Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-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.

The Burau determinant formula for a closed braid and its axis

Statement

Assume the Axiom of Choice (The Axiom of Choice) for the Alexander-module convention. Let n≥2, let β∈Bn with closure β^⊂S3 and braid axis A⊂S3, so that β^∪A is an oriented link (The closure of a geometric braid); let B‾β(t1,…,tn) be the coloured reduced Burau matrix of The coloured reduced Burau matrix and put Bβ(t):=B‾β(t,…,t) for its equal-label specialisation. Then:

(1) [Morton] the multivariable Alexander invariant of β^∪A satisfies Δβ^∪A(t1,…,tn,x)≐det⁡(I−x B‾β(t1,…,tn)) with the identifications tπ(j)=tj forced by the permutation π of β, where x is the axis variable and ≐ means equality up to multiplication by a unit of the Laurent ring Z[t1±1,…,tn±1,x±1];

(2) [deletion of the axis] with the same identifications tπ(j)=tj, the Torres--Fox deletion of the axis gives the multivariable invariant of the closed braid, Dβ^mv(t1,…,tn)≐det⁡(I−B‾β(t1,…,tn))1−t1t2⋯tn, and in the one-variable specialisation t1=⋯=tn=t the one-variable Alexander polynomial of The Alexander polynomial from the zeroth elementary ideal satisfies Δβ^(t)≐(1−t)det⁡(I−Bβ(t))1−tn; Here Dmv is Morton's multivariable invariant, a fraction for a knot. Its equal-label specialization satisfies Dβ^mv(t,…,t)≐Δβ^(t)/(1−t) for any number of components; when there is more than one component it differs from the library's Dβ^=Δβ^ by the factor 1−t. For a knot (π an n-cycle) the polynomial equals det⁡(I−Bβ(t))/(1+t+⋯+tn−1) up to units, and for a link with k components the identifications leave one variable per cycle and the same formulas hold, with the library's Alexander invariant Dβ^=Δβ^/(1−t) for a knot and Dβ^=Δβ^ for k>1. All formulas are stated up to multiplication by a unit ±tm (and ±tim in the multivariable case) of the corresponding Laurent ring.

Facts & Assumptions

Given: AC and an integer n≥2, a braid β∈Bn with closure β^ and braid axis A, the coloured reduced Burau matrix B‾β(t1,…,tn), and the equal-label specialisation Bβ(t). AC is inherited from the Alexander module; the finite Fox-determinant manipulations use no further choice.

[F1]

Literature input: diagram presentation. Use the bottom meridians u1,…,un of Morton's Figure 2 and their reverse partial products g0=1, gi=ui⋯u1. For a crossing, let Γi send ui to ui+1 and ui+1 to ui+1uiui+1−1, fixing the other meridians. For the word β=σi1ε1⋯σilεl read from the top, successive substitutions express the top meridians in the bottom generators by Tβ=Γilεl∘⋯∘Γi1ε1. Gluing the two disk slices gives the complement presentation with generators g1,…,gn,c and relations Tβ(gi)=c−1gic, where c is the axis meridian; φ(uj)=tj and φ(c)=x. This diagram presentation is the quoted topological input of Morton's proof of Theorem 1, printed pp. 4–6 (van Kampen, Seifert–van Kampen identifies the fundamental group with a group pushout). The substitutions here use moving disk slices; they are not the ordinary-composition automorphism of Artin automorphisms of the free group. No equality between those two word actions is assumed.

[F2]

Fox calculus. For a free basis g1,…,gn, ∂gk/∂gj=δkj, ∂(vw)/∂gj=∂v/∂gj+v ∂w/∂gj, and ∂(v−1)/∂gj=−v−1∂v/∂gj. Writing J(U)ij=∂U(gi)/∂gj, ordinary function composition satisfies J(U∘V)=U(J(V))J(U), with U applied entrywise to group-ring coefficients. These are the free-derivative rules used in Morton's proof, printed pp. 5–6. They give a product in word order for the successive-substitution action of [F1], not for the library's ordinary Artin word action.

[F3]

The deficiency-one Fox rule of The deficiency-one Fox calculus rule for the Alexander invariant: deleting the column of a generator c with φ(c)≠1 and dividing the determinant of the remaining square matrix by 1−φ(c) gives the evaluation of the Alexander invariant, up to a unit of Z[H].

[F4]

The coloured reduced Burau matrix of The coloured reduced Burau matrix is the matrix product of the C‾i(ar)±1 along an Artin word, with ar the label of the undercrossing string at crossing r; each factor is invertible with determinant −a if the label is a. At equal labels t1=⋯=tn=t the specialisation is the standard reduced Burau matrix of The reduced Burau representation up to the fixed basis change of the coloured matrix; in particular the characteristic polynomials agree (Morton, Remark (1)).

[F5]

The one-variable Alexander polynomial ΔL and the Alexander invariant DL=ΔL for a link with more than one component, DL=ΔL/(1−t) for a knot, of The Alexander polynomial from the zeroth elementary ideal, defined from the Alexander module of The one-variable Alexander module of an oriented link; the one-variable module is the cover classified by the total linking homomorphism.

