MathOverflow is a question and answer site for professional mathematicians. 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

Let $\mathbb{R}^{2n}$ be endowed with the canonical symplectic structure $(\omega, J)$, where $\omega$ the usual nondegenerate symplectic form and $J$ the usual almost complex structure (ie. $\omega(\cdot, J\cdot)$ gives the usual dot product).

Suppose we have $n$ 2-planes $P_i$ which give an $\omega$-orthogonal decomposition $\mathbb{R}^{2n}=\oplus_{i=1 \ldots n} P_i$ (ie. $\omega(P_i, P_j)=0$ for $i \neq j$). Then both $SO(P_1) \times \ldots \times SO(P_n)$ and $SL(P_1) \times \ldots \times SL(P_n)$ acts as symmetries summandwise.

We also have two symmetry groups $SO(2n), Sp_{2n}$ acting on $\mathbb{R}^{2n}$ defined relative to the fixed structure $(\omega, J)$. The question is then: determine the intersections $SO(2n) \cap. SO(P_1) \times \ldots \times SO(P_n)$ and $Sp_{2n} \cap. SL(P_1) \times \ldots \times SL(P_n)$.

share|cite|improve this question

[upd: this answers the old version of the question, which has since been changed.]

The intersection is $SO(2)^n=SO(2)\times SO(2)\times\cdots\times SO(2)$, as perhaps expected. The inclusion $SO(2)^n\subset (SL_2)^n\cap SO(2n)$ is clear. The other inclusion follows from this observation: If a linear map that preserves a vector subspace is to be orthogonal, it must preserve the metric restricted to the subspace.

share|cite|improve this answer
Actually I don't see the first inclusion as clear at all: this because $SO_2^n$ (as embedded in $SL_2^n \subset Sp_{2n}$) is not contained in $SO_{2n}$. Let us remind ourselves that $Sp_{2n} \cap SO_{2n}=U_n$--the unitary group. – J. Martel Jan 30 '12 at 20:25
forgive that last comment. Under all the usual embeddings we certainly do have $SO_2^n$ contained in $SO_{2n}$. What I obected to was the idea that a linear mapping which restricts to an isometry on certain transverse 2-planes must be an isometry of the ambient space. I should clarify the situation i had in mind: – J. Martel Jan 30 '12 at 20:43
Let us have a decomposition of a symplectic vector space $(V,\omega)$ into $\omega-orthogonal$ 2-planes $\oplus_{i=1 \ldots n} P_i$. Then we have $n$ copies of $SO_2$---one for each 2-plane $P_i$. What I wanted to consider was the quotient of these 2-plane isometries $SO_2^n$ modulo the isometries of the ambient space. To be totally explicit I was asking about the intersection of $\Pi_i SO(P_i)$ with $SO(V)$. Actually it's the quotient $\Pi_i SO(P_i) / SO(V) $ i'm interested in. – J. Martel Jan 30 '12 at 20:51
jmart -- I was presuming that the metric that $SO(2n)$ preserves is the direct sum of the standard metrics on each on the 2-planes in the decomposition $\mathbb{R}^{2n}=\mathbb{R}^2\oplus\cdots \mathbb{R}^2$. If this is the case, then $SO(2n)$ definitely contains $SO(2)^n$. If not, then could you please specify what the metric is and also what skew-linear form you use to define $Sp_n$. – algori Jan 30 '12 at 20:51
jmart -- another confusing point: do you take an embedding of a cartesian power of $SL_2$, or the diagonal embedding of $SL_2$ itself? – algori Jan 30 '12 at 20:53

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.