No label defined (Q3691902)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3914531
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 3914531 |
Statements
1985
0 references
(E,M)-factorization
0 references
torsion theories on the category of groups
0 references
reflective subcategories
0 references
factorization systems
0 references
abelian categories
0 references
pointed categories
0 references
localization
0 references
torsion subcategories in non additive categories
0 references
diagonal property
0 references
prefactorization system
0 references
torsion theory
0 references
stable units
0 references
reflexions
0 references
pointed category
0 references
preradicals
0 references
hereditary torsion theories
0 references
0.9333072
0 references
0.9232987
0 references
0.9159669
0 references
0.9125679
0 references
0.90983707
0 references
0.9074249
0 references
In this long and beautiful paper the authors deeply describe various types of reflective subcategories by means of factorization systems on the whole category. There are thus obtained a lot of results which are discussed and applied to additive and abelian categories, categories with a zero-object (i.e. pointed categories), groups and pointed spaces. Among these results we mention the characterization of a localization in terms of the associated factorization system as well as the deep analysis of reflective and coreflective factorization systems on a category with initial and final objects (respectively on a pointed category), analysis which leads up to the foundations for a theory of torsion subcategories in non additive categories. The paper contains eight sections. Sections 2 and 6 set the general definitions and constructions which are in the core of the paper. A morphism f has the diagonal property to (resp. from) g if whenever \(uf=gv\) (resp. \(fu=vg)\) there exists a unique w such that \(wf=v\) and \(gw=u\) (resp. \(fw=v\) and \(wg=u)\). A prefactorization system (we abbreviate it by pf-s) F on a category A consists of a pair (E,M) of subclasses of Mor A, each of them containing all isomorphisms and closed under composition and such that E (resp. M) is the class of all maps which have the diagonal property to (resp. from) each map in M (resp. E). F is a factorization system (abbreviated by f-s) if in addition every map of A has an E-M decomposition. Each subclass N of Mor A produces a pf-s (E,M) in which E is the class of all maps with the diagonal property to each map in N and M is the class of all maps with the diagonal property from each map in E. Every pf-s F produces a full subcategory \(\Psi_ 1F\) of A consisting of all X with \(X\to 1\) in M (1 denotes a final object of A); \(\Psi_ 1F\) is reflective provided: (*) each map \(X\to 1\) in A has an F-decomposition. Conversely each reflective subcategory B of A produces a pf-s \(\Phi_ 1B\), namely that which is given by Mor B. \(\Psi_ 1,\Phi_ 1\) are order preserving maps, \(\Psi_ 1\) is a left inverse of \(\Phi_ 1\) and \(\Phi_ 1\Psi_ 1F\leq F\) for each pf-s. F with (*) (the pf-s are ordered by \(F\leq F'\) if \(E\supset E'\) resp. \(M\subset M')\). Those pf-s F with (*) for which \(\Phi_ 1\Psi_ 1F=F\) are called reflective pf-s and are characterized by the fact that whenever fg and f are in E, g is too. Thus in a category with a terminal object, \(\Phi_ 1\) and \(\Psi_ 1\) induce a bijection between full reflective subcategories and reflective pf-s. Moreover \(\Phi_ 1\Psi_ 1F\) is called the interior \(\overset\circ F\) of F. Dually, if A has an initial object 0, one obtains maps \(\Phi_ 0,\Psi_ 0\) between coreflective subcategories and pf-s. In this case \(\Phi_ 0\Psi_ 0F\) which is ''greater'' than F is called the closure \(\bar F\) of F. The interior and closure operators are succesively applied to an f-s. F produces six different systems, and under the factorization hypothesis three reflective subcategories and three coreflective subcategories. In this context the authors define a torsion theory on A as a pair (C,B) of subcategories for which there exists a reflective F on A such that \(\bar F\) is an f-s, \(B=\Psi_ 1F=\Psi_ 