A matrix is diagonalizable iff it has a basis of eigenvectors. Now, why is this satisfied in case of a real symmetric matrix ?

  • 9
    $\begingroup$ See Proof 2: maecourses.ucsd.edu/~mdeolive/mae280a/lecture11.pdf $\endgroup$
    – Amzoti
    Commented Sep 3, 2013 at 1:14
  • $\begingroup$ IIRC, this is proved in Sheldon Axler's Down with Determinants, a very nice article. I would link to it, but I'm on my phone. $\endgroup$
    – user641
    Commented Sep 3, 2013 at 5:19
  • 3
    $\begingroup$ An important condition is missing in the question: the stated property only holds for real symmetric matrices (and diagonalisation is then possible over the real numbers, not requiring complex eigenvalues or coordinates for eigenvectors). $\endgroup$ Commented Aug 22, 2019 at 7:44
  • $\begingroup$ Possible duplicate of math.stackexchange.com/questions/82467/… $\endgroup$ Commented Apr 19, 2021 at 6:45
  • 2
    $\begingroup$ closing this question seems unnecessary to me. the question may be short, but it is precise and to the point. $\endgroup$
    – Uzu Lim
    Commented Sep 23, 2021 at 11:07

4 Answers 4


Suppose the ground field is $\mathbb C$. It is immediate then that every square matrix can be triangulated. Now, symmetry certainly implies normality ($A$ is normal if $AA^t=A^tA$ in the real case, and $AA^*=A^*A$ in the complex case). Since normality is preserved by similarity, it follows that if $A$ is symmetric, then the triangular matrix $A$ is similar to is normal. But obviously (compute!) the only normal triangular matrix is diagonal, so in fact $A$ is diagonalizable.

So it turns out that the criterion you mentioned for diagonalizability is not the most useful in this case. The one that is useful here is: A matrix is diagonalizable iff it is similar to a diagonal matrix.

Of course, the result shows that every normal matrix is diagonalizable. Of course, symmetric matrices are much more special than just being normal, and indeed the argument above does not prove the stronger result that symmetric matrices are orthogonaly diagonalizable.

Comment: To triangulate the matrix, use induction of the order of the matrix. For $1\times 1$ it's trivial. For $n\times n$, first find any arbitrary eigenvector $v_1$ (one such must exist). Thinking of the matrix as a linear transformation on a vector space $V$ of dimension $n$, write $V$ as $V=V_1\oplus W$, where $V_1$ is the subspace spanned by $v_1$. Then $W$ is $n-1$-dimensional, apply the induction hypothesis to $A|_{W}$ to obtain a base $v_2,\ldots, v_n$ in which $A|_W$ is triangular. It now follows that in the base $v_1,\ldots, v_n$ $A$ is triangular.

  • 1
    $\begingroup$ @IttayWeiss Could you (or anyone else) comment on why it's immediate that every square matrix can be triangulated? I'm not sure that part is as obvious to me as it should be. $\endgroup$
    – littleO
    Commented Sep 3, 2013 at 5:46
  • 1
    $\begingroup$ @IttayWeiss thanks! $\endgroup$
    – littleO
    Commented Sep 3, 2013 at 5:56
  • 2
    $\begingroup$ This argument is rather problematic; notably there is no such thing as restricting a matrix to a subspace. If the subspace is stable under the linear transformation given by the matrix (for which there seems to be no reason here) then one can choose a basis of the subspace and express the restriction of the linear transformation on that basis, giving a smaller square matrix. However that matrix highly depends on the choice of basis. Also, there is no reason why if $A$ is symmetric the smaller matrix would also need to be so, so applying induction fails rather miserably. $\endgroup$ Commented Aug 21, 2019 at 15:03
  • 3
    $\begingroup$ I admit I forgot that your last paragraph is about square, not symmetric or normal, matrices; excuse me. However what I said about restriction (being not defined) still applies. Also there is more wrong with this answer than would fit in my comment. Notably normality is not preserved under similarity: $-1~1\choose~0~~1$ is a reflection (so similar to a diagonal, normal matrix) but not normal. Symmetry does not imply normality (only works for real matrices). And normality directly implies diagonalisable with orthogonal eigenspaces, not need to triangularise before applying that. $\endgroup$ Commented Aug 22, 2019 at 7:41
  • 1
    $\begingroup$ @Ittay Weiss What if $A(W)\not\subset W$? $\endgroup$ Commented Sep 13, 2020 at 16:20

Here's some intuition (but not a rigorous proof).

If $A$ is hermitian (with entries in $\mathbb C$), you can easily show that the eigenvalues of $A$ are real and that eigenvectors corresponding to distinct eigenvalues are orthogonal.

