debate whether the words should be written reuse or whether re-use is also an accepted writing (IMHO it is, as hyphenation is an accepted tool in the English language), I think that simply adding diaeresis should not be accepted, as reüse simply is not an English word (albeit recognizable as what it should mean), and SE has an English-only policy.

Fixing grammar and spelling has always been an accepted and encouraged way to edit posts and make them more readable.

Given that my native tongue uses diaeresis, this writing gives me a headache since the sounds of those vowels are completely different then the word would be spoken in English. This is unnecessarily distracting and does actively hinder readability.

How does the Space SE community want to handle this? Allow different spellings? Then what is acceptable and what is not? Only allow proper English? And what is defined as proper English?

For reference, here are the three forms being discussed:

  • reëntry The diaeresis (¨) indicates that that the two vowels are voiced separately. This form is the one being discussed.
  • re-entry The dash (-) indicates that that the two vowels are voiced separately.
  • reentry No explicit indication that that the two vowels are voiced separately.

Note that Ngram shows that while "re-entry" and "reentry" are both used, "reëntry" does not appear in published works in the last 300 years, similarly it shows "reüse" seems not to be used in English literature, while reuse and re-use are.

Similarly, Ngram shows that "cooperate" and "co-operate" are the predominantly used forms of the word, with "coöperate", virtually non-existent in literature.
On the other hand, it also shows that "naive" is being rapidly replaced by "naïve", a spelling that it claims was virtually never used before the late 1990s. In this case, its a word imported from french, though.

The question remains whether SE should be a place where new spellings that are not broadly accepted (or to say it, appear to have been virtually non-existent until the late 1990s and were still rare as of 2008) should be allowed or if we want to use the commonly accepted forms of the English language.

Similar matters have been discussed before:

  • $\begingroup$ "Given that my native tongue uses diaeresis, this writing gives me a headache" This is a good argument. Why not post that as an answer? $\endgroup$ – Alex Hajnal Dec 29 '18 at 14:07
  • $\begingroup$ @AlexHajnal No need to add an IMHO, this is backed by all major english dictionaries (Merriam-Webster, Oxford, Collins et al.), and logic dictates that the one making an extraordinary claim has to back it up. $\endgroup$ – Polygnome Dec 29 '18 at 14:14
  • $\begingroup$ Actually OED (not the "Living Language" version) only lists the dash form (at least in the copy I have) with the other two versions not even mentioned as variants. It does list "reuse" but solely as an archaic form of the verb "roose". $\endgroup$ – Alex Hajnal Dec 29 '18 at 14:25
  • $\begingroup$ @AlexHajnal Then use that as an argument in your answer. $\endgroup$ – Polygnome Dec 29 '18 at 14:26
  • $\begingroup$ I may do so. And can we please try to keep things civil. I think this is a topic worthy of discussion but have no interest in engaging in a flame war. I would appreciate it if you kept the question as neutral and free of opinion as possible; IMO answers are the place for opinion, not questions. $\endgroup$ – Alex Hajnal Dec 29 '18 at 14:32
  • $\begingroup$ I do think that some of your arguments are quite persuasive. However, your (in my perception) confrontational attitude is not working in your favor. $\endgroup$ – Alex Hajnal Dec 29 '18 at 14:38
  • 4
    $\begingroup$ @AlexHajnal Please do not edit my question to include arguments which should go into your own answer, especially when misrepresenting what the ngram shows. "co - operate" is the ngram search term used for "co-operate" due to the way ngram works. $\endgroup$ – Polygnome Dec 29 '18 at 19:00
  • $\begingroup$ @AlexHajnal plus this isn't about BE vs AmE. $\endgroup$ – Polygnome Dec 29 '18 at 19:00
  • 1
    $\begingroup$ Ngram shows inconsistent results for - versus ` - `. And I agree, arguments should go into answers, not questions. If they are to go into the question they should show as many perspectives as possible. I would say doing so improves the question since it shows there is a current lack of consensus that ought to be debated (e.g. in the answers). $\endgroup$ – Alex Hajnal Dec 29 '18 at 23:05
  • $\begingroup$ When you have a chance could you put your arguments into answers so that they can be discussed? Thanks. $\endgroup$ – Alex Hajnal Dec 29 '18 at 23:20
  • $\begingroup$ They don't even go on German loanwords anymore, so: NO. $\endgroup$ – Mazura Dec 31 '18 at 20:49

In 30 years of reading and writing in English a lot, I've never come across words with a diaeresis. Words that are spelled with accents in other languages (French, German, Dutch) invariably lose them in English.

In English, the diaeresis was considered obsolete by 1960.

it is far less commonly used in words such as coöperate and reënter except in a very few publications—notably The New Yorker.

