Let R be a ring and C a class of right R-modules closed under finite direct sums. If we suppose that C has a set of representatives, that is, a set V(C) ⊆ C such that every M ∈ C is isomorphic to a unique element [M] ∈ V(C), then we can view V(C) as a monoid, with the monoid operationRecent developments in the theory of commutative monoids (e.g., [4,15]) suggest that one might obtain useful insights on decompositions of modules by considering the monoids V(C).Suppose, for example, that R is a semilocal ring and C = proj-R (see Section 1 for definitions and notation). Then V(C) is a positive normal monoid, equivalently [4, Exercise 6.4.16], it is isomorphic to the submonoid of N t (t ∈ N) consisting of solutions to a finite system of homogeneous linear equations with integer coefficients. (These monoids have also been called Diophantine monoids in the literature.) The first author and Herbera [13] showed, conversely, that given a positive normal monoid M, there is semilocal ring R such that M ∼ = V(proj-R). Thus the direct-sum behavior of finitely generated projective modules over a semilocal ring is precisely delineated by the structure of positive normal monoids.For another example, let R be a commutative Noetherian local ring and A a finitely generated R-module. Let add(A) denote the class of R-modules that are isomorphic to direct summands of direct sums of finitely many copies of A. Then E := End R (A) is a semilocal ring, and V(add(A)) is naturally isomorphic to V(proj-E).