How do i use a cyrillic glyph/symbol/character in a math formula?

Question

I am trying to use a cyrillic glyph as a symbol for a proportionality factor. It's the ӄ and its unicode number is 04C4 (see unicode page). So i crawled the english web for information on this, and after about 3 hours i managed to include the glyph in my text with

\usepackage[T2B, T1]{fontenc}   <- so it selects T1 as standard, but has T2B available
\usepackage{lmodern}            <- don't know if this is needed, but i guessed it might me useful for having a font with cyrillic support.

within the document:
`{\fontencoding{T2B}\selectfont \CYRKHK}`

This prints the glyph in ordinary text. In math mode it's possible to get it with a nifty workaround: enclose the above in \text{} (see here. This solution works somehow, but it is really crappy. For instance it doesn't work for lots of other good looking cyrillic letters, like ҟ U+049F (-> "Undefined control sequence." error). And i don't even want to think about, how agonising it would be, to include Japanese glyphs as well.

So the real question is: is there a simple and robust way to include any kind of glyph, i can find in unicode? Say a Katakana like ャ U+30E3, next to a cyrillic glyph, next to a Bengali ন U+09A8. Typing this into stackexchange without any problem raises an obvious second question: is Firefox actually better at typesetting unicode symbols, than latex? O.o

What i've tried, that didn't work somehow:

• using \DeclareUnicodeCharacter{<number>}{\<command to place the thing>} with \usepackage[utf8]{inputenc} -> \DeclareUnicodeCharacter isn't defined for some reason (beyond my latex horizon). There was another method with \newunicodecharacter or something like that (can't remember). That did nothing at all.

• using XeTeX -> no change, exept that other stuff broke. I returned to pdflatex. So this would not work.

• using the package mathtext -> nothing happend

• using the package cyrillic -> fails to load, even though it is installed with TeXlive.

My current configuration:

\RequirePackage{fix-cm}
\documentclass[12pt,index=totoc,bibliography=totoc,ngerman, BCOR=12mm]{scrreprt}

%header
\usepackage[T2B, T1]{fontenc}
\usepackage{lmodern}
\usepackage[utf8]{inputenc}
\usepackage[ngerman]{babel}

\usepackage[reqno]{amsmath}
\usepackage{amssymb}

\usepackage[svgnames,table,hyperref]{xcolor}
\definecolor{dblue}{rgb}{0 0.447 0.741}
\usepackage{lipsum}
\usepackage{graphicx}

\usepackage{listing}
\usepackage{matlab-prettifier}
\usepackage{rotating}
\usepackage{flafter}
\usepackage{subfig}
\usepackage{makeidx}
\usepackage[german,noprefix]{nomencl}
\usepackage[format=hang,labelformat=simple,font={small,sf},labelfont={small,sf},textfont={small,sf}]{caption}
\usepackage{booktabs}
\usepackage{setspace}
\usepackage{subfloat}
\usepackage[colorlinks,linkcolor=dblue,pagebackref]{hyperref}

Answer 1

If you're not short of math symbol fonts, you can define a new one:

\documentclass{article}
\usepackage[X2,T1]{fontenc}

\DeclareSymbolFont{cyrillic}{X2}{cmr}{m}{n}
\SetSymbolFont{cyrillic}{bold}{X2}{cmr}{bx}{n}

\DeclareMathSymbol{\khk}{\mathord}{cyrillic}{139}

\begin{document}

$a+\khk(b)_{\khk(b)}$

\end{document}

The advantage of the X2 encoding is that it covers all (well, almost all) glyphs in the various Cyrillic encodings.

If you're short of symbol fonts, a less efficient solution is to use \text.

\documentclass{article}
\usepackage[X2,T1]{fontenc}
\usepackage{amsmath}

\DeclareRobustCommand{\khk}{%
  \mathord{\text{\usefont{X2}{cmr}{m}{n}\symbol{139}}}%
}

\begin{document}

$a+\khk(b)_{\khk(b)}$

\end{document}

You can change the type of the math atom, if it is not \mathord, into something else (\mathrel, if it should be a relation symbol).

You can also use directly the letter in the formulas, with UTF-8. This is for the second solution, the first one will work the same.

\documentclass{article}
\usepackage[X2,T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{amsmath}

\DeclareUnicodeCharacter{04C3}{\khk}

\DeclareRobustCommand{\khk}{%
  \mathord{\text{\usefont{X2}{cmr}{m}{n}\symbol{139}}}%
}

\begin{document}

$a+Ӄ(b)_{Ӄ(b)}$

\end{document}

Sorry, no Japanese or Bengali, this way. Think to your readers who would have a hard time in deciphering those symbols.

Answer 2

Got it to work with XeLaTeX and a unicode-supporting font -- while I prefer libertinus (and it does beautiful math-mode Cyrillic and some Middle-Eastern scripts), it does not have good Asian coverage so for this I used my fallback Unifont. All you need to do get your glyph to display in math mode is to wrap it in \hbox (or even, apparently, \mathrm -- does someone know why?). To make it size properly with subscripts, we provide definitions with \mathchoice.

\documentclass{article}
\usepackage{fontspec}
\setmainfont{Unifont} % \usepackage{libertinus} has no Asian

\newcommand{\MM}[1]{\mathchoice{\hbox{#1}}{\hbox{#1}}{\hbox{\scriptsize{#1}}}{\hbox{\tiny{#1}}}}

\begin{document}

blah $\hbox{x}$ blah $a\MM{ャ}T = 2নy+3\hbox{ж}^5-x_{\MM{ж}}$ blah ャনж

\begin{equation}
  d_0 \mathbf{i}\mathrm{নT} = \forall 2x \MM{ャ}
\end{equation}

\end{document}

I also briefly tried the package unicode-math but 1. It only works with OT math fonts, of which 2. None seem to have Asian support.