Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

This is an elementary question which is do-able by hand but I am actually looking for suggestions or book references since I am sure that someone did this somewhere:

suppose $$ A = \left( \begin{array}{cc} a_{11} & a_{12} \\ a_{21} & a_{22} \\ \end{array} \right) \in M_2(\mathbb{C}). $$

Find $g=(g_{ij})\in GL(2,\mathbb{C})$ so that $gAg^{-1}$ is upper triangular.

One method is to explicitly write down $gAg^{-1}$ and set the function in 2nd row, 1st column equal to zero (which is $-a_{12} g_{21}^2 + g_{22} (a_{11} g_{21} - a_{22} g_{21} + a_{21} g_{22}) = 0 $) and attempt to find $g$ this way, while a second method is to find the eigenvalues of $A$ (the two eigenvalues may or may not be distinct) and find their eigenvectors.

Wiki recommends Linear Algebra Done Right by Sheldon Axler and I think Sheldon proves that any $A\in M_n(\mathbb{C})$ can be put into an upper triangular form using induction.

Either of the methods that I mentioned above seems to be quite messy if I want to explicitly write down such $g$ for any $A\in M_2(\mathbb{C})$, or even for any $A\in M_n(\mathbb{C})$.

Do you have any recommended approach or references because I would like to explicitly write down $g\in GL(n,\mathbb{C})$ so that $gAg^{-1}$ is upper triangular, where $A\in M_n(\mathbb{C})$.

share|cite|improve this question
For just the $2$-by-$2$ case, note that $B\in M_2$ is upper triangular if and only if $e_1$ is an eigenvector of $B$. So if you find a single eigenvector $v$ for $A$, and let $g^{-1}$ be any invertible $2$-by-$2$ matrix whose first column is $v$, then $gAg^{-1}$ is upper triangular. – Jonas Meyer Jun 30 '12 at 3:06
There is no reason for the resulting $g$ not to be messy... For example, one of the columns (or rows, I don't know) of $g$ is an eigenvector of $A$, and writing an eigenvector in terms of the entries of a generic matrix is a rather complicated business! – Mariano Suárez-Alvarez Jun 30 '12 at 3:07
Thanks Jonas and Mariano! I'll just bite my tongue and crank through for the case $n>2$... I'm just curious how bad (or nice) this $g$ can get. – math-visitor Jun 30 '12 at 3:12
(A little reflection shows that if you could find a $g$ whose entries are, say, rational functions on the entries of a generic $n$-by-$n$ matrix $A$, then you could then find formulas for the roots of polynomials of degree $n$ which are rational functions on the coefficients of the polynomial, and so on. We know such a feat is simply not possible) – Mariano Suárez-Alvarez Jun 30 '12 at 3:13
up vote 1 down vote accepted

An algorithm for finding $g$, with the added condition of taking $g$ to be unitary, is given in Hogben's Handbook of linear algebra.

share|cite|improve this answer
Thanks Jonas! I'll take a look at that. =) – math-visitor Jun 30 '12 at 3:14

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.