Typically, all the eigenvalues of $A$ are distinct. (It is in some sense a huge coincidence if two eigenvalues turn out to be equal.) So, typically $A$ has an orthonormal basis of eigenvectors.

Even if $A$ has some repeated eigenvalues, perturbing $A$ slightly will probably cause the eigenvalues to become distinct, in which case there is an orthonormal basis of eigenvectors. By thinking of $A$ as a limit of slight perturbations of $A$, each of which has an ON basis of eigenvectors, it seems plausible that $A$ also has an ON basis of eigenvectors.

  • 7
    $\begingroup$ I really love this idea, however it seems very tricky to turn it into a readable proof: 1) the traditional perturbation technique i.e. $+\epsilon I$ doesn't work here and it seems rather murky to have any simple perturbation path that can keep all eigenvalues distinct; 2) eigenvectors being orthogonal isn't a topological concept and what would be a nice way to associate it continuously with entries of the matrices? $\endgroup$
    – Vim
    Commented Sep 17, 2017 at 2:19
  • 1
    $\begingroup$ The orthogonality of eigenvectors seems to be the wrong thing to focus on when there are non-simple eigenvalues, as there is too much freedom in choosing eigenvectors. Rather consider orthogonality of distinct eigenspaces: that is easy to establish in general,, and once this is done it suffices to choose an orthonormal basis in each eigenspace (which obviously can be done) to get an orthonormal basis of eigenvectors. $\endgroup$ Commented May 25, 2021 at 6:33

This question is about the spectral theorem for (finite dimensional) real Euclidean spaces, which says that in such a space any self-adjoint operator is diagonalisable (over the real numbers) with mutually orthogonal eigenspaces (so that orthonormal bases of eigenvectors exist). This is of course a classic result that should be proved in any course on the subject, so you can look this up in any textbook. However there are quite a few questions on this site that more or less touch on this matter, yet no really satisfactory answers, so I thought this is as good a place as any for me to state what seems a good proof to me (the one I teach in my course).

Before starting, I would like to note this is a subtle result, as witnessed by the fact that it becomes false if one replaces the real numbers either by the rational numbers or by the complex numbers (the latter case can be salvaged by throwing in some complex conjugation, but I want to avoid the suggestion that the validity of the stated result somehow depends on that). Nonetheless there are a number of relevant considerations that are independent of the base field, which easy stuff I will state as preliminaries before we get to the meat of the subject.

First off, the matrix formulation in the question is just a restatement, in terms of the matrix of the operator with respect to any orthonormal basis, of the result I mentioned: under such expression the adjoint operator gets the transpose matrix, so a self-adjoint operator gets represented by a symmetric matrix. Since the basis used for such an expression bears no particular relation to our operator or the problem at hand (trying to find a basis of eigenvectors) it will be easier to ignore the matrix and reason directly in terms of the operator. The translation by expression in terms of matrices is easy and I will leave this aside; for reference, the claim I am proving translates to the claim that for every real symmetric matrix$~A$, there is an orthogonal matrix $P$ (whose columns describe an orthonormal basis of eigenvectors) such that $P^{-1}AP=P^tAP$ is a diagonal matrix. It is worth noting that the converse is true and obvious: if $D$ is diagonal and $P$ orthogonal, then $A=PDP^t$ is symmetric (since $D^t=D$).

A basic fact about adjoints is that for any operator $\phi$ on a Euclidean vector space$~V$, whenever a subspace $W$ is stable under$~\phi$, its orthogonal complement $W^\perp$ is stable under its adjoint$~\phi^*$. For if $v\in W^\perp$ and $w\in W$, then $\langle w\mid \phi^*(v)\rangle=\langle \phi(w)\mid v\rangle=0$ since $\phi(w)\in W$ and $v\in W^\perp$, so that $\phi^*(v)\in W^\perp$. Then for a self-adjoint operator $\phi$ (so with $\phi^*=\phi$), the orthogonal complement of any $\phi$-stable subspace is again $\phi$-stable.

Now our focus will be on proving the following fact.

Lemma. Any self-adjoint operator $\phi$ on a real Euclidean vector space$~V$ of finite nonzero dimension has an eigenvector.

Assuming this for the moment, one easily proves our result by induction on the dimension. In dimension$~0$ the unique operator is diagonalisable, so the base case it trivial. Now assuming $\dim V>0$, get an eigenvector $v_1$ by applying the lemma. The subspace $W=\langle v_1\rangle$ it spans is $\phi$-stable by the definition of an eigenvector, and so $W^\perp$ is $\phi$-stable as well. We can then restrict $\phi$ to a linear operator on $W^\perp$, which is clearly self-adjoint, so our induction hypothesis gives us an orthonormal basis of $W^\perp$ consisting of eigenvectors for that restriction; call them $(v_2,\ldots,v_n)$. Viewed as elements of $V$, the vectors $v_2,\ldots,v_n$ are eigenvectors of$~\phi$, and clearly the family $(v_1,\ldots,v_n)$ is orthonormal. It is an orthonormal basis of eigenvectors of$~\phi$, and we are done.

