Let $E$ be a vector space endowed with a non-degenerate symmetric bilinear form. Show $\dim F+\dim F^{\perp}=\dim E=\dim\left(F+F^{\perp}\right)+\dim\left(F\cap F^{\perp}\right)$

Lang uses this formula in proposition 1.2, page 573 of his book "Algebra" (Graduate).

I did not find a straight forward argument to show it, the only argument I found was really long, and not very enlightening. But I assume if he stated it without proof

Here is what I want to show:

Let $E$ be a finite-dimensional vector space over a field K.
Let $g$ be a symmetric/alternating/hermitian form (it suffices to me if you prove it for symmetric and hopefuly I can go from there).
Assume g is non-degenerate.
Let F be a subspace of E.
I want to show the following: $ \dim F+\dim F^{\perp}=\dim E=\dim\left(F+F^{\perp}\right)+\dim\left(F\cap F^{\perp}\right) $
It may be that not all hypothesis are necessary, he just uses this formula in a proof but doesn't derive it.

Thanx to anyone who helps! cheers

  • 3
    $\begingroup$ In $\LaTeX$, you may type \dim for $\dim$. $\endgroup$
    – user1551
    Nov 27, 2013 at 19:04

2 Answers 2


For arbitrary subspaces $U,V$ of an arbitrary vector space $W$, we have the dimension formula

$$\dim U + \dim V = \dim (U+V) + \dim (U\cap V).$$

(Take a basis $B_1$ of $U\cap V$, and extend it by systems $B_2$ to a basis of $U$, and $B_3$ to a basis of $V$. Then $B_1 \cup B_2 \cup B_3$ is a basis of $U+V$.)

That yields the

$$\dim F + \dim F^\perp = \dim (F + F^\perp) + \dim (F\cap F^\perp)$$

part. It remains to see that $\dim F + \dim F^\perp = \dim E$. Let's denote the non-degenerate bilinear form by $\beta$. $\beta$ induces a linear map $\Phi \colon E \to E^\ast$ ($E^\ast$ is the dual space of $E$, the space of all linear maps $E\to K$) via

$$\Phi(x)(y) = \beta(x,y).$$

The non-degeneracy of $\beta$ is equivalent to the injectivity of $\Phi$. Since the spaces are finite-dimensional, we have $\dim E^\ast = \dim E$, and thus $\Phi$ is an isomorphism. For a subspace $F \subset E$, the image of $F^\perp$ under $\Phi$ is the annihilator of $F$,

$$\Phi(F^\perp) = F^0 = \{ \lambda \in E^\ast : F\subset \ker\lambda\}.$$

If you already know that $\dim F + \dim F^0 = \dim E$ for all subspaces of finite-dimensional spaces, that's it. Otherwise, to see that, choose a basis $B_0 = \{v_1,\dotsc, v_f\}$ of $F$, extend it to a basis $B = B_0 \cup B_1 = \{v_1,\dotsc,v_f,v_{f+1},\dotsc,v_e\}$ of $E$, and consider the dual basis $B^\ast = \{\lambda_1,\dotsc,\lambda_e\}$ of $E^\ast$, defined by

$$\lambda_i(v_j) = \delta_{ij} = \begin{cases}1 &, i = j\\ 0 &, i \neq j. \end{cases}$$

It is then easy to see that $F^0 = \operatorname{span} \{\lambda_{f+1},\dotsc,\lambda_e\}$, whence $\dim F^0 = e - f = \dim E - \dim F$.

  • $\begingroup$ Thank you for this proof. Does the form really need to be symmetric? It does not seem to get used. $\endgroup$
    – JDZ
    Nov 12, 2016 at 23:30
  • $\begingroup$ @JDZ The form does indeed not need to be symmetric. Only the non-degeneracy is needed (for the $\dim F + \dim F^\perp = \dim E$ part). $\endgroup$ Nov 13, 2016 at 9:36

Have you read about complements of a vector space? Finite-dimensional vector spaces have the good property that every subspace is of finite dimension and has a complement. If you have a subspace $F \subseteq E$, then you get the complement $F^\perp$ why considering all elements of $E$ that are ortogonal with respect to your bilinear form, ie. all $x \in E$ with $$\langle x,y \rangle = 0 \text{ for all } y \in F$$ Then you get a decomposition of $E$ with respect to your subspace $F$ and your bilinear form $\langle,\rangle$ into $$ E = F \oplus F^\perp$$ So the dimension forumlar should be clear. If the bilinear form degenerates, the sum coudl be not direct so you have to take the dimension of "the rest" $F \cap F^\perp$ into account.

You can read the Wikipedia article on orthogonal complements to get informed about this subject as well.

Hope, that helps you!

  • $\begingroup$ I believe the sum could be not direct even if the form is non-degenerate, that g is not-degenerate only means that its kernel is 0. But not that the kernel of a restriction into any one subspace is 0. So it should be possible to have a non-degenerate form but still have a subspace which intersects its own orthogonal complement. Maybe I should have been clearer, g is non-degenerate as a bilinear form on E. $\endgroup$
    – fiftyeight
    Nov 27, 2013 at 12:47
  • $\begingroup$ No, you're actually wrong about that. If $g$ is non-degenerate and $F$ finite-dimensional, the orthogonal complement $F^\perp$ is a direct complement: $E = F \oplus F^\perp$ as you can see by choosing an orthonormal basis of $F$ and adding linearly independent orthonmal vectors to a full basis of $E$. Then the span of those added vectors is actually equal to $F^\perp$. $\endgroup$
    – BIS HD
    Nov 27, 2013 at 13:40
  • 1
    $\begingroup$ Over finite fields we can have a symmetric non-degenerate bilinear form and a subspace $C$ such that $C\cap C^\perp$ is not zero, in fact we could have $C\subseteq C^\perp$. The dimension formula still holds, in fact it's true for any pair of subspaces $E$ and $F$ (where $E=F^\perp$ above). $\endgroup$ Nov 27, 2013 at 13:46
  • $\begingroup$ Thanks for clearing that! :-) I was (wrongly) assuming $K$ to be a real/complex vector space. $\endgroup$
    – BIS HD
    Nov 27, 2013 at 13:47
  • $\begingroup$ Even for real spaces. Consider $\langle (x,y), (u,v)\rangle = xu - yv$, and the subspace $C = \mathbb{R}\cdot (1,1)$. Then $C^\perp = C$. $\endgroup$ Nov 27, 2013 at 19:11

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.