Startseite The average character degree of finite groups and Gluck’s conjecture
Artikel Öffentlich zugänglich

The average character degree of finite groups and Gluck’s conjecture

  • Alexander Moretó ORCID logo EMAIL logo
Veröffentlicht/Copyright: 17. Januar 2023

Abstract

We prove that the order of a finite group 𝐺 with trivial solvable radical is bounded above in terms of acd ( G ) , the average degree of the irreducible characters. It is not true that the index of the Fitting subgroup is bounded above in terms of acd ( G ) , but we show that, in certain cases, it is bounded in terms of the degrees of the irreducible characters of 𝐺 that lie over a linear character of the Fitting subgroup. This leads us to propose a refined version of Gluck’s conjecture.

1 Introduction

In the landmark paper [4], R. Guralnick and G. Robinson studied the commuting probability in finite groups. As pointed out in [4], if 𝐺 is a finite group, we may endow G × G with the structure of a probability space by assigning the uniform distribution. The probability that a randomly chosen pair of elements of 𝐺 commutes is k ( G ) / | G | , where k ( G ) is the number of conjugacy classes of 𝐺 (see [5]). This number is called the commuting probability of 𝐺. Note that its inverse | G | / k ( G ) is the average size of the conjugacy classes of 𝐺. We will write acs ( G ) = | G | / k ( G ) . The main results in [4] can be reformulated in terms of acs ( G ) , and we will do so when we refer to results in [4] in this note.

Similarly, we write

acd ( G ) = χ Irr ( G ) χ ( 1 ) k ( G )

to denote the average degree of the irreducible characters of 𝐺. This invariant was introduced in [10], motivated by a conjecture of K. Magaard and H. Tong-Viet [16]. Since then, acd ( G ) has been studied in a number of papers (see, for instance, [20, 22, 15, 7, 6, 8]), but several fundamental questions remain open. The work previously done shows that working with the average character degree tends to be more difficult than working with the average class size. For instance, it was proved in [4, Theorem 11] (see also [13, 14]) that if 𝐺 is a finite group and acs ( G ) < 12 , then 𝐺 is solvable. This theorem does not depend on the classification of finite simple groups. The analogous result for the average character degree is that if acd ( G ) < 16 / 5 , then 𝐺 is solvable. This could only be proved in [20], improving on earlier results in [16, 10], and uses the classification of finite simple groups. Note that if A 5 is the alternating group on 5 letters, then acs ( A 5 ) = 12 and acd ( A 5 ) = 16 / 5 . In this article, we obtain a version for acd ( G ) of another of the results in [4]. We also point out some results that do not admit an analog for acd ( G ) , discuss connections with Gluck’s conjecture on the largest character degree and mention some questions that, unfortunately, we have not been able to solve. All groups in this paper will be finite. Our notation follows [9].

By [4, Theorem 8], which does not depend on the classification of finite simple groups, if 𝐺 is a group with trivial solvable radical, then | G | is bounded from above in terms of acs ( G ) . The following is our main result.

Theorem A

Let 𝐺 be a finite group with trivial solvable radical. Then | G | is bounded from above in terms of acd ( G ) .

Using the CFSG, the bound in [4, Theorem 8] was improved in [4, Theorem 9] to | G | acs ( G ) 2 . This theorem actually asserts that if 𝐺 is any finite group, then

| G : sol ( G ) | acs ( G ) 2 ,

where sol ( G ) is the solvable radical of 𝐺. The proof of this result quickly reduces to the case of groups with trivial solvable radical because, by [4, Lemma 2 (ii)], acs ( G / N ) acs ( G ) for any finite group 𝐺 and N G . As we discuss in Section 5, the corresponding question for the average character degree seems more complicated.