1\bar F\) and \(C=\Psi_ 0E=\Psi_ 0\overset\circ E\). Thus, on one hand a reflective subcategory B produces a torsion theory (C,B') if \(\overline{\Phi_ 1B}\) together with its interior are f-s. If (C,B) is a torsion theory then from each X of C there exists exactly one map to each Y of B, and C (resp. B) are mutually determined by the above property (and this is just the definition of a torsion theory in the abelian case). Now the main theme of sections 3-4 is when \(\Phi_ 1B\) is an f-s. This holds for all reflective subcategories B of A provided A has all finite limits and all intersections of strong monomorphisms (A is f.w.c. in the authors' terminology). For such A there is more to say as the authors show: for a right adjoint T: A'\(\to A\), \(\Phi_ 1(TA')\) is an f-s and its E-part is just the class of all maps inverted by the left adjoint of T. If the condition on the existence of intersections of strong monos in A is dropped, then \(\Phi_ 1B\) will be an f-s only for special types of B. The authors define and characterize some of such types of reflexions \(r: A\to B\) (with \(\Phi_ 1B=(E,M)\) and \(\rho\) : id\({}_ A\to r\) the unit adjunction arrow). Thus r is simple if M consists exactly of all maps \(f: X\to Y\) for which r(f) \(\rho\) \({}_ X=\rho_ y f\) is a pullback; is semi-left-exact if by pulling back elements of E along maps of M one gets elements of E; and has stable units if by pulling back every \(\rho_ X\) along any map of A one gets elements of E. Moreover it is proved the following nice result: r is a localization (i.e. left exact) iff E is stable in C at pullbacks. Also, apart from other characterizations of simple, semi-left-exact or with stable units reflexions, the authors prove the chain: ''localization'' \(\Rightarrow\) ''with stable units'' \(\Rightarrow\) ''semi-left-exact'' \(\Rightarrow\) ''simple'' \(\Rightarrow\) \(''\Phi_ 1B\) is a reflective f-s'' and also they show (by examples) that none of the above reversed arrows is true. In section 5 it is studied the closure B' under subobjects of a reflective subcategory B in a category A when A has mono-strong epi decompositions. Thus, if \(\Phi_ 1B=(E,M)\) is an f-s then \(\Phi_ 1B'\) is the interior of \((E_*,M_*)\) which is an f-s, where \(E_*=E\cap \quad strong\quad epis\) and \(M_*\) consists of all m i with m in M and i in \(E\cap \quad monos.\) In section 7, reflexions of a pointed category A and their relations to the associated coreflection (via \(\Psi_ 1B)\) are studied. The associated coreflection is obtained by iterating the kernel of the given reflection; when this kernel is idempotent the reflection is called a normal one, and it is shown that every simple reflection is normal. Section 8 studies torsion theories on a pointed category A. It is proved that if A and \(A^{op}\) are f.w.c., then (C,B) is a torsion theory iff \(C^{\to}=B\) and \(B^{\leftarrow}=C\) where \(D^{\to}\) denotes the subcategory of all X for which each Hom(Y,X) with Y in D has only one element \((D^{\leftarrow}\) is dually defined). Moreover for such A any full subcategory D generates a torsion theory \((D^{\leftarrow \to}, D^{\to})\). A torsion theory (C,B) with the reflection \(\rho\) : \(1\to r\) onto B and the coreflection \(\sigma\) : \(s\to 1\) onto C is called left normal (right normal, resp. normal) if the sequence \(s\to^{\sigma}1\to^{\rho}r\) is left exact (right exact, resp. exact). A normal reflection r onto B produces a left normal torsion theory. There are studied and characterized types of torsion theories, and reflections on a pointed category A under various exactness conditions on A. The authors obtain thus that any torsion theory on the category of all groups is a normal one, any reflective subcategory B of groups gives a torsion theory (C,B') such that the reflection r' onto B' has stable units and the reflection onto B is normal iff B is the reflective part of a torsion theory. The last section defines notions as preradicals, radicals and hereditary torsion theories on a finitely complete and cocomplete category A, and studies their relations to localizations.
0 references