Alphabeta Math
Pipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

✓ 3 results · all verified · 1 also independently AI-judged
Every result on this page is machine-checked by a proof checker and read in full and owner-audited; the judge is an additional, independent cross-model AI review of the proofs. The 2 not AI-judged were verified by owner audit (typically over a confirmed judge false positive), not failures.

Kazhdan–Lusztig Bases, Polynomials, and Cells — Examples

1 · Prerequisites

2 · Summary

This companion collects finite computations for kazhdan-lusztig-bases-polynomials-and-cells. The first example computes the complete bar and Kazhdan–Lusztig bases in S2 and S3, including a generator product with a lower correction term. The ten-element interval [1324,3412] in S4 exhibits P1324,3412=1+q, a non-cover μ-pair, the R-recursion, and signed inverse coefficients. The RSK example works through all insertions in S3 and groups all 24 permutations of S4 by their recording tableaux, then checks how the cell classification and right descents appear in these tables.

3 · Logical flowchart

4 · Definitions, theorems and proofs

None yet.

5 · Examples, counterexamples and false statements

ExampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passaudited 2026-10-08Open item page →

The Kazhdan–Lusztig bases of S2 and S3

Facts & Assumptions

Given: The one-based permutation groups S2,S3, with composition as functions; left si swaps values i,i+1. Write α=v−v−1 and Cw=H‾w.

[F1]

The standard elements form a basis, Hsi2=1−αHsi, and the normalized bar assignment sends Hsi to Hsi+α (The normalized type-A Hecke algebra and its bar involution).

[F2]

The bar assignment descends to a multiplicative semilinear involution of the quotient Hecke algebra (The Hecke bar involution is well defined).

[F3]

A bar-fixed element in Hw+∑y<wvZ[v]Hy is uniquely Cw=H‾w, and these elements form a basis (Existence and uniqueness of the Kazhdan–Lusztig basis).

[F4]

The coefficients satisfy py,w=vℓ(w)−ℓ(y)Py,w(v−2) and μ(y,w)=[v]py,w (Kazhdan–Lusztig polynomials in the classical q-normalization).

[F5]

Multiplication by Cs is the descent scalar or the ascent sum with lower s-descent terms (Multiplication by a generator in the Kazhdan–Lusztig basis).

[F6]

Bruhat order has the reduced-subword characterization and is graded by inversion length (Basic properties of the Bruhat order on Sn).

[F7]

The R-coefficient ry,w is the coefficient of Hy in the standard-basis expansion of Hw‾ (Bruhat intervals and the R-coefficients).

Example

Work in the normalization of The normalized type-A Hecke algebra and its bar involution, write permutations in one-line notation, and put q=v−2 as in Kazhdan–Lusztig polynomials in the classical q-normalization. (a) In S2: rid,21=v−v−1, H‾id=Hid, H‾21=H21+vHid, Pid,21(q)=1, μ(id,21)=1 and H‾212=(v+v−1)H‾21. (b) In S3 the bar images of the standard basis are H132‾=H132+(v−v−1)H123,H213‾=H213+(v−v−1)H123, H231‾=H231+(v−v−1)(H132+H213)+(v2−2+v−2)H123, H312‾=H312+(v−v−1)(H132+H213)+(v2−2+v−2)H123, H321‾=H321+(v−v−1)(H231+H312)+(v2−2+v−2)(H132+H213)+(v3−2v+2v−1−v−3)H123, the Kazhdan–Lusztig basis is H‾123=H123,H‾132=H132+vH123,H‾213=H213+vH123, H‾231=H231+v(H132+H213)+v2H123,H‾312=H312+v(H132+H213)+v2H123, H‾321=H321+v(H231+H312)+v2(H132+H213)+v3H123, all Py,w(q) with y≤w equal 1, and μ(y,w)=1 exactly when y⋖w is a cover of the Bruhat order on S3 (the eight covers (123,132), (123,213), (132,231), (132,312), (213,231), (213,312), (231,321), (312,321) in one-line notation). (c) Multiplication checks: H‾212=(v+v−1)H‾21 and, in the case sw>w with one μ-edge, H‾132H‾231=H‾321+H‾132 (here s2⋅231=321 and s2⋅132=123<132, so the sum in Multiplication by a generator in the Kazhdan–Lusztig basis has the single term μ(132,231)=1).

