Suppose $M$ and $N$ are non-isomorphic $R$-modules (where $R$ is a commutative ring with a unit element).Can we conclude that $R \oplus M \not\simeq R \oplus N$ ? If not in this most general setup, is there an affirmative answer in some special circumstances ?

Motivation : I am reading Proposition 9(1.4) in Dale Husemoller's book 'Fibre Bundles' which reads :

If $u,v : \theta^1 \to \xi^k$ are two monomorphisms of the trivial line bundle over $B$ into $\xi^k$ (a k-dimensional vector bundle over $B$) , such that $n \leq ck-2$, then coker $u$ and coker $v$ are isomorphic over $B$.

Using the fact that short exact sequences of finite dimensional vector bundles over a paracompact space split, and replacing vector bundles over $B$ by $R$-modules, I get the question that I am asking here.

Essentially I am just trying to see if the theorem about vector bundles can be generalized to an arbitrary abelian category.


Two $R$-modules $M$ and $N$ are said to be "stably isomorphic" (see Stacks for example) if $M \oplus R^n$ and $N \oplus R^n$ are isomorphic for some $n \in \mathbb{N}$. It is possible for two non-isomorphic modules to be stably isomorphic. If $M'$ and $N'$ are such modules, let $n$ be the smallest (necessarily positive) integer such that $M' \oplus R^n \cong N' \oplus R^n$; then $M = M' \oplus R^{n-1}$ and $N = N' \oplus R^{n-1}$ satisfy your condition.

In The K-book: an introduction to algebraic K-theory, Weibel gives an example (Chapter I, Example 1.2.2) of a stably free module which is not free. The ring is $R = \mathbb{R}[x,y,z]/(x^2+y^2+z^2=1)$ and the module is $P = \ker \sigma$, where $\sigma : R^3 \to R$ is $(P,Q,R) \mapsto xP+yQ+zR$. Then $P \oplus R \cong R^3 = R^2 \oplus R$ as $R$-modules, but $P$ is not isomorphic to $R^2$. (Crazily enough, the reason comes from algebraic topology: it's the hairy ball theorem!)

In fact if you go through the definition, you will see that two finitely generated projective modules $M$, $N$ over $R$ are stably isomorphic iff they are equal in $K_0(R)$, the zeroth algebraic K-theory group of $R$. (See also the K-book, Chapter 2.)


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.