Abstract
The Möbius function of the subgroup lattice of a finite group has been introduced by Hall and applied to investigate several questions. In this paper, we consider the Möbius function defined on an order ideal related to the lattice of the subgroups of an irreducible subgroup G of the general linear group \(\textrm{GL}(n,q)\) acting on the n-dimensional vector space \(V=\mathbb {F}_q^n\), where \(\mathbb {F}_q\) is the finite field with q elements. We find a relation between this function and the Euler characteristic of two simplicial complexes \(\Delta _1\) and \(\Delta _2\), the former raising from the lattice of the subspaces of V, the latter from the subgroup lattice of G.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In this paper, motivated by recent and less recent results about the Möbius function \(\mu \) for the subgroup lattice \(\mathcal {L}(G)\) of a finite group G, we give a result which relates the Möbius function for a subgroup G of \(\textrm{GL}(n,q)\) to two simplicial complexes: one defined from the lattice of the subspaces fixed by a reducible subgroup \(H\le G\) and the second from the lattice \(\mathcal {L}(G)\) of the subgroups of G.
We will introduce in the Preliminaries all the definitions and details useful for reading the paper.
In his PhD thesis [8], Shareshian considers the problem of computing \(\mu (1,G)\) for several finite classical groups G; the idea is to approximate \(\mu (1,G)\) through a good function \(f_{G,n,p}(u,1)\,\), such that:
Here, \(G=G(n,p^u)\) denotes a family of finite classical groups with the same defining classical form, which act in a natural way on the vector space V of finite dimension n over the finite field \(\mathbb {F}_q\) of order \(q=p^u\,\). If \(\,\mathcal {C}_1,\dots ,\mathcal {C}_8,\mathcal {C}_9\,\) are the Aschbacher classes of maximal subgroups of a finite classical group (see [4]), \(\mathcal {C}_9\) is the class of almost-simple groups not belonging to the first 8 classes of “geometric type” and the function \(f_{G,n,p}(u,1)\,\) provides an estimate of \(\mu (1,G)\) with respect to the contributions given by the subgroups of G which belong to the classes \(\mathcal {C}_i\,\), for \(i\in \{1,\dots ,8\}\).
Actually, Shareshian’s approach focuses on the first class \(\mathcal {C}_1(G)\), that is, the class of reducible subgroups of G.
In particular, the reducible subgroups of G contribute to \(f_{G,n,p}(u,1)\) through the computation of the Möbius function of
which is obtained by adjoining the maximum G to the order ideal
that is,
and
In this paper, we will consider irreducible subgroups G of the general linear group \(\textrm{GL}(n,q)\), that is, groups of linear automorphisms of a vector space V of dimension n over the finite field \(\mathbb {F}_q\) with q elements, which fix no non-trivial subspace of V. In this hypothesis, we will take a reducible subgroup H of G (that is, H fixes some proper subspace of V) and we will work on the analogue of \(\mu _{\widehat{\mathcal {I}}_1(G)}(1,G)\), namely \(\mu _{\widehat{\mathcal {I}}(G,H)}(H,G)\), so that
(see Sects. 2 and 3 for all precise definitions).
The subject of this paper is somehow motivated by the following conjecture:
Conjecture 1.1
(Mann, [6]) Let G be a PFG group and \(\mu \) the Möbius function on the lattice of open subgroups of G. Then, \(|\mu (H,G)|\) is bounded by a polynomial function in the index |G : H| and the number of subgroups of G of index m with \(\mu (H,G)\ne 0\) grows at most polynomially in m.
Indeed, although the problem is still open in its general setting, it was reduced by Lucchini in [5] to the study of similar growth conditions for finite almost-simple groups.
The following theorem is the main result of the present paper:
Theorem 4.5
Consider a vector space V of finite dimension over \(\mathbb {F}_q\,\). Let G be an irreducible subgroup of \(\,\textrm{GL}(V)\) and \(H\le G\). Then,
\({\widehat{\mathcal {I}}(G,H)}\), \({\Psi '(G,H)}\), \({\Psi (G,H)}\), \(\Delta _i\), and \(\tilde{\chi }(\Delta _i)\) (for \(i=1,\,2\)) are defined in Sect. 3, also for irreducible subgroups H. This will allow us to avoid the restriction to only reducible subgroups H of G in the statement of the Theorem.