I prefer the form that doesn't use diaeresis or hyphen, and that seems to be the most common form.

  • 2
    $\begingroup$ This is outside the context of the question but apropos to your answer: Some loanwords in English are properly ("résumé") or frequently ("naïve") spelled with accented letters, i.e. they retain their original accents. $\endgroup$ – Alex Hajnal Dec 29 '18 at 17:08
  • $\begingroup$ Also note that the question is not about loanwords but about a construct native to English. $\endgroup$ – Alex Hajnal Dec 29 '18 at 23:19
  • $\begingroup$ On a related note, I'd love to see a pronunciation engine get aphelion right if it's never seen it before. A human has a chance of determining the correct root and working backwards from there. The loss of the diaeresis is due to the typewriter. But computers can handle if it we cared. $\endgroup$ – Joshua Jan 3 at 19:23
  • 1
    $\begingroup$ Dutch is a weird language to mention here. The only diacritic it uses is the diaeresis, and it uses it in the very way that is being discussed in this question. Though it is used (mostly) consistently, there aren't that many words that use it (and often it's only plural or a certain tense), so I'm not sure if it is in loan words at all. It is definitely in some words that have shared roots with an English word: naïef, coöperatie. $\endgroup$ – Jasper Jan 6 at 19:47

Search-ability is a highly desirable characteristic for posts.

Non-standard fancy stuff that could potentially interfere with the use of some search engines or using SE's internal search can lead to someone not seeing a useful post, and that's bad.

This is related to why MathJax in titles? is strongly discouraged for example. But searches go beyond titles and look into the post it self in some cases.

The "related" question about British/US spelling differences is not related, searches have synonyms and can handle a large, common bifurcation like that that results in what usually looks like a tiny spelling error. That doesn't compare to something as rare and unusual as what's ask about here.

So please go back and remove them all, restore all posts to standard, conventional English.

  • 2
    $\begingroup$ StackExchange search ignores accents. See, e.g., this search and this one. They both return the same results (though the highlighting does differ). Interestingly, even graphemes are handled, i.e. "æronautics" returns results for "aeronautics". $\endgroup$ – Alex Hajnal Dec 29 '18 at 23:55
  • 1
    $\begingroup$ @AlexHajnal software is a funny thing. Something like that should be 1) tested thoroughly and exhaustively and 2) discussed in a meta post where people familliar with SE can review your conclusions. A random spot check like that doesn't mean much. $\endgroup$ – uhoh Dec 30 '18 at 0:01
  • $\begingroup$ @uhoh: SE uses ElasticSearch for its site search, and I am quite certain that ElasticSearch is more than mature enough to get something as simple as accent normalization correct. This is a solved problem. $\endgroup$ – Nathan Tuggy Dec 31 '18 at 9:56
  • $\begingroup$ @NathanTuggy that sounds important enough to post as an answer, together with supporting links, for better visibility. It's not exactly an answer to the question, but this is meta... $\endgroup$ – uhoh Dec 31 '18 at 10:35
  • 1
    $\begingroup$ You’re the one making the unsupported assertion that it hinders search, @uhoh. $\endgroup$ – Russell Borogove Jan 2 at 19:54
  • $\begingroup$ @RussellBorogove "... that could potentially interfere with the use of some search engines or using SE's internal search..." is not an assertion. It's a concern. NathanTuggy suggested that part of that concern may be unwarranted, so I suggested he elevate that information to a place with greater visibility. $\endgroup$ – uhoh Jan 2 at 23:25
  • 1
    $\begingroup$ Google likewise matches reëntry when reentry is searched for, for the record, and no, I'm not writing an answer to that effect. $\endgroup$ – Russell Borogove Jan 3 at 1:08
  • $\begingroup$ @RussellBorogove will google always behave that way? Will it when people are using google from any location? Google returns different search results depending on where you are and which language it thinks you are most comfortable speaking. A spot check at one point in time at one location in one language isn't enough. I'll stand by my recommendation of caution. See BECO baby carrier and comments below it for example. $\endgroup$ – uhoh Jan 3 at 5:32
  • $\begingroup$ @uhoh A Google search for "reëntry" with the locale set to either United States or Deutschland returns the Wiktionary definition of "reëntry" as the first result. It also gives a handy definition as "the return of a spacecraft or missile into the earth's atmosphere" for the former locale and "Wiedereintritt eines Raumflugkörpers in die Erdatmosphäre" for the latter. So it appears that Google, at least, has no problems with that spelling even in locales where it could be ambiguous/have different usage. $\endgroup$ – Alex Hajnal Jan 7 at 2:15
  • 1
    $\begingroup$ @AlexHajnal you don't understand how google search works (neither do I of course). Your spot check does not demonstrate that google search will return Stack Exchange results reliably under all possible conditions. Google uses a huge adaptive algorithm and results can change for many reasons and does change over time. Search google images for "idiot" and you get pictures of a president. That never happened before. You can't use a few spot checks to define "how google works" or predict results for all users under all conditions. and not everybody on Earth uses only google! $\endgroup$ – uhoh Jan 7 at 2:20
  • $\begingroup$ FWIW, here's some sample results for Space.SE: "reentry" and "reëntry" return essentially the same results while "re-entry" and reentry each return totally different results. $\endgroup$ – Alex Hajnal Jan 7 at 2:48
  • $\begingroup$ @AlexHajnal i.stack.imgur.com/tbUQb.png $\endgroup$ – uhoh Jan 7 at 2:52
  • 1
    $\begingroup$ @uhoh :^) Touché $\endgroup$ – Alex Hajnal Jan 7 at 2:54
  • $\begingroup$ @uhoh Google returns pictures of Trump for the search "idiot" because the internet has changed, not because Google has. If enormous numbers of people repeatedly posted on the internet that I'm an idiot, my picture would appear instead. $\endgroup$ – David Richerby Jan 7 at 10:25
  • $\begingroup$ @DavidRicherby I'm torn between saying "google IS the internet" and "google is constantly changing", so I'll just say Google IS the ever-changing internet. $\endgroup$ – uhoh Jan 7 at 10:27

