Specht module

Intuition [Beginner]

A Specht module is the irreducible representation of the symmetric group $S_n$, built explicitly out of tableaux instead of through abstract characters. Wilhelm Specht's 1935 construction takes a Young diagram of shape $\lambda ⊢n$, fills its boxes with $1$ through $n$ in every possible way, and treats those fillings as basis vectors. A clever antisymmetrisation along columns then carves out the irreducible representation $S^{\lambda }$ as a concrete subspace.

Specht's construction has one advantage that Frobenius's character theory and Young's symmetriser construction lack: it works in any characteristic. Over the rational numbers (or complex numbers), the Specht modules $\{S^{\lambda }{\}}_{\lambda ⊢n}$ are exactly the irreducible representations of $S_n$ — the same list given by Frobenius and Young. Over a field of characteristic $p$ where $p$ divides $n!$, Maschke's theorem fails (representations stop being completely reducible), and the Specht modules are no longer always irreducible — but they remain the right starting point. The irreducible modular representations are obtained as the heads $D^{\lambda }=S^{\lambda }/\mathrm{rad}(S^{\lambda })$ for $p$-regular partitions $\lambda$.

This characteristic-flexible construction is why Specht modules are the workhorse of modern symmetric-group representation theory: every modern textbook (Sagan, James, James-Kerber, Mathas) builds the theory through Specht modules rather than through Young symmetrisers.

Visual [Beginner]

A polytabloid: take a Young diagram, fill it with $1$ through $n$, then antisymmetrise along columns by adding signed sums over column-permuted versions. The result lives inside the Specht module — the irreducible representation of $S_n$.

Worked example [Beginner]

Take $\lambda =(2,1)$ for $S_3$. The Young diagram has $3$ cells: two in the top row and one in the bottom. Fill the diagram with $1,2,3$ in some order — say $T=\begin{array}{cc}1& 2\\ 3& \end{array}$ — and form the tabloid $\{T\}$ by ignoring row order. Two tabloids are equal iff their row contents (as sets) match.

Tabloids of shape $(2,1)$: there are exactly $3$ tabloids, indexed by which of $\{1,2,3\}$ ends up in the bottom cell. Call them $\{T_1\},\{T_2\},\{T_3\}$ where $T_i$ has $i$ on the bottom.

The polytabloid of $T=\begin{array}{cc}1& 2\\ 3& \end{array}$ is

$$e_T=\{T\}-\{T^{\prime }\},$$

where $T^{\prime }$ is the tableau with the column $\{1,3\}$ swapped: $T^{\prime }=\begin{array}{cc}3& 2\\ 1& \end{array}$. So $e_T=\{T_3\}-\{T_1\}$ in our indexing.

The Specht module $S^{(2,1)}$ is spanned by all polytabloids as $T$ varies. With three columns to antisymmetrise (one column has two cells, the other has one), one can check $S^{(2,1)}$ is $2$-dimensional, matching the dimension of the standard representation of $S_3$. What this tells us: the Specht module realises the abstract irreducible representation as a concrete vector space with a concrete basis you can compute with on paper.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Fix a field $k$ and a partition $\lambda ⊢n$.

Definition (Tabloid). A Young tableau of shape $\lambda$ is a bijection $T:[\lambda ]\to \{1,2,\dots ,n\}$ — i.e., a filling of the Young diagram $[\lambda ]$ with each of $1,\dots ,n$ used exactly once. Two tableaux are row-equivalent if one is obtained from the other by permuting entries within each row. An equivalence class is a tabloid, written $\{T\}$.

The number of tabloids of shape $\lambda$ is

$$\left(\genfrac{}{}{0ex}{}{n}{\lambda }\right):=\frac{n!}{{\lambda }_1!{\lambda }_2!\cdots {\lambda }_k!}.$$

The set of tabloids carries a left $S_n$-action by permuting entries: $\sigma \cdot \{T\}=\{\sigma T\}$.

Definition (Permutation module). The permutation module on tabloids is the $k{S}_n$-module

$$M^{\lambda }:=k\{\{T\}:T\text{ tableau of shape }\lambda \},$$

a free $k$-module of rank $\left(\genfrac{}{}{0ex}{}{n}{\lambda }\right)$. It is the induced module ${\mathrm{Ind}}_{S_{\lambda }}^{S_n}(\mathrm{triv})$ where $S_{\lambda }:=S_{{\lambda }_1}\times S_{{\lambda }_2}\times \cdots \times S_{{\lambda }_k}$ is the Young subgroup.

