Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08
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.

The right weak order interval below a fully commutative element is the lattice of order ideals of its heap

Statement

Let (S,m), W, ℓ be as in Coxeter matrices, the presented Coxeter group, reduced words, length, and standard parabolic subgroups, let ≤R be the right weak order with intervals [u,v]R and covers ⋖R (The right and left weak orders, intervals, covers, and meets and joins of subsets, Weak order is a partial order with finite graded intervals; covers and the inversion-set criterion), and let fully commutative elements, reduced words and heaps be as in Words, heaps, linear extensions, commutation classes, and fully commutative elements and Labeled linear extensions of a heap are exactly the words in its commutativity class, and heaps classify commutativity classes. Let w∈W be fully commutative, let s=(s1,…,sk)∈R(w), and let P:=Ps be the heap of w, with lattice of order ideals J(P) (Lattices, distributive lattices, and order ideals, The order ideals of a finite poset form a distributive lattice under union and intersection).

(1) The ideal of an element of the interval. For u∈S write Cu={i:si=u}, a chain in P, with elements u(1)≺u(2)≺⋯≺u(nu), and for a word s′ let ν(u,s′) be the number of occurrences of u in s′. If x≤Rw and s′∈R(x), then ν(u,s′)≤nu for all u, and I(s′):={u(1),…,u(ν(u,s′)):u∈S}⊆P is an order ideal of P; it does not depend on the choice of s′∈R(x). Writing I(x) for this common ideal, one has ℓ(x)=∣I(x)∣, I(1)=∅ and I(w)=P.

(2) Order isomorphism. The map x↦I(x) is an order isomorphism from [1,w]R onto J(P) ordered by inclusion.

(3) Lattice structure. Consequently [1,w]R, as a subposet of (W,≤R), is a finite distributive lattice: for all x,y≤Rw the meet x∧y and the join x∨y exist in [1,w]R and satisfy I(x∧y)=I(x)∩I(y),I(x∨y)=I(x)∪I(y); the least element is 1 and the greatest element is w.

(4) Caveats. This identifies the right weak order interval [1,w]R with J(P), for a fully commutative w; no identification of the Bruhat order interval below w with J(P) is made, and no claim is made about the intervals of elements that are not fully commutative.

Facts & Assumptions

Given: A finite Coxeter matrix (S,m), the presented group W with length ℓ, a fully commutative element w∈W, a reduced word s=(s1,…,sk)∈R(w) with heap P=Ps, and the right weak order ≤R.

[F1]

The group W is presented with relators s2 and (st)m(s,t); R(x) is the set of reduced words of x, of common length ℓ(x), and concatenation of words represents the product of the represented elements (Coxeter matrices, the presented Coxeter group, reduced words, length, and standard parabolic subgroups).

[F2]

The right weak order is defined by u≤Rv if and only if v=ux with ℓ(v)=ℓ(u)+ℓ(x); intervals are [u,v]R={z:u≤Rz≤Rv}, and u⋖Rv means that u<Rv with no element strictly between (The right and left weak orders, intervals, covers, and meets and joins of subsets, clauses (1)-(2)).

[F3]

For all u,v∈W one has u≤Rv  ⟺  ℓ(v)=ℓ(u)+ℓ(u−1v) (length identity), and u≤Rv if and only if some reduced expression of v has a reduced expression of u as its initial segment (prefix property) (The length identity, the prefix property, left translation, and interval translation for weak order, clauses (1)-(2)).

[F4]

Covers in ≤R are exactly the pairs v=us with s∈S and ℓ(v)=ℓ(u)+1, and every u≤Rv is joined by a chain of covers u=u0⋖Ru1⋖R⋯⋖Rur=v (Weak order is a partial order with finite graded intervals; covers and the inversion-set criterion, clause (2)).

[F5]

In the heap Ps, any two positions with equal or noncommuting labels are comparable by the defining relation, so each same-label set Cu={i:si=u} is a chain; an element w is fully commutative when R(w)=C(s) for one (equivalently every) s∈R(w) (Words, heaps, linear extensions, commutation classes, and fully commutative elements, clauses (2), (6)).

