Find all eigenvalues associated to the linear transformation $T \in L(\mathbb{K}^n)$ defined by $$T(x_1,\ldots,x_n)=(x_1 + \cdots + x_n, \ldots, x_1 + \cdots + x_n)$$ So, to find all eigenvalues one would normally only find the matrix associated to the linear transformation, and easily find the eigenvalues from there, right? Well I have no idea how to "put" this linear transformation into matrix form.

  • 2
    $\begingroup$ Try to clarify the definition of $T$. What are $x_1+$ and $+x_n$? $\endgroup$ – user66407 Oct 17 '15 at 17:13
  • $\begingroup$ I believe that there are a bit too many commas there $\endgroup$ – AdLibitum Oct 17 '15 at 17:33
  • $\begingroup$ Yes, there were two commas that had no business being there. Thanks. $\endgroup$ – SharkFin Oct 17 '15 at 17:34
  • 1
    $\begingroup$ I don't think I see how this is linear when you have that $c$ there. Doesn't this have $T(0,0,\dots,0)=(c,0,\dots,0) \neq (0,0,\dots,0)$? $\endgroup$ – Ian Oct 17 '15 at 18:17
  • $\begingroup$ @Ian This is a typo introduced by the latest edit (not by OP). I removed it. OP please check that it is correct now. $\endgroup$ – Winther Oct 17 '15 at 18:21

The transformation matrix of $T$ is $$ A=\pmatrix{1\cdots1\\\hspace{-9 mm}\vdots&\hspace{-11 mm}\vdots\\1\cdots1}=\pmatrix{1\\\vdots\\1}\pmatrix{1\cdots1}=vv^T $$ Then the characteristic polynomial of $A$ is $$ p(\lambda)=|\lambda I-A|=|\lambda I-vv^T|=\lambda^n-Tr(vv^T)\lambda^{n-1}=\lambda^n-n\lambda^{n-1}=\lambda^{n-1}(\lambda-n) $$ for $vv^T$ is a Rank-$1$ matrix and all principal minors above $2$ are $0$.

Thus eigenvalues of $A$ are $0$ with multiplicity of $n$ and $n-1$ with multiplicity of $1$.


Denote $e_i$ the $i$-th vector of the standard basis $\cal E$. Then $$ T(e_i)=(1,...,1)=e_1+\cdots +e_n $$ for all $i=1,...,n$. Thus the matrix associated to $T$ with respect to the basis $\cal E$ is the matrix which has $1$ in every place.

  • $\begingroup$ So, the only eigenvalue would be 1, with mulitplicity n? Can one assume that? $\endgroup$ – SharkFin Oct 17 '15 at 17:41
  • $\begingroup$ Not till you convince yourself that the characteristic polynomial of that matrix is $(1-\lambda)^n$. $\endgroup$ – AdLibitum Oct 17 '15 at 17:43
  • $\begingroup$ HINT: clearly it is not $\endgroup$ – AdLibitum Oct 17 '15 at 17:43
  • $\begingroup$ HINT#2: Ask yourself what the rank of that matrix is, and when you have answered that, ask yourself what the dimension of $\ker(T)$ is. $\endgroup$ – AdLibitum Oct 17 '15 at 17:46
  • $\begingroup$ So, the the rank would be 1 and the dimension of ker(T) is n-1? $\endgroup$ – SharkFin Oct 17 '15 at 18:15

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.