Journal of Algebra 228, 633᎐642 Ž2000. doi:10.1006rjabr.2000.8290, available online at http:rrwww.idealibrary.com on A Characterization of the Common Multiples of the Degrees of the Absolutely Irreducible Representations of a Semisimple Algebra and Applications Manfred Leitz Fachbereich Informatik und Mathematik, Fachhochschule Regensburg, Postfach 12 03 27, D-93025 Regensburg, Germany E-mail: manfred.leitz@mathematik.fh-regensburg.de Communicated by Walter Feit Received June 29, 1998 1. INTRODUCTION In the following, the term semisimple algebra always means a semisimple finite dimensional algebra over the field C of the complex numbers. The aim of this paper is to give a proof of the following theorem which characterizes the set of the common multiples of the degrees of the irreducible characters of a semisimple algebra. Moreover, concrete applications of this theorem shall be demonstrated. 1.1. THEOREM. Let A be a finite dimensional semisimple algebra o¨ er the field C of the complex numbers, and let 1 be the unit element of A. The set of the irreducible characters of A is denoted by IrrŽ A., and denotes the regular character of A. Furthermore, let bi 4 and bXi 4 be dual bases of A with respect to such that Ž bXi bj . s ␦ i j Ž Kronecker delta.. Define the element s g A by s [ Ý bXi bXj bi bj . i, j If m g N is any fixed positi¨ e integer, then m2 ny1 ⭈ Ž sn . g alg.int. Ž C . ᭙ n g N m Ž 1 . ¬ m ᭙ g Irr Ž A . . Ž Here alg.int.ŽC. denotes the ring of all algebraic integers.. Proof. See Section 3. 633 0021-8693r00 $35.00 Copyright 䊚 2000 by Academic Press All rights of reproduction in any form reserved. 634 MANFRED LEITZ Remarks. Ži. Note that by the above equivalence the set of the common multiples of the set c.d.Ž A. [ Ž1. : g IrrŽ A.4 is on the whole independently of terms of representation theory characterized. ŽThe term regular character may be associated with linear algebra as well as with representation theory.. Žii. It is always true that Ž sn . is a positive rational number Žsee Proposition 3.5Žii.., and hence in Theorem 1.1 the set alg.int.ŽC. could be replaced by the set N of the positive integers. Since we are mainly interested in applications of the direction « , it will be better to retain the formulation such as given. 2. APPLICATIONS OF THEOREM 1.1 Throughout this section, let G be a finite group with the identity element e. A familiar result says that the degree of each irreducible complex projecti¨ e representation of G divides the order of G Žsee w1, Theorem 53.16; 6, Corollary 11.18x.. The main point of the standard proof of this result is the reduction of the problem to ordinary irreducible representations by introduction of so-called representation groups, due to Schur Žsee w1, Theorem 53.7; 11x., which requires considerable effort. Another proof which avoids representation groups has been given by Fossum w4, Sect. 4x, but this proof contains a lot of algebraic number theory. We will give a new proof by means of Theorem 1.1. Let ␣ : G = G ª C_ 04 be a so-called factor set Žsee w6, Definition 11.4x., thus ␣ Ž xy, z . ⭈ ␣ Ž x, y . s ␣ Ž x, yz . ⭈ ␣ Ž y, z . for all x, y, z g G. Furthermore, let A [ Cw G, ␣ x s [g g G C ⭈ bg , bx ⭈ by s ␣ Ž x, y . ⭈ bx y Ž x, y g G . be the so-called twisted group algebra with respect to the factor set ␣ . As