I'm not going to use the dash forms. It feels too much like bowing to the spellcheck.

Personally I feel like using diaeresis only to disambiguate. reentry doesn't have a second possibility that actually exists, and neither does cooperate.

But coöp and coop are two different words.


It looks like the use of diaeresis is making several people uncomfortable. In fact, so uncomfortable that we are now discussing this at length. That means it's not about whether the usage is correct English or not, or other technical reasons affecting the site, but whether using the unexpected (for several readers) spelling is disrupting to a large enough number of users. And that does indeed seem to be the case.

Since more people seem to have a problem with it than who prefer it, I think we should discourage the usage. So far there seems to be only one proponent but the topic really hits a nerve with several people.

It's likely to continue to be a topic that would be edited and discussed in the future if the diaeresis are continued to be used, even if we would say we're okay with it here on meta. By contrast, it only takes one user (as far as I can see) to stop using them to avoid having unrelated spelling discussions on their questions/answers. So I think discouraging the use is the better option in the long run and hope said user still wants to contribute to our community.

  • $\begingroup$ Welcome to Meta! It's more a case of having something to point at and say "yeah it's weird, but while we wouldn't recommend it we can put up with it" if people start editing them out of other's posts (or even editting them in) $\endgroup$ – JCRM Jan 16 at 7:54

In English there are three ways that words with adjacent, separately voiced vowels can be written: ¨ over second letter, split by a dash, or unmarked.

I prefer using the first form since its intent is clear (that's the symbol's purpose in English).

IMO, the second form (split by a dash) is also fairly good in that it clearly separates the syllables; what I don't like about it is that the symbol is also used to indicate compound words ("on-orbit") and alternate stems ("pre- and post-flight"). This form is also prone to splitting words at the ends of lines.

I am personally strongly against the third form (unmarked) since it gives no indication that the syllables are separately voiced.

  • 2
    $\begingroup$ fyi up and down voting in meta are simply a quick way to tally levels of agreement and disagreement. Don't take them personally. $\endgroup$ – uhoh Dec 30 '18 at 0:04
  • $\begingroup$ @uhoh Yea, I get that. What's annoying me a bit though is that people are downvoting without commenting. It's clear that there's disagreement but it would be good to know precisely why. $\endgroup$ – Alex Hajnal Dec 30 '18 at 0:07
  • 3
    $\begingroup$ In cases where "hey let's start doing things differently" is proposed, silent down votes are common. SE has evolved over a decade and millions of posts and thousands of moderators. Most aspects are the way they are because they've found an equilibrium point and are what is felt to work best. $\endgroup$ – uhoh Dec 30 '18 at 0:10
  • $\begingroup$ Another answer about How to English, +1 $\endgroup$ – Mazura Dec 31 '18 at 21:25
  • 5
    $\begingroup$ I suspect most of the disagreement - including my own - is because modern English simply does not use the diaeresis (except in a few loan words like naive and, even then, it's mostly just when auto-correct puts it in there.) Many (most?) native English speakers who don't also speak another language that still actively uses it probably don't even know what it means, so the claim that it's 'clear' doesn't really hold water, IMO. "Reentry" or "re-entry" are, at the very least, equally clear and likely more clear to most speakers of English. And they are also far more standard in modern usage. $\endgroup$ – reirab Jan 2 at 16:52
  • 2
    $\begingroup$ The diaeresis is so foreign to modern English usage that normal English keyboards (not the 'international' ones) don't even include a way to add a diaeresis to a letter. $\endgroup$ – reirab Jan 2 at 16:54
  • 1
    $\begingroup$ Your argument seems to be based on the misconception that English spelling is uniquely decodable into pronunciation. $\endgroup$ – David Richerby Jan 6 at 17:32

