What is the smallest value of $n$ for which there are two non-conjugate $n\times n$ matrices which have the same minimal and characteristic polynomials and eigenspaces of equal dimension?

I first thought that the answer would be 4, as we know that for $n<4$ having the same minimal and characteristic polynomial implies similarity, but since we have the additional condition that they must have eigenspaces of equal dimension, I am not sure.

  • $\begingroup$ $7.\phantom{ }$ $\endgroup$
    – user1551
    Oct 22, 2021 at 10:21

1 Answer 1


Let $k$ be an algebraically closed field.

Facts: Let $M$ be a matrix with coefficients in $k$. Then $M$ is conjugate to a block-diagonal matrix where each block is a Jordan block. These blocks, up to reordering, classify the conjugacy classes. Moreover, let $\lambda$ be an eigenvalue of $M$. Then

  • the number of Jordan blocks corresponding to $\lambda$ is the dimension of the eigenspace;
  • the size of the largest block corresponding to $\lambda$ is the multiplicity of $\lambda$ in the minimal polynomial;
  • the sum of the sizes of the blocks corresponding to $\lambda$ is the multiplicity of $\lambda$ in the characteristic polynomial.

Therefore, it is enough to look for the smallest integer $N$ such that there are an integer $k$, and two different $k$-tuples of integers $(n_1,\cdots,n_k)$ and $(m_1,\cdots,m_k)$ such that:

  • $n_1\geq \cdots \geq n_k \geq 1$ and $m_1\geq \cdots \geq m_k \geq 1$;
  • $\sum^k_{i=1} n_i = \sum^k_{i=1} m_i = N$;
  • $n_1 = m_1$;

and to form two matrices with only one eigenvalue having Jordan blocks of sizes $(n_1,\cdots,n_k)$ and $(m_1,\cdots,m_k)$.

It is easily seen that the smallest $N$ is $7$, and that examples of tuples are $(3,3,1)$ and $(3,2,2)$.

An example of a couple of two matrices is the following:

$ \begin{pmatrix} 1&1&0&0&0&0&0\\ 0&1&1&0&0&0&0\\ 0&0&1&0&0&0&0\\ 0&0&0&1&1&0&0\\ 0&0&0&0&1&1&0\\ 0&0&0&0&0&1&0\\ 0&0&0&0&0&0&1\\ \end{pmatrix}$ and

$\begin{pmatrix} 1&1&0&0&0&0&0\\ 0&1&1&0&0&0&0\\ 0&0&1&0&0&0&0\\ 0&0&0&1&1&0&0\\ 0&0&0&0&1&0&0\\ 0&0&0&0&0&1&1\\ 0&0&0&0&0&0&1\\ \end{pmatrix}.$


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.