MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

It is well known that, for square matrix $x$ and $y$, we have $\operatorname{tr}(xy)=\operatorname{tr}(yx)$. Here of course the trace of a matrix is just the sum of the elements of the diagonal.

The notion of trace has a lot of generalization. As I know, the most general definition is the following: let $(\mathcal C, \otimes, 1, ^\vee)$ be a rigid symmetric monoidal category, $X$ an object of $\mathcal C$ and $f$ an endomorphism of $X$. Then $\operatorname{tr}(f) \in \operatorname{End}(1)$ is defined by the following composition $$ 1 \longrightarrow X^\vee \otimes X \stackrel{\operatorname{id}_{X^\vee} \otimes f}{\longrightarrow} X^\vee \otimes X \longrightarrow X \otimes X^\vee \longrightarrow 1 $$ So my questions is: it is true, in this generality, that $\operatorname{tr}(f\circ g)=\operatorname{tr}(g \circ f)$, for $f$ and $g$ in $\operatorname{End}(X)$?


share|cite|improve this question
Have you tried drawing the appropriate string diagrams? – Qiaochu Yuan Jul 6 '11 at 21:33
Why specify "square" matrices in particular? The occasion for use of this identity with which I am most familiar depends crucially on the matrices not being square. – Michael Hardy Jul 6 '11 at 21:56
To Qiaochu: I'm completely new to string diagram, so it is possible that my question becomes trivial in that language. I just don't know. To Micheal: what is the trace of a non square matrix? I've always seen the trace of an endomorphism! – Ricky Jul 6 '11 at 22:12
If $X$ is $m \times n$ and $Y$ is $n \times m$, then $\mathrm{Tr}(XY) = \mathrm{Tr}(YX)$. – David Speyer Jul 6 '11 at 22:46
It is also true for Ribbon categories, which are in general not symmetric, but just "braided", see eg. Kassel - Quantum Groups. – Marcel Bischoff Jul 7 '11 at 3:30
up vote 13 down vote accepted

Yes. The string diagram chase can be found on page 8 of this paper by Kate Ponto and Mike Shulman.

share|cite|improve this answer
This is a great paper, but I think the chase given obscures the underlying topological idea a little: you take two wires, feed them into each other, then rotate everything by $180^{\circ}$. I'm not completely sure what structure is necessary to make that particular argument valid, though. – Qiaochu Yuan Jul 7 '11 at 4:38
I think they could have done it that way, by first introducing movie moves which express extranaturality, e.g., one which gives an equation where $g \otimes 1_{x^\ast}$ followed by the counit $x \otimes x^\ast \to I$ equals $1_y \otimes g^\ast$ followed by the counit $y \otimes y^\ast \to I$; here I'm using $g: y \to x$ and $g^\ast: x^\ast \to y^\ast$ denotes the mate of $g$. This interprets the move where you slide the bead labeled $g$ through the bottom critical point, after which $g^\ast$ pops up on the other side. There's probably a lemma or two that needs to be proven first. – Todd Trimble Jul 7 '11 at 10:13

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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