Is there any command for typesetting constants in mathematical formulas? I would imagine something similar to \operatorname or \DeclareMathOperator, but for constants.

I want the constant to be typeset in roman font as in the text mode (with proper spacing and ligatures - this concerns constants denoted by a couple of letters).

  • 3
    the usual advice is to use \mathrm{...}. – barbara beeton Aug 6 '14 at 14:40
  • Use siunitx package. – skpblack Aug 6 '14 at 15:23
  • @DCh -- that's funny. with this test, the two commands produce exactly the same result, with ligatures: \[ x\ \mathrm{iff}\ y\ \textup{iff}\ z_{\textup{iff}}^{\mathrm{iff}} \]. \mathrm is also what's documented in lamport's latex manual. (it can be used only in math mode, but it behaves the same as text roman; there's a reason for using \math... commands instead of \text... commands in math mode. \text... commands should be used in math only within \text{...} according to all the documentation i've read.) – barbara beeton Aug 6 '14 at 16:30
  • @barbara Interesting, maybe depends on typeface used.. The following demonstrates different outputs (pdflatex): \DeclareSymbolFont{AMSb}{U}{msb}{m}{n} % fix [noamsfonts] bug \documentclass[noamsfonts]{amsart} \usepackage[bitstream-charter]{mathdesign} \begin{document} \[ x\ \mathrm{fin}\ y\ \textup{fin} \] \end{document} – DCh Aug 6 '14 at 16:50

(updated to incorporate the comments by @egreg, barbara beaton, and @DCh)

I assume your document uses a roman (serif) font rather than a sans-serif font for mathematics. To ensure that constants are typeset consistently using upright Roman letters, it's handy to create a dedicated macro named, say, \ct that uses the following macros in a nested fashion: \text (from the amsmath package), \rmfamily(just in case the surrounding material is non-roman), and \upshape.

\newcommand\ct[1]{\text{\rmfamily\upshape #1}}

Then, use this macro to typeset an equation such as


Suppose, furthermore, that your documents contains two frequently-occurring constants named ab-cd and fi-fi. (You did say that the names of the constants might contain ligatures...) To help speed up typing, you could define two macros \abcd and \fifi as follows:


The result of a full MWE:

enter image description here

\documentclass[noamsfonts]{amsart}  \usepackage[bitstream-charter]{mathdesign}

%% use the \ct macro to define math constants
\newcommand\ct[1]{\text{\rmfamily\upshape #1}}

%% define two math constants with rather contrived names...

\sffamily  % switch to sans-serif for main text font


not in math mode: ab-cd, fi-fi
| improve this answer | |
  • \mathrm, not \textup. – egreg Aug 6 '14 at 16:53
  • In case of a sans serif or small caps context, \textup would respect it, while \mathrm wouldn't. – egreg Aug 6 '14 at 16:59
  • @egreg -- try the test provided in the second comment by DCh to the question. (i just did.) indeed, as he claims, the results are different -- i suspect he's right that it may depend on what fonts are being used, because i can't think of any other reason. (i tried it with both amsart and article to make sure the former wasn't a factor.) – barbara beeton Aug 6 '14 at 17:03
  • 1
    @egreg -- but the ligatures are made with \mathrm when computer modern is used. that may be unexpected, but when i checked several other strings, including "ffl", the ligatures are clearly present. – barbara beeton Aug 6 '14 at 17:17
  • 2
    @DCh - Instead of "reopening" this four-year-old question -- sadly, very few people will notice that you did so (mainly by revoking the previously-awarded "accept" badge, right?) -- it would have been better if you had posted a new question. The issue you've encountered appears to be sufficiently distinct to merit being raised in a new question; that way, far more people will get to see it and have a chance to think about providing an answer. – Mico Aug 17 '18 at 21:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.