Definition (Polytabloid). For a tableau $T$ of shape $\lambda$, the column stabiliser $C(T)\subseteq S_n$ is the subgroup of permutations preserving each column of $T$ as a set. The polytabloid associated to $T$ is

$$e_T:=\sum _{\sigma \in C(T)}\mathrm{sgn}(\sigma )\cdot \{\sigma T\}\in M^{\lambda }.$$

Definition (Specht module). The Specht module of shape $\lambda$ over $k$ is

$$S^{\lambda }:=k\{e_T:T\text{ tableau of shape }\lambda \}\subseteq M^{\lambda },$$

the $k$-span of all polytabloids. It is a $k{S}_n$-submodule of $M^{\lambda }$.

Theorem (Specht 1935; James 1976).

(1) Over $k$ of characteristic $0$ (or characteristic not dividing $n!$), the Specht modules $\{S^{\lambda }{\}}_{\lambda ⊢n}$ are exactly the irreducible $k{S}_n$-modules.

(2) Over $k$ of characteristic $p$, define $D^{\lambda }:=S^{\lambda }/\mathrm{rad}(S^{\lambda })$ where $\mathrm{rad}(S^{\lambda })$ is the radical of an explicit symmetric bilinear form on $S^{\lambda }$. Then $D^{\lambda }\ne 0$ iff $\lambda$ is $p$-regular (no part is repeated $p$ times), and $\{D^{\lambda }:\lambda ⊢n,p\text{-regular}\}$ are exactly the irreducible $k{S}_n$-modules.

Standard basis (Specht-James-Kerber). The polytabloids $\{e_T:T\text{ standard}\}$ — those with $T$ a standard Young tableau, increasing along rows and down columns — form a $\mathbb{Z}$-basis of $S^{\lambda }$ (over any base ring including $\mathbb{Z}$). Hence

$${\dim }_k{S}^{\lambda }=f^{\lambda }=\frac{n!}{{\prod }_{(i,j)\in [\lambda ]}h(i,j)},$$

independently of the field $k$, recovering the hook length formula from 07.05.02.

Counterexamples to common slips.

• $S^{\lambda }$ is not a permutation module — the polytabloids are signed sums of tabloids, not single tabloids.
• $D^{\lambda }$ for $p$-regular $\lambda$ is not the same as $S^{\lambda }$ — the former is the irreducible head, the latter is generally reducible in modular characteristic.
• $p$-regularity means no part repeated $p$ times; $\lambda =(3,3,3)$ is not $3$-regular even though $|\lambda |=9$ has no special divisibility issue.

Key theorem with proof [Intermediate+]

Theorem (James submodule theorem, 1976). Let $U\subseteq M^{\lambda }$ be a $k{S}_n$-submodule. Then either $S^{\lambda }\subseteq U$ or $U\subseteq (S^{\lambda }{)}^{\perp }$, where the orthogonal is taken with respect to the natural symmetric bilinear form $⟨\{T\},\{T^{\prime }\}⟩={\delta }_{\{T\},\{T^{\prime }\}}$ on $M^{\lambda }$.

This dichotomy is the structural backbone of modular representation theory of $S_n$: it forces the radical of the Specht module to be $S^{\lambda }\cap (S^{\lambda }{)}^{\perp }$, identifying $D^{\lambda }=S^{\lambda }/(S^{\lambda }\cap (S^{\lambda }{)}^{\perp })$ as the irreducible head.

Proof sketch.

Step 1 (Garnir relations). For a tableau $T$ and a pair of subsets $X\subseteq \mathrm{column}(T,j),Y\subseteq \mathrm{column}(T,j+1)$ with $|X|+|Y|>$ length of column $j$, the Garnir element $G_{X,Y}:={\sum }_{\sigma }\mathrm{sgn}(\sigma )\sigma$ (sum over coset representatives of $S_X\times S_Y$ in $S_{X\cup Y}$) annihilates $\{T\}$ in $M^{\lambda }$. The Garnir relations express how column antisymmetrisation interacts with row contents.

Step 2 (Standard basis). Using Garnir relations, every polytabloid $e_T$ can be reduced to a $\mathbb{Z}$-linear combination of standard polytabloids $e_{T_0}$ for $T_0$ a standard Young tableau. So $\{e_T:T\text{ standard}\}=\{e_T:T\in \mathrm{SYT}(\lambda )\}$ spans $S^{\lambda }$, and a careful count (or an inductive dimension argument) shows it is a basis.