is well known, A s Cw G, ␣ x is semisimple because the regular character induces on A a nondegenerate symmetric associative bilinear form Žsee w5, V Hilfssatz 2.6x.. The following theorem is the equivalent of the theorem on the degrees of the irreducible projective representations of G mentioned above Žsee w6, p. 177x.. 2.1. THEOREM Žm w1, Theorem 53.16x.. Let A s Cw G, ␣ x be the twisted group algebra of the group G relati¨ e to the factor set ␣ , and let 1 be the unit element of A. Then the degree Ž1. of each irreducible character of A di¨ ides the order of G. Before giving the proof of Theorem 2.1, we state some elementary facts Žsee w3, pp. 136᎐137; 5, p. 632x.. ŽWe omit the evident proofs.. 2.2. LEMMA. Let A s Cw G, ␣ x be the twisted group algebra of G relati¨ e to the factor set ␣ with the basis bg : g g G4 Ž bx ⭈ by s ␣ Ž x, y . ⭈ bx y for all 635 DEGREES OF REPRESENTATIONS x, y g G .. Note 1 denotes the unit element of A, and e denotes the identity element of G. Then: Ži. ␣ Ž g, e . s ␣ Ž e, g . s ␣ Ž e, e . Ž g g G .. Žii. ␣ Ž gy1 , g . s ␣ Ž g, gy1 . Ž g g G .. Žiii. 1 s ␣ Ž e, e .y1 ⭈ be . Živ. by1 s Ž ␣ Ž gy1 , g . ⭈ ␣ Ž e, e ..y1 ⭈ bgy1 Ž g g G .. g 2.3. LEMMA. Let notations be as in Lemma 2.2, and let denote the regular character of A. Ži. Then Ž bg . s ␦ g , e ⭈ ␣ Ž e, e . ⭈ < G < Ž g g G . , Žii. Then Ž 1. s < G < . If g g G, let bXg [ < G <y1 ⭈ by1 s Ž< G < ⭈ ␣ Ž gy1 , g . ⭈ ␣ Ž e, e ..y1 ⭈ bgy1 . g Ž bXx ⭈ by . s ␦ x , y Ž x, y g G . , i.e., bg 4 and bXg 4 are dual bases with respect to . Proof of Theorem 2.1. In accordance with Theorem 1.1, we regard in A s Cw G, ␣ x the element s [ bXx bXy bx by . Ý Ž x , y .gG=G Then by Lemma 2.3Žii., we have w [ < G < 2 ⭈ s s Ý y1 by1 x b y bx b y . Ž x , y .gG=G According to Theorem 1.1, we have to show that the number < G < 2 ny1 ⭈ Ž sn . s < G <y1 ⭈ Ž wn . s Ý Žx , y.gG n=G n < G <y1 ⭈ n ž y1 Ł by1 x b y bx b y is1 i i i i / is an algebraic integer, for all n g N. From now on, we regard an arbitrary fixed n g N and an arbitrary fixed y1 . Ž y1 y1 . pair Žx, y. g G n = G n. If Ž xy1 1 y 1 x 1 y 1 ⭈⭈⭈ x n yn x n yn / e, then, by n y1 . Lemma 2.2Živ. and Lemma 2.3Ži., it follows that ŽŁ is1 bx i by1 y i bx i b y i s 0. 636 MANFRED LEITZ y1 . Ž y1 y1 . Ž . Ž . If Ž xy1 1 y 1 x 1 y 1 ⭈⭈⭈ x n yn x n yn s e, then, by Lemma 2.2 iii and iv , we have n x, y [ y1 Ł by1 x b y bx b y is1 i i i i s c ⭈ 1, Ž ). where c is a complex number, and by Lemma 2.3Ži., < G <y1 ⭈ Ž x, y . s c. It suffices to show that c is an algebraic integer. Let D : A ª Hom C Ž A, A. be the left regular representation of A, and let Det : Hom C Ž A, A. ª C be the determinant. Application of the composed mapping Det(D to Eq. Ž). yields c <G < s 1. Remark. Note that in the above proof, it was not necessary to know that ␣ is equivalent to a factor set  whose values are roots of unity Žsee w2, Lemma 11.38; 3, Theorem 25.3x.. Still let G be a finite group with the identity e. Up to the end of this section, let A s Cw G x be the ordinary group algebra, and let IrrŽ G . be the set of all irreducible complex characters of G. By the following theorem, the common divisors of the co-degrees < G <r Ž e . Ž g IrrŽ G .. are independently of terms of representation theory characterized. 