Is there a way to type an empty combining diacritic in LaTeX or Word or Indesign?

Sylph's picture
charles ellertson's picture

Actually, what you want is a spacing modifier, see

Enter it the way you'd enter any Unicode character. There is a difference between the *characters* in spacing modifiers and combining diacriticals, even though they are often the same *glyph* -- a concept sort of at the core of Unicode.

Sylph's picture

I've actually tried that, but it didn't turn out well.

What I need is an inverted breve as a modifier and not a combining diacritic, but I'm not sure it even exists.

Sylph's picture

Perhaps I can combine the inverted breve with an empty space, but I don't know what is the width of the horizontal space between two words, such as in this line, for example.

LexLuengas's picture

In Word or InDesign just copy and paste the unicode character using a font which supports it: ̑.

In LaTeX there are multiple ways to achieve this:

  • LaTeX has a series of built-in diacritical commands. Leaving the argument empty produces the diacritical mark. For instance, \^{} renders as ˆ. Here is a list of such commands.

If you use a XeTeX (putting \usepackage{fontspec} in your preamble):

  • Type the character directly.
  • Type the unicode (hexadecimal) codepoint into the symbol command, with a preceding double apostrophe ''. To produce an inverted breve you would type \symbol{''0311}.
charles ellertson's picture

Yes, I don't know if there is an inverted breve (might also be called "roundcap") in Unicode that isn't ALSO a combining diacritic.

As has been suggested, easy enough to get one to print -- just use the combining diacritic over, say, a thin space between word spaces, but odd things happen to space characters in some applications, such as EPUB.

If the mark can be longer, as with a scansion mark, you could use one of those over a word space.

Not sure why you need it, but things can be found in odd places in Unicode -- like the glottal stops for Guatemalan Mayan in a section generally devoted to Chinese...

It could be that if you described the need, one of the people more familiar with the Cyrillic alphabets could help. I forget which, but either Brill or Gentium is pretty good with extended Cyrillic characters.

Otherwise, while most of us are never in favor of using Unicode characters for a purpose outside their intent -- like a combining diacritic for a spacing modifier -- in this case, you have have no choice.

Michel Boyer's picture

In LaTeX, you can use phantom letters: they are not in the text but the text is typeset as if they were there. You can thus put you combining diacritic on a fictitious letter. Here is an example:


на првоме слогу и са \phantom{a}\symbol{"0311}\, на другоме.


and here is the output:

The \setlength command doubles the interword spacing, so as to look more like your picture; a phantom "a" is used for the combining diacritic and the command \, puts a thin space after because otherwise, it looked ill centered.

Sylph's picture

Wow, thank you, Lex, Charles and Michel!

I knew there was a way to insert symbols via their Unicode numbers! That's what I wanted!


This sort of code

Vin.\ i zvat.\ mn.\ imaju akcente kao i im.\ mn.: \emph{zȗbe : zȗbi}, \emph{knȅzove : knȅzovi}, \emph{púževe} i t.\ d.

\phantom{a}\symbol{"0311} na ˊ u svima padežima osim onih koji će se na po se spomenuti, n.\ p.\ \emph{strȋc} ima u rod.\ jed.\ \emph{stríca}, dat.: \emph{strícu}, vin.: \emph{stríca}, tvor.: \emph{strícem}, skaz.: \emph{strícu}; tako \emph{dvȏr} ima u im.\ mn.\ \emph{dvóri} i \emph{dvórovi}, i t.\ d.


Why the such a big space between the paragraphs?

Michel Boyer's picture

I don't know the why and I usually use \phantom in equations without problems. If you write your phantom so that the letter be a mathematical letter the problem seems to disappear. With the input

Vin.\ i zvat.\ mn.\ imaju akcente kao i im.\ mn.: \emph{zȗbe : zȗbi}, \emph{knȅzove : knȅzovi}, \emph{púževe} i t.\ d.

$\phantom{i}$\symbol{"0311} na ˊ u svima padežim [\ldots]

blabla [\ldots]

I get

\phantom{a} was giving a "left side bearing" that was too large; with Cambria, $\phantom{i}$\symbol{"0311} also looked fine between two words.

For the whys, y0u may try

charles ellertson's picture

Michael, IIRC, isn't there a (sort of dangerous) "squash" command in TeX? It's not named \squash (and it's been over 10 years), but we use to use something like that to hide vertical height.

Michel Boyer's picture

I guess it is \smash. Here is a grab from personal notes (which also reminds us that the alignment of "accents" is culture dependent).

Added: I used an explicit kern to fix the l squared. Here is the source

Let us try with \verb+\vphantom{\othertheta}+ and next with \verb+\rdot+
L\ = \frac{ml^{\kern0.05em2}}{2}\left(\dot{\othertheta}^2 + \sin^2\othertheta \cdot \dot{\vphantom{\othertheta}\othervarphi}^2\right) + mgl\,\cos \othertheta
L\ = \frac{ml^{\kern0.05em2}}{2}\big(\rdot{\othertheta}^2 + \sin^2\othertheta \cdot \rdot{\othervarphi}^2\big) + mgl\, \cos \othertheta

The math font is Fourier whose latin letters come from Utopia

Syndicate content Syndicate content