We know that in general, if $X$ is a vector space and $Y, Z, W$ are subspaces such that $X=Y\oplus Z$ and $X=Y\oplus W$ then it may not be true that $Z=W$.

But if we are given further that $Z\subseteq W$, then this should be true. A proof I can think of goes like this: Let $w\in W$. As $X=Y\oplus Z$, we get $w = y+z$ for some $y \in Y$ and $z \in Z \subseteq W$. But also $X=Y\oplus W$ and $w=0+w$, so the uniqueness of the decomposition forces $y=0,z=w$, whence $w=z \in Z$. So $W \subseteq Z$, and equality follows.

Somehow, I feel that there should perhaps be a cleaner and more straightforward way to see that $W=Z$, but I haven't been able to work one out.

My question is: Is there any other way this could be seen more directly?

Edit: Perhaps I should have added that these spaces could be infinite-dimensional.


Your argument is sound, but it can be streamlined a bit.

What you need to show is that $W\subseteq Z$. So take $w\in W$; since $X=Y\oplus Z$, you can uniquely write $$ w=y+z $$ with $y\in Y$ and $z\in Z$; however $z\in W$, so $y=w-z\in W$ and therefore $y=0$.


We can use dimension if $Z$ is of finite dimension, we have $dim(Z)=dim(X)-dim(Y)=dim(W)$ and $Z\subset W$ so $$ Z=W $$

  • $\begingroup$ What if they are infinite dimensional? $\endgroup$ – Math Maniac Feb 23 '16 at 23:15
  • $\begingroup$ you can see the egreg answer :) $\endgroup$ – Hamza Feb 23 '16 at 23:18
  • $\begingroup$ You should state the finite-dimension restriction in your answer. $\endgroup$ – Rory Daulton Feb 23 '16 at 23:36

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.