Verification

1.1F1F2F3F4F7algebra

Rank one. By [F1, F2], H21‾=H21+α and Hid‾=Hid; by [F7], this gives rid,21=α=v−v−1. Since α+v−1=v, C:=H21+vHid is bar-fixed. It is triangular below H21, so uniqueness [F3] gives C21=C and Cid=Hid. From H212=1−αH21 in [F1], C212=(H21+v)2=1+(v+v−1)H21+v2=(v+v−1)(H21+v). Its off-diagonal coefficient is pid,21=v, so [F4] gives Pid,21=1 and μ(id,21)=1.

1.2F1F2algebra

Every bar image in S3. Put X=H213 and Y=H132. The reduced words give H231=XY, H312=YX and H321=YXY. By [F1, F2], their bar images are computed by expanding (X+α)(Y+α) and its reverse for the two length-two images, and (Y+α)(X+α)(Y+α) for the length-three word, replacing Y2 by 1−αY. The latter gives H321+α(H231+H312)+α2(X+Y)+(α3+α)H123. Since α2=v2−2+v−2 and α3+α=v3−2v+2v−1−v−3, these are exactly all the displayed bar images. Multiplicativity computes the images directly; it does not assert that Hw‾ itself is bar-fixed.

1.3F1F2F3F6algebra

The six KL elements. Put A=X+v and B=Y+v. By [F1, F2], both are bar-fixed, and AB=H231+v(H213+H132)+v2H123,BA=H312+v(H132+H213)+v2H123. A direct expansion gives BAB=H321+v(H231+H312)+v2H213+(1+v2)H132+(v+v3)H123, so subtracting B=H132+vH123 gives exactly the displayed C321. Each of 1,A,B,AB,BA,BAB−B is bar-fixed and has top coefficient 1 with lower coefficients in vZ[v]; all lower indices are below its top by [F6]. Thus uniqueness [F3] identifies all six elements.

2.1F4F6step 1.3algebra

Polynomials and covers. Reading the expansions in step 1.3 gives py,w=vℓ(w)−ℓ(y) for every y≤w, so [F4] gives Py,w=1 and μ(y,w)=[v]py,w is 1 exactly at length difference one. By the reduced-subword criterion [F6], the Bruhat ranks in S3 are {123}, {132,213}, {231,312}, and {321}, and each element in one of these layers is below every element in the next layer. Hence the covers are precisely the two edges from 123, the four edges from rank one to rank two, and the two edges into 321; these are exactly the eight pairs listed. Grading rules out other covers.

3.1F5F6step 1.1step 2.1algebra∎

The ascent multiplication check. The subword criterion [F6] gives [id,231]={123,132,213,231}. Left s2 sends 231 to 321, whereas it sends 132 to 123 and 213 to 312; thus among the strict lower indices only 132 has a left s2-descent. Step 2.1 gives μ(132,231)=1, so [F5] yields C132C231=C321+C132. For comparison, left s1 sends 231 to 132<231, and the descent formula gives C213C231=(v+v−1)C231. The rank-one square was already proved in step 1.1.

ExampleConstruction: AI-generatedVerification: AI-adaptedprecheck passaudited 2026-10-08Open item page →

The R- and Kazhdan–Lusztig recursions on a small singular interval

Facts & Assumptions

Given: One-based S4, b=1324=s2, w=3412=s2s1s3s2, and q=v−2. Put α=v−v−1.

[F1]

Bruhat order is graded by inversion length, has the reduced-subword characterization and prefix-rank criterion, and satisfies the lifting implication: if y≤z, sy>y, and sz<z, then y≤sz (Basic properties of the Bruhat order on Sn).

[F2]

The R-coefficient ry,z is defined as the coefficient of Hy in Hz‾ (Bruhat intervals and the R-coefficients).

[F3]

Py,z has constant term 1 on comparable pairs, degree at most (ℓ(z)−ℓ(y)−1)/2 for y<z, and py,z=vℓ(z)−ℓ(y)Py,z(v−2); μ vanishes for even length differences (Kazhdan–Lusztig polynomials in the classical q-normalization).

[F4]

The KL left descent recursion uses c=1 when sy<y, with correction indices y≤u≤sz having su<u and μ(u,sz)≠0 (The Kazhdan–Lusztig polynomial descent recursion).

[F5]

The chain-defined inverse coefficients satisfy qx,z′=−∑x≤u<zqx,u′pu,z for x<z, with diagonal 1; their matrix is the two-sided inverse of (px,z) (Inverse Kazhdan–Lusztig polynomials, The Kazhdan–Lusztig inversion formula).

[F9]

The R-coefficients vanish unless y≤z, have diagonal rz,z=1, obey the left descent recursion, and have leading and trailing terms vℓ(z)−ℓ(y) and sgn⁡(y)sgn⁡(z)v−(ℓ(z)−ℓ(y)) on comparable pairs (The R-coefficient recursion, support, degree bounds and inversion).

Example

In S4, written in one-line notation, let b:=s2=1324 and w:=s2s1s3s2=3412 (a reduced word of length 4; the two middle generators commute). (a) The interval [b,w] has exactly ten elements: 1324; 1342,1423,2314,3124; 1432,2413,3142,3214; and 3412. (b) For comparable pairs in this interval the only Kazhdan–Lusztig polynomial different from 1 is Pb,w(q)=1+q; so μ(b,w)=1, and (b,w) is a μ-pair with ℓ(w)−ℓ(b)=3>1: μ-pairs need not be covers. All other μ-pairs inside the interval are covers, and all ry,z for y,z∈[b,w] are the corresponding Laurent polynomials read off from The R-coefficient recursion, support, degree bounds and inversion. (c) The descent recursion of The Kazhdan–Lusztig polynomial descent recursion at y=b, w and s=s2 (a left descent of w, with sw=2413 and sb=1234=id, so c=1) reads Pb,w(q)=Pid,2413(q)+qPb,2413(q)− ⁣ ⁣ ⁣∑b≤z≤2413s2z<z, μ(z,2413)≠0 ⁣ ⁣ ⁣μ(z,2413) q(4−ℓ(z))/2Pb,z(q); the sum is empty because [b,2413]={1324,1423,2314,2413} and its only element with s2z<z is 1324, for which μ(1324,2413)=0; since Pid,2413=Pb,2413=1, the recursion returns Pb,w=1+q, in agreement with (b). (d) The inverse Kazhdan–Lusztig polynomial of Inverse Kazhdan–Lusztig polynomials is qb,w′=−v−v3, while qb,1342′=qb,1423′=qb,2314′=qb,3124′=−v and qb,1432′=qb,2413′=qb,3142′=qb,3214′=v2; the matrix identity ∑zqx,z′pz,w=δx,w of The Kazhdan–Lusztig inversion formula holds on the ten-point interval. In particular the inverse coefficients are not all nonnegative even though all py,z are.

Verification

1.1F1algebra

The full interval. The displayed word for w has inversion length 4, so it is reduced. Its reduced subwords of lengths 0,1,2,3,4 give respectively 1234; 1324,2134,1243; 1342,1423,2314,3124,2143; 1432,2413,3142,3214; and 3412. The four excluded elements 1234,2134,1243,2143 have no reduced subword s2, so they are not above b; the other ten are. For an explicit order check, write a=1342=s2s3, c=1423=s3s2, d=2314=s1s2, f=3124=s2s1, and A=1432=s2s3s2, C=2413=s1s3s2, D=3142=s2s1s3, F=3214=s2s1s2. The reduced-subword criterion gives the intermediate covers a<A,D; c<A,C; d<C,F; f<D,F; each rank-two element is also above b=s2, and each rank-three element is below w. These are all cover incidences between adjacent ranks, so all other comparisons are their transitive consequences.

2.1F1F3step 1.1algebra

The correction interval. Left s2 gives s2w=2413 and s2b=1234. Step 1.1 gives [b,2413]={b,1423,2314,2413}. Their left s2 products are respectively 1234,1432,3214,3412, of lengths 0,3,3,4, whereas the original lengths are 1,2,2,3. Thus only b has that descent, and μ(b,2413)=0 by its even length gap. In particular 1342 is excluded: R1342(3,2)=1<2=R2413(3,2).

2.2F1F2F9step 1.1algebra

All the R-coefficients. By [F9], noncomparable pairs have ry,z=0, diagonal entries are 1, and every comparable coefficient is nonzero because its leading term is vℓ(z)−ℓ(y). For a cover, the degree range and parity in [F9] leave only the terms v and −v−1, so ry,z=α. For a comparable pair of gap two, induct on ℓ(y) and choose a left descent s of z. If sy<y, the recursion gives ry,z=rsy,sz; this coefficient is nonzero, so [F9] implies sy≤sz, and induction gives ry,z=α2. If sy>y, then rsy,sz=0: its indices have equal length, and equality would force y=z. By the lifting implication in [F1], y≤sz, so the other recursion term is αry,sz=α2. Thus every gap-two pair in S4 has coefficient α2. The only gap-three pair in [b,w] is (b,w). Since s2 is a left descent of both, rb,w=r1234,2413. For 2413, s1 is a left descent, so r1234,2413=r2134,1423+αr1234,1423. The first term is zero because 1423=s3s2 has no reduced subword s1; the second is α⋅α2. Hence rb,w=α3=v3−3v+3v−1−v−3. This determines every coefficient for pairs in the displayed interval.