In the final section, we will use Theorem 4.5 to deal with \(\mu (H,G)\) in some particular case. In [1], Theorem 4.5 is used to attack Conjecture 1.1 for some class of subgroups H of linear and projective groups.
We thank Andrea Lucchini and Johannes Siemons for many useful discussions.
2 Preliminaries
In this paper, all the groups and sets are finite.
For main results about posets and lattices, we refer to [9]. Here, we just recall some basic fact, useful for reading the paper.
Definition 2.1
Let \(\mathcal {P}\) be a finite poset. The Möbius function associated with \(\mathcal {P}\) is the map \(\,\mu _{\mathcal {P}}:\mathcal {P}\times \mathcal {P}\rightarrow \mathbb {Z}\) satisfying
and defined recursively for \(x\le y\,\) by
Notation
If \(\mathcal {P}\) is the subgroup lattice \(\mathcal {L}(G)\) of G, we will write \(\mu (H,K)\) instead of \(\mu _{\mathcal {L}(G)}(H,K)\).
Definition 2.2
An (abstract) simplicial complex \(\Delta \) on a vertex set T is a collection \(\Delta \) of subsets of T satisfying the two following conditions:
-
if \(t\in T\), then \(\{t\} \in \Delta \);
-
if \(F\in \Delta \) and \(G\subseteq F\), then \( G\in \Delta \).
An element \(F\in \Delta \) is called a face of \(\Delta \), and the dimension of F is defined to be \(|F|-1\). In particular, the empty set \(\emptyset \) is a face of \(\Delta \) (provided \(\Delta \ne \emptyset \)) of dimension \(-1\).
Definition 2.3
The Euler characteristic \(\chi (\Delta )\) of the simplicial complex \(\Delta \) is so defined:
where \(\mathcal {F}_i\) denotes the number of the i-faces that is of the faces of dimension i, in \(\Delta \).
We recall here the definition of an order ideal:
Definition 2.4
Let \((\mathcal {P},\le )\) be a poset. An order ideal of \(\mathcal {P}\) is a subset \(I\subseteq \mathcal {P}\) such that
In particular, if A is a subset of \(\mathcal {P}\), then the set
is the order ideal of \(\mathcal {P}\) generated by A.
Observe that a simplicial complex on a set of vertices T is just an order ideal of the Boolean algebra \(B_T\) that contains all one element subsets of T.
Definition 2.5
Let \(\Delta \) be a simplicial complex, and \(\chi (\Delta )\) be the Euler characteristic of \(\Delta \). The reduced Euler characteristic \(\tilde{\chi }(\Delta )\) of \(\Delta \) is defined by \(\tilde{\chi }(\emptyset )=0\), and \(\tilde{\chi }(\Delta )=\chi (\Delta )-1\) if \(\Delta \ne \emptyset \). See also Equation (3.22) in [9].
It is possible to build a simplicial complex from a poset \(\mathcal {P}\) in the following way:
The order complex \(\Delta (\mathcal {P})\) of \(\mathcal {P}\) is defined as the simplicial complex whose vertices are the elements of \(\mathcal {P}\) and whose k-dimensional faces are the chains \(a_0\prec a_1\prec \cdots \prec a_k\) of length k of distinct elements \(a_0,\ldots ,a_k\in \mathcal {P}\).
Now, denote by \(\widehat{\mathcal {P}}\) the finite poset obtained from \(\mathcal {P}\) by adjoining a least element \(\hat{0}\) and a greatest element \(\hat{1}\). The Möbius function \(\mu _{\widehat{\mathcal {P}}}(\hat{0},\hat{1})\) is related to \(\tilde{\chi }(\Delta (\mathcal {P}))\) by a well-known result by Hall in [3] about the computation of \(\mu _{\widehat{\mathcal {P}}}(\hat{0},\hat{1})\) by means of the chains of even and odd length between \(\hat{0}\) and \(\hat{1}\).
Proposition 2.6
(see [9], Proposition 3.8.6) Let \(\mathcal {P}\) be a finite poset. Then,
3 The ideal \(\mathcal {I}(G,H)\) and the complexes \(\Delta _i\) in Theorem 4.5
In this section, and in the rest of the paper, G is an irreducible subgroup of the general linear group \(\textrm{GL}(n,q)\) over the vector space \(V=\mathbb {F}_q^n\) of finite dimension n over the finite field \(\mathbb {F}_q\) with q elements. We consider the natural action of G on the set of subspaces of V.
