Suppose VV and WW are both finite-dimensional. Prove that there exists an injective linear map from VV to WW if and only if dimVdimW\dim V \le \dim W.


First suppose there exists TL(V,W)T \in \mathcal L(V,W) such that TT is injective.
3.16 implies dimnull T=0\dim \text{null }T = 0, then apply 3.22 to get dimV=dimrange T\dim V = \dim \text{range }T.
Now because range T\text{range }T is a subspace of WW it's dimension is less, meaning

dimV=dimrange TdimW\dim V = \dim \text{range }T \le \dim W

This completes the forward direction.

Now suppose dimVdimW\dim V \le \dim W, Let v1,,vnv_1,\dots,v_n be a basis of VV and let w1,,wmw_1,\dots,w_m be a basis for WW. Define Tvj=wjTv_j = w_j for 1jn1 \le j \le n and Tvj=0Tv_j = 0 for n<jmn < j \le m. Clearly null T={0}\text{null } T = \{0\} since each vjv_j is mapped to a nonzero wjw_j, thus 3.16 implies TT is injective.