Multiple diacritics on one character

Question

As linguistics student, I have to put more than one thing on a character every now and then.

For example, I need to type ā̆. I thought I could simply do that by typing \={\u{a}}, but then the macron is placed above an empty space in front of the character.

\documentclass{article}\begin{document}\={\u{a}}\end{document}

How can I solve this?

I use MiKTeX 2.9, running on Windows 8.

Answer 1

I suggest you to use XeLaTeX with UTF-8 encoding and fontspec for direct character inputing. Charis SIL and Brill fonts are enhanced so as to handle diacritic stacking. SIL distributes an IPA keyboard which lets you enter diacritics or you can use the Characters Map in windows.

Here's a code sample with output.

\documentclass{article}
\usepackage{fontspec}
   \setmainfont{Charis SIL}

\begin{document}

ẵ̊́ b̥̩̬̪

\end{document}

Answer 2

Using the tipa package, \textipa{\u=a} yields

Answer 3

This is well suited for a stack. Here, I set a \={} atop a \u{a}, such that the baseline of the former is 0.4pt above the nominal baseline. You may tweak that value, as I have done in the second combination, if you want separation between the diacritics.

Other diacritic combinations pose no problem, using the same technique.

\documentclass{article}
\usepackage{stackengine}
\begin{document}
\renewcommand\stacktype{L}
\stackon[0.4pt]{\u{a}}{\={}}\ 
\stackon[2pt]{\u{a}}{\={}}
\end{document}  

If one wanted the macron on the bottom, the one would stack \u{} atop \={a}.

Answer 4

Using @Bernard’s hint, I checked the covington package, and its documentation says: “LaTeX doesn’t give you a convenient way to put two accents on the same letter” and introduces some macros for the purpose. There are special macros for combining acute, grave, or circumflex with macron, and a more general macro twoacc for combining two accents, with a peculiar syntax: “its arguments are in square brackets, not curly brackets, and are separated by |. The first argument is the upper accent (only) and the second argument is the letter with the lower accent indicated.” So in your case you could use, after

\usepackage{covington}

the command

\twoacc[\=|\u{a}]

or have the accents differently nested with

\twoacc[\u|\={a}]

This works, but especially the first one does not give a very impressive result (the diacritics are too far from each other):

(In Latin grammars and dictionaries, the latter form seems to be more common, when indicating that a vowel may be either long or short. And the result for it looks actually better than in my old Latin books.)

Answer 5

Here is a better solution, though not perfect: il you compile with xelatex or lualatex and load the fontspecpackage, you can "nest" the accents command. The only restriction is that the corresponding character does exist in the font. If it doesn't exist a workaround is to use the analog commands for accents in mathmode (e.g. \u for text mode becomes \breve in mathmode. As math letter are set in italic by default, you also have to use the \mathrmcommand – unless you want to use italic.

Below is a code that demonstrates all this:

\documentclass[a4paper, 12pt] {article}%

\usepackage[no-math]{fontspec}%
\setmainfont{Heuristica}

\usepackage{fourier}
\usepackage{covington}

\begin{document}

\begin{tabular}{@{}cc}
Math mode: & Covington: \\[6pt]
  $ \breve{\bar{\mathrm a}} $ & \twoacc[\u|\={a}]
 \end{tabular}\\[6pt]

 With fontspec + Heuristica:\\[6pt]
\verb+\={\u a}+ work:\enspace \={\u a},\\
\verb+\~{\^a}+ and \verb+\~{\u a}+ work: \enspace\~{\^a},\ \~{\u a}\\
as well as \verb+ \^{\~ a}+ and \verb+\u{\~ a}+:\enspace \^{\~ a},\ \u{\~ a}.\\

 \fontspec{Minion Pro}
 With fontspec + Minion Pro:\\[6pt]
\verb+\={\u a}+ doesn’t work: \enspace \={\u a}\enspace,\\
 \verb+\~{\^a}+ and \verb+\~{\u a}+ work:\enspace \~{\^a},\ \~{\u a}\\
but \verb+ \^{\~ a}+ and \verb+\u{\~ a}+ don’t:\enspace\^{\~ a}\enspace, \u{\~ a}~\enspace (the order matters)

\end{document}