Proposition (P1)

Let RR be a commutative ring, and let F=RnF = R^{\oplus n} be a finitely generated free RR-module. Then HomRMod(F,R)F.\Hom_{R-\mathrm{Mod}}(F,R) \cong F.

Proof

Let ff be any map in HomRMod(F,R)\Hom_{R-\mathrm{Mod}}(F,R). Take the set of “one-hot” nn-tuples j1,,jnF.j_1,\ldots,j_n \in F. Note that any element vv of FF can be written down as a sum

v=k=1nrkjkv = \sum_{k=1}^n r_kj_k

for some set of r1,,rnR.r_1,\ldots,r_n \in R.

We can find f(v)f(v) by using the fact that it’s a homomorphism, since

f(v)=f(k=1nrkjk)=k=1nf(rkjk)=k=1nrkf(jk),\begin{align*}f(v) &= f\left(\sum_{k=1}^n r_kj_k\right) \\ &=\sum_{k=1}^n f(r_kj_k) \\&=\sum_{k=1}^n r_kf(j_k),\end{align*}

that is, ff is completely determined by the values it takes on j1,,jn.j_1,\ldots,j_n.

Note also that every possible assignment of an element of RR to each jkj_k gives rise to a unique homomorphism in HomRMod(F,R).\Hom_{R-\mathrm{Mod}}(F,R). This follows from the universal property of a free object in RModR\mathrm{-Mod}.

Consider the map τ:HomRMod(F,R)F\tau : \Hom_{R-\mathrm{Mod}}(F,R) \to F defined by

fk=1nf(jk)jk.f \mapsto \sum_{k=1}^n f(j_k)j_k.

We have already established that τ(f)\tau (f) contains sufficient data to recover ff uniquely (injectivity) and that this mapping is surjective, so we only need to show that τ\tau is an RR-module homomorphism and this is straightforward. Let gHomRMod(F,R)g \in \Hom_{R-\mathrm{Mod}}(F,R) and uR,u \in R, then

τ(f+g)=k=1n(f+g)(jk)jk=k=1n(f(jk)jk+g(jk)jk)=k=1nf(jk)jk+k=1ng(jk)jk=τ(f)+τ(g),\begin{align*} \tau (f + g) &= \sum_{k=1}^n (f + g)(j_k)j_k \\ &= \sum_{k=1}^n (f(j_k)j_k + g(j_k)j_k) \\ &= \sum_{k=1}^n f(j_k)j_k + \sum_{k=1}^n g(j_k)j_k \\ &= \tau(f) + \tau(g), \end{align*}

and

τ(uf)=k=1n(uf)(jk)jk=uk=1nf(jk)jk=uτ(f).\begin{align*} \tau(uf) &= \sum_{k=1}^n (uf)(j_k)j_k \\ &= u \sum_{k=1}^n f(j_k)j_k \\ &= u \tau (f). \end{align*}

\blacksquare

NOTE

We need commutativity of RR to have an RR-module structure on HomRMod(F,R).\Hom_{R-\mathrm{Mod}}(F,R). Aluffi clarifies this towards the end of §5.2


Proposition (P2)

There exists a ring RR and a non-zero RR-module MM such that HomRMod(M,R)=0.\Hom_{R-\mathrm{Mod}}(M,R) = 0.

Proof

Indeed, set R=ZR=\mathbb{Z} and M=Z/2Z.M = \mathbb{Z}/2\mathbb{Z}. An abelian group homomorphism from MM to RR must send 1M1_M to an element of finite order in R,R, but the only such element is 0R,0_R, forcing every homomorphism MRM \to R to be trivial.

\blacksquare