We define the order ideal \(\mathcal {I}(G,H)\) and the simplicial complexes \(\Delta _i\) (for \(i=1,2\)) that we consider in Theorem 4.5. In Remark 2, we will explicitly observe that the two complexes \(\Delta _i\) are not the order simplicial complex rising from \(\mathcal {I}(G,H)\).
Given a subgroup G of \(\textrm{GL}(n,q)\) and a subgroup H of G, put
and
Definition 3.1
The reducible subgroup ideal in \(\mathcal {L}(G)_{\ge H}\) is the order ideal of \(\mathcal {L}(G)_{\ge H}\) generated by \(\mathcal {C}(G,H)\,\). Namely,
Remark 1
If H is reducible, that is, H fixes some non-trivial subspace W of V, then \(H\in \mathcal {I}(G,H)\). Otherwise, if H is irreducible, \(H\notin \mathcal {I}(G,H)\) and \(\mathcal {I}(G,H)\) is empty. The subspaces fixed by H are said to be H-invariant.
Definition 3.2
If H is reducible, we set
by adjoining the maximum G to \(\mathcal {I}(G,H)\), which has minimum H. Otherwise, if H is irreducible, we set \(\widehat{\mathcal {I}}(G,H):=\{H,G\}\) by adjoining the minimum H and the maximum G to the empty poset \(\emptyset \).
Remark 2
Here, we just note that the poset \(\mathcal {I}(G,H)\) has already a minimum if H is reducible, so that \(\mu _{\widehat{\mathcal {I}}(G,H)}(H,G)\) is not, in general, the reduced Euler characteristic of the order complex \(\Delta ({\mathcal {I}}(G,H))\) of \(\mathcal {I}(G,H)\).
To define the simplicial complexes \(\Delta _i\) of Theorem 4.5, we begin with fixing some more notation. We denote by \(\mathcal {S}(V,H)\) the lattice of H-invariant subspaces of V and define
Moreover, given an irreducible group \(G\le \textrm{GL}(V)\), and \(H\le G\), we will consider the following three sets:
-
(a)
\(\Psi (G,H):=\{X\subseteq \mathcal {C}(G,H)\,\mid \,\bigcap _{M\in X}M\ne H\,\}\,\);
-
(b)
\(\Psi (G,H)^\complement :=\{Y\subseteq \mathcal {C}(G,H)\,\mid \, \bigcap _{M\in Y}M= H\,\}\,\);
-
(c)
\(\Psi '(G,H):=\{E\subseteq \mathcal {S}(V,H)^*\,\mid \,\bigcap _{W\in E}\textrm{stab}_G(W)\ne H\}.\)
Observe that \(\emptyset \in \Psi (G,H)\) and \(\emptyset \in \Psi '(G,H)\), but \(\emptyset \notin \Psi (G,H)^\complement \,\).
Remark 3
If H is an irreducible subgroup of G, then \(\mathcal {S}(V,H)^*=\emptyset \) and \(\Psi '(G,H)=\{\emptyset \}=\Psi (G,H)\). Since for irreducible H we have that \(\widehat{\mathcal {I}}(G,H)=\{H,G\}\), then Theorem 4.5 is trivially verified in this case.
Definition 3.3
The simplicial complexes \(\Delta _i\) of Theorem 4.5 are so defined:
\(\Delta _1\):
-
The set of vertices \(T_1\) is given by the subspaces \(W\in \mathcal {S}(V,H)^*\) for which \(H\ne \textrm{stab}_G(W)\);
-
the set of faces of \(\Delta _1 \) is given by \(\Psi '(G,H)\).
\(\Delta _2\):
-
The set of vertices \(T_2\) is given by the subgroups \(M\in \mathcal {C}(G,H)\) such that \(H\ne M\);
-
the set of faces of \(\Delta _2 \) is given by \(\Psi (G,H)\).
We explicitly observe what happens in the special case when, for some proper non-trivial subspace W of V, the subgroup \(H=\textrm{stab}_G(W)\) is maximal with respect to the property of being a stabilizer of a proper non-trivial subspace of V in G. In this case, we note that, by definition, \(T_1=T_2=\emptyset \) and the set of faces of \(\Delta _1\) and \(\Delta _2\) is \(\{\emptyset \}\). Then, Theorem 4.5 is trivially verified.
4 Computing \(\mu _{\widehat{\mathcal {I}}(G,H)}(H,G)\)
In order to prove Theorem 4.5, we need Proposition 4.2 that gives a link between \(\Psi (G,H)\) and \(\Psi '(G,H)\) and also shows that the reduced Euler characteristics of the complexes \(\Delta _i\) coincide. The proof of Proposition 4.2 follows at once from Lemma 4.1.
Lemma 4.1
Let T be a subgroup of a finite group L acting on a finite set X, and let \(X'\subseteq X\) be a subset such that \(T\le L_x\,\) for all \(x\in X'\). (As usual, \(L_x\) denotes the stabilizer of x in L.) Set
-
\(\mathcal {L}:=\{L_x\,\mid \,x\in X'\}\,\);
-
\(\mathcal {R}:=\{E\subseteq {\mathcal {L}}\,\mid \, \bigcap _{K\in E}K\ne T\}\,\);
-
\(\mathcal {S}:=\{Q\subseteq X'\,\mid \, \bigcap _{x\in Q}L_x\ne T\}\,\).
Then,
Proof
If \(Q\in \mathcal {S}\) and \(E\in \mathcal {R}\), set
It is immediate to realize that
is the disjoint union of all the \(\mathcal {S}_E\). So, it suffices to show that for each \(E\in \mathcal {R}\) the following identity is verified:
For \(E=\emptyset \), the identity (7) is trivially true. Now, fix a non-empty \(E\in \mathcal {R}\), and for each \(K\in E\) define
Let \(Q\in \mathcal {S}_E\) and observe that Q can be represented as the following disjoint union:
where \(\,Q_K=\{x\in Q\,\mid \, L_x=K\}\subseteq X'_K\,\) and \(Q_K\ne \emptyset \). (This property characterizes the elements Q of \(\mathcal {S}_E\).) Since \(\sum _{\emptyset \ne Q_K\subseteq X'_K}(-1)^{|Q_K|}{=-1}\) (see Remark 4), we get
and we obtain the identity (7). \(\square \)
Proposition 4.2
Let V be a vector space of finite dimension over \(\mathbb {F}_q\,\) and consider a subgroup \(H\le G\le \textrm{GL}(V)\). We have:
Equivalently, \(\tilde{\chi }(\Delta _1 )=\tilde{\chi }(\Delta _2)\).
Proof
Consider the natural action of G on the set of subspaces of V. By Lemma 4.1, the equality follows at once from the definitions of \(\mathcal {S}(V,H)^*\), \(\mathcal {C}(G,H)\), \(\Psi (G,H)\) and \(\Psi '(G,H)\,\).
With previous notation, we take \(T=H\), \(L=G\), \(\mathcal {R}=\Psi (G,H)\), \(\mathcal {S}=\Psi '(G,H)\). \(\square \)
To prove Theorem 4.5, we need Proposition 4.4 which is achieved through Theorem 4.3 (Crosscut Theorem, see [9, Corollary 3.9.4]) and Remark 4.
Remark 4
For every finite set A of cardinality \(n>0\), we have
Theorem 4.3
(Crosscut Theorem) Let L be a finite lattice with minimum \(\hat{0}\) and maximum \(\hat{1}\), so that \(\hat{0}\ne \hat{1}\). Let M be the set of all coatoms in L. Let \(X\subseteq L\) be a subset such that \(M \subseteq X\) and \(\hat{1} \notin X\).
Given \(\mathcal {Y}:= \{Y\subseteq X\,\mid \, Y\ne {\emptyset }\, \text { and }\, \bigcap _{Y\in \mathcal {Y}}Y=\hat{0}\}\), the following equality holds:
Proposition 4.4
Let V be a vector space of finite dimension over \(\mathbb {F}_q\,\). Let \(H\le G\le \textrm{GL}(V)\). Then, we have that
Proof
We observe that \(\widehat{\mathcal {I}}(G,H)\subseteq \mathcal {L}(G)_{\ge H}\) is a lattice because the join of two subgroups \(K_1,K_2\in \mathcal {I}(G,H)\) is either in \(\mathcal {I}(G,H)\) or equal to G. The meet of \(K_1,K_2\in \mathcal {I}(G,H)\) is \(K_1\cap K_2\,\). Hence, \(\widehat{\mathcal {I}}(G,H)\) is a finite lattice, whose set of coatoms is contained in \(\mathcal {C}(G,H)\), and its maximum \(G\notin \mathcal {C}(G,H)\), because G is assumed to be irreducible. Since
by Theorem 4.3 we immediately obtain (9). \(\square \)
Now, observe that the disjoint union \(\Psi (G,H)\cup \Psi (G,H)^\complement \) is the power set of \(\mathcal {C}(G,H)\), so that by Remark 4 we have
If we put together Eqs. (8), (9), and (10), we get:
Theorem 4.5
Let V be a vector space of finite dimension over \(\mathbb {F}_q\,\). Let \(H\le G\le \textrm{GL}(V)\). Then,
Proof
By (10), we have
Then, by Proposition 4.2 and Proposition 4.4,
\(\square \)
5 Final remark
Going back to Conjecture 1.1, we observe that the knowledge of
coming from (11) can be exploited to estimate the value \(\mu (H,G)\) of the Möbius function \(\mu \) of G for \(H\le G\), at least in some particular case. Here, we just give an example for a particular reducible H, taking \(G=\textrm{GL}(n,q)\).