So that was the easy stuff, the harder stuff is proving the lemma. As said, we need to use that the base field is the real numbers. I give two proofs, one in purely algebraic style, but which is based on the fundamental theorem of algebra and therefore uses the complex numbers, albeit in an indirect way, while the other uses a bit of topology and differential calculus but avoids complex numbers.

My first proof of the lemma is based on the fact that the irreducible polynomials in $\def\R{\Bbb R}\R[X]$ all have degree at most $2$. This comes from decomposing the polynomial$~P$ into a leading coefficient and monic linear factors over the complex numbers (as one can by the fundamental theorem of algebra); any monic factor not already in $\R[X]$ must be $X-z$ with $z$ a non-real complex number, but then $X-\overline z$ is relatively prime to it and also a divisor of$~P$, as is their product $(X-z)(Z-\overline z)=X^2-2\Re z+|z|^2$, which lies in $\R[X]$.

Given this we first establish (without using self-adjointness) in the context of the lemma the existence of a $\phi$-stable subspace of nonzero dimension at most$~2$. Start taking any monic polynomial $P\in\R[X]$ annihilating$~\phi$; the minimal or characteristic polynomial will do, but we really only need the existence of such a polynomial which is easy. Factor $P$ into irreducibles in $\R[X]$, say $P=P_1P_2\ldots P_k$. Since $0=P[\phi]=P_1[\phi]\circ\cdots\circ P_k[\phi]$, at least one of the $P_i[\phi]$ has nonzero kernel (indeed the sum of the dimensions of their kernels is at least $\dim V>0$); choose such an$~i$. If $\deg P_i=1$ then the kernel is a nonzero eigenspace, and any eigenvector in it spans a $\phi$-stable subspace of dimension$~1$ (and of course we also directly get the conclusion of the lemma here). So we are left with the case $\deg P_i=2$, in which case for any nonzero vector $v$ of its kernel, $v$ and $\phi(v)$ span a $\phi$-stable subspace of dimension$~2$ (the point is that $\phi^2(v)$ lies in the subspace because $P_i[\phi](v)=0$).

Now that we have a $\phi$-stable subspace$~W$ of nonzero dimension at most$~2$, we may restrict $\phi$ to $W$ and search an eigenvector there, which will suffice. But this means it suffices to prove the lemma with the additional hypothesis $\dim V\leq2$. Since the case $\dim V=1$ is trivial, that can be done by showing that the characteristic polynomial of any symmetric real $2\times2$ matrix has a real root. But that can be shown by showing that its discriminant is non-negative, which is left as an easy exercise (it is a sum of squares).

(Note that the final argument shows that for $\phi$ self-adjoint, the case $\deg P_i=2$ actually never occurs.)

Here is a second proof of lemma, less in the algebraic style I used so far, and less self-contained, but which I find more intuitive. It also suggests a practical method to find an eigenvector in the context of the lemma. Consider the real function$~f:V\to\R$ defined by $f:x\mapsto \langle x\mid \phi(x)\rangle$. It is a quadratic function, therefore in particular differentiable and continuous. For its gradient at $p\in V$ one computes for $v\in V$ and $h\in\R$: $$ f(p+hv)=\langle p+hv\mid \phi(p+hv)\rangle =f(p)+h\bigl(\langle p\mid \phi(v)\rangle +\langle v\mid \phi(p)\rangle\bigr) +h^2f(v) \\=f(p)+2h\langle v\mid \phi(p)\rangle+h^2f(v) $$ (the latter equality by self-adjointness), so that gradient is $2\phi(p)$. Now the point I will not prove is that the restriction of $f$ to the unit sphere $S=\{\,v\in V\mid\langle v\mid v\rangle=1\,\}$ attains its maximum somewhere on that sphere. There are many ways in analysis of showing that this is true, where the essential points are that $f$ is continuous and that $S$ is compact. Now if $p\in S$ is such a maximum, then every tangent vector to$~S$ at$~p$ must be orthogonal to the gradient $2\phi(p)$ of $f$ at$~p$ (or else one could increase the value of $f(x)$ near $p$ by varying $x$ along $S$ in the direction of that tangent vector). The tangent space of$~S$ at $p$ is $p^\perp$, so this statement means that $\phi(p)\in(p^\perp)^\perp$. But we know that $(p^\perp)^\perp=\langle p\rangle$, and $\phi(p)\in\langle p\rangle$ (with $p\neq 0$) means precisely that $p$ is an eigenvector of$~\phi$. (One may check it is one for the maximal eigenvalue of$~\phi$.)