Step 3 (Submodule dichotomy). Suppose $U\subseteq M^{\lambda }$ is a submodule and $u\in U$ is non-zero. Express $u={\sum }_T{c}_T\{T\}$. Apply the signed column sum operator ${\kappa }_T={\sum }_{\sigma \in C(T)}\mathrm{sgn}(\sigma )\sigma$ to $u$. By a calculation involving Garnir relations and the row-stabiliser action,

$${\kappa }_T\cdot \{T^{\prime }\}=\{\begin{array}{ll}\pm e_T& \text{if }\{T^{\prime }\}=\{T\}\\ 0\text{ or torsion}& \text{otherwise (depends on characteristic).}\end{array}$$

If for some $T$ the coefficient $c_T$ in $u$ is non-zero, then ${\kappa }_T\cdot u$ is a non-zero scalar multiple of $e_T$. So $e_T\in U$ for some $T$, and by $S_n$-equivariance, $S^{\lambda }=⟨e_T:T⟩\subseteq U$. Otherwise, $u\in (S^{\lambda }{)}^{\perp }$ for the natural pairing. Hence the dichotomy: $S^{\lambda }\subseteq U$ or $U\subseteq (S^{\lambda }{)}^{\perp }$. $\square$

The proof in characteristic $0$ follows directly: the symmetric bilinear form is non-degenerate (since $|S_n|$ is invertible), so $S^{\lambda }\cap (S^{\lambda }{)}^{\perp }=0$ and $S^{\lambda }$ is irreducible. In characteristic $p$ dividing $n!$, the form may degenerate on $S^{\lambda }$, and the radical $S^{\lambda }\cap (S^{\lambda }{)}^{\perp }$ produces $D^{\lambda }=S^{\lambda }/\mathrm{rad}$ as the irreducible head. ^{[Sagan The Symmetric Group Ch. 2; James LNM 682; Fulton-Harris §4; Mathas Iwahori-Hecke Algebras]}

Bridge. The construction here builds toward later units of the strand, where the same pattern is taken up at higher structure. The defining pattern appears again in those units in a sharpened form, where the local data is glued or quotiented. Putting these together, the foundational insight is that the data of this unit gives the structural signature that the rest of the strand reads off.

Exercises [Intermediate+]

Lean formalization [Intermediate+]

lean_status: partial — Mathlib has Equiv.Perm.cycleType, Combinatorics.Young.YoungDiagram, and basic representation theory, but not the polytabloid-based Specht module construction or its modular structure.

[object Promise]

Advanced results [Master]

Decomposition matrices and the Carter conjecture. For the $p$-modular representation theory of $S_n$, the decomposition matrix $D=(d_{\lambda \mu })$ records the multiplicities $[S^{\lambda }:D^{\mu }]$ of irreducible composition factors of Specht modules. Key facts:

• $d_{\lambda \mu }=0$ unless $\lambda ⊴\mu$ in dominance order ($\mu$ dominates $\lambda$); $d_{\lambda \lambda }=1$ for $p$-regular $\lambda$.
• The $p$-modular characters of $S_n$ are determined by the decomposition matrix, but computing it for general $n$ and $p$ remains an open problem.
• James's conjecture (1990) predicted the decomposition numbers in terms of adjustment matrices and the LLT algorithm; resolved positively for $p\ge h(\lambda )\cdot e$ where $e$ is the quantum characteristic.

Hecke algebras and $q$-Schur algebras. The Iwahori-Hecke algebra $H_n(q)$ deforms $\mathbb{C}[S_n]$ via parameter $q$; the $q$-Schur algebra $S_q(n,r)$ generalises Schur-Weyl duality to the $q$-deformed setting. At roots of unity, $H_n(q)$ behaves analogously to ${\mathbb{F}}_p[S_n]$ in modular representation theory, and the (quantum) Specht modules over $H_n(q)$ provide the analogous building blocks. The LLT algorithm (Lascoux-Leclerc-Thibon 1996) computes decomposition numbers of Hecke algebras at roots of unity via crystals of affine ${\mathfrak{sl}}_e$.

Categorification. The categorical structures of Khovanov-Lauda-Rouquier (KLR algebras / quiver Hecke algebras, mid-2000s) categorify the Hecke algebra and reveal the modular representation theory of $S_n$ as a piece of an affine ${\mathfrak{sl}}_p$-action on the direct sum ${⨁}_n\mathrm{Rep}(S_n)$. This affine Lie-theoretic structure had been conjectured by Lascoux-Leclerc-Thibon and Ariki and was made categorical by Khovanov-Lauda and Rouquier.

Schur-Weyl duality in characteristic $p$. The classical Schur-Weyl duality $V^{\otimes n}\cong {⨁}_{\lambda }{S}^{\lambda }V\otimes V^{\lambda }$ extends to characteristic $p$ but with subtleties: the Schur algebra $S(n,r):={\mathrm{End}}_{S_r}(V^{\otimes r})$ is not always semisimple, the polynomial $\mathrm{GL}$-representations and the $S_n$-Specht modules form a highest-weight category with quasi-hereditary structure (Cline-Parshall-Scott 1988), and the decomposition matrix factors through both rep theories.

Cellular algebras (Graham-Lehrer 1996). The group algebra $k{S}_n$ is a cellular algebra with cell modules the Specht modules. This abstract framework subsumes Specht modules, Brauer algebras, and Hecke algebras, and provides a uniform framework for the structure of cell modules: each cell module has a natural bilinear form whose radical is the largest proper submodule, and the simple heads exhaust the irreducibles.

Plancherel measure and asymptotics. For random $\lambda ⊢n$ chosen with probability $(f^{\lambda }{)}^2/n!$ (Plancherel measure), the rescaled diagram converges to the Vershik-Kerov-Logan-Shepp limit shape; the longest row scales as $\sqrt{n}$ with Tracy-Widom fluctuations (Baik-Deift-Johansson 1999). These deep probabilistic results connect symmetric-group rep theory to integrable probability and random matrix theory.

Synthesis. This construction generalises the pattern fixed in 07.05.01 (symmetric group representation), with the symmetric data replaced by its skew or twisted analogue. Read in the opposite direction, the construction is dual to the metric story: complements and orthogonality are taken with respect to the bilinear datum of this unit, not a metric, and the resulting category of subobjects is the one the rest of the strand classifies. The central insight is that this datum identifies algebra with geometry: functions become vector fields, subspaces become quotients, and invariants become cohomology classes — and that identification is the engine driving every theorem downstream.

Full proof set [Master]

The Specht module construction and the submodule theorem are presented in Sagan The Symmetric Group §2.3-§2.4 (about 25 pages of careful exposition with all proofs), James Representation Theory of the Symmetric Groups (LNM 682, 1978) §4-§7 (the canonical modular treatment), and James-Kerber Encyclopedia of Mathematics Vol. 16 (1981) Ch. 7 (encyclopaedic). The Garnir relations and standard basis theorem are in Sagan §2.6, James §8. The Mullineux conjecture is proved in Ford-Kleshchev "A proof of the Mullineux conjecture" Math. Z. 226 (1997) 267-308 and independently in Bessenrodt-Olsson "On Mullineux symbols" J. Combin. Theory A 81 (1998). The branching rule is in Sagan §2.8 and Fulton-Harris §4.2. Cellular algebra structure is developed in Graham-Lehrer "Cellular algebras" Invent. Math. 123 (1996) 1-34, with the Specht-as-cell-module case worked through. The Schur algebra perspective is in Donkin The $q$-Schur Algebra (LMS Lecture Notes 253, 1998) and Mathas Iwahori-Hecke Algebras and Schur Algebras of the Symmetric Group (AMS 1999). Categorification via KLR algebras is in Khovanov-Lauda "A diagrammatic approach to categorification of quantum groups I-III" (2008-2012) and Brundan-Kleshchev "Blocks of cyclotomic Hecke algebras and Khovanov-Lauda algebras" Invent. Math. 178 (2009).

Connections [Master]

• 07.05.01 (symmetric group representation) is the parent unit; Specht modules are the characteristic-flexible realisation of the irreducibles classified by Frobenius-Young. 07.05.02 (Young diagrams) provides the combinatorial scaffolding — every Specht module has an explicit basis indexed by standard Young tableaux. 07.01.01 (group representation) is invoked structurally; Specht modules realise abstract $S_n$-modules concretely. 07.01.02 (Schur's lemma) is invoked at the irreducibility step in characteristic $0$. The categorification picture connects to 07.06.06 (Verma modules) via the Lascoux-Leclerc-Thibon affine ${\mathfrak{sl}}_p$ action: the Specht modules of $S_n$ over ${\bar{\mathbb{F}}}_p$ form a level-$1$ integrable representation of $\widehat{{\mathfrak{sl}}_p}$, and the modular branching rule is an instance of crystal-theoretic branching. Through Schur-Weyl duality, modular Specht modules connect to modular representations of ${\mathrm{GL}}_d({\bar{\mathbb{F}}}_p)$ — the basis of Steinberg's tensor product theorem and modular highest-weight theory.

