Hook Length Formula¶
Count standard Young tableaux of a partition shape by dividing n! by the product of its cell hook lengths.
Core Idea¶
The hook length formula counts the standard Young tableaux of any ordinary partition shape λ with n cells. Fill the diagram with the labels 1 through n, each once, increasing along every row and column. For each cell u, its hook consists of u and the cells directly to its right and below. If h(u) is the number of cells in that hook, the count is f^λ = n! / ∏_{u∈λ} h(u). The formula turns a global order-constrained enumeration into one factorial divided by local shape-dependent factors.[1][2]
The theorem is about a family of partition shapes, not one named numerical answer. Its quotient also gives the dimension of the irreducible symmetric-group representation indexed by λ through the standard-tableau correspondence. Plancherel measure uses \((f^\lambda)^2/n!\), not \(f^\lambda\) itself. A rooted-tree increasing-labeling hook formula is related but defines different hooks and labelings; it is not automatically an instance of this exact cell rule.[1][2][3]
Structural Signature¶
Sig role-phrases:
- Partition shape — A straight, left-justified Young diagram with n cells supplies the object whose admissible fillings are counted. A skew shape is outside this simple theorem's stated premise.[1]
- Standard-tableau order rule — Labels 1,…,n appear once each and strictly increase rightward and downward. These are globally coupled constraints, not independent cell tests.[1]
- Cell hook length — For each cell, count itself plus cells to its right and below in the same diagram. The hook depends on shape and cell location.[1]
- Global hook-product quotient — Multiply all local hook lengths and divide n! by the product to obtain the exact integer f^λ.[1][2]
Condensed: partition shape → standard row/column order → per-cell hooks → n! divided by their product.
What It Is Not¶
- Not a claim that hook conditions are independent. Writing a product of reciprocal hook lengths can suggest independent local probabilities, but standard-tableau row and column orders constrain one another. The product is a theorem requiring proof, not an independence assumption.[2]
- Not a formula for every tableau. Semistandard tableaux permit repeated entries and have different counts. A skew shape also needs a different rule or additional terms, not this unmodified quotient.[1]
- Not identical to every “hook formula.” Rooted-tree posets have an analogous product with hooks defined by rooted subtrees rather than right-and-below diagram cells.[3]
- Not Plancherel probability by itself. The probability of a shape under the standard Plancherel distribution is \((f^\lambda)^2/n!\), a different expression built from this count.[2]
Scope of Application¶
Within enumerative combinatorics, the formula gives an efficient exact count for standard tableaux across ordinary partition shapes. One can compute hooks directly from the diagram instead of enumerating every permitted filling. Stanley's example for shape (4,2,2) gives 8! divided by its eight hook lengths and yields 56 tableaux.[1]
The count also has a representation-theoretic meaning: for each partition λ of n, f^λ is the dimension of the corresponding irreducible representation of the symmetric group S_n. That is a genuine second interpretation of the same number, not a license to use the product on an arbitrary group representation. Separate hook-formula analogues in poset theory must restate the counted objects and the hook definition.[1][3]
Clarity¶
The formula separates three objects often blurred together: the diagram shape, one valid filling of that shape, and the number of all valid fillings. The hook lengths depend only on the shape; they are not entries placed in a tableau. A four-cell shape may have several tableaux, and the product produces their count without listing them.[1]
It also shows precisely which local geometric data control the global count. “Hook” does not mean any connected neighborhood; it is the cell plus its rightward row and downward column in the straight partition diagram. This definition is what distinguishes the original theorem from analogues.[1][3]
Manages Complexity¶
Directly testing n! label permutations against row and column inequalities becomes cumbersome as n grows. The hook formula compresses that search into n local integer factors and one quotient. It keeps the shape's asymmetry visible: cells near the upper left usually have larger hooks, while terminal cells have hook length one.[1]
The compression should not be overread. It evaluates an exact enumerative result; it does not by itself generate all tableaux, prove independence among hook events, or solve counts for shapes outside the theorem. A different formula may be required when premises change.[2]
Abstract Reasoning¶
Given an ordinary partition λ, first check that the objects to be counted are standard tableaux: labels 1 through n without repetition, increasing across rows and down columns. Mark each cell's hook and compute its length. Multiply the lengths, divide n!, and interpret the result as f^λ. If a computation produces a noninteger, that is a strong warning of an incorrectly calculated hook or an invalid input claim, because f^λ counts a finite set.[1]
For shape (3,1), the four hooks have lengths 4, 2, 1 and 1, yielding 4!/8 = 3. For shape (2,2), they have lengths 3, 2, 2 and 1, yielding 4!/12 = 2. The change of shape changes local factors and hence the global count, even though n is the same.[1]
Knowledge Transfer¶
Literal reuse within the theorem covers every ordinary partition shape, including shapes far larger than the examples. Its result then transfers through the proven tableau–representation correspondence to dimensions indexed by those partitions. It can also feed later constructions such as Plancherel probabilities, provided the extra squaring and normalization are kept explicit.[1][2]
The rooted-tree hook formula illustrates a wider family resemblance: Gessel and Seo define a vertex hook length as the number of descendants including that vertex and derive n!/∏ h(v) for increasing labelings of a fixed rooted forest. Those factors are rooted-subtree sizes, not Young-diagram right-and-below cell hooks. The matching quotient shape is a related theorem, not literal transfer of this exact formula.[3] A separate paper gives bijective proofs for the rooted-tree formula.[4]
Examples¶
Four-cell shape (3,1)¶
The diagram has three cells in its first row and one beneath the first. Its hook lengths are 4 at the upper left, 2 and 1 across the rest of the top row, and 1 below. Thus the formula gives 4!/(4·2·1·1)=3 standard fillings.[1]
Mapped back: shape = straight partition (3,1); order = labels 1–4 increasing right and down; hooks = 4, 2, 1, 1; quotient = three valid tableaux.
