About a question concerning \(\mathbb{Q}\)-groups (Q1295696)

A \(\mathbb{Q}\)-group or rational group \(G\) is a finite group all of whose characters are rational-valued. The reviewer proved that the order of a solvable rational group with the property that each of its complex irreducible characters has Schur index 1 over the real numbers is of the form \(2^a3^b\). He also asked if it is the case that each irreducible character of such a solvable rational group \(G\) actually has Schur index 1 over the rational numbers. The author of the paper under review claims to have given a positive answer to this question. The reviewer is not convinced by the arguments presented. To be specific, on p.~132 of the paper, the author constructs certain induced characters \(\theta_i\) of \(G\), which are the characters of real representations, and are certainly rational-valued under the given hypothesis. He then considers an irreducible constituent \(\Phi\) of \(\theta_i\) and argues that as \(\Phi\) has Schur index 1 over the real numbers (by the hypothesis on \(G\)), it has even multiplicity in \(\theta_i\). It is completely impossible to make such a claim: it is only true if \(\Phi\) has Schur index 2 over the real numbers, something ruled out by hypothesis. Then the conclusion that \(\theta_i(a)\) is even for all \(i\), where \(a\) is the generator of a certain cyclic group, is manifestly insupportable and the whole argument collapses. Indeed, most of the elaborate previous argument developed in the paper appears to be totally irrelevant if we accept the truth of the author's logical fallacy.