It follows from [4, Theorem 4 and Theorem 8] that in fact | G : F ( G ) | is bounded from above in terms of acs ( G ) , where F ( G ) is the Fitting subgroup of 𝐺. Actually, | G : F ( G ) | acs ( G ) 2 by [4, Theorem 10]. It is not difficult to see that | G : F ( G ) | is not bounded in terms of acd ( G ) . This is perhaps surprising in view of the many parallel results between conjugacy class sizes and character degrees. If 𝐺 is the Frobenius group of order ( p 1 ) p , where 𝑝 is any odd prime, then | G : F ( G ) | = p 1 is arbitrarily large but acd ( G ) < 2 . One could think that, perhaps, this is due to the fact that these groups have arbitrarily many linear characters and just one nonlinear irreducible character. We will see in Example 4.1 that | G : F ( G ) | cannot be bounded from above in terms of the average of the degrees of the nonlinear irreducible characters either. But what if we just consider certain irreducible characters?

As usual, given a finite group 𝐺, N G , and λ Irr ( N ) , we write Irr ( G | λ ) to denote the set of irreducible characters of 𝐺 that lie over 𝜆. We set

acd ( G | λ ) = χ Irr ( G | λ ) χ ( 1 ) | Irr ( G | λ ) | .

With this notation, we have the following result.

Theorem B

Let 𝐺 be a solvable group. Then there exists λ Irr ( F ( G ) ) linear such that | G : F ( G ) | acd ( G | λ ) α , where α < 2.596 .

If 𝐺 is a finite group, b ( G ) is the largest degree of the irreducible characters of 𝐺. Gluck’s conjecture [3] asserts that if 𝐺 is a solvable group, then

| G : F ( G ) | b ( G ) 2 .

D. Gluck [3] proved that | G : F ( G ) | b ( G ) 13 / 2 . There has been a series of improvements on this bound (see, for instance, [21, 2, 25]), but the conjecture remains open. Clearly, acd ( G | λ ) b ( G ) , and it is not difficult to find examples where acd ( G | λ ) < b ( G ) for every λ Irr ( F ( G ) ) , so Theorem B suggests the following refinement of Gluck’s conjecture [3]. We remark, however, that our proof of Theorem B just mimics the arguments of known results on Gluck’s conjecture.

Conjecture C

Let 𝐺 be a solvable group. Then there exists λ Irr ( F ( G ) ) linear such that | G : F ( G ) | acd ( G | λ ) 2 .

We close this introduction with thanks to N. N. Hung for many helpful conversations on this paper. In particular, it was he who suggested that, perhaps, the index of the Fitting subgroup of a group could be bounded in terms of the average degree of the irreducible characters of 𝐺 lying over some irreducible character of some normal subgroup.

2 Proof of Theorem A

In this section, we prove Theorem A. We start by recalling a well-known consequence of the Krull–Schmidt theorem. We will write S n to denote the symmetric group on 𝑛 letters.

Lemma 2.1

Let F = N 1 × × N t , where N i is the direct product of u i 1 copies of a nonabelian simple group S i for every 𝑖. Assume that S i S j if i j . Then Aut ( F ) Aut ( N 1 ) × × Aut ( N t ) . Furthermore, Aut ( N i ) Aut ( S i ) S u i .

We need to use cohomology to prove the following extendability criterion.

Lemma 2.2

Let M , N G with M N = 1 . Put K = M × N , where 𝑀 and 𝑁 are perfect. Let α Irr ( M ) and β Irr ( N ) be 𝐺-invariant. Assume that 𝛽 extends to 𝐺. Then α × β extends to 𝐺 if and only if 𝛼 extends to 𝐺.

Proof

Let χ Irr ( G ) be an extension of 𝛽. Then χ K extends 𝛽 and χ ( 1 ) = β ( 1 ) . Since 𝑀 is perfect, χ K = 1 M × β , so β ~ = 1 M × β extends to 𝐺.

Suppose first that 𝛼 extends to 𝐺. Arguing as before, we deduce that α ~ = α × 1 N extends to 𝐺. Note that α × β = α ~ β ~ . By [11, Corollary 6.4],

[ α × β ] G / K = [ α ~ ] G / K [ β ~ ] G / K .

By [9, Theorem 11.7], [ β ~ ] G / K = 1 = [ α ~ ] G / K , so [ α × β ] G / K = 1 . By [9, Theorem 11.7], α × β extends to 𝐺.

