Complete objects in categories (Q2664586)

The subject of the article under review is a categorical approach to the concept of a \emph{complete} object, meant to extend the group-theoretical notion from [\textit{R. D. Carmichael}, Introduction to the theory of groups of finite order. Ginn and Company: Ginn and Company Boston (1937; JFM 63.0860.01)]: a~group with a trivial center such that all automorphisms are \emph{inner} (i.e., given by conjugation with a fixed element of the group). In a first approach, the following equivalent characterization due to [\textit{R. Baer}, Bull. Am. Math. Soc. 52, 501--506 (1946; Zbl 0061.02702)] is more easily understood from a categorical perspective. We may call an object \(X\) in a pointed category \emph{complete} if every normal monomorphism with domain \(X\) admits a retraction (= is a split monomorphism). For groups this agrees with the original definition, while in any abelian category such are precisely the injective objects. On the other hand, in [\textit{B. J. Gardner}, Contemp. Math. 9, 277--283 (1982; Zbl 0488.16007)] it was explained that an object in a category of associative (or alternative, or autodistributive) algebras is complete if and only if it is unitary, and an object in a category of Lie algebras is complete precisely when its annihilator is trivial and each derivation is inner. The ultimate aim of this article is to recover Baer's equivalence mentioned above by means of categorical methods, so that it becomes valid in a general setting, where it for instance unifies the cases of groups and Lie algebras. This development depends on a subtle variation of degree of completeness, depending on which kind of normal monomorphism is supposed to split. This is the subject of Section~4, where various strengths of completeness are introduced and studied, leading to the main Theorem 4.25. In the final Section 5, this unified viewpoint is used to provide a common categorical explanation for two a priori unrelated well-known facts: a general result from which we may deduce at once that both the derivation algebra of a perfect Lie algebra with trivial center and the automorphism group of a characteristically simple group are complete. In preparation for all this, the article's Sections 2 and 3 provide a rather exhaustive view of the essential categorical-algebraic background knowledge: the definitions of actions, normal subobjects, etc., are recalled in detail, and the existence of generic split extensions for objects with trivial center / monomorphisms with trivial centralizer are analyzed. This makes the article self-contained and quite accessible even to non-specialists.