raising math symbols

1522 views
Skip to first unread message

Joris

unread,
May 3, 2009, 11:11:39 AM5/3/09
to
Hello,

I'm using the \intercal symbol for transposition, but it is typeset a
bit too low. Is there an equivalent of \raisebox in math mode? Using
something like x^{\raisebox{0.25ex}{$\intercal$}} does not e.g. use
superscript size.

Thanks,

J.

Joris

unread,
May 3, 2009, 11:22:04 AM5/3/09
to

I stuck in an extra \scriptsize and it works fine. But is there
something more elegant?

Enrico Gregorio

unread,
May 3, 2009, 11:52:04 AM5/3/09
to
Joris <pin...@gmail.com> wrote:

I don't find \intercal particularly pleasing in that context; maybe
\mathsf{T} would do better; the standard method to avoid explicit
math style declarations is

\newcommand\T{{\mathpalette\raiseT\intercal}}
\newcommand\raiseT[2]{\raisebox{0.25ex}{$#1#2$}}

$x^\T$

Ciao
Enrico

Joris

unread,
May 3, 2009, 12:57:47 PM5/3/09
to

That works, thanks. And I learned \mathpalette; very useful!

Dan Luecking

unread,
May 4, 2009, 2:39:59 PM5/4/09
to
On Sun, 3 May 2009 08:11:39 -0700 (PDT), Joris <pin...@gmail.com>
wrote:

\intercal is a binary operator, and its vertical position
is designed for that use. If you want to use something
to denote transposition, choose some form of T. In a
pinch, \top also looks very T-like and is certainly
higher than \intercal.

$A^{\mathsf{T}}$ also works fine and looks more like
what you seem to want.

Personally, I use a $A^T$ in my linear algebra course
because that's what the textbook uses, and I have no
problem with it. In my papers I prefer $A^t$ because
that is what I was raised on, and editors don't seem to
care.


Dan
To reply by email, change LookInSig to luecking

Dan Luecking

unread,
May 4, 2009, 3:11:52 PM5/4/09
to

Probably the amount to raise should depend on the style.
In practice, the difference is barely visible (except
in nonscript usage $\T$).

I would start by raising it by its depth. This could be
measured automatically or, for simplicity, just measure
them and hard code the values (converted to ex in case
font size is not 10pt):

Hard coded:
\newcommand\T{%
{\mathchoice
{\raisebox{.45ex}{$\displaystyle{\intercal}$}}
{\raisebox{.45ex}{$\textstyle{\intercal}$}}
{\raisebox{.30ex}{$\scriptstyle{\intercal}$}}
{\raisebox{.23ex}{$\scriptscriptstyle{\intercal}$}}}
}

Automatic:


\newcommand\T{{\mathpalette\raiseT\intercal}}

\newcommand\raiseT[2]{%
\setbox0\hbox{$#1{#2}$}\raise\dp0\box0}

>$x^\T$

Enrico Gregorio

unread,
May 4, 2009, 3:21:48 PM5/4/09
to
Dan Luecking <Look...@uark.edu> wrote:

> On Sun, 3 May 2009 08:11:39 -0700 (PDT), Joris <pin...@gmail.com>
> wrote:
>
> >Hello,
> >
> >I'm using the \intercal symbol for transposition, but it is typeset a
> >bit too low. Is there an equivalent of \raisebox in math mode? Using
> >something like x^{\raisebox{0.25ex}{$\intercal$}} does not e.g. use
> >superscript size.
>
> \intercal is a binary operator, and its vertical position
> is designed for that use. If you want to use something
> to denote transposition, choose some form of T. In a
> pinch, \top also looks very T-like and is certainly
> higher than \intercal.
>
> $A^{\mathsf{T}}$ also works fine and looks more like
> what you seem to want.

That's exactly what I suggested in the first place. The \intercal
symbol has rounded bars which are not suitable for that meaning.

> Personally, I use a $A^T$ in my linear algebra course
> because that's what the textbook uses, and I have no
> problem with it. In my papers I prefer $A^t$ because
> that is what I was raised on, and editors don't seem to
> care.

The only problem with $...^T$ is that with some letters it comes too
far from them; a simple $...^{\!T}$ is usually sufficient.

Ciao
Enrico

Joris

unread,
May 4, 2009, 5:19:17 PM5/4/09
to
On May 4, 3:21 pm, Enrico Gregorio <grego...@math.unipd.it> wrote:

Thank you and Dan for your responses. The reason I don't want to use
T or t (or prime) for transposition is that those symbols have another
meaning in the paper. Thanks for pointing out the issues with
\intercal. It looks fine now, but I'll try \top, also.

Best,

J.

Reply all
Reply to author
Forward
0 new messages