Conversely, assume that α × β extends to 𝐺. Write α × β = α ~ β ~ , where α ~ = α × 1 N is the unique extension of 𝛼 to 𝐾. By [9, Theorem 11.7] and [11, Corollary 6.4],

1 = [ α × β ] G / K = [ α ~ ] G / K [ β ~ ] G / K = [ α ~ ] G / K .

Using [9, Theorem 11.7] again, α ~ , and hence 𝛼, extends to 𝐺, as wanted. ∎

This result is false if we remove the hypothesis that 𝑀 and 𝑁 are perfect. A counterexample can be found in G = SmallGroup ( 32 , 2 ) (see [26]). I thank G. Navarro for pointing this out.

Our next lemma is an immediate consequence of Jordan’s theorem on linear groups. We will write m ( S ) to denote the smallest degree of the nonlinear irreducible characters of 𝑆.

Lemma 2.3

Let 𝑆 be a nonabelian simple group. Then | S | is bounded from above in terms of m ( S ) .

Proof

This follows from Jordan’s theorem. Using the classification of finite simple groups, sharp bounds were obtained by M. Collins [1], building on work of Weisfeiler. ∎

In Theorem A, we have the hypothesis that the solvable radical of 𝐺 is trivial. This implies that the generalized Fitting subgroup of 𝐺 coincides with the socle of 𝐺 and is the direct product of the (nonabelian) minimal normal subgroups of 𝐺 (see [12, 6.5.5], for instance). Recall that if we write 𝐹 to denote the generalized Fitting subgroup of 𝐺, then C G ( F ) F by [12, Theorem 6.5.8]. It follows that if the solvable radical of 𝐺 is trivial, then 𝐺 is isomorphic to a subgroup of Aut ( F ) , and by Lemma 2.1, we know the structure of this group. We will use this fact repeatedly.

We need to introduce some notation referring to the irreducible characters of 𝐹. Let

F = S 11 × × S 1 u 1 × × S t 1 × × S t u t

be a direct product of nonabelian simple groups with S i j S i k S i for every j , k and S i S j for i j . Let

θ = α 11 × × α 1 u 1 × × α t 1 × × α t u t Irr ( F ) .

We will say that each of the characters α i j is a factor of 𝜃. Each factor α i j of 𝜃 is an irreducible character of a simple group S i j . Given a factor 𝛼 of 𝜃, we will write F α F to denote the direct product of the direct factors of 𝐹 whose factor in 𝜃 is 𝛼. Note that

F = F α (direct product) ,

where the product runs over the factors 𝛼 of 𝜃. Write Γ = Aut ( F ) . Note that, for every factor 𝛼 of 𝜃, F α is normal in I Γ ( θ ) . We will refer to the decomposition F = F α as the 𝜃-decomposition of 𝐹.

For every simple group S i , we fix α i Irr ( S i ) nonprincipal such that α i extends to Aut ( S i ) . Note that such a character exists by [19, Lemma 4.2]. If 𝜃 has “many” principal factors associated to any given simple group S i , we define a character θ Irr ( F ) by means of the following rules.

Step 1: For every 𝑖, if the number of factors α i j of 𝜃 that are 1 S i is bigger than u i / 2 , then we replace 1 S i by α i all the times. We write θ ¯ to denote this new character.

Step 2: In this case, if α i was already a factor of 𝜃, then we replace the α i factors of 𝜃 by 1 S i . We write θ to denote this new character.

Note that the 𝜃-decomposition, the θ ¯ -decomposition and the θ -decompositions of 𝐹 coincide.

If the number of factors α i j in 𝜃 that are 1 S i is at most u i / 2 for every 𝑖, we do not define θ and we say that θ is not defined. Note that if θ is not defined, then at least one half of the factors corresponding to the copies of each S i is nonprincipal, so

θ ( 1 ) i = 1 t m ( S i ) u i / 2 .

In the next result, we collect some key properties of 𝜃 and θ when θ is defined.

Lemma 2.4

Let 𝐺 be a group with trivial solvable radical. Let 𝐹 as before be the generalized Fitting subgroup of 𝐺. Let θ Irr ( F ) and assume that θ is defined. The characters 𝜃 and θ satisfy the following properties.

  1. θ ( 1 ) i = 1 t m ( S i ) u i / 2 .