2.4. THEOREM Žcompare w8, Satz 1x.. For x, y g G let w x, y x [ xy1 yy1 xy be the commutator of x and y, and for positi¨ e integers n g N let L nŽ e . denote the set of solutions of the equation Ž within the set G n = G n . w x 1 , y 1 x ⭈ w x 2 , y 2 x ⭈⭈⭈ w x n , yn x s e. Let d g N di¨ ide the order of G. Then d 2 ny1 L n Ž e . ᭙ n g N m d <G< Ž e. ᭙ g Irr Ž G . . Proof. We use as a pair of dual bases Žwith respect to . G and < G <y1 ⭈ gy1 : g g G4 . Then, in the notation of Theorem 1.1, we have s s < G <y2 ⭈ ÝŽ x, y.g G=G w x, y x. Apparently, the identity <G< ž / d 2 ny1 ⭈ Ž sn . s Ln Ž e . d 2 ny1 Ž n g N. holds, and now we are done by Theorem 1.1. Remark. If N is a normal abelian subgroup of G, then it can be shown without use of further representation theory that < N < 2 ny1 ¬ < L nŽ e .< holds for all n g N Žsee w8, p. 278x., and in consequence of this, we obtain a well known classical result, due to Ito ˆ Žsee w6, Theorem 6.15x.. DEGREES OF REPRESENTATIONS 637 3. PROOF OF THEOREM 1.1 AND RELATED RESULTS In this section, we will prove a generalization ŽTheorem 3.1. of Theorem 1.1 and a related result ŽTheorem 3.7. which can be done without additional effort. First of all, we fix frequently used notations. Notations. N, Z, Q, C R A 1 bi 4 IrrŽ A. Ž1. e set of positive rational integers, rational integers, rational numbers, and complex numbers R [ alg.int.ŽC., ring of all algebraic integers semisimple finite dimensional algebra over the field C unit element of A basis of A regular character of A set of the irreducible characters of A degree of the character g IrrŽ A. [ Ž1.y1 ⭈ Ž g IrrŽ A.. central primitive idempotent associated with g IrrŽ A. Let g Hom C Ž A, C. be a symmetric linear function on A, i.e., Ž xy . s Ž yx . for all x, y g A. Then, as is well known, can be expressed in a unique way as a linear combination of the set IrrŽ A. Žwith coefficients in C., s Ý g IrrŽ A. c ⭈ . The symmetric associative bilinear form Ž x, y . g A = A ª Ž xy . g C associated with is nondegenerate if and only if c / 0 holds for all g IrrŽ A. Žsee w2, Lemma 9.7x.. Then the dual basis bXi 4 Žwith respect to . associated with the basis bi 4 of A exists such that Ž bXi bj . s ␦ i j . 3.1. THEOREM. Let s Ý g IrrŽ A. f ⭈ be a character of A, where f g N holds for all g IrrŽ A.. ŽThus it is f / 0 for all g IrrŽ A... Furthermore, let bi 4 and bXi 4 be dual bases of A with respect to , i.e., Ž bXi bj . s ␦ i j . Define the element s g A by s [ Ý bXi bXj bi bj . i, j If m g N is a fixed positi¨ e integer, then m2 ny1 ⭈ Ž sn . g R ᭙ n g N m f ¬ m ᭙ g Irr Ž A . . The proof of Theorem 3.1 which will be given after the proof of Lemma 3.6 results from a combination of Proposition 3.5Žii. and of Lemma 3.6 638 MANFRED LEITZ below. Preparatory to the proof of Proposition 3.5, we need Proposition 3.3 and Lemma 3.4. 3.2. DEFINITION Žsee w1, p. 481x.. Let notations and assumptions be as in Theorem 3.1. The so-called Gaschutz᎐Ikeda operator Žwith respect to ¨ . is the C-linear mapping : A ª A, where Ž x . [ Ý bXi xbi Ž x g A. . i We are sure that the following proposition is completely known though we could not find Žiii. and Živ. exactly in full generality in the literature Žsee w1, p. 481; 10, Satz 5 Ž1., Ž4.x.. 3.3. PROPOSITION. Let notations and assumptions be as in Theorem 3.1 and in Definition 3.2. For the Gaschutz᎐Ikeda operator the following ¨ holds: Ži. respect to Žii. Žiii. Živ. The map does not depend on the choice of dual bases with . Ž A. : centerŽ A.. s ( Ž composition of mappings.. Ž x . s Ý g IrrŽ A. fy1 ⭈ Ž x . ⭈ e Ž x g A.. Proof. For Ži. and Žii., see w1, pp. 481, 482x. Žiii. For each y g A, we have y s Ý j Ž ybXj . ⭈ bj . Replacing y by xbi yields xbi s Ý j Ž xbi bXj . ⭈ bj for all i, so by definition of the regular character, Ž x. s Ý Ž xbi bXi . s Ý Ž bXi xbi . s Ž Ž x . . . i i Živ. In advance, we mention the equations zs Ý gIrr Ž A . Ž z . ⭈ e s Ý Ž 1. gIrr Ž A . y1 ⭈ Ž z . ⭈ e Ž u . s fy1 ⭈ Ž ue . , Ž ue . s Ž 1 . ⭈ Ž u . Ž u . ⭈ e s Ž ue . Ž z g center Ž A . . , Ž u g A, g Irr Ž A . . , Ž u g A, g Irr Ž A . . . ŽThe first of the above equations follows from Schur’s Lemma, the second from Ž ue . s 0 if / g IrrŽ A., and the third from e g centerŽ A... From these equations and Proposition 3.3Žiii. which has just been shown 639 DEGREES OF REPRESENTATIONS follows Ž x . s Ý Ž 1. Ý Ž 1. Ý Ž 1. Ý Ž 1. y1 gIrr Ž A . s y1 ⭈ fy1 ⭈ Ž Ž x . ⭈ e . ⭈ e y1 ⭈ fy1 ⭈ Ž Ž xe . . ⭈ e y1 ⭈ fy1 ⭈ Ž xe . ⭈ e s gIrr Ž A . s gIrr Ž A . s ⭈ Ž Ž x . . ⭈ e gIrr Ž A . fy1 ⭈ Ž x . ⭈ e . Ý gIrr Ž A . 3.4. LEMMA Žsee w2, Proposition 9.17x.. The following equation holds, e s f ⭈ Ý Ž bXi . ⭈ bi Ž g Irr Ž A . . . i Proof. e s Ý i Ž e bXi . ⭈ bi s Ý i f ⭈ Ž bXi . ⭈ bi . Special cases of the following Proposition 3.5Ži. can be found in w7, Theorem 1; 8, Lemma; 10, p. 42x. 3.5. PROPOSITION. Let notations and assumptions be as in Theorem 3.1. The element s s Ý i, j bXi bXj bi bj is in the center of the algebra A, and the following equations hold. Ž i. Ž ii . s s m 1 Ý f2 gIrr Ž A . 2 ny1 ⭈ Ž sn ⭈ e , .s Ý sn s 1 Ý gIrr Ž A . Ž 1. ⭈ gIrr Ž A . m f2 n ⭈ e Ž n g N. . 2 ny1 Ž m, n g N . . ž / f Proof. In the following calculation, we use Definition 3.2, Proposition 3.3Živ., and Lemma 3.4, s s Ý bXi bXj bi bj s Ý Ž bXj . ⭈ bj s Ý i, j s j Ý gIrr Ž A . fy2 ⭈ f ⭈ ž j Ý Ž bXj . ⭈ bj j / ž Ý gIrr Ž A . ⭈ e s fy1 ⭈ Ž bXj . ⭈ e ⭈ bj / Ý gIrr Ž A . fy2 ⭈ e . 640 MANFRED LEITZ Taking the nth power yields the equations in Ži.. Especially, we have s g centerŽ A.. Multiplication of sn by m2 ny1 and application of to the resulting equation yields the equation in Žii.. 3.6. LEMMA Žsee w9, Lemma 1x.. Let qi Ž i g I . be finitely manyᎏnot necessarily differentᎏpositi¨ e rational numbers. Then all the qi Ž i g I . are integers if and only if for each n g N the power sum sum n [ Ý i qin is an integer. Proof. Let sum n g N for all n g N. We write the numbers qi as fractions qi s m irt with suitable numbers m i g N and the least common denominator t g N. If t ) 1, then there is a prime p, where p ¬ t. We choose a positive integer r, where p r ) < I < and set n [ Ž p y 1. ⭈ p ry1. Then we have n G r and t n ⭈ sum n ' 0 Žmod p r .. If I p [ i g I : p ¦ m i 4 , we have on the other hand by Euler t n ⭈ sum n s Ý i m ni ' < I p < Žmod p r .. At all < I p < ' 0 Žmod p r ., and hence I p s ⭋ Žbecause of p r ) < I < G < I p <.. But, then t cannot be the least common denominator. Proof of Theorem 3.1. Ž¥. This is evident by Proposition 3.5Žii.. Ž«. Multiplication of the equation in Proposition 3.5Žii. by the positive integer mX [ l.c.m. f : g IrrŽ A.4 yields for m, n g N X m ⭈m 2 ny1 ⭈ Ž sn .s Ž 1. ⭈ m ⭈ Ý gIrr Ž A . mX f ⭈ m 2 žž / / f ny 1 . Ž )). If m2 ny1 ⭈ Ž sn . g R, then, by the above equation, we have mX ⭈ m2 ny1 ⭈ Ž sn . g R l Qqs N. If the last holds for a fixed m g N and all n g N, then, by Lemma 3.6, Ž mrf . 