Fitting subgroup is defined as the subgroup generated by all normal and nilpotent subgroups of a group G. If G is a finite group, we have that Fitting subgroup is nilpotent. If G is infinite not necessarily this occurs. We are looking for examples of infinite groups such that Fitting subgroup is non-nilpotent. Or some criterion to Fitting subgroup to be nilpotent in infinite case. In general, any answer in this direction (possibly with references) is welcome!

  • 4
    $\begingroup$ A direct product of finite groups $G_n$ for $n \ge 0$, where $G_n$ is nilpotent of class $n$ is an infinite group in which the Fitting subgroup (the whole group) is not nilpotent. $\endgroup$ – Derek Holt Nov 6 '19 at 8:16
  • $\begingroup$ "Direct product" in Derek's comment can both be interpreted as "restricted direct product", and "unrestricted direct product". $\endgroup$ – YCor Nov 7 '19 at 15:19
  • $\begingroup$ Thanks for the example, but how can i show that this direct product is not nilpotent? For instance, is true that $\gamma_{m}(G_{1} \times G_{2} \times \dots ) = \gamma_{m}(G_{1}) \times \gamma_{m}(G_{2}) \times \dots $ , where $\gamma_{m}$ is the mth term of lower central series? $\endgroup$ – Inácio Nov 8 '19 at 23:32

There's an obvious criterion: the Fitting subgroup of a group $G$ is nilpotent if and only if there exists $s$ such that every nilpotent normal subgroup $N$ of $G$ is $n$-step nilpotent (i.e., satisfies $N^{s+1}=\{1\}$).


  • to find examples it's enough to produce groups with suitably prescribed nilpotent normal subgroups (as in Derek's comment). Alternatively, the group of upper unipotent (=upper triangular with 1 on diagonal) with entries indexed by $\mathbf{N}$ matrices, over an arbitrary nonzero commutative associative unital ring, has such normal subgroups.
  • if for some reason it is known that there is a bound on the step of nilpotency of (normal) subgroups of $G$, then one gets that the Fitting subgroup of $G$ is nilpotent. This applies if $G$ is linear (i.e., isomorphic to a subgroup of $\mathrm{GL}_d(K)$ for some field $K$ and $d<\infty$).

Probably examples of finitely generated groups with non-nilpotent Fitting subgroup are known but this seems much harder to produce.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.