Four-cell shape (2,2)¶
A two-by-two Young diagram has hook lengths 3, 2, 2 and 1. The result is 4!/(3·2·2·1)=2. The same value is the dimension of the irreducible S_4 representation indexed by (2,2), through the established representation-theoretic correspondence.[1]
Mapped back: shape = straight partition (2,2); order = standard row/column increase; hooks = 3, 2, 2, 1; quotient = two tableaux and corresponding representation dimension.
Near miss: a skew diagram¶
Removing a cell from an inner corner of a Young diagram produces a skew shape. A viewer can still draw rightward and downward neighbors, but the simple straight-shape quotient has lost its premise. One must use an appropriate skew-tableau counting result rather than assume the original formula survives unchanged.
Structural Tensions¶
The exact hook-length identity has no intrinsic two-sided design tradeoff in a valid straight-shape standard-tableau instance. One computes the shape's hooks and the theorem supplies the count; there is no competing goal whose cost must be balanced. Two cautions concern proof and scope, not tension: the product is not justified by pretending cell placements are independent, because row/column admissibility couples them; and similar-looking hook products for skew shapes or rooted trees require their own theorems. Diagnostic: is this a straight partition shape with standard increasing fillings and right-and-below hooks, or has an appealing quotient been transferred beyond its hypotheses?[1][2][3]
Structural–Framed Character¶
The hook length formula is highly structural within algebraic combinatorics: a partition shape and a precise row/column order determine local hooks and an exact integer count. Its vocabulary also travels to symmetric-group representation theory by theorem, but importing it to an arbitrary diagram or poset is not recognition of the same rule. Human practice sets notation and proof conventions, while the result is not an institutional or aesthetic judgment. Evaluative weight is low for a well-posed calculation; correctness is mathematical. The historical Frame–Robinson–Thrall naming identifies a theorem, not one organization's policy. A broader local-to-global counting skeleton may recur, yet the named formula remains tied to Young diagrams and standard fillings. Its character: strongly structural, with narrow mathematical premises rather than broad prime-level portability.
Structural Core vs. Domain Accent¶
The skeleton is a global count expressed through a product of local factors. The accent is exact and irreducible: a straight partition shape, standard row/column-increasing fillings and right-and-below cell hooks. The same quotient pattern in rooted trees uses a different partial order and hook, so the resemblance is not enough to make this formula a prime. The live Formal Theorem node is the strict genus; a nearer enumerative-formula theorem may be admitted later. A topical edge to Kostka Number would still be false.
Instantiates / Related Primes¶
This entry is a kind of Formal theorem.
Strict parent: Formal theorem. The proved hook-length product counts standard tableaux of straight partition shape under specified hook definitions; many theorems lack this differentia. Kostka Number remains a neighbor counting semistandard tableaux of specified shape and weight. Rooted-tree hook products and Plancherel probability are separate results, not aliases or additional edges.
Relationships to Other Abstractions¶
Current abstraction Hook Length Formula Domain-specific
Parents (1) — more general patterns this builds on
-
Hook Length Formula is a kind of Formal theorem Domain-specific
The hook-length formula is a formal theorem.Its proved product count has specified straight-shape standard-tableau and hook hypotheses; many formal theorems do not.
Hierarchy paths (2) — routes to 2 parentless roots
- Hook Length Formula → Formal theorem → Formal System → Formalization → Representation → Abstraction
- Hook Length Formula → Formal theorem → Formal System → Formalization → Transformation → Function (Mapping)
Neighborhood in Abstraction Space¶
Hook Length Formula sits in a sparse region of the domain-specific corpus (89th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Graph Structures & Combinatorial Objects (44 abstractions)
Nearest neighbors
- Kostka number — 0.82
- Polygon — 0.80
- Finite subdivision rule — 0.80
- Euler Characteristic — 0.80
- Wilf Equivalence — 0.80
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
Do not conflate hook length with a tableau label, the tableau count with Plancherel probability, or this Young-diagram product with every rooted-tree or skew-tableau analogue. The exact recognition test is: ordinary partition shape, labels 1 through n used once, row/column increase, then n! divided by the product of each cell's right-and-below hook length.[1][2][3]
References¶
[1] Richard P. Stanley, Topics in Algebraic Combinatorics, Chapter 8, Theorem 8.1, including historical notes on tableaux and representation dimensions. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s
[2] Greene, Nijenhuis and Wilf, original probabilistic hook-formula research reprint, hosted on Wilf's university site. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i
[3] Ira M. Gessel and Seunghyun Seo, “A Refinement of Cayley's Formula for Trees,” Electronic Journal of Combinatorics 11(2) (2006), #R27, §2 p. 3, defines vertex hooks by descendant count and gives the increasing-labeling quotient in Lemma 2.1 with S=V(F). registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g
[4] “Bijective Proofs of the Hook Formula for Rooted Trees,” Ars Combinatoria 106 (2012), pp. 483–494, publisher abstract only; cited for the existence of separate bijective proofs, not the specific descendant-count identity. registry ↩