Let $R$ be a ring, $f: R^k \to R^m$ aswell as $g: R^l \to R^n$ module homomorphisms and $M = coker(f)$, $N = coker(g)$. Let $p_M: R^m \to M$ and $p_N: R^n \to N$ be the natural projections.

I now want to show that for each module homomorphism $\Phi: M \to N$, there exist module homomorphisms $\psi$ and $\phi$, so that the following diagram is commutative, and I want to concretely construct a formula for $\phi$ and $\psi$:

$$\begin{array} 0R^k&\stackrel{f}{\longrightarrow}&R^m& \stackrel{p_M}{\longrightarrow} &M\\ \downarrow{\psi}&&\downarrow{\phi}&&\downarrow{\Phi}\\ R^l&\stackrel{g}{\longrightarrow}&R^n& \stackrel{p_N}{\longrightarrow} &N\\ \end{array}$$

Thanks in advance. I'm not very used to these kind of constructions.


Let $\{e_1,...,e_n\}$ denote the standard basis of $R^m$. For each $1 \leq i \leq m$, let $\phi(e_i)$ be any vector which is a representative of the coset $\Phi(e_i+Im(f))\in N$. This defines $\phi$ on the basis $\{e_1,...,e_m\}$ for $R^m$, so extend $\phi$ to all of $R^m$ by linearity. By construction, we have that $p_N(\phi(e_i)) = \phi(e_i)+Im(g) = \Phi(e_i+Im(f)) = \Phi(p_M(e_i))$, so by linearity we have that $p_N \circ \phi = \Phi \circ p_M$.

I claim that $\phi(Im(f)) \subset Im(g)$. This is true because if $m \in R^k$, then $p_N(\phi(f(m)))=\Phi(p_M(f(m))) = \Phi(0) = 0$, so $\phi(f(m)) \in \ker(p_N)=Im(g)$.

Next, let $\{a_1,...,a_k\}$ denote the standard basis of $R^k$. For $1\leq j \leq k$, define $\psi(a_j)$ to be any element of $R^l$ for which $g(\psi(a_j))=\phi(f(a_j))$. We can do this since $\phi(Im(f))\subset Im(g)$. We have defined $g$ on a basis for $R^k$, so just extend by linearity as before.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.