  2. I G ( θ ) = I G ( θ ) .

  3. | Irr ( G | θ ) | = | Irr ( G | θ ) | .

  4. If γ Irr ( F ) is such that γ is defined (possibly γ = θ ), then 𝜃 and γ are not 𝐺-conjugate. Furthermore, if 𝜃 and 𝛾 are not 𝐺-conjugate, then θ and γ are not 𝐺-conjugate.

Proof

(i) It suffices to note that θ has been defined so that the number of nonprincipal irreducible factors corresponding to the copies of each S i is at least one half of the number of copies of S i .

(ii) Let Γ = Aut ( F ) . It follows from Lemma 2.1 and the fact that the characters α i are Aut ( S i ) -invariant that I Γ ( θ ) = I Γ ( θ ) . Since 𝐺 is a subgroup of Γ, the result follows.

(iii) Put T = I G ( θ ) = I G ( θ ) . By Clifford’s correspondence [9, Theorem 6.11], it suffices to see that | Irr ( T | θ ) | = | Irr ( T | θ ) | . These numbers are known to be the number of 𝜃-special classes and the number of θ -special classes, respectively [9, Problem 11.10]. By [9, Problem 11.9], it suffices to see that if H T , then 𝜃 extends to 𝐻 if and only if θ extends to 𝐻.

Let H T and assume that 𝜃 extends to 𝐻. Our task is to see that θ extends to 𝐻. Put N = Ker θ . Note that 𝑁 is the direct product of the simple groups S i j whose corresponding factor is the principal character. Let 𝑀 be the direct product of the remaining direct factors of 𝐹 so that F = M × N . Note that, since 𝜃 is 𝑇-invariant, 𝑀 and 𝑁 are normal in 𝑇. In particular, 𝑀 and 𝑁 are normal in 𝐻 too. Write θ = θ ~ × 1 N . Since 𝜃 extends to 𝐻, θ ~ extends to 𝐻, and by Lemma 2.2, so does the character obtained after the first step in the transformation from 𝜃 to θ . (Note that θ ¯ = θ ~ × φ for some character 𝜑 all of whose factors are either principal characters or α i for some 𝑖. Since α i extends to Aut ( S i ) , it follows from [18, Lemma 1.3] that 1 M × φ extends to its inertia group in Γ. In particular, 𝜑 extends to 𝐻.) If we need to replace some factors of this character θ ¯ by principal characters, then we can apply Lemma 2.2 for a second time to deduce that θ extends to 𝐻.

Analogously, one can see that if θ extends to 𝐻, then 𝜃 extends to 𝐻 too. This completes the proof.

(iv) Again, this follows from Lemma 2.1 and the fact that 𝐺 is a subgroup of Aut ( F ) . ∎

If 𝑁 is a normal subgroup of a group 𝐺 and T Irr ( N ) , we write

Irr ( G | T ) = φ T Irr ( G | φ ) .

Also, we write

acd ( G | T ) = χ Irr ( G | T ) χ ( 1 ) | Irr ( G | T ) |

for the average of the degrees of the irreducible characters of 𝐺 that lie over characters in 𝒯. Now, we complete the proof of Theorem A.

Theorem 2.5

Let 𝐺 be a group with trivial solvable radical. Then | G | is bounded (from above) in terms of acd ( G ) .

Proof

Let 𝐹, as before, be the generalized Fitting subgroup of 𝐺. Since 𝐺 is isomorphic to a subgroup of Aut ( F ) , it suffices to bound | F | . For this, we want to see that | S i | and u i is bounded in terms of acd ( G ) for every 𝑖.

Let

Δ = { θ 1 , , θ r } { θ 1 , , θ r } { θ r + 1 , , θ s }

be a complete system of representatives of the 𝐺-orbits on Irr ( F ) , where θ i is not defined for i { r + 1 , , s } . Note that such a complete system of representatives exists by Lemma 2.4 (iv). We have

Irr ( G ) = i = 1 r ( Irr ( G | θ i ) Irr ( G | θ i ) ) j = r + 1 s Irr ( G | θ j ) ,

and these subsets form a partition of Irr ( G ) .

