← All problems

Farjoun's questions on group localization

Let LL be a localization functor on the category of groups, with localization map ηG ⁣:G→L(G)\eta_G\colon G\to L(G). Determine whether the following assertions hold: (1) for every nilpotent group GG, the group L(G)L(G) is nilpotent; and (2) for every finite pp-group GG, where pp is prime, the localization map ηG ⁣:G→L(G)\eta_G\colon G\to L(G) is surjective.

References

Source: arXiv

Progress summary

Refreshed
Open · solution reported

A preprint dated October 12, 2026 claims that two expected properties of group localization fail, but its construction has not been independently checked.

Farjoun asked whether localization preserves nilpotence and whether localization maps from finite pp-groups are surjective. The new construction claims negative answers to both questions.

Known results

  • Right exact localization preserves nilpotence, with no increase in nilpotence class, and maps from finite pp-groups are surjective (2019).
  • A separate 2021 result showed that the cokernel of F2→LF2F_2\to L F_2 can contain a free subgroup of countable rank, refuting Farjoun’s third conjecture.
  • A 2014 Vogel–Levine example showed localization can produce a group that is not transfinitely nilpotent, but it is not the finite-pp-group case.

October 12, 2026 counterexample

Sergei O. Ivanov and Georgii Kadantsev claim that, for every odd prime pp, there are a finite pp-group GG, an infinite non-nilpotent group HH, and a local embedding G↪HG\hookrightarrow H with Lf(G)≅HL_f(G)\cong H. This would show that localization need not preserve nilpotence and that G→Lf(G)G\to L_f(G) need not be surjective. The preprint is unrefereed and unverified.

Current status (as of October 2026): The two targeted questions have claimed negative answers via the Ivanov–Kadantsev construction, but the construction remains mathematically unverified.

Sources

Solutions 0

No solutions have been posted yet.