Left and right subscript / superscript

Are you doing this in math mode?

${}_{t}p_{x}$

should work.

The best solution I have seen so far was:

${}_{2}x_{1}$ 

For the more general case of left and right superscript and subscript, one writes:

${}^{1}_{2}x^{3}_{4}$ 

You could use the leftidx package which is capable of producing left and right subscripts and superscripts.

In contrast to using simply consecutive subscripts or superscripts such as in $_{t}p_{x}$ it takes care of adjusting indices to the argument using \vphantom. From its source:

\newcommand\leftidx[3]{% {\vphantom{#2}}#1#2#3% } 

The tensor package will do exactly what you require, and more. For your particular case, you need only type

\tensor[_t]{p}{_x} 

Regards

The amsmath package has a \sideset command which lets you put indices at all four corners of a symbol (even if this symbol itself has limits above/below):

\[ \sideset{_t}{_x}p = \sideset{^*_*}{^*_*}\prod_*^* \] 

\sideset is intended to be used with large operator symbols such as \sum and \prod. The amsmath documentation says: this command is not designed to be applied to anything other than sum-class symbols. _{(Thanks to Stefan Kottwitz for this additional remark.)} For other symbols, have a look at the other answers here.

A simple method I apply - which does away with the need for extra packages - is the following:

~^o\!G_{Al}^i 

Which produces the following:

I had the same problem with the subscript and superscript, sooo I made a package to make my life easier and to other. It's a common problem with actuarial symbol, and it help align the sub and superscript. Plus, I’ve add some shortcut to save time.

The project page and the CTAN.

All you need is the actuarialsymbol package.

At the beginning of the code you have to write

\usepackage{actuarialsymbol} 

For the sub/superscript

\actsymb['subscripLeft']['superscriptL']{<middle>}{'subscriptR'}{'superscriptR'} 

Example of output:

Example of shortcut for actuarial symbol :

I am dissatisfied with the existing options, including \prescript from mathtools, since they do not even make an attempt at fixing the spacing between the symbol and the left superscript. Therefore, based on egreg’s excellent answer here, I recently added the package leftindex to CTAN. It provides the command

\leftindex^{<left superscript>}_{<left subscript>} {<symbol>} 

This measures the negative indentation of the right subscript and indents the left superscript by the same value. This works well for most symbols, with a few exceptions, such as A and L (and P and \Gamma in some fonts other than Computer Modern). Therefore, \leftindex takes a couple of optional arguments. The only one we shall need right now is

\leftindex[<slanting phantom>]^{<left superscript>} _{<left subscript>} {<symbol>} 

If the command receives a <slanting phantom>, it will use this symbol instead of <symbol> to calculate the indentation of the left superscript.

\documentclass{article} \usepackage{leftindex} \begin{document} Good results: $ \leftindex^a_b {f}, \leftindex^* {M}, \leftindex_q^p {H}, \displaystyle \leftindex_{-\infty}^{\infty} {\int} f(x)\, dx $ Bad results: $ \leftindex^* {A}, \leftindex^a_b {L} $ With slanting correction: $ \leftindex[T]^* {A}, \leftindex[I]^a_b {L} $ \end{document} 

If you grow tired of having to write slanting phantoms manually, you can have a look at my other package, SemanTeX, which provides semantic, keyval-based mathematics where notation can be controlled centrally from the preamble. Its left index support is based on leftindex, but slanting phantoms can now be chosen once and for all from the preamble:

\documentclass{article} \usepackage{semantex} \NewVariableClass\MyVar[output=\MyVar] % No slanting phantom needed: \NewObject\MyVar\vf{f} \NewObject\MyVar\vM{M} \NewObject\MyVar\vH{H} % Slanting phantom needed: \NewObject\MyVar\vL{L}[slanting phantom=I] \NewObject\MyVar\vGamma{\Gamma}[slanting phantom=] % Slanting phantom and a \!\! to the % right of the superscript needed: \NewObject\MyVar\vA{A}[ slanting phantom=I, post upper left=\!\!, ] \begin{document} $ \vf[upper left=a,lower left=b] $, $ \vM[upper left=*] $, $ \vH[upper left=p,lower left=q] $, $ \vA[upper left=*] $, $ \vL[upper left=a,lower left=b] $ \end{document} 

A solution I adopt, whenever I'm working in environments where extra packages are unavailable, is to manually adjust the index position by adding spaces...

For example

{}_{a+b}^c Z 

would result in

So, you could do

{}_{a+b}^{\quad c} Z 

which gives

But you have to adjust them "case by case" till you find it satisfactory, for example

{}_{ab+\prod x_i}^{\quad\quad\ c} Z 

Definitely, punk-LaTex, not ideal and not recommended, but sometimes useful when no better solution is available...

I made one little improvement on Henrik L's answer: I replace the ~ by {}

This removed the extra space given by ~

For my usage, this was important as I improved

$\Sigma _{~^{\wt M} \! M^{(m)} }^{0,+}$ 

which appears as

to

$\Sigma _{{}^{\wt M} \! M^{(m)} }^{0,+}$ 

which appears as

(ok, it's a complicated expression, but it's necessary, and this version makes it prettier!).

I forget where I got this. But Chemistry people have the same problem because they do need left superscript and subscript.

So use chemistry

\ce{^a\log b} 

This actually link the left superscript with the log after it. Also the spacing is reduced.

How to reduce the spacing in chemical symbols?

It works just fine here https://saxarona.github.io/mathjax-viewer/?input=%5Cce%7B%5Ea%5Clog+b%7D

But somehow it didn't work in tex stackexchange. I have no idea why