Note that if χ Irr ( G | θ j ) for some j > r , then

χ ( 1 ) θ j ( 1 ) i = 1 t m ( S i ) u i / 2 ,

where we have used the inequality that precedes Lemma 2.4. On the other hand, the average degree of the irreducible characters of 𝐺 that lie over T i = { θ i , θ i } (for every i r ) is

acd ( G | T i ) θ i ( 1 ) + θ i ( 1 ) 2 1 2 i = 1 t m ( S i ) u i / 2 ,

where we have used Lemma 2.4 (iii) together with Clifford theory in the first inequality and Lemma 2.4 (i) in the second inequality.

Now, we have

acd ( G ) min { { acd ( G θ i ) i = r + 1 , , s } { acd ( G T i ) i = 1 , , r } } 1 2 i = 1 t m ( S i ) u i / 2 .

If we now take acd ( G ) to be a fixed number, we deduce that 𝑡, u i and m ( S i ) are bounded in terms of acd ( G ) . Since | S i | is bounded in terms of m ( S i ) by Lemma 2.3, the result follows. ∎

Note that our proof of Theorem A depends on the classification of finite simple groups by means of [19, Lemma 4.2]. It would be interesting to find a classification-free proof.

3 Bounding the index of the Fitting subgroup

In this short section, we prove Theorem B. The following is the large orbit theorem that we will use. In the following,

α = log ( 6 ( 24 ) 1 / 3 ) log 3 2.595 .

Recall that if a group 𝐺 acts on a module 𝑉, b ( G , V ) is the size of the largest 𝐺-orbit on 𝑉.

Theorem 3.1

Suppose that a solvable group G > 1 acts faithfully and completely reducibly on a module 𝑉 of possibly mixed characteristic. Then | G | < b ( G , V ) α .

Proof

This is [25, Theorem 3.4]. ∎

Now, we show how to complete the proof of Theorem B. In the following, we use the standard reasoning to get bounds for Gluck’s conjecture from a large orbit theorem. We will write Φ ( G ) to denote the Frattini subgroup of 𝐺.

Theorem 3.2

Let 𝐺 be a solvable group. Then there exists λ Irr ( F ( G ) ) linear such that | G : F ( G ) | acd ( G | λ ) α .

Proof

By Gaschütz’s theorem [17, Theorem 1.12], G / F ( G ) acts faithfully and completely reducibly on V = F ( G ) / Φ ( G ) . By [17, Proposition 12.1], the same holds for the action of G / F ( G ) on Irr ( V ) . Applying Theorem 3.1 to this action, we deduce that there exists λ Irr ( V ) such that the size of the G / F ( G ) -orbit of 𝜆 is

| G : I G ( λ ) | | G : F ( G ) | 1 / α .

By Clifford’s correspondence [9, Theorem 6.11], all the characters in Irr ( G | λ ) are induced from irreducible characters of I G ( λ ) . In particular, if χ Irr ( G | λ ) , then

χ ( 1 ) | G : I G ( λ ) | | G : F ( G ) | 1 / α .

It follows that

acd ( G | λ ) | G : F ( G ) | 1 / α ,

as desired. ∎

The proof of the following strong form of Conjecture C for odd order groups is essentially identical.

Theorem 3.3

Let 𝐺 be an odd order group. Then | G : F ( G ) | acd ( G | λ ) 1.643 .

Proof

Mimic the proof of Theorem 3.2 using [24, Theorem 3.2] rather than Theorem 3.1. ∎

As we have just seen, all known results on Gluck’s conjecture can be quickly adapted to give similar results on Conjecture C. Note also that we could add the condition that 𝜆 has square-free order, i.e. λ Irr ( F ( G ) / Φ ( G ) ) , in Conjecture C and in all the results in this section.

4 Examples

If N G , Irr ( G | N ) stands for the set of irreducible characters of 𝐺 whose kernel does not contain 𝑁. We write acd ( G | N ) to denote the average of the degrees of the characters in this set, i.e.

acd ( G | N ) = χ Irr ( G | N ) χ ( 1 ) | Irr ( G | N ) | .

