MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

$M$ compact manifold. Let $\lambda$ be an eigenvalue for the Dirac operator of multiplicity greater than 2. I'm interested in showing the existence of two linearly independant eigenspinors $u$ and $v$ of eigenvalue $\lambda$, with $\|u\|_2=\|v\|_2=1$ and so that there exists a point $x_0 \in M$ for which $\|u(x_0)\|_{x_0} \neq \|v(x_0)\|_{x_0}$. Trying to show this by contradiction, I assumed that $\|u(x)\|_x = \|v(x)\|_x$ for all $x \in M$. I can prove the desired result provided there be two linearly independant eigenspinors of eigenvalue $\lambda$ with $\langle u,v \rangle_2 \neq \frac{\langle u(x),v(x) \rangle_x}{|u(x)|_x^2}$ for some $x \in M$.

My question is:

Is it possible that any two normalized linearly independant eigenspinors $u$,$v$ in such $\lambda$ eigenspace satisfy both $$\|u(x)\|_x = \|v(x)\|_x \quad \quad \text{and} \quad \quad\langle u,v \rangle_2 =\frac{\langle u(x),v(x) \rangle_x}{|u(x)|_x^2} \quad \text{for all } x \in M \;?$$

share|cite|improve this question

The Dirac operator on $M$ is self-adjoint (or skew-adjoint, depending on your conventions). So the $\lambda$-eigenspace must be spanned by eigenvectors. Does this not solve your problem straight away?

share|cite|improve this answer
I don't see how. I'm already assuming that they are eigenvectors. – user16436 Jan 27 '12 at 4:04
What do you understand by "multiplicity"? Normally one would mean the dimension of the (generalized) eigenspace. Then you're done. – Bob Yuncken Jan 27 '12 at 12:19
Let's see. If $u$ and $v$ are two orthonormal linearly independant eigenvectors of Dirac, then $\|u\|_2=\|v\|_2$ and $\langle u,v \rangle _2=0$. If in addition I ask that for every point in the manifold the norms in the fiber are equal ($\|u(x)\|_x=\|v(x)\|_x)$), it won't be always true that $$\langle u(x), v(x) \rangle_x =0 \quad \forall x\in M.$$ Indeed, I expect it to be very unlikely. – user16436 Jan 27 '12 at 15:37
Thank you very much for giving some thought to my problem. – user16436 Jan 27 '12 at 15:38

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.