Suppose VV and WW are finite-dimensional. Let vVv \in V. Let

E={TL(V,W):Tv=0}.E = \{T \in \mathcal L(V,W) : Tv = 0\}.
  1. Show that EE is a subspace of L(V,W)\mathcal L(V,W).
  2. Suppose v0v \ne 0. What is dimE\dim E?

Obviously (T1+T2)v=T1v+T2v=0(T_1+T_2)v = T_1v+T_2v = 0 so EE is closed under addition, likewise (λT)v=λ(Tv)=0(\lambda T)v = \lambda(Tv) = 0 making it closed under addition and scalar multiplication, hence it's a subspace. This shows (1)

For (2) let v1=vv_1 = v and extend to a basis v1,,vnv_1,\dots,v_n of VV. Pick any basis w1,,wmw_1,\dots,w_m of WW.
Because M(Tv1)=M(T)M(v1)=0\mathcal M(Tv_1) = \mathcal M(T)\mathcal M(v_1) = 0 the first column of M(T)\mathcal M(T) must be zero (since multiplying by v1v_1 just extracts the first column)
The other columns can be anything, therefor dimE=(n1)m=(dimV1)(dimW)\dim E = (n-1)m = (\dim V - 1)(\dim W) (simply count the parameters that can be anything)


todo: make rigorous, cite theorems from 3.d