3

The following is a minimal example of code that gives me an Argument of \multline* has an extra } error.

\documentclass{article}
\usepackage{amsmath}   
\begin{document}
 \begin{tabular}{c}
  \begin{minipage}{200pt} 
   \begin{multline*}
    \begin{pmatrix}
     a 
     &  
    \end{pmatrix}
   \end{multline*}
  \end{minipage}
 \end{tabular}
\end{document}

Removing the tabular environment kills the error, removing the ampersand kills the error, and changing the multline* environment to $$ kills the error. Is this a known limitation of the environments involved? Something related to having things too deeply nested?

Corentin
  • 9,981
Chris Grant
  • 33

1 Answers1

3

It's not clear why you'd want that level of nesting (eg a p column is rather more natural than a minipage in a c column) but anyway an extra {} gets you back on track:

\documentclass{article}
\usepackage{amsmath}   
\begin{document}
 \begin{tabular}{c}
  {\begin{minipage}{200pt} 
   \begin{multline*}
    \begin{pmatrix}
     a 
     &  
    \end{pmatrix}
   \end{multline*}
  \end{minipage}}
 \end{tabular}
\end{document}
David Carlisle
  • 757,742
  • Replacing the c column + minipage with a p column still gives an error: "Forbidden control sequence found while scanning use of \multline*." The workaround with the braces does work, though. Thanks for the tip. – Chris Grant Jan 08 '14 at 00:07
  • 1
    @ChrisGrant yes, it would:-) (but still it seemed worth saying:-) ams could have defined multline so it would nest, just as tabular can nest but probably they just thought it would never be needed (it would complicate the code somewhat). – David Carlisle Jan 08 '14 at 00:10