Let $M_n(K)$ denote the space of all $n×n$ matrices with entries in a field $K$. Fix a non-singular matrix $A=(A_{ij})\in M_n(K)$ and consider the linear map $T:M_n(K)→M_n(K)$ given by:



  1. $trace(T)=n\sum_{i=1}^nA_{ii}$
  2. $trace(T)=\sum_{i=1}^n\sum_{j=1}^nA_{ij}$
  3. $Rank(T)=n^2$ I am confused that in what sense options 1 and 3 are given correct? And then why not option 2 is correct in that sense? Please help.

For (c), since $A$ is non-singular hence $T$ is also non-singular, so nullity$(T)=0$. By rank-nullity theorem rank$(T)=n^2$ which is the dimension of $M_n(K)$.

(a) can be proved by considering a $2\times2$ case and then generalized.

(b) is false because trace involves only diagonal entries.

Another view for (c)

Suppose $$T(X)=T(Y)$$ $$AX=AY$$ $$A^{-1}AX=A^{-1}AY$$ $$X=Y$$ therefore $T$ is one-one, hence rank$(T)$=dim$(M_n(K))$


Consider the matrices $E_{ij}$, having coefficients always vanishing except for coefficient at row $i$ and column $j$ which has value equal to $1$.

$$\mathcal{E}=(E_{1,1}, \dots , E_{1,n},E_{2,1} \dots, E_{2,n},\dots,E_{n,1}, \dots, E_{n,n})$$ is a basis of $M_n(K)$.

You'll verify that for $1 \le i \le n , 1 \le j \le n$, you have $$T(E_{i,j})=\sum_{k=1}^n A_{k,i} E_{k,j}.$$ This is no more no less the way to get the matrix of $T$ in the basis $\mathcal{E}$ where $A=(A_{i,j})$.

To get (1), you can now apply the definition of the trace. The diagonal of the matrix of $T$ in the basis $\mathcal{E}$ is $$(A_{1,1},A_{1,1}, \dots,A_{1,1},A_{2,2},\dots,A_{2,2}, \dots, A_{n,n},\dots, A_{n,n})$$ proving that $\text{tr}(T)=n \sum_{i=1}^n A_{i,i}=n \text{tr}(A)$.

Obviously (2) is conflicting with (1)...

Regarding (3), what is the kernel of $T$? A matrix $X$ is in the kernel of $T$ if and only if $$T(X)=AX=0$$ which means that $A^{-1} A X=X=0$ as $A$ is supposed to be invertible. That means that $T$ is a non-singular linear map of $M_n(K)$ which is of dimension $n^2$. Hence $\text{rank}(T)=n^2$.


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.