Following the idea suggested by Shareshian in [8], one could write
where \(f_{G,n,q}(H)\) depends on the classes \(\mathcal {C}_i(G,H)\), for \(i=1,\dots ,8\), in Aschbacher’s classification. In some lucky case, H is not contained in maximal subgroups belonging to classes \(\mathcal (C_i),\, i\ne 1,\, 9\). This happens, for example, in the following case:
Let \(V\simeq \mathbb {F}_q^n\,\) be a vector space of finite dimension n over \(\mathbb {F}_q\,\) and fix the following basis of \(V\,\):
so that
If \(W=\langle w_1,\dots ,w_{m}\rangle \), H is the subgroup of \(\textrm{GL}(n,q)\) acting as \(\textrm{GL}(m,q)\) on W and fixing all the elements \(v_{m+1},\dots , v_n\).
We do not give in this context the details of the proof of the following theorem, which needs many technical arguments. In [1], all the details are given.
Theorem 5.1
([2]) Let \(G=\textrm{GL}(n,q)\), and let \(H\le G\) be such that
Let \(q=p\,\) be an odd prime, and let the dimension n be prime. If \(\,n-m+1\,\) is prime, then
so that
In general, we do not have much information about the ninth class. Just to give an example, we considered the groups of low dimension studied by Schröder in her PhD thesis ([7]), and we saw that in dimension \(n=13\) also class \(\mathcal {C}_9(G,H)\) is empty for \(p>5\). In this case, \(\mu (H,G)=0\).
All the details and data needed to prove Theorem 5.1 are contained in [2] and are available if requested.
References
Di Gravina, L.: A closure operator on the subgroup lattice of \({\rm GL}(n, q)\) and \({\rm PGL}(n, q)\) in relation to the zeros of the Möbius function. J. Group Theory 27, 275–296 (2024)
Di Gravina, L.: Some questions about the Möbius function of finite linear groups, Ph.D. Thesis, Università degli Studi di Milano-Bicocca (2022)
Hall, P.: The Eulerian functions of a group. Q. J. Math. 7, 134–151 (1936)
Kleidman, P., Liebeck, M.: The Subgroup Structure of the Finite Classical Groups. Cambridge University Press, Cambridge (1990)
Lucchini, A.: On the subgroups with non-trivial Möbius number. J. Group Theory 13, 589–600 (2010)
Mann, A.: A probabilistic zeta function for arithmetic groups. Internat. J. Algebra Comput. 15, 1053–1059 (2005)
Schröder, A.K.: The maximal subgroups of the classical groups in dimension 13, 14 and 15, Ph.D. Thesis, University of St Andrews (2015)
Shareshian, J.: Combinatorial properties of subgroup lattices of finite groups, Ph.D. Thesis, Rutgers - The State University of New Jersey (1996)
Stanley, R.P.: Enumerative Combinatorics, Volume I, second ed., Cambridge University Press, Cambridge, 2012 (first ed. published by Wadsworth & Brooks/Cole, 1986)
Funding
Open access funding provided by Università degli Studi di Milano - Bicocca within the CRUI-CARE Agreement.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors are members of the Italian National Group for Algebraic and Geometric Structures and their Applications (GNSAGA - INdAM) and thank INdAM for the support. The second author is also a member of the research training group GRK 2240: Algebro-Geometric Methods in Algebra, Arithmetic and Topology, funded by DFG.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Dalla Volta, F., Di Gravina, L. Möbius function of the subgroup lattice of a finite group and Euler characteristic. J Algebr Comb (2024). https://doi.org/10.1007/s10801-024-01329-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10801-024-01329-8