Many texts and internet resources use certain symbols to prefix correct, sometimes correct, questionable, and incorrect examples:

  • ○ (U+25CB): correct (converts with まる in Windows IME)
  • △ (U+25B3): sometimes correct (converts with さんかく in Windows IME)
  • ? (U+FF1F): questionable (full-width question mark)
  • × (U+00D7): incorrect (converts with ばつ in Windows IME)

Ignoring the fact that these look terrible with the default SE font (at least on my Windows box), should we adopt a similar convention?

  • Maybe we should go with ◎. :) Maybe I got a bit too much Japanified, but △ just looks to me like we don't value them too much - though otherwise it would look the nicest of them all. Maybe we should look for some signs that don't already have a widely accepted meaning in Japanese?
    – Boaz Yaniv
    Commented Jun 9, 2011 at 19:20
  • As long as they're easy to input somehow, otherwise they probably won't gain wide acceptance.
    – Troyen
    Commented Jun 9, 2011 at 19:26
  • @Troyen: Actually IME makes most of them quite easy to input. Try to convert えっくす, まる or さんかく for instance. :)
    – Boaz Yaniv
    Commented Jun 10, 2011 at 7:43
  • U+2573 (Box Drawings Light Diagonal Cross; ╳) may not be visible on some systems. At least no fonts on fileformat.info can show it. I think that U+00D7 (Multiplication Sign; ×) is more common in electronic text in Japanese, although it may be a slight abuse. Commented Jun 10, 2011 at 23:40
  • but what about our users who can only read romaji? They probably don't have the IME installed. Commented Jun 11, 2011 at 0:57
  • @Mark: Not sure what you mean. Computers can display these symbols without Japanese IME installed. Do you mean that some users may not have fonts necessary to display these characters? I guess that such cases are rare, but I am not completely sure. Commented Jun 11, 2011 at 12:40
  • @ito: Not being able to create those fonts easily because they don't have IME installed. Of course those characters can be accessed via the character mapping tool, but I don't think we should make it such that people have to go hunting around in the character map for those characters to annotate their answers. You are correct in that it shouldn't be a display issue in the browser. Commented Jun 11, 2011 at 15:51
  • @Mark: That's definitely a problem, but probably not a major one. Users without IME probably wouldn't be able to comfortably write Japanese example sentences in the first place.
    – Boaz Yaniv
    Commented Jun 11, 2011 at 17:42
  • @Boaz They might write romaji example sentences, though.
    – Amanda S
    Commented Jun 11, 2011 at 18:57
  • @Mark: I do not think that we are not talking about enforcing the use of these symbols. If the symbols are hard to enter, users can post without using them. Commented Jun 11, 2011 at 19:29
  • 2
    @Tsuyoshi: I think any system that can display Japanese characters can display these symbols (they are basic UTF additions). / To chime in with the rest: assuming users have systems that support Japanese characters sounds like a very reasonable assumption on such a site (not sure you can get much out of any of the current questions or answers without that most basic ability).
    – Dave
    Commented Jun 12, 2011 at 2:49
  • @ito: thanks for the clarification, if it is not a rule, but something more as a standardization I'm all for it :D Commented Jun 12, 2011 at 3:21
  • Interestingly enough, the EL&U Site FAQ has a section explaining the notation they use for questionable or incorrect (? and *) examples. Whatever we pick, it looks like we would be able to explain the system in our FAQ as well, so discoverability isn't as much of a concern.
    – Troyen
    Commented Aug 25, 2011 at 19:37

6 Answers 6


I think this is a worthy convention to establish. Thank you for bringing this up for discussion, Derek.

Some kind of a symbol before the sentence would be the best way to indicate the type of example it is. I was originally going to suggest some sort of a colour-coding format but this will not work if the reader is colour blind.

I suppose this suggestion kind of goes hand-in-hand with deceze's request for adding additional buttons in the tool bar.


