TheoremComplete

The Fundamental Theorem of Vassiliev Invariants

Theorem7.2Fundamental Theorem (1-Cocycle Characterization)

Let vv be a knot invariant with values in an abelian group AA. The following are equivalent:

  1. vv is a Vassiliev invariant of type nn: vv vanishes on all singular knots with >n> n double points.
  2. The restriction of vv to singular knots with exactly nn double points satisfies the 4-term (4T) relation and depends only on the chord diagram.
  3. There exists a weight system wAn(A)=Hom(An,A)w \in \mathcal{A}_n^*(A) = \mathrm{Hom}(\mathcal{A}_n, A) such that for any singular knot KK with nn double points and chord diagram D(K)D(K): v(K)=w(D(K))v(K) = w(D(K)).

Moreover, the sequence 0Vn1VnϕAn00 \to \mathcal{V}_{n-1} \to \mathcal{V}_n \xrightarrow{\phi} \mathcal{A}_n^* \to 0 is exact (where ϕ(v)=wv\phi(v) = w_v is the associated weight system). That is, Vn/Vn1An\mathcal{V}_n/\mathcal{V}_{n-1} \cong \mathcal{A}_n^*.


Proof

Proof

(1)(2)(1) \Rightarrow (2): Type-nn invariants define weight systems.

Already established: if vv is type nn, then for singular knots with exactly nn double points, v(K)v(K) depends only on the chord diagram D(K)D(K) (since crossing changes in the non-singular part produce singular knots with n+1n+1 points, on which vv vanishes). The 4T relation follows from considering a specific family of singular knots.

4T relation derivation. Consider a singular knot KK with n1n-1 double points and a region where two strands are close but do not intersect. By introducing one additional double point in four different ways (corresponding to four chord diagrams that differ in how the new chord is placed relative to the two strands), the four resulting singular knots satisfy a linear relation coming from the type-nn condition applied to a singular knot with n+1n+1 points. Explicitly: the resolution of a triple-point singularity (three strands meeting) gives the 4T relation.

(2)(3)(2) \Rightarrow (3): Trivial, since a function on chord diagrams satisfying 4T is precisely an element of An\mathcal{A}_n^*.

(3)(1)(3) \Rightarrow (1): Weight systems lift to Vassiliev invariants.

This is the deep direction, requiring the Kontsevich integral. Given wAnw \in \mathcal{A}_n^*, define vw(K)=w(Zn(K))v_w(K) = w(Z_n(K)) where ZnZ_n is the degree-nn part of the Kontsevich integral. Then:

  • vwv_w is a knot invariant (since ZZ is).
  • For singular knots with n+1n+1 points: Zn(K×)Z_n(K_\times) involves only chord diagrams from the resolutions, and the alternating sum over >n>n double points produces zero in An\mathcal{A}_n.
  • Therefore vwv_w is type nn, and its associated weight system is ww.

Exactness. We need to verify:

  • ϕ\phi is surjective: every weight system ww equals ϕ(vw)\phi(v_w) by the Kontsevich construction.
  • kerϕ=Vn1\ker\phi = \mathcal{V}_{n-1}: if ϕ(v)=0\phi(v) = 0, then vv vanishes on all singular knots with nn double points, so vv is type n1n-1.

This completes the proof. \square


ExampleDimension Computations

The theorem reduces the study of Vassiliev invariants to combinatorics of chord diagrams. The dimensions dn=dimAn=dim(Vn/Vn1)d_n = \dim\mathcal{A}_n = \dim(\mathcal{V}_n/\mathcal{V}_{n-1}) can be computed by counting chord diagrams modulo 4T. Known values: d0=1d_0=1, d1=0d_1=0, d2=1d_2=1, d3=1d_3=1, d4=3d_4=3, d5=4d_5=4, d6=9d_6=9, d7=14d_7=14, d8=27d_8=27, d9=44d_9=44, d10=80d_{10}=80. The cumulative dimensions: dimVn=1,1,2,3,6,10,19,33,60,104,184,\dim\mathcal{V}_n = 1,1,2,3,6,10,19,33,60,104,184,\ldots These grow exponentially: dimVnCαn\dim\mathcal{V}_n \sim C \cdot \alpha^n for some constants C>0C > 0 and α>1\alpha > 1.

RemarkThe Integration Problem

While the fundamental theorem asserts that every weight system lifts to a Vassiliev invariant (via the Kontsevich integral), finding explicit combinatorial formulas for the invariant corresponding to a given weight system is the integration problem -- generally unsolved except for low degrees. The Kontsevich integral provides an existence proof but involves difficult iterated integrals. For specific weight systems (those from Lie algebras), quantum group RR-matrices provide alternative, more computable constructions.