I'm interested in results about functorially-defined subgroups (in a loose sense), especially in the non-abelian case, and would like to know about references I may have missed.

The question, it seems, comes up in its simplest form when noticing a number of common subgroups (the center, commutator subgroup, Frattini subgroup, etc) are characteristic. The characteristicity can be justified by the fact that the object mappings that define of those subgroups give rise to subfunctors of the identity functor, on the core category of Grp.

Hence I'm interested in functors F in Grp (or a carefully chosen sub-category) such that

∀ A F(A) ⊆ A and ∀ A,B ∀ f ∈ hom(A, B), f(F(A)) ⊆ F(B)

Does that ring a bell ?

The topic was mentioned a couple of months ago on Mathoverflow, and can be traced back to (at least) 1945, where Saunders MacLane explains it in some detail in the third chapter of *A General Theory of Natural Equivalences*.

In between, it seems that those functors have been baptised radicals, pre-radicals or subgroup functorials, and studied mostly in the framework of ring theory, notably by A. Kurosh. Among a number of not-so-recent (and therefore quite-hard-to-find) papers mostly dealing with rings, semigroups, or abelian groups, I came across a single reference mentioning the non-abelian case, by B.I. Plotkin : *Radicals in groups, operations on classes of groups, and radical classes*. Connections seem have been made with closure operators¹, but do not focus much on Grp.

**Do you have ideas of connections from those functors to other parts of algebra or category theory, other than (pre-)radicals ?****Do you have some pointers to material I may have missed, specially if they mention non-abelian groups ?**

¹: Categorical structure of closure operators with applications to topology By N. Dikranjan, Walter Tholen, p.51