What about ∴? It looks quite nice to me as an example marker. Using it with quote markdown looks ok:

∴ 色は匂へど散りぬるを

But we can also just 例, which I've seen used in many Japanese textbooks:

例: 色は匂へど散りぬるを

  • 1
    I do like the ∴ symbol, but I think we're more interested in adding information about the example's grammaticalness--it's usually clear already that a sentence is an example.
    – Amanda S
    Commented Jun 11, 2011 at 18:47
  • 3
    (1) Is ∴ used to denote an example? I only know this symbol as “therefore” in mathematical writing (usually followed by a mathematical formula), and this use looks strange to me. (more) Commented Jun 11, 2011 at 20:32
  • (Cont’d) (2) I am not sure about the use of the Japanese word “例:”. If the post is in Japanese, it is perfectly fine to use “例:”, but if the post is in English, I cannot see why anyone wants to write “例” instead of “ex.” (a shorthand for example/examples). Commented Jun 11, 2011 at 20:34
  • 1
    @Tsuyoshi: I don't know of any special markers for examples in English (besides "ex."). @Amanda: Generative linguists use asterisks to mark grammaticality of language specimens. AFAIK, the more asterisks an example has, the less natural it is. I'm not a great fan of this method, but we can use a single asterisk to mark a "false" example, since those are also given sometimes.
    – Boaz Yaniv
    Commented Jun 12, 2011 at 14:54

Interesting to me that Derek, whom I believe is a native English speaker, is proposing the Japanese style:

○ (U+25CB): correct

△ (U+25B3): sometimes correct

? (U+FF1F): questionable

× (U+00D7): incorrect

I have several questions:

  1. I know that what '×' (and maybe '◯' too) means in English and Japanese is totally the opposite. In English, '×' means the same as check-marking, meaning acceptance or some positive reaction, whereas in Japanese, this means rejection or some negative reaction. If you see this mark on an examination sheet graded by a teacher, they mean completely the opposite in English and in Japanese. It is a famous story that in the video game consoles like PlayStation, the commands "accept" and "cancel" are assigned to the two buttons '◯' and '×' in completely the opposite way depending on the geographical region it is sold in order to accomodate with this culture difference. Do you really think the proposed notation does not cause confusion for people who are not much familiar with Japanese culture?

  2. I am not sure how computers without Japanese language component handle these marks. Do they display these characters correctly, and without mojibake? For macintosh computers, this should not be a problem. Windows is the most discriminative among the major OSs with respect to language.

  3. (As Boaz Yaniv points out,) in (generative) linguistics, it is a worldwide convention to use

* : ungrammatical

?: does not sound completely natural but should not be considered totally ungrammatical

%: grammaticality depends on the informant

#: grammatically (syntactically) correct, but strange if you consider the meaning (at the pragmatic level), and hence not expected to be useful if used

This is an established convention, and will cause less confusion. I am not proposing that all of these should be used here, but maybe you can use * and ?.

