Suppose TT is a function from VV to UU. The graph of TT is the subset of V×WV \times W defined by

graph of T={(v,Tv)V×W:vV}.\text{graph of }T = \{(v,Tv) \in V \times W : v \in V\}.

Prove that TT is a linear map if and only if the graph of TT is a subspace of V×WV \times W.


Suppose TT is a linear map, then the graph of TT is a subspace of V×WV \times W because (v1,Tv1)+(v2,Tv2)=(v1+v2,T(v1+v2))(v_1,Tv_1) + (v_2,Tv_2) = (v_1+v_2,T(v_1+v_2)) is in the graph, likewise for scalar multiplication.

Now suppose the graph is a subspace of V×WV \times W, because the graph is closed under addition we have (for all v1,v2Vv_1,v_2 \in V)

(v1,Tv1)+(v2,Tv2)=(v1+v2,Tv1+Tv2)graph of T\begin{aligned} (v_1,Tv_1) + (v_2,Tv_2) &= (v_1+v_2, Tv_1+Tv_2) \in \text{graph of }T \end{aligned}

Since the graph of TT is defined as all pairs (v,Tv)(v,Tv) we must have T(v1+v2)=Tv1+Tv2T(v_1+v_2)=Tv_1+Tv_2 for all v1,v2v_1,v_2. Thus TT is additive, a similar argument works for scalar multiplication. Thus TT is a linear map