A commutative version of the group ring. (Q376021)

Let \(G\) be a finitely presented group, \(\kappa\) a field of characteristic 0. This is an extremely interesting paper in which combinatorial group theory and group algebra theory join. It provides a method to attack some well-known problems in combinatorial group theory, which is demonstrated by reproving the result of \textit{S. Boyer} [J. Pure Appl. Algebra 51, No. 3, 217-229 (1988; Zbl 0649.20026)], namely, that the free product of two nontrivial cyclic groups may not be normally generated by a proper power. The Scott-Wiegold conjecture (see \textit{V. D. Mazurov} (ed.), [Kourovka notebook. 12th ed., Novosibirsk: Institut Matematiki SO RAN (1992; Zbl 0831.20003)]) states that the free product of three nontrivial cyclic groups may not be normally generated by a single word. This problem is reduced to a problem of proving a statement in commutative ring theory. There is a hope to continue this approach to attack the free product of four nontrivial cyclic groups which would mean solving several hard problems such as the Wiegold and the Relation Gap problems, see \textit{K. W. Gruenberg} and \textit{P. A. Linnell} [J. Group Theory 11, No. 5, 587-608 (2008; Zbl 1160.20002)]. The approach is to construct a functor sending a group \(G\) to a commutative ring \(\kappa[G]^\#\) and a homomorphism \(f\colon G\to H\) to a homomorphism \(f^\#\colon\kappa[G]^\#\to\kappa[H]^\#\) (preserving surjectivity). Let \(*\) be the classical involution of the group algebra and denote by \(\kappa[G]^*\) the subspace of symmetric elements, by \(S\) the ideal generated by \([\kappa[G]^*,\kappa[G]]\). Let \(A_G\) be the factor algebra \(\kappa[G]/S\). The involution \(*\) induces an involution on \(A_G\) also denoted by \(*\), let \(\kappa[G]^\#\) and \(\Lambda_G\) be the subspaces of symmetric and skew-symmetric elements of \(A_G\). Then \(A_G\) is the (module) direct sum of these two subspaces, \(\kappa[G]^\#\) is a \(\kappa\)-algebra central in \(A_G\) and \(\Lambda_G\) is a \(\kappa[G]^\#\)-module. For \(x\in\kappa[G]\) let \(x+S=[x]\), \(\overline x={1\over2}(x+x^*)\), \(\widetilde x={1\over2}(x-x^*)\), hence \([\overline x]\in\kappa[G]^\#\), \([\widetilde x]\in\Lambda_G\). Let \(B\subseteq A_G\) be a generating system for the \(\kappa[G]^\#\)-module \(A_G\), \(L\subseteq G\), \(L^\#\) the ideal of \(\kappa[G]^\#\) generated by all \([\overline{bl}]-[\overline b]\) (\([b]\in B\), \(l\in L\)). Let \(f\colon G\to H\) be an epimorphism with kernel \(K\). The key result is that if \(L\subset K\) normally generates \(K\) then \(\text{Ker}(f^\#)=L^\#\), which provides the key method to prove that another set \(N\subset K\) does not normally generate \(K\), namely, it suffices to prove that \(N^\#\neq L^\#\) (there is also a slight refinement of this tool). Furthermore, the author discusses the following. Two operations, bracket and dot are introduced on \(\kappa[G]^\#\), by means of which several identities in \(\kappa[G]^\#\) are found. Using these, a complete description of \(\kappa[G]^\#\) is given through generators and generating relations, and generators are given for \(A_G\) as a \(\kappa[G]^\#\)-module. Applying these results finding finite number of generators for the ideals \(L^\#\) and \(N^\#\) becomes possible, and one may try to establish that the ideals are indeed distinct.