  • $\begingroup$ Thanks a lot for this complete answer. Regarding the proof of lemma I was wondering if the following argument is also valid: by the FTA the characteristic polynomial of $\phi$ splits over $\mathbb{C}$. Since $\phi$ is self-adjoint, all roots of the characteristic poylnomial must be real. Hence the characteristic polynomial of $\phi$ splits over $\mathbb{R}$. $\endgroup$
    – Alphie
    Commented Sep 13, 2021 at 15:39
  • $\begingroup$ @Alphie: Yes, you can invoke the complex numbers, develop some theory of Hermitian vector spaces (including self-adjoin operators), and invoke the Fundamental Theorem of Algebra. This avoids some technicalities. But as I said in the answer: I want to avoid the suggestion that the validity of the stated result somehow depends on that. But note that my first proof of the lemma does invoke FTA, so it is close (but does not involve complex vector spaces). $\endgroup$ Commented Sep 14, 2021 at 3:52
  • $\begingroup$ Shouldn't it be $\langle p+hv\mid \phi(p+hv)\rangle= f(p) +\cdots + h^2 f(p)$, not $\phi(p)+\cdots +h^2\phi(v)$? $\endgroup$
    – user760
    Commented Apr 22, 2023 at 17:32
  • $\begingroup$ Also, when $\dim V=2$, the characteristic polynomial of symmetric $A_{2\times 2}=\begin{pmatrix} a & c\\ c & b \end{pmatrix}$ with real entries is $$p_A(\lambda)=\det (A-\lambda)=\det \begin{pmatrix} a-\lambda & c\\ c & b-\lambda \end{pmatrix}=\left(\lambda-\frac{a+b}{2}\right)^2-\frac{(a-b)^2}{4}-c^2.$$ So it splits. But $\det A$ is not a sum of squares. $\endgroup$
    – user760
    Commented Apr 23, 2023 at 2:46
  • 1
    $\begingroup$ @user760 Thanks for the correction suggested in the first comment, which I have applied. As for the second comment, it is not $\det A$ that is a sum of squares, but the discriminant of the characteristic polynomial (in your notation it is $(a-b)^2+(2c)^2$). $\endgroup$ Commented Apr 23, 2023 at 18:35

Let's assume, for simplicity, that our $n\times n$ symmetric matrix $A$ has $n$ distinct eigenvalues.

Step 1:

The eigen-vectors are orthogonal.

Say $v$ and $w$ are two eigen-vectors with distinct eigen-values $\alpha$ and $\beta$. We have

$$ v'Aw = v'\beta w = \beta v'w $$

and the above is also equal to

$$ v'Aw = (Aw)'v = w'A'v = w'Av = w' \alpha v = \alpha v'w $$

hence $\beta v'w = \alpha v'w$ and, since by assumption $\alpha$ and $\beta$ are distinct, this implies $v'w=0$

Step 2:

Notice that, if $v$ is an eigen-vector of $A$, then also $k v$ for $k\in \mathbb{R}$ is.

We can thus assume without loss of generality that the eigen-vectors $v^1, \dots, v^n$ are orthonormal and $P=(v^1, \dots, v^n)$ is orthogonal.

Consider the diagonal matrix of the eigen-values

$$ D = \text{diag}(\alpha^1, \dots, \alpha^n) = \begin{pmatrix} \alpha^1 & 0 & \dots & 0 \\ 0 & \alpha^2 & \dots & 0 \\ 0 & 0 & \ddots & 0 \\ 0 & 0 & \dots & \alpha^n \\ \end{pmatrix} $$

Since $Av^j = \alpha^j v^j$, the $i$th component is $ (Av^j)_i = \alpha^j v_i^j $. We therefore have

$$ (AP)_{ij} = \sum_k A_{ik}P_{kj} = \sum_k A_{ik}v_{k}^j = (Av^j)_i = \alpha^j v_i^j $$

On the other hand, the matrix $PD$ is given by

$$ (PD)_{ij} = \sum_k P_{ik}D_{kj} = P_{ij}\alpha^j = \alpha^j v_i^j $$

This proves that $AP = PD$. By orthogonality we know $P'=P^{-1}$, thus multiplying on the right by $P'$ we get

$$ A = PDP' $$

  • 6
    $\begingroup$ If you assume, there are n distinct eigen values, shouldn't you also explain what happens when there are repeated eigen values? $\endgroup$ Commented Jun 24, 2020 at 2:38
  • $\begingroup$ @NagabhushanSN: of course! this is not a rigorous prove, rather an exercise based on a simplifying assumption. The objective is to convey the intuition for the result in the easiest possible way $\endgroup$
    – Abramo
    Commented Jun 26, 2020 at 17:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .