Do orthogonal projections preserve norms?

840 Views Asked by At

I'm working on a different problem, but I encountered this question, which would basically solve my problem for me.

The problem I'm working on requires me to show that $\langle P_Mx,y\rangle=\langle P_M x,P_M y\rangle=\langle x,P_My\rangle,$ where $P$ is the projection operator onto a closed subspace, $M$, of an inner product space. If projections do in fact preserve norms, then I can just use the polarization identity along with the fact that projections are idempotent.

But I feel like I have to be misunderstanding something. The claim that $T$ is orthogonal iff $||Tx||^2=||x||^2$ seems patently false. For example, if $x=(1,1)\in\mathbb{R}^2$, and we let $T$ be the projection onto the $x$ axis, then $Tx = (1,0)$. So $||Tx||^2=1$, but $||x||^2 = 2$.

Any clarification on what I'm missing would be helpful.

Thanks in advance

1

There are 1 best solutions below

1
On BEST ANSWER

Unless you are projecting onto the whole space they do not preserve the norm. What you need is decompose $y$ as $$ Y=P_M y+ P_{M^\perp }y $$ In your first expression. When you distribute the product the second piece vanishes because it is an inner product of vectors in orthogonal subspaces so you get your second term etc.