Our first example shows that | G : F ( G ) | is not bounded from above in terms of acd ( G | F ( G ) ) . This was our first attempt to refine Gluck’s conjecture along the lines followed in this paper.

Example 4.1

Let G = S 3 × F p , where 𝑝 is any odd prime and F p is the Frobenius group of order ( p 1 ) p . We have that

V = F ( G ) = C 3 × C p and G / F ( G ) C 2 × C p 1 = H .

Note that G = H V . Let λ , μ Irr ( V ) with o ( λ ) = 3 and o ( μ ) = p . The action of 𝐻 on Irr ( V ) has three nontrivial orbits, with representatives λ × 1 , 1 × μ and λ × μ . Their inertia groups in 𝐻 are, respectively, C p 1 , C 2 and 1. Since all of them are cyclic, any character in Irr ( V ) extends to its inertia group by [9, Corollary 11.22]. Using Clifford’s correspondence and Gallagher’s theorem [9, Corollary 6.17], we deduce that 𝐺 has p 1 irreducible characters of degree 2 lying over 𝜆, 2 irreducible characters of degree p 1 lying over 𝜇 and one irreducible character of degree 2 ( p 1 ) lying over λ × μ . These are all the characters in Irr ( G | V ) . We deduce that

acd ( G | V ) = ( p 1 ) 2 + 2 ( p 1 ) + 2 ( p 1 ) p + 2 < 6 ,

but | G : F ( G ) | = 2 ( p 1 ) is arbitrarily large. Note that Irr ( G | V ) coincides with the set of nonlinear irreducible characters of 𝐺.

As we have seen, the proof of Theorem B is a consequence of a large orbit theorem. It is known that Gluck’s conjecture cannot be proven as an immediate consequence of a large orbit theorem. More precisely, there are examples of even order solvable groups 𝐺 acting faithfully and completely reducibly on finite modules 𝑉 without orbits of size at least | G | 1 / 2 (see [23, Example 13], for instance). Any counterexample to Gluck’s conjecture, or to Conjecture C, should involve these actions. We show that the group in [23, Example 13] is not a counterexample to Conjecture C.

Example 4.2

Let G = S 4 S 3 . Recall that S 4 is the semidirect product of S 3 acting faithfully and irreducibly on W = C 2 × C 2 so that V = F ( G ) is elementary abelian of order 2 6 . Write G = H V for some subgroup H G / F ( G ) . As in [23, Example 13], the 𝐺-orbits in Irr ( V ) have size 1, 9, 27 and 27. Let μ Irr ( W ) be nonprincipal so that λ = μ × μ × μ Irr ( V ) lies in a 𝐺-orbit of size 27. Note that I H ( λ ) = C 2 S 3 and I G ( λ ) = Q 8 S 3 . By [9, Problem 6.18], 𝜆 extends to its inertia subgroup in 𝐺. Using that every character of the base subgroup of I H ( λ ) extends to its inertia subgroup, it is easy to check that I H ( λ ) = I G ( λ ) / V has 4 linear characters, 2 irreducible characters of degree 2 and 4 irreducible characters of degree 3. In particular, its average character degree is 2. Using Gallagher’s theorem and Clifford theory, we conclude that

| Irr ( G | λ ) | = 10 and acd ( G | λ ) = 2 27 = 2 3 3 .

Thus

| G : F ( G ) | = 2 4 3 4 < 2 2 3 6 = acd ( G | λ ) 2 ,

as desired.

On the other hand, if φ 1 , φ 2 and φ 3 are the three nonlinear irreducible characters of S 4 (two of them of degree 3 and the other one of degree 2), then

χ = ( φ 1 × φ 2 × φ 3 ) G Irr ( G ) .

We have that χ ( 1 ) = 108 = b ( G ) , while acd ( G | λ ) = 54 . This example shows that, even in the case when Gluck’s conjecture does not follow from a large orbit theorem, there is perhaps room for improvement in the bound predicted by Gluck’s conjecture.

5 Further remarks and questions

We start with a very fundamental question on the average character degree. It is even surprising that some results on acd ( G ) have been obtained without an answer to it. As mentioned in the introduction, an affirmative answer to this question, together with Theorem A, would imply that | G : sol ( G ) | is bounded from above in terms of acd ( G ) .