3.1F1F3F4step 1.1step 2.1algebra

The sole nonconstant polynomial. By [F3], all comparable pairs of gap at most two have P=1, so the only possibly nonconstant pair within the interval is (b,w). To evaluate P1234,2413, apply [F4] with left s1 to 2413=s1s3s2, obtaining lower top s3s2=1423. No element below 1423 has left s1-descent: its subwords are 1234,1243,1324,1423, with no inversion between the values 1,2. Also s1≰1423. Hence P1234,2413=qP2134,1423+P1234,1423=0+1. Now use [F4] at (b,w,s2): step 2.1 makes its correction sum empty, c=1, and Pb,2413=1, so Pb,w=1+q. Therefore pb,w=v3+v and μ(b,w)=1. Every other comparable distinct pair has py,z=vℓ(z)−ℓ(y), so its nonzero μ occurs exactly on covers.

4.1F1F3F5step 1.1step 3.1algebra∎

Inverse entries and both matrix products. Step 3.1 gives diagonal p=1, cover entries v, and gap-two entries v2. Each gap-two interval in step 1.1 has two intermediate elements. Thus [F5] gives inverse entries 1,−v,v2 at gaps 0,1,2. At the sole gap-three pair there are four elements at each intermediate rank, so qb,w′=−(v3+v)−4(−v)v2−4v2v=−v3−v. This gives every entry of the inverse matrix, including every displayed value in part (d). For Q′P, the off-diagonal entries at gaps one and two are −v+v=0 and v2−2v2+v2=0; at gap three the entry is (v3+v)−4v3+4v3−(v3+v)=0. For PQ′, these entries are v−v=0, v2−2v2+v2=0, and −(v3+v)+4v3−4v3+(v3+v)=0. Diagonal entries are 1 and noncomparable entries are zero by support, so both matrix products are the identity. Although all p-entries in this finite example are nonnegative, its cover inverse entries and qb,w′ are negative. Every calculation is finite and uses no choice principle.

ExampleConstruction: Literature-sourcedVerification: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

RSK cells in S3 and S4

Facts & Assumptions

Given: The row-insertion and recording-tableau conventions for one-line permutations in S3 and S4, and the corresponding Kazhdan–Lusztig cell relations.

[F1]

Row insertion replaces the leftmost entry strictly greater than the carried letter, bumps that entry to the next row, and stops by appending at the right end of a row; the recording tableau places label k in the new box created by inserting the kth letter (Row insertion and the bumping route, The Robinson-Schensted correspondence).

[F2]

The type-A cell theorem identifies left cells with Q-fibers, right cells with P-fibers, and two-sided cells with common RSK-shape fibers (Kazhdan–Lusztig cells of type A are classified by RSK tableaux).

[F3]

The right descent set is R(w)={si:ℓ(wsi)<ℓ(w)}, where wsi swaps positions i,i+1 in one-line notation; right descents are constant on a left cell (L-, R- and two-sided Kazhdan–Lusztig preorders and cells).

[F4]

The length ℓ(w) is the number of inversions of the one-line word, and the standard tableaux in the RSK pairs are increasing along rows and columns (Permutation Weyl group and inversion length, Tableaux and standard tableaux, Partitions, English diagrams, and conjugation).

Statement

Use the RSK correspondence of The Robinson-Schensted correspondence for the one-line word (insertion tableau P, recording tableau Q; the row insertion is Row insertion and the bumping route), and write a standard tableau as its rows separated by bars. (a) In S3: 123↦(123,123), 132↦(12∣3,12∣3), 213↦(13∣2,13∣2), 231↦(13∣2,12∣3), 312↦(12∣3,13∣2), 321↦(1∣2∣3,1∣2∣3) (pairs (P,Q)). (b) The left cells of S3 are the four Q-fibers {123}, {132,231}, {213,312}, {321}; the right cells are the four P-fibers {123}, {132,312}, {213,231}, {321}; the two-sided cells are the three shape fibers {123}, {132,213,231,312}, {321}. (c) In S4 the ten left cells are the ten Q-fibers: [1234]:{1234}; [12∣34]:{2413,3412}; [123∣4]:{1243,1342,2341}; [124∣3]:{1324,1423,2314}; [13∣24]:{2143,3142}; [134∣2]:{2134,3124,4123}; [12∣3∣4]:{1432,2431,3421}; [13∣2∣4]:{3241,4132,4231}; [14∣2∣3]:{3214,4213,4312}; [1∣2∣3∣4]:{4321}; these are in bijection with the ten standard tableaux of size 4. (d) Right descent sets are constant on left cells but do not determine them: in S4 the permutations 1324 and 2413 both have right descent set {s2}, while Q(1324)=[124∣3] and Q(2413)=[12∣34], so 1324 and 2413 lie in different left cells.

Proof

technique · apply the row-insertion rule to the listed permutations and then read the cell equivalences from the type-A classification
1.1F1

The six RSK pairs in S3. Repeated insertion using [F1] gives wP(w)Q(w)123[123][123]132[12∣3][12∣3]213[13∣2][13∣2]231[13∣2][12∣3]312[12∣3][13∣2]321[1∣2∣3][1∣2∣3]. For example, 132 inserts 1, then appends 3, then inserts 2 in place of 3 and bumps 3 to a new second row; the third recording label is therefore in row two. The same leftmost-greater rule gives the other displayed pairs.

1.2F1

All RSK pairs in S4, grouped by Q. Applying [F1] to each of the 24 one-line words gives Q(w)(w,P(w))[1234](1234,[1234])[12|34](2413,[13|24]),(3412,[12|34])[123|4](1243,[123|4]),(1342,[124|3]),(2341,[134|2])[124|3](1324,[124|3]),(1423,[123|4]),(2314,[134|2])[13|24](2143,[13|24]),(3142,[12|34])[134|2](2134,[134|2]),(3124,[124|3]),(4123,[123|4])[12|3|4](1432,[12|3|4]),(2431,[13|2|4]),(3421,[14|2|3])[13|2|4](3241,[14|2|3]),(4132,[12|3|4]),(4231,[13|2|4])[14|2|3](3214,[14|2|3]),(4213,[13|2|4]),(4312,[12|3|4])[1|2|3|4](4321,[1|2|3|4]). As a nontrivial check on the convention, insertion of 2413 first gives rows [2,4], then bumps 2 below when 1 is inserted, and finally bumps 4 below 3; thus P(2413)=[13∣24] and Q(2413)=[12∣34].

2.1F2step 1.1

Cells in S3. By [F2] and step 1.1, grouping by equal Q gives the four left fibers in part (b), grouping by equal P gives the four right fibers, and grouping by the common shape gives the three two-sided fibers. The displayed RSK pairs contain all six permutations, so there are no omitted elements in any fiber.

2.2F1F2F4step 1.2

Left cells in S4. By [F2], each row label Q in step 1.2 indexes exactly one left cell. The possible shapes of size four are (4),(3,1),(2,2),(2,1,1),(1,1,1,1); their standard tableaux are respectively [1234]; [123∣4],[124∣3],[134∣2]; [12∣34],[13∣24]; [12∣3∣4],[13∣2∣4],[14∣2∣3]; and [1∣2∣3∣4]. These are exactly the ten distinct Q-labels in the table. The listed fibers contain 1+2+3+3+2+3+3+3+3+1=24 permutations, so every element of S4 occurs and the table proves part (c) and the claimed bijection.

3.1F2F3F4step 1.2∎

Equal right descents do not determine the left cell. In one-line notation, 1324 has length 1 and right products 1324s1=3124, 1324s2=1234, 1324s3=1342 of lengths 2,0,2. Thus R(1324)={s2}. The word 2413 has length 3 and right products 2413s1=4213, 2413s2=2143, 2413s3=2431 of lengths 4,2,4, so R(2413)={s2} as well. But step 1.2 gives Q(1324)=[124∣3]≠[12∣34]=Q(2413), so [F2] places them in different left cells. This proves the counterexample while [F3] records that descent sets are constant within each left cell.

The calculations concern only S3 and S4; no empty or singleton group case is asserted. All insertion procedures are finite and deterministic, so no choice principle is used.

Remarks

The classification use in [F2] is exactly its preserved Q-, P- and shape-fiber interface: step 2.1 uses all three in S3, step 2.2 uses only the Q-fiber clause in S4, and step 3.1 uses that clause to separate the two recording tableaux. The tables and descents are computed locally; the supplier's sole cited shape-invariance implication is not replaced by a new source assumption here.

Sources