On subfunctors of the identity (Q1096724)

The author decomposes the notion of exact subfunctor of the identity on a category of (unital) R-modules, where R is a commutative ring with unit into two notions: F is called multiplicative if for every submodule B of an R-module A, \(F(A/B)=(B+FA)/B\) and is called hereditary if, under the same conditions, \(FB=B\cap FA.\) Theorem 4. There is a bijective correspondence between: (1) Multiplicative functors. (2) Classes of modules closed under submodules, quotients and direct products. (3) Families of ideals of R satisfying (a) \(I\in {\mathcal F}\), \(J\supset I\Rightarrow\) \(J\in {\mathcal F}\), (b) \(\{I_ k\}_{k\in K}\subset {\mathcal F}\Rightarrow\cap_{k\in K}I_ k\in {\mathcal F}.\) That is, in a sense, a dual bijective correspondence to the known one between hereditary functors, classes of modules closed under submodules, quotients and direct sums, and pretopologies of R [see \textit{B. Stenström}, Rings of Quotients (1975; Zbl 0296.16001)]. Proposition 9. - For a subfunctor F of the identity, the following are equivalent: (1) F commutes with direct limits. (4) F is hereditary and multiplicative. (5) F is exact. In the last section, the exact sequence \(E: 0\to B\to A\to C\to 0\) is called F-exact (resp. pure F-exact), for a subfunctor of the identity, if the induced sequence FE: \(0\to FB\to FA\to FC\to 0\) is exact (resp. pure- exact). If T is the torsion functor, it is proved that pure-exactness is equivalent to pure-T-exactness. Finally, if F is a hereditary functor and \({\mathcal F}\) is the corresponding pretopology then pure-F-exactness is equivalent to \({\mathcal F}\)-purity.