8
$\begingroup$

I think this question's answer(s) will be of profound use to the future generation of human beings who happen to stumble upon the website math.stackexchange.com.


What are the differences between $^t$, $^\dagger$, $^*$, $^H$, and $^T$. Further, to which is which associated, and are there some cases where one is used for both?


I'd like to see definitions for each group of superscript symbols and how $T$ the transformation and $A$ the matrix are related. I feel it would be good to have this as an answer here for all eyes to see.

$\endgroup$
  • $\begingroup$ I thing generally $t$ and $T$ are for transpose, $*$ and $H$ are for the conjugate transpose or adjoint, and the dagger for the pseudo-inverse. $\endgroup$ – copper.hat Jun 4 '13 at 3:05
  • $\begingroup$ Is what you're saying is that $A^*$ and $T^*$ convey the same information? $\endgroup$ – Tyquan Pesik Jun 4 '13 at 4:13
  • $\begingroup$ Don't forget ${}^t A$. $\endgroup$ – GEdgar Jun 4 '13 at 12:01
9
$\begingroup$

When $A$ is a matrix,

  • $A^t$ and $A^T$ stand for the transpose
  • $A^*$ (and possibly $A^H$, but I don't remember seeing it used this way) stand for Hermitian adjoint. The Wikipedia article says that $\dagger$ is also used for this purpose, by the kind of people who say bra and ket.
  • $A^\dagger$ stands for the pseudoinverse
$\endgroup$
  • 1
    $\begingroup$ I have seen people using $A^{*}$ for complex conjugate of $A$ and $A^{\dagger}$ for Hermitian adjoint. They even appear in same article. What is what is usually clear in the context. $\endgroup$ – achille hui Jun 4 '13 at 3:10
  • 3
    $\begingroup$ For pseudoinverse, $A^\dagger$ was popular in earlier linear algebra literature, but $A^+$ seems to be more widespread in recent years. Also, some physicists seem to use $A^\dagger$ for Hermitian adjoint. Transposes can be denoted by $A^T$ (most common), $A^t,\,A^\top$ and $\phantom{}^tA$. $\endgroup$ – user1551 Jun 4 '13 at 3:48
  • 1
    $\begingroup$ ... and $A^\prime$. (A habit initiated by Matlab-programmers?) $\endgroup$ – Myself Jun 4 '13 at 4:37
  • 1
    $\begingroup$ You really want $*$ for the adjoint, because if $H$ is a complex Hilbert space, then $B(H)$ is a $C^*$-algebra where the star operation is precisely "take the adjoint". $\endgroup$ – kahen Jun 4 '13 at 10:46
3
$\begingroup$

If you use $A^\sf T$ and $A^\sf H$ respectively for the transpose and Hermitian conjugate of $A$ (which I prefer), then you are free to use $t,T,H,$ and * for other things, which is useful. The notation $A^+$ for the pseudo-inverse is aesthetically in line with this and more modern than $A^\dagger$.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.