2 g N. Thus mrf g N for all g IrrŽ A.. ŽIn applying Lemma 3.6, note that in Eq. Ž)). the summand corresponding to the index of summation again can be written as a sum of Ž1. ⭈ m ⭈ mXrf equal powers.. It may be of interest to obtain theorems like Theorem 1.1 or Theorem 3.1, where not all the g IrrŽ A. are under consideration. 3.7. THEOREM. Let s Ý g IrrŽ A. f ⭈ be a generalized character of A Ž i.e., f g Z ᭙ ., where f / 0 for all g IrrŽ A., and let bi 4 and bXi 4 be dual bases of A with respect to Ž Ž bXi bj . s ␦ i j .. Define the element s g A by s [ Ý bXi bXj bi bj . i, j 641 DEGREES OF REPRESENTATIONS Furthermore, let s Ý g IrrŽ A. f˜ ⭈ be a character of A and Irr Ž A. [ g IrrŽ A. : f˜ / 04 . If m g N is a fixed positi¨ e integer, then žŽm 2 ⭈ s . n / g R ᭙n g N m f ¬ m ᭙ g Irr Ž A . . Proof. The equation in Proposition 3.5Ži. remains correct with the more general assumptions concerning . Multiplication of this equation by m2 n and application of to the resulting equation yields žŽm 2 ⭈ s . n /s Ý gIrr Ž A . Ž 1 . ⭈ f˜ ⭈ m 2 n žž / / f Ž m, n g N . . Ž ))). From this, the assertion follows by Lemma 3.6. The following Corollary 3.8 is related more or less to w12, Theorems 1 and 3x. 3.8. COROLLARY. Let H be a subgroup of the finite group G with the identity e. For x, y g G, let w x, y x [ xy1 yy1 xy denote the commutator of x and y, and for positi¨ e integers n g N, let LHn denote the set of solutions Ž in the set G n = G n = G = H . of the equation zy1 ⭈ w x 1 , y 1 x ⭈ w x 2 , y 2 x ⭈⭈⭈ w x n , yn x ⭈ z s h. Let d g N di¨ ide the order of G. Then < H < ⭈ d 2 n < LHn < ᭙ n g N m d <G< Ž e. ᭙ g Irr Ž G . , where Ž 1GH , . G / 0. Ž Note that 1GH is the character of G induced from the principal character 1 H of H.. Proof. In Theorem 3.7, choose s Žregular character ., s 1GH , and m s < G <rd. The carrying out of the technical details is left to the reader. REFERENCES 1. C. W. Curtis and I. Reiner, ‘‘Representation Theory of Finite Groups and Associative Algebras,’’ Interscience, New YorkrLondonrSidney, 1962. 2. C. W. Curtis and I. Reiner, ‘‘Methods of Representation Theory,’’ Vol. 1, WileyrInterscience, New York, 1981. 3. L. Dornhoff, ‘‘Group Representation Theory,’’ Part A, Dekker, New York, 1971. 4. T. V. Fossum, Characters and centers of symmetric algebras, J. Algebra 16 Ž1970., 4᎐13. 642 MANFRED LEITZ 5. B. Huppert, ‘‘Endliche Gruppen I,’’ Springer-Verlag, BerlinrNew York, 1967. 6. I. M. Isaacs, ‘‘Character Theory of Finite Groups,’’ Dover, New York, 1976. 7. P. Kellersch and K. Meyberg, On a Casimir element of a finite group, Comm. Algebra 25 Ž1997., 1695᎐1702. 8. M. Leitz, Kommutatoren und Ito’s ˆ Satz ¨uber Charaktergrade, Arch. Math. 67 Ž1996., 275᎐280. 9. M. Leitz, Casimir-Operator und ein Satz von Reynolds ¨ uber Charaktergrade, Arch. Math. 71 Ž1998., 358᎐367. 10. F. Roesler, Darstellungstheorie von Schur-Algebren, Math. Z. 125 Ž1972., 32᎐58. ¨ 11. I. Schur, Uber die Darstellung der endlichen Gruppen durch gebrochene lineare Substitutionen, J. Reine Angew. Math. 127 Ž1904., 20᎐50. 12. S. P. Strunkov, On the characters of minimal p-defect in finite groups, So¨ iet Math. Dokl. 42 Ž1991., 689᎐692.