Why do we call it trace?

For any module $P$, we define $\mathrm{tr}(P)=\sum \mathrm{im}(f)$, where $f$ ranges over all elements of $\mathrm{Hom}(P,R)$， and call it trace.

Why does it have such a name? Does it have any relation to the trace of matrix？

• Why do we call a trace of the matrix trace, btw? – Alexei Averchenko Mar 11 '11 at 12:16
• Like many things in Algebra, the name came from German, "die Spur". It entered English usage [in a 1922 translation of Hermann Weyl's ](jeff560.tripod.com/t.html) Raum, Zeit, Materie. Maybe someone else can provide insight on why the German "Spur" is chosen for the trace of a matrix. – Willie Wong Mar 11 '11 at 12:51

I have always viewed the term trace as used to name $$\displaystyle \operatorname{tr}(P)=\sum_{f:P\to R}f(P)$$ as a reference to the fact that that ideal is the part of $R$ which you can reach from $P$... whatever that may mean :) One can check, for example, that $P$ is a generator of the category of modules iff $\operatorname{tr}(P)=R$, and that makes sense (to me!)
A more serious connection is the following. Suppose $V$ is a finite dimensional vector space over a field $k$, and let $\hom_k(V,k)$ be its dual space. Then there is a natural isomorphism $$\phi:V\otimes\hom_k(V,k)\cong\hom(V,V).$$ On the other hand, we have the usual trace map $\operatorname{tr}:\hom(V,V)\to k$, so we can consider the composition $$\operatorname{Tr}=\operatorname{tr}\circ\phi:V\otimes\hom_k(V,k)\to k,$$ which is also sometimes called a trace map. If you now replace $k$ by ring, $V$ by a left $R$-module, then what you wrote $\operatorname{tr}(P)$ is the image of my $\operatorname{Tr}$: it follows that the trace ideal is the image of the trace map.