I have this numbering problem for days, checking this website to find these https://tex.stackexchange.com/a/25212/34618 and https://tex.stackexchange.com/a/5694/34618

but these are not what I want really, because their directions would means to renew many commands for user-defined theorems, lemmas, exercises and problems, not to mention other default setting that \thechapter hack carries over to equations, figures and tables.

Here is my question.

In the chapter running head, I have

I. Chapter for heading
and what follows are many sections in the chapter body.

II. Chapter for heading
and many sections ...

III. Chapter for heading
and many sections ...

In TOC, I would like the following style.

A. Chapter name,
1.1 section,
1.2 section,

B. Chapter name,
2.1 section,
2.2 section,

C. Chapter name,
3.1 section,
3.2 section,

In short, all numbering styles (Romanic, alphabetic, numeral) are different. I can figure out how to tackle the heading problem by redefining command @makechapterhead but don't know how to deal those in TOC. Flipping through definition of l@chapter, numberline, etc, never hit me with how, thus needing guru helps.

I can foresee solving this problem may benefit different numbering styles available for non-English languages.

PS. I also know some prevailing packages may achieve this goal, but those are not what I want either.


1 Answer 1


At first, I considered this numbering schema as confusing since the same object is being denoted using three different representations: Roman, Alphabetic, Arabic. However, in a comment to my answer, Jesse has explained that this is for a Chinese text in which this schema is logically sound.

Here's a possible solution:


\def\@chapter[#1]#2{\ifnum \c@secnumdepth >\m@ne
  {\parindent \z@ \raggedright \normalfont
    \ifnum \c@secnumdepth >\m@ne
    \vskip 30\p@




\chapter{First test chapter}
\section{Test section}
\section{Test section}
\section{Test section}
\chapter{Second test chapter}
\section{Test section}
\section{Test section}
\section{Test section}


An image of the resulting TOC:

enter image description here

An image of a chapter in the document body:

enter image description here

  • 1
    The initial notions was intended to use (Roman, alphabetic, arabic) to represent (壹,壹,1.1) in Chinese context, but I expanded a little bit to (壹,一,1.1) to generalize the notions when posted my question. Literally in Chinese 壹 = 一 = 1. Furthermore, in order not to confuse the audience with Chinese characters, I use notational expression (I,A,1.1) instead. Yes, it sounds odd to English, but in Chinese, it is logically sound. Sorry for the confusion. Indeed, curiosity lures me to ask because I spent days in vain, eager to known an answer. Thanks for the speedy response
    – Jesse
    Aug 5, 2013 at 0:48
  • @Jesse Ah, I see. It's confusing not only in English but also in Spanish (my native tongue); thanks for explaining that this was for a Chinese text and that, in this context, this is logically sound. I'll add a little note about this to my answer, so other readers can also understand the context. Aug 5, 2013 at 0:52
  • I achieved the intended objective (壹,一,1.1) with xeCJK and zhnumber packages easily, thus, whishing to express my thanks toward you for sharing the info and the litte note. Thank you very much.
    – Jesse
    Aug 5, 2013 at 4:44

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .