Question #23571

Let R be a k-algebra where k is a field, and M,N be left Rmodules, with dimkM <∞. It is known that, for any field extension K ⊇ k, the natural map θ : (HomR(M,N))K → HomRK(MK,NK) is an isomorphism of K-vector spaces. Replacing the hypothesis dimkM < ∞ by “M is a finitely presented R-module,” give a basis-free proof for the fact that θ is a K-isomorphism.

Expert's answer

First consider the special case when MRnM \sim R^n . In this case, we have the canonical isomorphisms

(HomR(M,N))K(HomR(Rn,N))K(nN)KnNK,\left(\operatorname{Hom}_R(M,N)\right)^K \sim \left(\operatorname{Hom}_R(R^n,N)\right)^K \sim (nN)^K \sim n \cdot N^K,

HomRK(MK,NK)HomRK((RK)n,NK)nNK.\operatorname{Hom}_{R^K}(M^K, N^K) \sim \operatorname{Hom}^{R^K}((R^K)^n, N^K) \sim n \cdot N^K.

From this, we can safely conclude that θ\theta is an isomorphism. (Some commutative diagrams must be checked, but it is mostly routine work.) Next we assume MM is a finitely presented RR -module, which means that there exists an exact sequence of RR -modules M1M2M0M1 \to M2 \to M \to 0 where M1=RnM_1 = R^n and M2=RmM_2 = R^m . Applying the left-exact Hom - functors into NN and into NKN^K , we have the following commutative diagram:

0 \rightarrow (HomR(M,N))K(HomR(M2,N))K\left(Hom_{R}\left(M,N\right)\right)^{K}\rightarrow \left(Hom_{R}\left(M_{2},N\right)\right)^{K} (HomR(M1,N))K\left(Hom_{R}\left(M_{1},N\right)\right)^{K}

θ\downarrow \theta

θ2\downarrow \theta_{2}

θ1\downarrow \theta_{1}

0 \rightarrow HomRK(MK,NK)HomRK(M2K,NK)HomRK(M1K,NK)Hom_{R^K}\left(M^K,N^K\right)\rightarrow Hom_{R^K}\left(M_2^K,N^K\right)\rightarrow Hom_{R^K}\left(M_1^K,N^K\right)

Since θ1,θ2\theta_{1},\theta_{2} are both isomorphisms, an easy diagram chase shows that θ\theta is also an isomorphism.

LATEST TUTORIALS
APPROVED BY CLIENTS