The diaeresis is an accepted part of the English language and should thus be allowed:

  • Merriam-Webster
    1: a mark ¨ placed over a vowel to indicate that the vowel is pronounced in a separate syllable (as in naïve or Brontë)

  • Collins
    1. the mark ¨, in writing placed over the second of two adjacent vowels to indicate that it is to be pronounced separately rather than forming a diphthong with the first, as in some spellings of coöperate, naïve, etc
    2. this mark used for any other purpose, such as to indicate that a special pronunciation is appropriate to a particular vowel

  • Oxford (Living Language edition)
    1 A mark (¨) placed over a vowel to indicate that it is sounded separately, as in naïve, Brontë.

  • Oxford (print edition)
    1. The division of one syllable into two, esp. by the separation of a diphthong into two parts.
    b[sic]. The sign [ ¨ ][sic] marking such a division, or, more usually, placed over the second of two vowels which otherwise make a diphthong or single sound, to indicate that they are to be pronounced separately.

(Non-relevant meanings, etymologies, etc. omitted from all of the above.)

  • $\begingroup$ To the downvoters: Can you elaborate on why you feel the dictionary definitions are invalid? $\endgroup$ – Alex Hajnal Dec 30 '18 at 0:02
  • 4
    $\begingroup$ If you are asking a question about the English language in English SE then this answer would be relevant. But the question here is not "is this proper English", it's "is it okay to start adding weird symbols within words that nobody else recognizes?" and to that question, this answer is irrelevant. $\endgroup$ – uhoh Dec 30 '18 at 0:36
  • 2
    $\begingroup$ They're appropriate to use when the context is about language, to provide pronunciation help. They do not belong in standard discourse - unless your Manual of Style says otherwise, which other than loanwords, I doubt. $\endgroup$ – Mazura Dec 31 '18 at 20:48
  • $\begingroup$ @Mazura The definitions above refer to standard English usage. Loanwords comprise only a very small subset of English words using diaeresis (the only one I can think of is "naïve"). For scholarly discussion about language one uses IPA which is a very different alphabet. $\endgroup$ – Alex Hajnal Dec 31 '18 at 21:01
  • $\begingroup$ I hate IPA. But naive is French and takes an 'acute' mark, no? Only the very small subset of loanwords words should use diaeresis, if even then, as IMO no word on this page is still a loanword. If schadenfreude had umlauts, we'd be leaving them out by now. $\endgroup$ – Mazura Dec 31 '18 at 21:12
  • $\begingroup$ @Mazura No, "naïve" takes a diaeresis. You're probably thinking of words like "résumé" or "détente". Keeping the acute accent in the English spelling is pretty common for loanwords from French. And yea, IPA is horrible. $\endgroup$ – Alex Hajnal Dec 31 '18 at 21:16
  • $\begingroup$ +1. I can't refute any of this (not that I like it). This is a question for Serious Language Enthusiasts: ELU.SE $\endgroup$ – Mazura Dec 31 '18 at 21:23
  • $\begingroup$ @AlexHajnal "The definitions above refer to standard English usage." They refer to standard English usage of the word "diaeresis", not to standard English usage of the diaeresis. The fact that the dictionaries tell you what "dinosaur" means doesn't imply that dinosaurs are still alive. $\endgroup$ – David Richerby Jan 6 at 17:31
  • $\begingroup$ @DavidRicherby If the definitions didn't refer to modern usage they would have explicitly said so. Take ſ for example. Collins says "a lower-case s, printed ʃ, formerly used in handwriting and printing" while Oxford (Living Language) says "An obsolete form of lower-case s, written or printed as ſ. It was used in initial and medial but not final position in a word, and was generally abandoned in English-language printing shortly before 1800." (emphasis mine) $\endgroup$ – Alex Hajnal Jan 7 at 1:52
  • $\begingroup$ @AlexHajnal The diaeresis is commonly used in "naïve" and occasionally in the names Chloë and Zoë. It is utterly obsolete in cases such as "reënter". $\endgroup$ – David Richerby Jan 7 at 10:21
  • $\begingroup$ If the dictionary defined a swimming arm band as a "inflatable toroidal device worn on the upper arm by swimmers" would you then declare all swimmers should wear them? $\endgroup$ – JCRM Jan 13 at 9:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .