Abstract
The link between modular functions and algebraic functions was a driving force behind the 19th century study of both. Examples include the solutions by Hermite and Klein of the quintic via elliptic modular functions and the general sextic via level 2 hyperelliptic functions. This paper aims to apply modern arithmetic techniques to the circle of “resolvent problems” formulated and pursued by Klein, Hilbert and others. As one example, we prove that the essential dimension at \(p=2\) for the symmetric groups \(S_n\) is equal to the essential dimension at 2 of certain \(S_n\)coverings defined using moduli spaces of principally polarized abelian varieties. Our proofs use the deformation theory of abelian varieties in characteristic p, specifically SerreTate theory, as well as a family of remarkable mod 2 symplectic \(S_n\)representations constructed by Jordan. As shown in an appendix by Nate Harman, the properties we need for such representations exist only in the \(p=2\) case. In the second half of this paper we introduce the notion of \(\mathcal {E}\)versality as a kind of generalization of Kummer theory, and we prove that many congruence covers are \(\mathcal {E}\)versal. We use these \(\mathcal {E}\)versality result to deduce the equivalence of Hilbert’s 13th Problem (and related conjectures) with problems about congruence covers.
Similar content being viewed by others
Notes
When the functions are understood, we denote an algebraic function simply by the cover \(\tilde{X}\rightarrow X\).
The reader may think of any version of the theory of padic analytic spaces they prefer (Tate, Raynaud, Berkovich, or Hüber’s adic spaces), as this will have no bearing on our arguments.
Mutatis mutandis, this follows by the same reasoning as in [21].
Recall that there are exceptional isomorphisms \(\text {Sp}_4(\mathbb {F}_2)\cong O_4^+(\mathbb {F}_3)\cong S_6\).
Recall that there is an exceptional isomorphism of \(O_5^+(\mathbb {F}_3)\) with the Weyl group of \(E_6\).
References
Abbot, R., Bray, J., Linton, S., Nickerson, S., Norton, S., Parker, R., Suleiman, I., Tripp, J., Walsh, P., Wilson, R.: Atlas of finite group representations, v.3, http://brauer.maths.qmul.ac.uk/Atlas/v3/ (2019)
Allcock, D., Carlson, J.A., Toledo, D.: The complex hyperbolic geometry of the moduli space of cubic surfaces. J. Alg. Geom. 11, 659–754 (2002)
Arnold, Vladimir: Shimura, Goro: Superpositions of algebraic functions. Proc. Sym posia in Pure Math. 28, 45–46 (1976)
Michael, J.J.: Barry, Large abelian subgroups of Chevalley groups. J. Austral. Math. Soc. Ser. A 27(1), 59–87 (1979)
Benson, Dave: Spin modules for symmetric groups. J. Lond. Math. Soc. 38(2), 250–262 (1988)
Brassil, Matthew: Reichstein, Zinovy: Hermite’s theorem via Galois cohomology. Arch. Math. 112, 467–473 (2019)
Brauer, Richard: On the resolvent problem. Ann. Mat. Pura Appl. 102, 45–55 (1975)
Brosnan, P., Fakrhuddin, N.: Fixed points in toroidal compact ifications of Shimura varieties and essential dimension of congruence covers. arXiv:2007.01752 (2020)
Buhler, J., Reichstein, Z.: On the essential dimension of a finite group. Compos. Math., 106 (2), 159–179, 1997
Buhler, J., Reichstein, Z.: On Tschirnhaus transformations. In: Topics in Number Theory (University Park, PA, 1997) 467(2) (1999), pp. 127–142
Burkhardt, H.: Grundzüge einer allgemeinen Systematik der hyperelliptischen Functionen I Ordnung. Math. Ann. 35 (1890) 198296
H. Burkhardt, Untersuchungen aus dem Gebiete der hyperelliptischen Modulfunctionen. Zweiter Teil. Math. Ann. 38 (1891), 161224
H. Burkhardt, Untersuchungen aus dem Gebiete der hyperelliptischen Modulfunctionen.III. Math. Ann. 41 (1893), 313343
Chebotarev, N.G.: Die Probleme der modernen Galoisschen Theorie. In: Proceedings of the ICM (1932)
Chebotarev, N.G., Nikolai Grigorievich Chebotarev: The problem of resolvents and critical manifolds. Izvestia Akad. Nauk SSSR 7, 123–146 (1943)
Coray, Daniel: Cubic hypersurfaces and a result of hermite. Duke Math. J. 54, 657–670 (1987)
Deligne, P.: Variétés de Shimura: interprétation modulaire, et techniques de con struction de modèles canoniques. In: Automorphic Forms, Representations and LFunctions (Proc. Sympos. Pure Math., Oregon State Univ., Corvallis, OR, 1977), Part 2 (1979), pp. 247–289
Dickson, L.E., Leonard Eugene Dickson: Representations of the general symmetric group as linear groups in finite and infinite fields. Trans. AMS 9(2), 121–148 (1908)
Dolgachev, I., Ortland, D.: Point Sets in Projective Spaces and Theta Functions (1988)
Duncan, A.., Reichstein, Z.: Pseudoreflection groups and essential dimension. J. Lond. Math. Soc. 90(3), 879–902 (2014)
Duncan, A.: Versality of algebraic group actions and rational points on twisted varieties. J. Alg. Geom. 24, 499–530 (2015)
Faltings, G., Chai, C.L.: Degeneration of abelian varieties, Ergebnisse der Mathematik und ihrer Grenzgebiete (3) [Results in Mathematics and Related Areas (3)], vol. 22 (SpringerVerlag, Berlin, 1990). With an appendix by David Mumford
Farb, Benson, Wolfson, J.: Resolvent degree, Hilbert’s 13th problem and geometry. L’Enseignement Math. 65(3), 303–376 (2019)
Farb, B., Kisin, M., Wolfson, J.: Essential dimension of congruence covers. Compositio Math. 157(11), 2407–2432 (2021)
Fricke, Robert: Lehrbuch der Algebra, vol. 2. Viewig und Sohn, Braunschweig (1926)
R. Fricke and F. Klein, Vorlesungen über die Theorie der automorphen Functionen. Teubner, Berlin, 1912
Garibaldi, S., Merkurjev, A., Serre, J.P.: Cohomological invariants in Galois cohomology. University Lecture Series, vol. 28 (American Mathematical Society, Providence, RI, 2003)
Gordan, Paul: Ueber Gleichungen siebenten Grades mit einer Gruppe von 168 Substi tutionen. Math. Ann. 20, 515–530 (1882)
Green, J.A.: The modular representation algebra of a finite group. Illinois Journal of Mathematics 6(4), 607–619 (1962)
Grove, L.C.: Classical groups and geometric algebra. In: Graduate Studies in Math., vol. 39, AMS (2000)
Hermite, Charles: Sur la résolution de l’equation du cinquième degré. Comptes rendus de l’Académie des Sciences 46, 508–515 (1858)
Hermite, C.: Sur l’invariant du 18e ordre des formes du cinquième degré et sur le rôle qu’il joue dans la résolution de l’équation du cinquième degré, extrait de deux lettres de M Hermite à l’éditeur. J. Reine Angew. Math. 59 (1861), 304–305
Hilbert, David: Uber die Gleichung neunten Grades. Math. Ann. 97(1), 243–250 (1927)
Hilbert, D.: Problems, Mathematical: Proceedings of the 1900 ICM, English translation reprinted in Bull. AMS 37(4), 407–436 (2000)
Hirzebruch, Friedrich: Hilbert’s Modular Group of the Field Q( v 5) and the Cubic Di agonal Surface of Clebsch and Klein. Russ. Math. Surv. 31(5), 153–166 (1976)
Hirzebruch, F.: The Ring of Hilbert Modular Forms for Real Quadratic Fields of Small Discriminant, Modular Functions of One Variable, VI (Proc. Second Int. Conf., Univ. Bonn, Bonn, 1976) Lecture Notes in Math., vol. 627, pp. 287323 (1977)
Hunt, B.: The Geometry of Some Special Arithmetic Quotients. Springer Lect. Notes in Math., vol. 1637 (1996)
G.D. James, The Representation Theory of Symmetric Groups, Lecture Notes in Mathematics, vol. 682. Springer, Berlin 1978
James, G.D., Peel, M.H.: Specht series for skew representations of symmetric groups. Journal of Algebra 56(2), 343–364 (1979)
Karpenko, Nikita A., Merkurjev, Alexander S.: Essential dimension of finite p groups. Invent. Math. 172(3), 491–508 (2008)
Klein, Felix: Ueber eine geometrische Repräsentation der Resolventen algebraischer Gleichungen. Math. Ann. 4, 346–358 (1871)
Klein, Felix: Ueber die Auflösung gewisser Gleichungen vom siebenten und achten Grade. Math. Ann. 15, 252–282 (1879)
Klein, Felix: Vorlesungen über das Ikosaeder und die Auflösung der Gleichungen vom fünften Grade (Lectures on the Icosahedron and the Solution of the Equation of the Fifth Degree). Tübner, Leipzig (1884)
Klein, F.: Zur Theorie der allgemeinen Gleichungen sechsten und siebenten Grades. Math. Ann. 28, 499–532 (1887)
Klein, Felix: Sur la resolution, par les fonctions hyperelliptiques de l’equation du vingtseptieme degre, de laquelle depend la determination des vingtsept droites d’une surface cubique. J. Math. Pure. Appl. 4, 169–176 (1888)
F. Klein, Lectures on Mathematics, MacMillan and Co., New York, 1894
Klein, Felix: Uber die Auflösung der allgemeinen Gleichungen fünften und sechsten Grades. Journal für die reine und angewandte Mathematik 129, 150–174 (1905)
Klein, F.: Gesammelte Mathematische Abhandlungen, Vol. 2, Springer, Berlin, 1922
Klein, F., Gesammelte Mathematische Abhandlungen, Vol. 3, Springer, Berlin, 1922
Klein, F.: Vorlesungen uber die Entwicklung der Mathematik im 19. Jahrhundert, Springer, Berlin (1926)
Klein, F., Fricke, R.: Vorlesungen über die Theorie der elliptischen Modul funktionen, vol. 1, 2. Teubner, Leipzig (1892)
A.S. Kleshchev, Branching rules for symmetric groups and applications. In: R.W. Carter and J. Saxl (eds.), Algebraic Groups and their Representations. Springer, Dordrecht, 1998, pp. 103130
Kondo, S.: The Segre Cubic and Borcherds Products. Arithmetic and Geometry of K3 Surfaces and CalabiYau Threefolds, vol. 67, pp. 549–565 (2013)
Kraft, Hanspeter: A result of Hermite and equations of degree 5 and 6. J. Algebra 297, 234–253 (2006)
Kronecker, Leopold: Ueber die Gleichungen fünften Grades. Journal für die reine und angewandte Mathematik 59, 306–310 (1861)
Landesman, A.: The Torelli map restricted to the hyperelliptic locus. arXiv:1911.02084 (2019)
Merkurjev, Alexander: Essential dimension. Bull. AMS 54(4), 635–661 (2017)
Merkurjev, Alexander: Suslin, Andrei: KCohomology of SeveriBrauer Varieties and the Norm Residue Homomorphism. Math. USSR Izv. 21(2), 307–340 (1983)
Meyer, Aurel: Reichstein, Zinovy: The essential dimension of the normalizer of a maximal torus in the projective linear group. Algebra & Number Theory 3(4), 467–487 (2009)
T. Okuyama, On a certain simple module and cohomology of the symmetric group over GF(2). Kyoto Univ. Dep. Bull. 1581 (2008), 5863
Prokhorov, Y.: Simple finite subgroups of the Cremona group of rank 3. J. Algebraic Geom. 21(3), 563–600 (2012)
Reichstein, Z.: Essential Dimension, Proceedings of the International Congress of Mathematicians (2010)
Reichstein, Zinovy: On a Theorem of Hermite and Joubert. Canad. J. Math. 51(1), 69–95 (1999)
Reichstein, Z., Youssin, B.: Essential dimensions of algebraic groups and a resolution theorem for Gvarieties. Can. J. Math. 52(5), 1018–1056 (2000). With an appendix by J. Kollár and E. Szabò
ShepherdBarron, Nick: Taylor, Richard: Mod 2 and Mod 5 Icosahedral Representa tions. J. Amer. Math. Soc. 10(2), 283–298 (1997)
Sheth, Jagat: Branching rules for two row partitions and applications to the inductive systems for symmetric groups. Communications in Algebra 27(9), 3303–3316 (1999)
Shimura, Goro: On purely transcendental fields of automorphic functions of several variables. Osaka Jour. Math. 1, 1–14 (1964)
D.E. Taylor, Geometry of the Classical Groups Sigma Series in Pure Math. Helderberg Verlag, Berlin, 1992
Uno, K.: The rank variety of a simple module of a symmetric group. RIMS Kokyuroku 1251, 8–15 (2002)
G. van der Geer, Hilbert modular surfaces, SpringerVerlag, Berlin 1988
Wagner, A.: The Faithful Linear Representation of Least Degree of S n and A n over a Field of Characteristic 2. Math. Z. 151, 127–137 (1976)
Wagner, A.: The Faithful Linear Representation of Least Degree of S n and A n over a Field of Odd Characteristic. Math. Z. 154, 103–114 (1977)
Weil, A.: Adeles and algebraic groups. In: Progress in Mathematics, vol. 23. Birkhäuser, Boston, MA (1982). With appendices by M. Demazure and T. Ono
Wolfson, Jesse: Tschirnhaus transformations after Hilbert. L’Enseignement Math. 66(3), 489–540 (2020)
Acknowledgements
The authors would like to thank Robert Guralnick for pointing out a mistake in an earlier version of this paper. We also thank Igor Dolgachev, Bert van Geemen, Bruce Hunt, Aaron Landesman, Curt McMullen, Zinovy Reichstein and Ron Solomon for helpful correspondence. We thank the anonymous referee for helpful comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Vasudevan Srinivas.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors are partially supported by NSF grants DMS1811772 and Jump Trading Mathlab Research Fund (BF), DMS1601054 (MK) and DMS1811846 (JW).
Appendix by Nate Harman Department of Mathematics, University of Michigan, Ann Arbor MI; Email nharman@umich.edu.
Appendices
Appendix by Nate Harman
On quadratic representations of \(S_n\)
1.1 Statement of results
Recall that any linear representation of a pgroup G over a field k of characteristic p contains a nonzero invariant vector, in particular this implies that the only irreducible representation of G over k is the trivial representation. This does not mean that all representations are trivial though, there are nonsplit extensions of trivial representations and understanding their structure is a central part of modular representation theory.
In a nonsemisimple setting, one basic invariant of a representation is its Lowey length. For representations of pgroups in characteristic p it can be defined as follows: Start with a representation V and then quotient it by its space of invariants to obtain a new representation \(V' = V / V^G\), then repeat this process until the quotient is zero. The Lowey length is the number of steps this takes.
In the above work Farb, Kisin, and Wolfson analyze certain special representations of symmetric groups in characteristic 2, the socalled Dickson embeddings. Typically denoted \(D^{(n1,1)}\) in the representation theory literature, these representations have the following key property: Let \(n = 2m\) or \(2m+1\), these representations have Lowey length 2 when restricted to the rank m (which is the maximum possible) elementary abelian 2subgroup \(H_n\) generated by \((1,2), (3,4), \dots , \text { and } (2m1,2m)\).
This motivates the following definition: We say that an irreducible representation of a \(S_n\) in characteristic p is quadratic with respect to a maximal rank elementary abelian psubgroup H if it has Lowey length 2 upon restriction to H. The purpose of this note is to prove first that this is only a characteristic 2 phenomenon, and second that these representations \(D^{(n1,1)}\) are the only representations which are quadratic with respect to some maximal rank elementary abelian psubgroup for n sufficiently large (\(n\ge 9\)).
In characteristic \(p > 2\), the maximal rank elementary abelian psubgroups in \(S_n\) are just those generated by a maximal collection of disjoint pcycles. Our first main theorem tells us that there are no quadratic representations in characteristic \(p >2\), and in fact we can detect the failure to be quadratic here by restricting to a single pcycle.
Theorem A.1
Any irreducible representation of \(S_n\) with \(n \ge p\) in characteristic \(p>2\) which is not a character has Lowey length at least 3 upon restriction to the copy of \(C_p\) generated by \((1,2,\dots , p)\), and therefore is not quadratic with respect to any maximal rank elementary abelian psubgroup.
Note that in any characteristic \(p >2\) the characters of \(S_n\) are just the trivial and sign representations.
In characteristic 2 things are a bit more complicated. While the subgroup \(H_n\) of \(S_{n}\) is a maximal rank elementary 2subgroup, it is no longer the unique such subgroup up to conjugation. Recall that in \(S_4\) there is the Klein four subgroup \(K = \{e, (12)(34), (13)(24), (14)(23) \}\), which is a copy of \(C_2^2\) not conjugate to \(H_4\).
We can construct other maximal rank elementary 2subgroups of \(S_{n}\) by taking products
and up to conjugacy though these are all the maximal rank elementary abelian 2subgroups inside \(S_{n}\).
\(S_8\) has a special irreducible representation \(D^{(5,3)}\) of dimension 8 which upon restriction \(A_8\) decomposes as a direct sum \(D^{(5,3)+} \oplus D^{(5,3)}\) of two representations of dimension 4. These representations realize the “exceptional" isomorphism \(A_8 \cong GL_4(\mathbb {F}_2)\), or rather they realize two different isomorphisms differing by either by conjugating \(A_8\) by a transposition in \(S_8\) or by the inversetranspose automorphism of \(GL_4(\mathbb {F}_2)\). Under this isomorphism the subgroup \(K\times K \subset A_8\) gets identified with the subgroup of matrices of the form
which is manifestly quadratic. Our second main theorem will be to show that there are no other quadratic representations other than the Dickson embedding once n is at least 9.
Theorem A.2
Suppose V is a nontrivial irreducible representation of \(S_n\) with \(n \ge 9\) over a field of characteristic 2 which is quadratic with respect to a maximal rank elementary abelian 2subgroup H. Then \(V \cong D^{(n1,1)}\), and H is conjugate to \(H_n\).
1.2 Proofs of main theorems
We will be assuming a familiarity with the modular representation theory of symmetric groups. A standard reference for this material the book [38] of James, which we will be adopting the notation from and referring to for all the basic results we need. The irreducible representations of \(S_n\) in characteristic p are denoted by \(D^\lambda \), for pregular partitions \(\lambda \) of n. These arise as quotients of the corresponding Specht modules \(S^\lambda \), which are well behaved reductions of the ordinary irreducible representations in characteristic zero.
1.2.1 Proof of Theorem A.1
First we will reduce the problem to just looking at representations of \(S_p\). For that we have the following lemma:
Lemma A.1

(1)
Every irreducible representation V of \(S_n\) with \(n \ge p\) in characteristic \(p > 3\) which is not a character has a composition factor when restricted to \(S_p\) which is not a character.

(2)
Every irreducible representation V of \(S_n\) with \(n \ge 4 \) in characteristic 3 which is not a character has a composition factor when restricted to \(S_4\) which is not a character.
Proof
For part (a) suppose V only has composition factors which are characters when restricted to \(S_p\). If we restrict this to the alternating group \(A_p\) all the composition factors must be trivial, as \(A_p\) only has the trivial character. If we further restrict to \(A_{p1}\) the whole action must be trivial because representations of \(A_{p1}\) are semisimple in characteristic p. However if the action of \(A_{p1}\) is trivial on V then so is the action of the entire normal subgroup generated by \(A_{p1}\) inside \(S_n\), which we know is all of \(A_n\) if \(n>3\). So V must be the trivial as a representation of \(A_n\), and is therefore a character of \(S_n\).
For part (b), let’s again suppose V only has composition factors that are characters when restricted to \(S_4\), which implies it only has trivial composition factors when restricted to \(A_4\). If we further restrict to the Klein four subgroup K the whole action must be trivial because representations of K are semisimple in characteristic \(p \ne 2\). As before we see V must be trivial for the normal subgroup of \(S_n\) generated by K, which we know is all of \(A_n\) for \(n > 4\). Therefore V is a character. \(\square \)
Remark: The modification for characteristic 3 is necessary because in characteristic 3 the only irreducible representations of \(S_3\) are the trivial and sign representations. Theorem A.1 holds vacuously in this case.
It is now enough to prove Theorem A.1 for \(S_p\) in characteristic \(p >3\), and for \(S_4\) in characteristic 3. Let’s first focus on the case where \(p>3\). If \(\lambda \) is a pcore, then Nakayama’s conjecture (which is actually a theorem, see [38] Theorem 21.11) tells us \(D^\lambda = S^\lambda \) is projective, and hence remains projective when restricted to \(C_p\) and therefore has Lowey length p. This leaves those irreducible representations corresponding to hook partitions \( \lambda = (pk, 1^k)\).
In the simplest case where \(\lambda = (p1,1)\) then \(D^\lambda \) is the \((p2)\)dimensional quotient of the standard \((p1)\)dimensional representation \(S^{(p1,1)}\) by its one dimensional space of invariants, and one can easily verify this forms a single \((p2)\)dimensional indecomposable representation of \(C_p\). Peel explicitly computed the decomposition matrices for \(S_p\) in characteristic p (see [38] Theorem 24.1), and it follows from his calculation that the remaining irreducible representations \(D^\lambda \) with \(\lambda = (pk, 1^k)\) for \(1 < k \le p2\) are just exterior powers \(\Lambda ^k D^{(p1,1)}\) of this \((p2)\)dimensional representation.
Since \(k < p\) we know that \(\Lambda ^k D^{(p1,1)}\) is a direct summand of \((D^{(p1,1)})^{\otimes k}\), which as a representation of \(C_p\) is just the unique \((p2)\)dimensional indecomposable representation tensored with itself k times. Tensor product decompositions for representations of cyclic groups are known explicitly ([29] Theorem 3), and in particular it is known that a tensor product of two odd dimensional indecomposable representations of \(C_p\) always decomposes as a direct sum of odd dimensional indecomposable representations. So we see \((D^{(p1,1)})^{\otimes k}\) and \(\Lambda ^k D^{(p1,1)} = D^{(pk, 1^k)}\) only have odd length indecomposable factors when restricted to \(C_p\). If it had Lowey length 1 when restricted to \(C_p\) that means the action is trivial, which implies the action of \(A_p\) must also be trivial as \(A_p\) is simple, but that would imply the original representation of \(S_n\) was a character.
In the characteristic 3 case there are only two irreducible representations of \(S_4\), they are the standard 3dimensional representation \(S^{(3,1)} = D^{(3,1)}\) and its sign twisted version \(S^{(2,1,1)} = D^{(2,1,1)}\). These are 3core partitions so again by Nakayama’s conjecture they are both projective and therefore remain projective when restricted to \(C_3\) and have Lowey length 3. \(\square \)
1.2.2 Proof of Theorem A.2
The overall structure of the proof will be to successively rule different classes of representations and maximal rank elementary abelian 2subgroups through a sequence of lemmas. The first such lemma will let us rule out those irreducible representations \(D^\lambda \) where \(\lambda \) is a 2regular partition with at least 3 parts.
Lemma A.2
If \(\lambda \) is a 2regular partition with at least 3 parts, then the irreducible representation \(D^\lambda \) of \(S_n\) contains a projective summand when restricted to \(S_6\).
Proof: Note that any 2regular partition \(\lambda \) with at least 3 parts can be written as \((3,2,1) + \mu = (\mu _1 +3, \mu _2 + 2 , \mu _3 + 1, \mu _4, \dots , \mu _\ell )\) for some partition \(\mu = (\mu _1, \mu _2, \dots , \mu _\ell )\). James and Peel [39] constructed explicit Specht filtrations for \(Ind_{S_6 \times S_{n6}}^{S_n}(S^{(3,2,1)} \otimes S^\mu )\), which have \(S^\lambda \) as the top filtered quotient. In particular this implies \(Ind_{S_6 \times S_{n6}}^{S_n}(S^{(3,2,1)} \otimes S^\mu )\) has \(D^\lambda \) as a quotient. However by Frobenius reciprocity we know that
So since the left hand side is nonzero, the right hand side is as well.
Now if we look at \(S^{(3,2,1)} \otimes S^\mu \) as a representation of \(S_6\) it is just a direct sum of \(\text {dim} (S^\mu )\) copies of \(S^{(3,2,1)}\), which we know is irreducible and projective by Nakayama’s conjecture. In particular the image under any nonzero homomorphism is also just a direct sum of copies of \(S^{(3,2,1)}\), so \(D^\lambda \) must contain at least one copy of \(S^{(3,2,1)}\) as a direct summand. \(\square \)
Corollary A.1
If \(\lambda \) is a 2regular partition with at least 3 parts, then \(D^\lambda \) is not quadratic with respect to any maximal rank elementary abelian 2subgroup of \(S_n\).
Proof. After conjugating we may assume that our maximal rank elementary abelian subgroup intersects \(S_6\) in an elementary abelian 2group of rank at least 2. The previous lemma says any such irreducible representation must contain projective summand when restricted to \(S_6\), and then this summand remains projective upon restriction to the intersection of \(S_6\) with our maximal rank elementary 2subgroup. Projective representations of \(C_2^2\) have Lowey length 3, so the Lowey length for the entire maximal rank elementary abelian 2subgroup of \(S_n\) must be at least that big. \(\square \)
This reduces the problem to understanding what happens for twopart partitions \(\lambda = (a,b)\). These representations are much better understood then the general case. For one thing, the branching rules for restriction are completely known in this case, although we’ll just need the following simplified version:
Lemma A.3
(See [52] Theorem 3.6, following [66]). If (a, b) is a twopart partition of n with \(ab > 1\) then \(D^{(a1,b)}\) appears as a subquotient with multiplicity one inside the restriction of \(D^{(a,b)}\) to \(S_{n1}\), and the other composition factors are all of the form \(D^{(a1+r,br)}\) with \(r>0\).
Recall that we defined \(H_{2k} \subset S_{2k}\) to be the elementary abelian 2subgroup of \(S_{2k}\) generated by the odd position adjacent transpositions \((2i1, 2i)\) for \(1 \le i \le k\), we will also consider \(H_{2k}\) as a subgroup of \(S_n\) for \(n>2k\) via the standard inclusions of \(S_{2k}\) into \(S_n\). The next lemma will be to settle for us exactly which representations are have Lowey length 2 when restricted to the standard maximal rank elementary abelian subgroup \(H_n\).
Lemma A.4
\(D^{(nk,k)}\) contains a projective summand when restricted to \(H_{2k}\).
Proof: We know from the branching rules (Lemma A.3) that \(D^{(nk,k)}\) contains a copy of \(D^{(k+1,k)}\) as a subquotient when restricted to \(S_{2k+1}\), so it is enough to verify it for \(D^{(k+1,k)}\). Moreover \(H_{2k} \subset S_{2k}\) so really this calculation is taking place inside \(M(2k) := \text {Res}^{S_{2k+1}}_{S_{2k}}D^{(k+1,k)}\).
These representations \(D^{(k+1,k)}\) and M(2k) are well studied. Benson proved \(D^{(k+1,k)}\) is a reduction modulo 2 of the socalled basic spin representation of \(S_{2k+1}\) in characteristic zero ([5] Theorem 5.1). Nagai and Uno ([69] Theorem 2, or see [60] Proposition 3.1 for an account in English), gave explicit matrix presentations for M(2k) and showed that they have the following recursive structure:
In particular since M(2) can easily be seen to be the regular representation of \(S_2 = H_2\), it follows by induction that M(2k) is projective (and just a single copy of the regular representation) for \(H_{2k}\). \(\square \)
Corollary A.2
The only nontrivial irreducible representation of \(S_n\) which is quadratic with respect to \(H_n\) is \(D^{(n1,1)}\).
Proof: Corollary A.1 tells us that if \(\lambda \) has at least 3 parts, \(D^\lambda \) has Lowey length at least 3 when restricted to \(H_n\). Then Lemma A.4 tells us that \(D^{(nk,k)}\) has Lowey length at least \(k+1\) as a \(H_n\) representation and is therefore not quadratic for \(k>1\). \(\square \)
To finish the proof of Theorem A.2 we need to show that for n at least 9 that there are no representations which are quadratic with respect to to any of these other maximal rank elementary abelian 2subgroups \(K^m \times H_{n4k}\) with \(m \ge 1\). Lemma A.1 rules out \(D^\lambda \) for \(\lambda \) of length at least 3, so again we will just need to address the case when \(\lambda \) is a length 2 partition.
We do this through a series of lemmas ruling out different cases, but first will state the following wellknown fact from the modular representation theory of symmetric groups:
Lemma A.5
([38] Theorem 9.3). If \(\lambda \) is a partition of n, then \(S^\lambda \) restricted to \(S_{n1}\) admits a filtration
such that the successive quotients \(M_i / M_{i1}\) are isomorphic to Specht modules \(S_\mu \), and \(S^\mu \) appears if and only if \(\mu \) is obtained from \(\lambda \) by removing a single box, in which case it appears with multiplicity one.
Lemma A.6
\(D^{(n1,1)}\), for \(n \ge 5\), contains a projective summand when restricted to K, and is therefore not quadratic with respect to any group containing K.
Proof: It suffices to prove it for \(D^{(4,1)}\) as every \(D^{(n1,1)}\) for \(n >5\) contains it as a composition factor upon restriction to \(S_5\) by Lemma A.3. This representation \(D^{(4,1)}\) is just the 4 dimensional subspace of \(\mathbb {F}_2^5\) where the sum of the coordinates is zero. If we restrict this representation to \(S_4\) this can be identified with the standard 4dimensional permutation representation via the map \((a,b,c,d) \rightarrow (a,b,c,d,abcd)\). The restriction of the standard action of \(S_4\) on a 4element set to K is simply transitive, so this representation is just a copy of the regular representation. \(\square \)
Lemma A.7
\(D^{(n2,2)}\) for \(n \ge 7\) and \(D^{(n3,3)}\) for \(n \ge 9\) each contain a projective summand when restricted to K, and are therefore not quadratic with respect to any group containing K.
Proof: It suffices to prove it for \(D^{(5,2)}\) and \(D^{(6,3)}\) as every \(D^{(n2,2)}\) for \(n >7\) contains \(D^{(5,2)}\) as a composition factor upon restriction to \(S_7\), and similarly every \(D^{(n2,2)}\) for \(n >7\) contains \(D^{(6,3)}\) as a composition factor upon restriction to \(S_9\) by Lemma A.3.