  • You have a point about mojibake, I'm not sure how a default Windows install without Japanese would display them. SE only supports fairly recent browsers though, and the system requirements for the latest browsers require modern operating systems, so that problem would sort itself out over time.
    – Troyen
    Commented Jul 12, 2011 at 17:31
  • 2
    The problem with the linguistic conventions is the same problem as the Japanese (and really any other) markers: some people won't know what they mean. The only time I've seen an * in front of a sentence is when it's a footnote. Whatever symbol is chosen, new people will potentially be confused. My personal opinion is the text of the answer should make it clear whether the example is grammatical or not. (Use ~Y like this: <example>, but not like this: <bad example>)
    – Troyen
    Commented Jul 12, 2011 at 17:35
  • 3
    Your point on × having different meanings in English and Japanese is incorrect; × is not the same as ✓ (check mark) and should never stand for acceptance. (○ can be ambiguous in some settings.) My main reason for proposing the Japanese style (I was not aware that there is more than one style until reading your and Boaz's posts) is my assumption that students of Japanese will tend to encounter this style more than the international style. In my own experience, every textbook I have uses the ○/△/?/× convention. Can you elaborate on why the international set "will cause less confusion"? Commented Jul 12, 2011 at 17:36
  • @Derek In English, when you fill in a form that has boxes to chose, some people puts a checkmark, and some people crosses it. Especially when you are filling in a form using a typewriter, the official way is to type the letter 'x' in the box. Am I wrong? The reason I think '*' does not cause confusion is because the difference among people is whether they know it or not, whereas with '×', it might mean different things. But if you think the convention used in Japanese text book does not cause confusion, I am not particularly against it.
    – user458
    Commented Jul 12, 2011 at 17:57
  • 1
    @sawa: The checkbox is a separate issue that simply requires the user to place some mark inside the box. The shape of the mark is immaterial. In the example you mentioned (an examination paper), × would most certainly indicate an incorrect answer in either cultural setting. Furthermore, in user interface design, × is universally used to mean "close", "cancel", "stop", and "no". So this point is invalid. However, if you have other reasons to believe the international marker set would cause less mental friction, by all means, please provide them. Commented Jul 12, 2011 at 18:18
  • @sawa: Also, lest my comments give rise to confusion, I am not against either set of marks. If the set accepted by linguists is demonstrably better, let's use that one. My goal with this post is consistency. :) Commented Jul 12, 2011 at 18:21
  • Maybe my description about the examination was not correct. To be more accurate, US teachers have tendency to mark the correct answers in some way, whereas Japanese teachers mark the incorrect answers. But for TV game consoles sold in US, the 'X' rather than '◯' button is used for select. I gave the only reason I have about possibility of confusion in my comment above; I have nothing to add. I am happy to go with either way.
    – user458
    Commented Jul 12, 2011 at 18:26
  • 1
    @sawa: "But for TV game consoles sold in US, the 'X' rather than '◯' button is used for select." is peculiar to game consoles, and strikes me as rather strange. In the US, 'X' unequivocally means "no". Commented Jul 16, 2011 at 3:34
  • 1
    As you did in this very post, an asterisk * has to be escaped to avoid the system from interpreting it as a special symbol in Markdown. Because of this, I do not want to use an asterisk to mark an incorrect example on this site. Commented Jul 16, 2011 at 21:41
  • @Tsuyoshi_Ito That makes sense.
    – user458
    Commented Jul 17, 2011 at 2:22
  • 1
    After reading the English Site's FAQ I'm altering my stance to be okay with * and ?, since those are in use there as well. Furthermore, it looks like we are able to have a section explaining whatever notation we pick in the site's FAQ, which I wasn't originally aware of.
    – Troyen
    Commented Aug 25, 2011 at 19:35
  • It's true that some people may not understand symbols like * even though they're relatively standard, but we can solve this problem by providing footnotes to explain exactly what a symbol means when we use it. People who already know can ignore the footnotes, and people who aren't familiar with symbols like * or # can still understand the answer.
    – user1478
    Commented Sep 25, 2013 at 21:23

I think that this is a good convention. It is a pity that some fonts render “○” very small compared to other symbols, but I guess that we have to live with that.

The only other problem I can think of is that the symbols may not make sense to everyone. So I suggest users who use this convention to link to this page.


I suggest the following convention:

  • correct: (no mark). Usually, an example sentence is intended to be grammatically correct. Hence there's no need to mark them.
  • questionable: '?' (U+003F QUESTION MARK)
  • incorrect: '*' (U+002A ASTERISK)


As stated in the other posts, "*" and "?" are a well-established convention, used universally in linguistics. I've seen these two frequently, but have rarely seen "%" and "#" as described in sawa's post.

In particular, I don't think it's worth distinguishing "sometimes correct", "informant-dependent", "semantically strange" in our symbol convention. They are all subsumed by "questionable". The simplicity of using only 2 symbols is clearly advantageous. (If you need to distinguish, just do so in prose.)

"?" will be intuitive for everyone. Admittedly, "*" is not immediately obvious, but I think using the more widely established convention is still preferable. (Anyway, the "Japanese textbook" convention isn't universally obvious either.) Probably a FAQ entry on "What are all those asterisks?" is warranted.

Formatting conventions

The marks go in front of the example text, with no intervening space, and outside any quotes or brackets. For example:

  • ?"Colorless green ideas sleep furiously." (syntactically correct, semantically self-contradictory)
  • */je/ (phoneme sequence that does not occur in modern Japanese)
  • "This is correct."
  • *[nʲi.hon] (foreigners' mispronunciation of the moraic nasal)
  • *⟨つずく⟩ (misspelling)

I think this is a good question and even if we pick something that is not easily typable for everybody it can be one of the jobs of the higher rep editors to add this formatting where they see it lacking.

Besides the geometric symbols we could also employ combinations of bold and italics, which are easy to enter in markdown.

An important part of this in my opinion is how to deal with kanji vs. kana vs. romaji for those who have a degree of proficiency in spoken Japanese but limited skill in kanji.

  • 2
    I am against italics. Because Japanese text is traditionally not written in italics, italic fonts for Japanese are usually hard to read. Bold is fine, but let’s not rely only on font. Commented Jun 10, 2011 at 23:29
  • Oh yes I definitely agree no italics for Japanese script but it could be useful for romaji. Commented Jun 12, 2011 at 12:59

You must log in to answer this question.

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