[F6]

Literature input (quoted). Torres--Fox deletion (Morton, Remark (2), printed p. 3): for a link L∪C with meridian of C replaced by 1, DL(t)=ΔL∪C(t,1)/(1−φ(c)), where φ(c) is the element represented by C in the complement of L; for the axis C=A one has φ(A)=t1t2⋯tn (Conway, Theorem 3.15 and its proof, where the same deletion is computed through the twisted chain complex).

[F7]

Literature input (quoted). Birman--Brendle, section 4.2 equation (15): for the closure b(X) of a braid X∈Bn the Alexander polynomial satisfies Δb(X)(t)=det⁡(ρˉ(X)−In−1)/(1+t+⋯+tn−1) up to the usual unit, where ρˉ is the reduced Burau representation; by Morton's Remark (1) the equal-label coloured matrix Bβ(t) is a matrix of that representation, so the same display reads Δβ^(t)≐(1−t)det⁡(I−Bβ(t))/(1−tn) up to sign. This one-variable normalization is the classical formula for knots and links and is quoted here; the identity between Bβ(t) and the matrix of the reduced Burau representation is verified in the next proposition on this page.

Proof

1.1F1givenalgebra

The diagram basis and elementary substitutions. The reverse partial products of [F1] are a free basis, since ui=gigi−1−1. Direct substitution gives Γi(gi)=gi+1gi−1gi−1 and Γi−1(gi)=gi−1gi−1gi+1; every other gj, including gn, is fixed. Thus [F1] supplies a deficiency-one presentation of the closed braid and axis, with n+1 generators and n relations. Deleting the column of c is admissible because φ(c)=x≠1.

2.1F1F2F4step 1.1algebra

The Jacobian product with transported labels. Put Γr=Γirεr and Sr=Γl∘⋯∘Γr, with Sl+1=id. Since Sr=Sr+1∘Γr, [F2] gives φ(J(Sr))=φ(Sr+1(J(Γr)))φ(J(Sr+1)). For a positive crossing, step 1.1 and the product rule give the exceptional row (gi+1gi−1,−gi+1gi−1,1), truncated at i=1; for a negative crossing the row is (1,−gi−1gi−1,gi−1gi−1). The suffix Sr+1 expresses the meridians immediately below crossing r in the bottom generators. Hence the positive coefficient is φ(Sr+1(ui+1))=ar, while the negative coefficient is φ(Sr+1(ui))−1=ar−1: these are precisely the undercrossing labels of [F4]. Their reduced blocks are C‾ir(ar)εr. Iterating the displayed recurrence therefore gives φ(J(Tβ))=B~β=(B‾βv01), with the leading factors in the defined word order. This calculation concerns Tβ of [F1].

3.1F1F2step 2.1algebra

The relation matrix. Differentiate Tβ(gi)−c−1gic with respect to the gj. The second term evaluates to x−1δij, so deleting the column of c leaves B~β−x−1In. Its block form in step 2.1 gives det⁡(B~β−x−1In)=(1−x−1)det⁡(B‾β−x−1In−1).

4.1F3step 1.1step 2.1step 3.1algebra

The Fox rule and the characteristic polynomial. Apply [F3] with the deleted generator c and the divisor 1−φ(c)=1−x, which cancels the explicit factor 1−x−1 up to the unit −x−1 of step 3.1: Δβ^∪A≐−x−1det⁡(B‾β−x−1In−1)≐det⁡(I−x B‾β), since det⁡(xB‾β−In−1)=(−1)n−1det⁡(I−xB‾β) and x is a unit. The variable identifications tπ(j)=tj are those of the closed braid: strings joined at the top and bottom carry the same meridian. This proves (1).

5.1F4F6F7step 4.1algebra

Deletion of the axis. Put x=1 and apply the Torres--Fox deletion of [F6] to the pair (β^,A) with φ(A)=t1t2⋯tn and the identifications tπ(j)=tj; part (1) at x=1 gives the multivariable identity Dβ^mv(t1,…,tn)≐det⁡(I−B‾β(t1,…,tn))/(1−t1t2⋯tn), the first display of (2). In the one-variable specialisation t1=⋯=tn=t the denominator becomes 1−tn and the coloured matrix becomes Bβ(t); the one-variable normalization Δβ^(t)≐(1−t)det⁡(I−Bβ(t))/(1−tn) is the quoted classical formula [F7], so the equal-label multivariable invariant is Δβ^(t)/(1−t), rather than the library's multi-component normalization D=Δ.

6.1F5F7step 5.1algebra∎

Knot and multi-component normalisations. Suppose first that π is an n-cycle, so that the closure is a knot. By [F5] the knot normalisation is D=Δ/(1−t), and using 1−tn=(1−t)(1+t+⋯+tn−1) gives Δβ^(t)≐det⁡(I−Bβ(t))/(1+t+⋯+tn−1), equivalently Dβ^(t)≐det⁡(I−Bβ(t))/(1−tn) up to units. If instead π has k disjoint cycles, the identifications tπ(j)=tj leave one variable per cycle, and the multivariable identity of step 5.1 is a k-variable statement; the one-variable formula of step 5.1 is the classical formula [F7] and requires no knot hypothesis, so it computes Δβ^ for a link as well, with Dβ^=Δβ^ for k>1 by [F5]. Thus for k>1 the latter is (1−t) times the equal-label specialization of Dmv, as explicitly stated; the same determinant formula for the polynomial retains its factor 1−t. This proves the displayed normalisations of (2).

Remarks

Depends on

Used by

Dependency tree · two levels

65 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