Question 5.1

Let 𝐺 be a finite group and N G . Is it true that acd ( G / N ) is bounded from above in terms of acd ( G ) ? Is it true that even acd ( G / N ) acd ( G ) ?

Recall that it was proved in [4, Theorem 9] that | G : sol ( G ) | acs ( G ) 2 . We propose the following question.

Question 5.2

Let 𝐺 be a finite group. Is it true that | G : sol ( G ) | acd ( G ) 4 ?

An interesting first step would be to achieve this bound for groups with trivial solvable radical. Using results from [2], it has been shown by N. N. Hung that this bound holds when 𝐺 is simple.

Gluck’s conjecture was extended to arbitrary finite groups in [2, Question 5], where it was asked whether | G : F ( G ) | b ( G ) 3 for any finite group. It is interesting to note that, as a consequence of the Guralnick–Robinson inequality | G : F ( G ) | acs ( G ) 2 , it was proved in [2] that | G : F ( G ) | b ( G ) 4 . As A 5 shows, it is not true that if 𝐺 is a finite group, then | G : sol ( G ) | acd ( G ) 3 . We do not know any counterexamples to the following question for arbitrary finite groups.

Question 5.3

Let 𝐺 be a finite group. Is it true that there exists λ Irr ( F ( G ) ) linear such that | G : F ( G ) | acd ( G | λ ) 4 ?

We have seen in Example 4.1 that, unlike for class sizes, the index of the Fitting subgroup of a solvable group cannot in general be bounded in terms of acd ( G ) . We have been unable to decide the answer to the following question, even in the 𝑝-group case. We refer the reader to [4, Theorem 12] for the corresponding result for class sizes.

Question 5.4

Let 𝐺 be a solvable group. Is it true that the derived length of 𝐺 is bounded in terms of acd ( G ) ?

Award Identifier / Grant number: PID2019-103854GB-I00

Funding source: Generalitat Valenciana

Award Identifier / Grant number: AICO/2020/298

Award Identifier / Grant number: CIAICO/2021/163

Funding statement: Research supported by Ministerio de Ciencia e Innovación (Grant PID2019-103854GB-I00 funded by MCIN/AEI/10.13039/501100011033) and Generalitat Valenciana AICO/2020/298 and CIAICO/2021/163.

Acknowledgements

I thank Gabriel Navarro for pointing out an inaccuracy in a previous version of Lemma 2.2 and the anonymous reviewers for their careful reading and many helpful comments.

  1. Communicated by: Olivier Dudas

References

[1] M. J. Collins, On Jordan’s theorem for complex linear groups, J. Group Theory 10 (2007), no. 4, 411–423. 10.1515/JGT.2007.032Suche in Google Scholar

[2] J. P. Cossey, Z. Halasi, A. Maróti and H. N. Nguyen, On a conjecture of Gluck, Math. Z. 279 (2015), no. 3–4, 1067–1080. 10.1007/s00209-014-1403-6Suche in Google Scholar

[3] D. Gluck, The largest irreducible character degree of a finite group, Canad. J. Math. 37 (1985), no. 3, 442–451. 10.4153/CJM-1985-026-8Suche in Google Scholar

[4] R. M. Guralnick and G. R. Robinson, On the commuting probability in finite groups, J. Algebra 300 (2006), no. 2, 509–528. 10.1016/j.jalgebra.2005.09.044Suche in Google Scholar

[5] W. H. Gustafson, What is the probability that two group elements commute?, Amer. Math. Monthly 80 (1973), 1031–1034. 10.1080/00029890.1973.11993437Suche in Google Scholar

[6] N. N. Hung, Characters of p -degree and Thompson’s character degree theorem, Rev. Mat. Iberoam. 33 (2017), no. 1, 117–138. 10.4171/RMI/929Suche in Google Scholar

[7] N. N. Hung and P. H. Tiep, Irreducible characters of even degree and normal Sylow 2-subgroups, Math. Proc. Cambridge Philos. Soc. 162 (2017), no. 2, 353–365. 10.1017/S0305004116000669Suche in Google Scholar

