My makeindex 2.12 changes all accented words (as they appear in Spanish) found the .idx file into strange words in the .ind file: all accented letters are changed into strange letters. In fact my editor even says that the file .idx contains "characters with unknown encoding".

I am using babel. What am I doing wrong? It should be something simple.

I am using pdflatex: "This is pdfTeX, Version 3.14159265-2.6-1.40.18 (TeX Live 2017)" I am using babel and 60 additional packages.

Extract from the .idx file:

\indexentry {La Montaña del Movimiento!propósito de la serie de libros|hyperpage}{7}

\indexentry {cubo de la física|hyperpage}{8}

\indexentry {cubo de Bronshtein|hyperpage}{8}

\indexentry {cubo!de Bronshtein|hyperpage}{8}

\indexentry {cubo!de la física|hyperpage}{8}

\indexentry {física!mapa de la|hyperpage}{8}

\indexentry {constante gravitacional $G$!física y|hyperpage}{8}

\indexentry {velocidad!de la luz $c$!física y|hyperpage}{8}

\indexentry {acción, quantum de, $\hbar $!física y|hyperpage}{8}

\indexentry {carga!elemental $e$, física y|hyperpage}{8}

\indexentry {constante de Boltzmann $k$!física y|hyperpage}{8}

\indexentry {desafíos!clasificación de los|hyperpage}{9}

The makeindex 2.12 command: makeindex -s motionmountain.ist -o MM-espanol-vol3.ind MM-espanol-vol3.idx

Nothing special in the .ist file.

The resulting .ind file (extracts):

  \item a posteriori \hyperpage{307}

  \item a priori \hyperpage{307}

    \subitem concepto \hyperpage{275}

  \item abeja

    \subitem melífera \hyperpage{121}

  \item abejas \hyperpage{42}

    \subitem y campo eléctrico \hyperpage{38}

  \item aberración

    \subitem crom√°tica \hyperpage{135}

    \subitem de la lente de Seide \hyperpage{179}

  \item absorción

    \subitem color y \hyperpage{252}

    \subitem cuerpo negro y \hyperpage{256}

    \subitem de la luz \hyperpage{252}

    \subitem de ondas de radio \hyperpage{251}

        \subitem resonancia del ciclotrón y \hyperpage{250}

      \item absorción saturable \hyperpage{255}
  • Please, do not ask as in a guess game. To answer what you're doing wrong, show us what you're doing exactly with a minimal working example (working=complete document, compilable as is, that produces the problem; minimal= remove all the code and text except the needed to still have a working document and still show the problem).
    – Fran
    Apr 19, 2020 at 8:27
  • Also say how you compile your document --- via tex, latex, pdflatex, context, ...? Apr 19, 2020 at 17:30
  • MakeIndex has only a dumb ascii sort. Therefore, it's necessary to feed it only the "original" style of input coding, using separately accented letters, such as \'o \~n, etc. xindy is able to handle pre-accented letters, but I'm not an xindy user so can only suggest that you refer to the program documentation to see whether there's any difference in the input format. If you stick with MakeIndex, potentially any index term containing an accented letter might need a sort field to ensure proper alphabetic arrangement. Apr 19, 2020 at 20:43
  • Thank you - another solution is esindex.
    – user64520
    Apr 20, 2020 at 10:59

1 Answer 1


I got a hint from a acquaintance. There is a package esindex that solves the problem. It is new and maintained. It is the simplest way to get an index in Spanish. It makes sure that all accents are properly processes in the idx file, so that makeindex can handle them.

Thank you for those who have helped.

Your Answer

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