Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

let 's assume we have a self-adjoint operator whose trace is known

$ \mathrm{Tr}(sT)= g(s) $ for a known function $ g(s) $.

My quesiton is, can we recover the operator simply by knowing the trace ?

share|cite|improve this question
What is $s$ in your notation? – Willie Wong Jun 13 '12 at 16:11

No, but we can recover quite a bit of information if we know the eigenvalues.

Any operator $F\colon V\to V$ acting on a vector space $V$, if it is nonsingular, has $n=\mathrm{dim}(V)$ vectors $\vec{v}_{1},\dots,\vec{v}_{n}\in V$ which are "stretched" by $F$: $$ F(\vec{v}_{i})=\lambda_{i}\vec{v}_{i} $$ where $\lambda_{i}$ are scalars.

We have $\mathrm{tr}(F)=\lambda_{1}+\dots+\lambda_{n}$ and $\det(F)=\lambda_{1}(\dots)\lambda_{n}$.

How can we reconstruct the operator? Well, we need the eigenvectors and eigenvalues. Eigenvalues alone don't cut it. This is the spectral theorem for finite dimensional vector spaces.

share|cite|improve this answer

The question will be much more meaningful if $s$ will denote an operator.

Trace of operator is well defined for compact operators. Hence I'll assume that $T\in S^p(H)$ for some $p\in[1,+\infty]$. By Schatten- Von-Neumann theorem there exist isometric isomorphism $$ I:S^p(H)\to S^q(H)^*: T\mapsto (s\mapsto\mathrm{Tr}(sT)) $$ Hence, your function $g$ is a an element of $S^q(H)^*$. The desired operator is $T=I^{-1}(g)$. Well this is not constructive proof, but it shows that such a method does exist.

If $s$ is just a positive constant, then you know very little about $T$, and your function $g$ is necessary linear. Indeed $$ g(s)=\mathrm{Tr}(sT)=s\mathrm{Tr}(T). $$ The only thing that you get from this function is its slope, i.e. $\mathrm{Tr}(T)$. Knowledge of trace of course is not sufficient for reconstructing $T$. See explanation in Alex Nelson's answer.

share|cite|improve this answer
what is $ I^{-1}(g) $ ?? is the inverse function of $ g(s) $ thanks – Jose Garcia Jun 13 '12 at 16:02
$I$ is the mapping defined (as in the displayed math line) by associating to the compact operator $T$ the function $g(s)$. So by definition $I^{-1}$ would map $g(s)$ to the operator $T$. – Willie Wong Jun 13 '12 at 16:06
Willie Wong, I'm not sure abut I remember that read it somwhere. Is it true there exist such a pairing between this classes? – Norbert Jun 13 '12 at 16:09
@Norbert: it is true (if I remember right) if $s$ is interpreted as an element of $S^q(H)$. I am however not sure whether that is the OP's intent. – Willie Wong Jun 13 '12 at 16:11
's' here is a real number $ s >0 $ – Jose Garcia Jun 13 '12 at 18:43

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.