[F6]

For every word q one has L(Pq,q)=C(q), the set of labeled words of linear extensions of Pq (Labeled linear extensions of a heap are exactly the words in its commutativity class, and heaps classify commutativity classes, clause (1)).

[F7]

Every finite poset Q has a linear extension; for every order ideal I of Q and every linear extension of the induced poset on I, there is a linear extension of Q whose first ∣I∣ entries are exactly the elements of I. A linear extension is as in Linear extensions of a finite poset (Linear extensions of a finite poset: existence, prescribed initial ideals, and adjacent-swap connectivity, clause (2)).

[F8]

An order ideal of a poset P is a subset closed downward under ⪯, and J(P) denotes the set of order ideals ordered by inclusion (Lattices, distributive lattices, and order ideals). Here an order isomorphism means an order-preserving bijection whose inverse is order-preserving.

[F9]

For a finite poset P, J(P) is a finite distributive lattice under inclusion, with meet intersection, join union, least element ∅ and greatest element P (The order ideals of a finite poset form a distributive lattice under union and intersection).

Proof

Given: A fully commutative element w∈W, a reduced word s∈R(w) and its heap P=Ps.

Proof technique: direct.

1.1givenF1F2F3F5F6F8

Clause (1), construction of the ideal. Let x≤Rw, in the right weak order of [F2], and let s′∈R(x). By the length identity F3, ℓ(w)=ℓ(x)+ℓ(x−1w); fix a reduced word u′ of x−1w. Then s′u′ represents x⋅x−1w=w and has length ℓ(x)+ℓ(x−1w)=ℓ(w), so s′u′∈R(w). Since w is fully commutative, R(w)=C(s)=L(P,s) by [F5] and [F6], so s′u′ is the labeled word of a linear extension π of P. In any linear extension the elements of the chain Cu occur in increasing order u(1)≺⋯≺u(nu), so the first ∣s′∣=ℓ(x) letters of s′u′, namely the letters of s′, are exactly u(1),…,u(ν(u,s′)) for each u. Hence ν(u,s′)≤nu, and I(s′) is the set of the first ℓ(x) entries of π, an initial segment of a linear extension; a prefix of a linear extension is downward closed, so I(s′) is an order ideal of P.

1.2givenF5F7

Clause (2), the heap of an ideal. Let I∈J(P). By [F7], choose a linear extension π of P whose initial block πI=(a1,…,am) lists exactly I, where m=∣I∣. Let sI be the word of labels on this block, let xI be its product in W, and define φ:[m]→I by φ(j)=aj. This bijection preserves labels. For any strict relation a≺Pb with a,b∈I, choose a chain a=z0≺z1≺⋯≺zr=b of maximal length between a and b (such a chain exists because P is finite). Every zi lies in I, since zi⪯b and b∈I. Each consecutive pair is a cover in P; it must be a generating pair of the heap, because a generating path with an intermediate element would contradict the cover property. Since πI is a linear extension, each such pair occurs in the same order in sI, and its labels are equal or noncommuting. Thus the corresponding positions are related in PsI, and transitivity shows that a≺Pb in I implies φ−1(a)≺PsIφ−1(b).

2.1givenF1F5F6step 1.1

Clause (1), well-definedness and basic properties. If s′′∈R(x) is a second reduced word, then with the same suffix u′ the word s′′u′ also has length ℓ(w) and represents w, so s′′u′∈R(w)=L(P,s); a word in L(P,s) is the labeled word of a linear extension of P, hence contains each u∈S exactly nu times, and this holds for s′u′ as well. Subtracting the common multiplicity ν(u,u′) of the suffix gives ν(u,s′)=ν(u,s′′) for every u, so I(x):=I(s′) is well defined. Moreover ℓ(x)=∣s′∣=∑uν(u,s′)=∣I(x)∣ because the sets {u(1),…,u(ν(u,s′))} are disjoint; I(1)=∅ because the empty word has ν(u,( ))=0; and I(w)=P because for s′=s one has ν(u,s)=nu.

