Groups whose nonlinear irreducible characters separate element orders or conjugacy class sizes. (Q2475575)

Consider a finite group \(G\) and assume that any two elements in distinct conjugacy classes have distinct orders. Such a group has to be the symmetric group of degree \(2\) or \(3\). This conjecture was around for a long time but was finally proved in 1985 by \textit{P. Fitzpatrick} [Proc. R. Ir. Acad., Sect. A 85, 53-58 (1985; Zbl 0558.20016)]. Related to this is a conjecture that the symmetric group of degree \(3\) is the only non-Abelian group in which every distinct conjugacy class has a distinct size. This is still a conjecture but progress has been made [\textit{J.-P. Zhang}, J. Algebra 170, No. 2, 608-624 (1994; Zbl 0816.20023) and \textit{R. Knörr, W. Lempken} and \textit{B. Thielcke}, Isr. J. Math. 91, No. 1-3, 61-76 (1995; Zbl 0836.20035)], Zhang shows that the result hold for soluble groups. The authors define an `order function' on a finite group to be a class function if it takes the same value on non-identity elements of the same order and `class-size function' on a finite group to be a class function if it takes the same value on non-identity elements lying in conjugacy classes of the same size. Thus the conditions above are equivalent to saying that either every character is an order function in the first case or that every character is a class-size function in the second. (This follows since if \(a,b\) have the same order and for every irreducible character \(\chi\) (being an order function) we have \(\chi(a)=\chi(b)\) then \(a\) and \(b\) are conjugate by the second orthogonality relation. So same order means conjugate.) The authors vary this in an interesting way. Let \(\varphi\) be an order function. If \(\varphi(x)=\varphi(y)\) implies that \(x\) and \(y\) have the same order whenever \(x,y\in G\setminus\{1\}\) then \(\varphi\) is said to be an `order separator'. If \(\varphi(x)=\varphi(y)\) implies that \(x\) and \(y\) are in conjugacy classes with the same order whenever \(x,y\in G\setminus\{1\}\) then \(\varphi\) is said to be a `class-size separator'. Given this the authors prove two theorems. Theorem A. Let \(G\) be a non-Abelian finite group, and assume that every nonlinear irreducible character of \(G\) is an order separator. Then one of the following holds: (a) \(G\) is isomorphic to a quaternion group of order \(8\). (b) \(G\) is isomorphic to \(S_3\). (c) \(G\) is isomorphic to \(\text{AGL}(1,2^n)\), the affine group of dimension \(1\) over the field with \(2^n\) elements, where \(2^n-1\) is a prime. In particular, \(G\) has a unique nonlinear irreducible character. Conversely, if \(G\) is as in (a), (b) or (c) then the unique nonlinear irreducible character of \(G\) is an order separator. Theorem B. Let \(G\) be a non-Abelian finite group, and assume that every nonlinear irreducible character of \(G\) is a class-size separator. Then one of the following holds: (a) \(G\) is an extraspecial \(2\)-group. (b) \(G\) is isomorphic to \(\text{AGL}(1,p^n)\), the affine group of dimension \(1\) over the field with \(p^n\) elements. In particular, \(G\) has a unique nonlinear irreducible character. Conversely, if \(G\) is as in (a) or (b) then the unique nonlinear irreducible character of \(G\) is a class-size separator. These are deduced form a more general result which uses the following ideas. Let \(\mathcal R\) an equivalence relation on the set \(G^\#=G\setminus\{1\}\), and \(\varphi\) a class function of \(G\). Then \(\varphi\) is said to be an \(\mathcal R\)-function if for every \(x,y\in G^\#\) the fact \(x\mathcal Ry\) implies \(\varphi(x)=\varphi(y)\); moreover, \(\varphi\) is said to be an \(\mathcal R\)-separator if for every \(x,y\in G^\#\) the fact \(x\mathcal Ry\) is equivalent to \(\varphi(x)=\varphi(y)\). Furthermore, the equivalence relation \(\mathcal R\) is called rational if any two elements of \(G^\#\) lie in the same \(\mathcal R\)-equivalence class whenever they generate the same subgroup of \(G\). In this setting the authors prove the following. Main Theorem. Let \(G\) be a non-Abelian finite group, \(\mathcal R\) a rational equivalence relation on \(G^\#\), and assume that every nonlinear irreducible character of \(G\) is an \(\mathcal R\)-separator. Then \(G'\) is a minimal normal subgroup of \(G\), and one of the following holds: (a) \(G\) is an extra-special \(2\)-group. (b) \(G\) is a Frobenius group with kernel \(G'\) (which is an elementary Abelian \(p\)-group for a suitable prime \(p\)) and cyclic complement.