# Linear Algebra Question

Let $V$ and $W$ be vector spaces over a field $F$, $T : V → W$ a linear map, $X$ a subspace of $V$ and $Y$ a subspace of $W$.

Show that:

a) $(T^{-1}(Y))^0=T^*(Y^0)$

b)$T(X) ⊆ Y$ if and only if $T^∗(Y^0) ⊆ X^0$

## Answer

a) Let $f\in T^{*}(Y^0)$. Then $f=T^*(g)$ where $g\in Y^0$. Now let $v\in T^{-1}(Y)$. Then $Tv \in Y$. So $$g(Tv)=0 \implies f(v)=(T^*(g))(v)=g(Tv)=0$$

Therefore $f\in (T^{-1}(Y))^0$. So $T^{*}(Y^0) \subset (T^{-1}(Y))^0$.

Now let $f\in (T^{-1}(Y))^0$. Note that $\ker T =T^{-1}(0)\subset T^{-1}(Y)$. So $f(\ker T)=\{0\}$. Now for any $w\in T(V)$ there is $v\in V$ so that $T^{-1}(w)=v+\ker T$. So $$ f(T^{-1}(w))=\{f(v)\}$$ Let us define $g\in W^*$ by $g(w):=f(u)$ where $u$ is some arbitrary element of $ T^{-1}(Pw)$ and $P$ is some fixed projection onto $T(V)$. Note that by the above relation $g$ is well-defined. It is easy to see that $g$ is linear because $f,P$ are linear and the set $T^{-1}(w)$ depends linearly on $w$. And for any $v\in V$ we have $$ g(Tv)=f(v) $$ since $v\in T^{-1}(Tv)=T^{-1}(PTv)$. So $f=T^*(g)$ and thus $T^{*}(Y^0) \supset (T^{-1}(Y))^0$.

---------------------------------------

The following proof of the above fact also works when both $V,W$ are finite dimensional:

Now we know that $ T|_{T^{-1}(Y)} :T^{-1}(Y)\to Y$ is one-to-one and onto, so by rank-nullity theorem we have $$ \dim Y + \dim \ker T = \mathrm{rank} T|_{T^{-1}(Y)} + \dim \ker T = \dim T^{-1}(Y)$$ Note that $\ker T \subset T^{-1}(Y)$ so $\ker T|_{T^{-1}(Y)} = \ker T$. So

$$ \dim (T^{-1}(Y))^0 =\dim V - \dim T^{-1}(Y)=\dim V - (\dim Y + \dim \ker T) \\ \;\\ = \dim V - (\dim Y + \dim V -\mathrm{rank} T)=\mathrm{rank} T-\dim Y =\mathrm{rank} T^* -\dim Y\\\;\\ =\mathrm{rank} T^* -(\dim W -\dim Y^0)=\mathrm{rank} T^* -\dim W^* +\dim Y^0\\\;\\ = -\dim \ker T^* +\dim Y^0= \dim T^{*}(Y^0)$$

Note that in the last two equalities we applied the rank-nullity theorem to the map $T^* :W^* \to V^*$ and the onto map $T^*|_{Y^0} :Y^0 \to T^*(Y^0)$. We also used the fact that the dimension of a subspace plus the dimension of its annihilator is equal to the dimension of the space, and the rank of a map and its adjoint are the same.

So we must have $T^{*}(Y^0) = (T^{-1}(Y))^0$ since the smaller subspace has the same dimension as the larger one.

b) Suppose $T(X)\subset Y$. Then $X\subset T^{-1}(Y)$. So $X^0 \supset (T^{-1}(Y))^0=T^{*}(Y^0)$. Conversely suppose $(T^{-1}(Y))^0=T^{*}(Y^0)\subset X^0$. Then $$T^{-1}(Y)\supset X\implies Y\supset T(X)$$ as desired.

- answered
- 1651 views
- $20.00

### Related Questions

- Show that the $5\times 5$ matrix is not invertable
- Consider the vector v = (3, 4, 5)^T, calculate the orthogonal projection
- Questions about using matrices for finding best straight line by linear regression
- Conjugate / Transpose - Matrix
- Linear Algebra Exam
- Hello! I Would like a proof detailed of the following question.
- [Linear Algebra] $T$-invariant subspace
- Decide if the following representations are linear representations.

What do you mean by Y^0 in this context?

Is it the dual of Y?

Hey. Annihilators.

T^* meaning adjoint.