Historical & philosophical context [Master]

Wilhelm Specht (1907-1980) was a German mathematician who completed his doctorate in Leipzig in 1932 under Heinrich Brandt, then habilitated in Königsberg in 1937. His 1935 paper Die irreduziblen Darstellungen der symmetrischen Gruppe, published in Mathematische Zeitschrift 39 (1935) 696-711, gave the polytabloid construction that now bears his name. The motivation was concrete: Specht wanted a construction of the irreducible representations of $S_n$ that did not depend on choosing a base field where Maschke's theorem applies. Young's 1900-1933 symmetriser construction worked over $\mathbb{C}$ but the proofs of irreducibility relied on full reducibility. Frobenius's 1900 character calculation gave the dimensions and characters but not an explicit module. Specht's polytabloid antisymmetrisation produced an actual module, defined over any base ring, whose irreducibility over $\mathbb{C}$ could then be verified directly via the bilinear form.

The deeper significance — that Specht modules are well-defined over any field, even when the resulting irreducibility theorem fails — was not fully appreciated until modular representation theory of finite groups developed in the 1940s-1960s through the work of Brauer (decomposition theory) and others. The systematic application of Specht modules to modular representation theory of $S_n$ began with Gordon James's 1976 Cambridge Tracts article and his definitive 1978 monograph in the Springer Lecture Notes series (LNM 682), which proved the submodule theorem, the standard basis theorem, and the classification of modular irreducibles via $p$-regular partitions and the heads $D^{\lambda }$. James-Kerber's two-volume Encyclopedia article (1981) consolidated the field.

Subsequent developments connected Specht modules to deep structures elsewhere in mathematics. The Lascoux-Leclerc-Thibon algorithm (1996) revealed an affine Lie-theoretic structure on the modular characters, with the decomposition numbers of $S_n$ over ${\bar{\mathbb{F}}}_p$ controlled by the canonical basis of the basic representation of $\widehat{{\mathfrak{sl}}_p}$. The Mullineux conjecture (Mullineux 1979, proved Ford-Kleshchev 1997 and Bessenrodt-Olsson 1998) gave the modular analogue of $V^{\lambda }\otimes \mathrm{sgn}\cong V^{{\lambda }^{\prime }}$. The categorification programme of Khovanov-Lauda (mid-2000s) and Rouquier elevated the Hecke algebra and quantum group structure to a fully categorical level, with Specht modules appearing as classes in a $2$-categorical action.

The Specht module construction is now the standard starting point for modular representation theory of $S_n$, with five major monographs (Sagan, James, James-Kerber, Mathas, Klesh­chev) covering the field at increasing levels of depth. Specht's original 1935 paper, only sixteen pages long and written in clear mathematical prose, remains a remarkable example of how a single careful construction — an antisymmetrised sum over a column stabiliser — can launch an entire research field that continues to develop ninety years later. ^{[Specht 1935; James 1976; James-Kerber 1981]}

Bibliography [Master]

• Specht, Die irreduziblen Darstellungen der symmetrischen Gruppe, Mathematische Zeitschrift 39 (1935) 696-711.
• James, The Representation Theory of the Symmetric Groups, Lecture Notes in Mathematics 682, Springer (1978).
• James-Kerber, The Representation Theory of the Symmetric Group, Encyclopedia of Mathematics and its Applications 16, Addison-Wesley (1981).
• Sagan, The Symmetric Group: Representations, Combinatorial Algorithms, and Symmetric Functions, Springer GTM 203 (2nd ed., 2001).
• Fulton-Harris, Representation Theory: A First Course, Springer GTM 129 (1991), §4.
• Mathas, Iwahori-Hecke Algebras and Schur Algebras of the Symmetric Group, AMS University Lecture Series 15 (1999).
• Mullineux, "Bijections of $p$-regular partitions and $p$-modular irreducibles of the symmetric groups," J. London Math. Soc. 20 (1979) 60-66.
• Ford-Kleshchev, "A proof of the Mullineux conjecture," Mathematische Zeitschrift 226 (1997) 267-308.
• Bessenrodt-Olsson, "On Mullineux symbols," Journal of Combinatorial Theory A 81 (1998) 201-219.
• Lascoux-Leclerc-Thibon, "Hecke algebras at roots of unity and crystal bases of quantum affine algebras," Communications in Mathematical Physics 181 (1996) 205-263.
• Graham-Lehrer, "Cellular algebras," Inventiones Mathematicae 123 (1996) 1-34.
• Kleshchev, Linear and Projective Representations of Symmetric Groups, Cambridge Tracts in Mathematics 163 (2005).
• Brundan-Kleshchev, "Blocks of cyclotomic Hecke algebras and Khovanov-Lauda algebras," Inventiones Mathematicae 178 (2009).