Re: TeX (was: apostrophe etc.)

From: Markus Kuhn (
Date: Thu Aug 19 1999 - 09:44:22 EDT

Michael Everson wrote on 1999-08-19 10:54 UTC:
> Ar 16:33 -0400 1999-08-18, scrobh John Cowan:
> >Contrary to your doubts on p. 2, TeX typists throughout the world always
> >have used, and continue to use, GRAVE ACCENT doubled to indicate
> >an opening double quotation mark, and APOSTROPHE doubled to
> >indicate a closing double quotation mark. ("Opening" and "closing"
> >here refer to Anglo-American practice.)
> So what are you saying? That TeX typists use an archaic inputting method?
> Or that TeX typists employ appalling typography?

There is nothing wrong with TeX and your paper. John's comment didn't
make sense to me either in relation to what you wrote in your paper.

For those not familiar with TeX (a system dedicated to the production of
scientific papers in the highest possible typographic quality):

TeX has a number of ASCII mnemonics that TeX's input stomach converts
with a little parser automatically into proper high-quality typographic

Examples are:

  - -> hyphen
  -- -> en dash
  --- -> em dash
  ` -> left single quote
  `` -> left double quote
  ' -> right single quote
  ' -> right double quote
  ~ -> no-break space

TeX also allows the user to switch between a normal mode and a
mathematical mode. Math mode starts and ends with a $ character.
In math mode, the following substitutions are used

  - -> minus
  ' -> prime
  '' -> double prime

TeX takes as much care as Unicode of good typography and has as far
as quotes and apostrophes are concerned a very comparable repertoire
of punctuation characters. TeX also has combining characters, and
it is even possible (like in full Unicode implementations) to stack
many combining characters in TeX:

  \"a ->
  \'a ->
  \`a ->
  ... and many more ...

Surely, the use of apostrophe as the input mnemonic for the acute
and right quote was clearly inspired by the glyph shapes that early
VDUs followed, but this has nothing to do with bad typography or
with any other problem. The apostrophe is as much just a mnemonic for
the correct character as --- is a mnemonic for the em dash.

TeX goes even beyond Unicode at least for one interesting special
character: TeX distinguishes between normal spaces and spaces after a
sentence. The latter are allowed to grow a bit larger during paragraph
formatting (unless the \frenchspacing option is activated which
prohibits this), and according to at least US typewriter rules are
represented by a double space on typewriters (not so Europe). TeX
handles this as follows: any space character between a full stop and a
following capital letter is by default interpreted as such an
end-of-sentence space, and in situations where this is not appropriate
(such as in "M. Kuhn"), the writer is supposed to use either "~"
(no-break space) or "\ " (explicit normal space) to tell TeX that this
is not an end-of-sentence space.

I wonder, whether it would be a good idea to add such an end-of-sentence
space to Unicode. I understand that it is primarily an artifact of US
typographic tradition and that Europeans tend not to make any
distinction between spaces that terminate sentences. I could imagine
smart-space algorithms (similar to the smart-quote algorithm) that help
editor users to conveniently select the right space type automagically.


  Donald E. Knuth: The TeXbook. Addison-Wesley, 1986. ISBN 0-201-13448-9.

Unicode gurus should especially be familiar with chapters 2 and 9 as
well as annex C and F of this book. A quick look at the second paragraph
of chapter 2 should convince you that Knuth was knowing what he was
doing and that he is equally familiar with the issues involved as is
Michael. Knuth explicitly warns the reader that the apostrophe character
that he uses as the mnemonic for right quotation mark can either look
like | or / on various video display terminals! Few of those who think
that TeX defines any practice of how the apostrophe should look like in
a font really remember what chapter 2 of the TeX bible had to say on
this originally ...


Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at,  WWW: <>

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:51 EDT