HR-length of a free group via polynomial functors (Q6103351): Difference between revisions

During the 1970's, \textit{A. K. Bousfield} [Topology 14, 133--150 (1975; Zbl 0309.55013)] introduced a fundamental notion of \(R\)-localisation of pointed topological spaces, where \(R\) is a nice ring (as a quotient or a localisation of \(\mathbb{Z}\)), with respect to the functor of \(R\)-homology \(H_*(-;R)\). This theory is very well-behaved and well-understood for \textit{nilpotent spaces}, but much harder in general; all the difficulty comes essentially from the \(\pi_1\) (see [\textit{A. K. Bousfield} and \textit{D. M. Kan}, Homotopy limits, completions and localizations. Berlin-Heidelberg-New York: Springer-Verlag (1972; Zbl 0259.55004)]). At the level of \(\pi_1\), \(R\)-localisation corresponds to \textit{\(HR\)-localisation} of groups, denoted by \(L_R\). The transfinite lower central series of the \(HR\)-localisation of a group \(G\) is fundamental in the theory: its quotients can be inductively build from suitable universal relative \(R\)-central extensions of groups (this is reminded in \S 5.1 of the article under review), and it \textit{stops}. The smallest ordinal where it happens is called the \textit{\(HR\)-length} of the group. In [\textit{S. O. Ivanov} and \textit{R. Mikhailov}, Isr. J. Math. 226, No. 2, 635--683 (2018; Zbl 1498.55011)], the authors studied \(H\mathbb{Z}\)-length of groups and showed in particular that it is at least \(\omega+2\) for a non-abelian free group. In the article under review, they prove that the \(H\mathbb{Q}\)-length (which is at most the \(H\mathbb{Z}\)-length) of such a free group is at least \(\omega+\omega\). They also study \(H\mathbb{Q}\)-length of some semi-direct products of abelian groups by using naturality properties and weight decomposition for polynomial endofunctors of \(\mathbb{Q}\)-vector spaces.