Splitting homotopy idempotents. II (Q687573)

[For Part I see \textit{H. M. Hastings} and the second author, in `Shape theory and geometric topology', Lect. Notes Math. 870, 23-36 (1981; Zbl 0471.55008).] This is a famous paper whose existence has been known for years. At last it is published. The problem it addresses arises in several settings. Here it is introduced as giving an example of a retract of a representable functor in unpointed homotopy, that is not itself representable. Another formulation is that there are free homotopy idempotents that do not split, i.e. there is a map \(f:X\to X\) such that \(f^ 2\simeq f\) but there are no \(g: X\to Y\), \(h:Y\to X\) such that \(hg=f\) and \(gh=\text{Id}_ Y\). The completeness of the answer given is beautiful, namely there is a universal example of a group with a conjugacy idempotent that does not split. The paper ends with a discussion of the main properties of this universal example.