[8] N. N. Hung and P. H. Tiep, The average character degree and an improvement of the Itô–Michler theorem, J. Algebra 550 (2020), 86–107. 10.1016/j.jalgebra.2019.12.022Suche in Google Scholar

[9] I. M. Isaacs, Character Theory of Finite Groups, Dover, New York, 1994. Suche in Google Scholar

[10] I. M. Isaacs, M. Loukaki and A. Moretó, The average degree of an irreducible character of a finite group, Israel J. Math. 197 (2013), no. 1, 55–67. 10.1007/s11856-013-0013-zSuche in Google Scholar

[11] I. M. Isaacs, G. Malle and G. Navarro, A reduction theorem for the McKay conjecture, Invent. Math. 170 (2007), no. 1, 33–101. 10.1007/s00222-007-0057-ySuche in Google Scholar

[12] H. Kurzweil and B. Stellmacher, The Theory of Finite Groups, Universitext, Springer, New York, 2004. 10.1007/b97433Suche in Google Scholar

[13] P. Lescot, Degré de commutativité et structure d’un groupe fini, Rev. Math. Spéciales (1988), 276–279. Suche in Google Scholar

[14] P. Lescot, Degré de commutativité et structure d’un groupe fini, Rev. Math. Spéciales (1989), 200–202. Suche in Google Scholar

[15] M. L. Lewis, Variations on average character degrees and 𝑝-nilpotence, Israel J. Math. 215 (2016), no. 2, 749–764. 10.1007/s11856-016-1393-7Suche in Google Scholar

[16] K. Magaard and H. P. Tong-Viet, Character degree sums in finite nonsolvable groups, J. Group Theory 14 (2011), no. 1, 53–57. 10.1515/jgt.2010.041Suche in Google Scholar

[17] O. Manz and T. R. Wolf, Representations of Solvable Groups, London Math. Soc. Lecture Note Ser. 185, Cambridge University, Cambridge, 1993. 10.1017/CBO9780511525971Suche in Google Scholar

[18] S. Mattarei, On character tables of wreath products, J. Algebra 175 (1995), no. 1, 157–178. 10.1006/jabr.1995.1180Suche in Google Scholar

[19] A. Moretó, Complex group algebras of finite groups: Brauer’s problem 1, Adv. Math. 208 (2007), no. 1, 236–248. 10.1016/j.aim.2006.02.006Suche in Google Scholar

[20] A. Moretó and H. N. Nguyen, On the average character degree of finite groups, Bull. Lond. Math. Soc. 46 (2014), no. 3, 454–462. 10.1112/blms/bdt107Suche in Google Scholar

[21] A. Moretó and T. R. Wolf, Orbit sizes, character degrees and Sylow subgroups, Adv. Math. 184 (2004), no. 1, 18–36. 10.1016/S0001-8708(03)00093-8Suche in Google Scholar

[22] G. Qian, On the average character degree and the average class size in finite groups, J. Algebra 423 (2015), 1191–1212. 10.1016/j.jalgebra.2014.11.005Suche in Google Scholar

[23] T. R. Wolf, Large orbits of supersolvable linear groups, J. Algebra 215 (1999), no. 1, 235–247. 10.1006/jabr.1998.7730Suche in Google Scholar

[24] Y. Yang, Arithmetical conditions of orbit sizes of linear groups of odd order, Israel J. Math. 237 (2020), no. 1, 1–14. 10.1007/s11856-020-1963-6Suche in Google Scholar

[25] Y. Yang, On Gluck’s conjecture, preprint (2020), https://arxiv.org/abs/2010.08073. Suche in Google Scholar

[26] The GAP Group, GAP – Groups, Algorithms, and Programming, Version 4.11.0, 2020, http://www.gap-system.org. Suche in Google Scholar

Received: 2022-07-20
Revised: 2022-11-13
Published Online: 2023-01-17
Published in Print: 2023-07-01

© 2023 Walter de Gruyter GmbH, Berlin/Boston

Heruntergeladen am 19.9.2025 von https://www.degruyterbrill.com/document/doi/10.1515/jgth-2022-0120/html
Button zum nach oben scrollen