For \(S_7\) and \(S_9\) the decomposition matrices are known explicitly and we have that \(D^{(5,2)} = S^{(5,2)}\) and \(D^{(6,3)} = S^{(6,3)}\) (see the appendix of [38]). For Specht modules the branching rules are given by Lemma A.5 and \(S^{(5,2)}\) and \(S^{(6,3)}\) both contain \(S^{(4,1)}\) as a subquotient upon restriction to \(S_5\). The result then follows from the previous lemma. \(\square \)
Lemma A.8
\(D^{(nk,k)}\) for \(k \ge 4\) and \(n \ge 2k+1\) is not quadratic when restricted to \(K^m \times H_{n4m}\) for any \(m \ge 1\).
Proof: We know by Lemma A.4 these are projective upon restriction to \(H_{2k}\), and are therefore projective when restricted to the intersection of \(H_{2k}\) with \(K^m \times H_{n4m}\). This intersection has rank at least 2 since \(k\ge 4\), and therefore projective objects have Lowey length at least 3. This completes the proof of Theorem A.2. \(\square \)
1.3 Modifications for \(A_n\)
We will now briefly describe what changes if we work with alternating groups instead of symmetric groups, but we will omit some of the details of the calculations. First a quick summary of the modular representation theory of alternating groups in terms of the theory for symmetric groups:
Upon restriction from \(S_n\) to \(A_n\), the irreducible representations \(D^\lambda \) either remain irreducible, or split as a direct sums \(D^\lambda \cong D^{\lambda +} \oplus D^{\lambda }\) of two irreducible nonisomorphic representations of the same dimension; all irreducible representations of \(A_n\) are uniquely obtained this way. We’ll note that in characteristic \(p > 2\) this is a standard application of Clifford theory, but in characteristic 2 it is a difficult theorem of Benson ([5] Theorem 1.1). Moreover it is known exactly which \(D^\lambda \) split this way, but we won’t go into the combinatorics here.
When \(p >2\) the maximum rank abelian pgroups in \(S_n\) all lie in \(A_n\), and the proof of Theorem A.1 goes through without modification to give the following theorem.
Theorem A.1’. Any nontrivial irreducible representation of \(A_n\) with \(n \ge p\) in characteristic \(p>2\) has Lowey length at least 3 upon restriction to the copy of \(C_p\) generated by \((1,2,\dots , p)\), and is therefore not quadratic with respect to any maximal rank elementary abelian psubgroup.
When \(p=2\), the difference is more dramatic. It is no longer true that every maximum rank abelian 2subgroup of \(S_n\) lies in \(A_n\), in particular \(H_n\) is not a subgroup of \(A_n\). Let \(\tilde{H}_n\) denote the intersection of \(H_n\) and \(A_n\), this has rank one less than \(H_n\). The maximal rank elementary abelian 2subgroups of \(A_n\) are as follows:
If \(n = 4b\) or \(4b+1\) then up to conjugacy the only maximal rank elementary abelian 2subgroup inside \(A_n\) is \(K^b\), and it is of rank 2b. If \(n = 4b+2\) or \(4b+3\) then all maximal rank elementary abelian 2subgroups in \(S_n\) still have maximal rank when intersected with \(A_n\), and up to conjugacy the maximal rank elementary abelian 2subgroups inside \(A_n\) are of the form:
and these have rank \(2b1\).
The appropriate modification to Theorem A.2 for alternating groups is the following:
Theorem A.2’. Suppose V is a nontrivial irreducible representation of \(A_n\) with \(n \ge 9\) over a field of characteristic 2 which is quadratic with respect to a maximal rank elementary abelian 2subgroup H. Then \(n \equiv 2 \text { or } 3\) modulo 4, \(V \cong D^{(n1,1)}\), and H is conjugate to \(\tilde{H}_n\).
The proof of Theorem A.2 mostly goes through in this case. Some additional care is needed to handle the representations \(D^{\lambda +}\) and \(D^{\lambda }\) which are not restrictions of irreducible representations of \(S_n\), however one simplifying observation is that since \(D^{\lambda +}\) and \(D^{\lambda }\) just differ by conjugation by a transposition, they are actually isomorphic to one another upon restriction to a maximal rank elementary abelian 2subgroup. We will omit the remaining details though.
Rights and permissions
About this article
Cite this article
Farb, B., Kisin, M. & Wolfson, J. Modular functions and resolvent problems. Math. Ann. 386, 113–150 (2023). https://doi.org/10.1007/s00208022023958
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00208022023958