2.2givenF1F5F6step 1.2

Conversely, each generating relation of PsI joins positions with equal or noncommuting labels. Their corresponding elements of I are comparable in P by [F5], and the order is the one in πI, so every generating relation of PsI respects the induced order on I. Together with 1.2, this proves that φ is a labeled poset isomorphism. If a different linear extension of I is used, transport its listing through φ−1 to a linear extension of PsI; the labels are unchanged, so its labeled word lies in L(PsI,sI) and [F6] makes it commutation-equivalent to sI, and [F1] says these interchanges preserve the product. Therefore xI depends only on I, and ψ(I):=xI is well defined.

2.3givenF1F3F5F6F7step 1.2

Clause (2), ψ is order-preserving. Let I⊆J be order ideals of P. Apply [F7] twice: extend a linear extension of I (with first ∣I∣ entries I) to a linear extension of the induced poset on J, which therefore has first ∣I∣ entries I and first ∣J∣ entries J, and extend that in turn to a linear extension π of P; then the first ∣I∣ entries of π are I and its first ∣J∣ entries are J. The full labeled word of π lies in L(P,s)=C(s)=R(w) by [F5] and [F6]. The word sI constructed in 1.2 is a prefix of the word sJ, and both are reduced: replacing either prefix by a shorter word for its product would shorten the full reduced word of w; by the prefix property F3 applied to sI and sJ we get ψ(I)≤Rψ(J).

3.1givenF4step 2.1

Clause (2), monotonicity of x↦I(x). If x⋖Ry, then by [F4] y=xs with s∈S and ℓ(y)=ℓ(x)+1; for s′∈R(x) the word s′s represents y and has length ℓ(x)+1=ℓ(y), hence lies in R(y), and ν(u,s′s)=ν(u,s′)+[u=s]≥ν(u,s′) for all u. By the well-definedness 2.1 the ideals may be computed from these words, so I(x)⊆I(y). For arbitrary x≤Ry≤Rw, [F4] joins x to y by a chain of covers and inclusion is transitive along that chain; hence x≤Ry implies I(x)⊆I(y).

3.2givenF1F3F5F6F8step 1.1step 2.1step 2.2

The full labeled word q of π belongs to L(P,s) by definition. By [F6] and full commutativity [F5], L(P,s)=C(s)=R(w), so q represents w and is reduced of length k=ℓ(w). Its prefix sI is also reduced, since a shorter expression for that prefix would shorten q as an expression of w. The prefix property F3 gives ψ(I)=xI≤Rw. Conversely, for any x≤Rw, step 1.1 gives a reduced word of x as the initial segment of a linear extension of P with initial ideal I(x); using that extension in the definition of ψ gives ψ(I(x))=x. Finally, because I is an order ideal and each Cu is a chain, I∩Cu is an initial segment of Cu with exactly ν(u,sI) elements; hence I(xI)=I. Thus ψ:J(P)→[1,w]R is a two-sided inverse of x↦I(x).

4.1givenF9step 3.1step 3.2step 2.3∎

Clauses (3) and (4). By steps 3.1 and 3.2, x↦I(x) is an order-preserving bijection with inverse ψ, and by step 2.3 the inverse is order-preserving; hence this is an order isomorphism [1,w]R→J(P). For x,y≤Rw, put m0:=ψ(I(x)∩I(y)) and j0:=ψ(I(x)∪I(y)). Since ψ preserves order and is inverse to I, m0≤Rx,y and x,y≤Rj0. If z∈[1,w]R satisfies z≤Rx,y, then monotonicity of I gives I(z)⊆I(x)∩I(y), so z=ψ(I(z))≤Rm0; therefore m0=x∧y. Dually, if z∈[1,w]R is a common upper bound, then I(x)∪I(y)⊆I(z), so j0≤Rz and j0=x∨y. Applying I gives the displayed intersection and union formulas. By [F9], J(P) is a finite distributive lattice with least element ∅ and greatest element P, so the order isomorphism transports this structure to [1,w]R, whose least and greatest elements are 1 and w. Clause (4) holds because the argument uses only the right weak order and its prefix property, and it assumes that w is fully commutative; no Bruhat-interval or non-fully-commutative claim is made.

Depends on

Used by

Dependency tree · two levels

41 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources