Knot Topology Explorer

Dr. Van Warren · Gemini Pro Extended · Claude 4.8
Embedding: Whitney Disk Correction
Boundary: T(2, 7)  ·  Seifert genus 3
Framing −1  ·  4D time 0.19π  (period 2π)
Drag: orbit · Wheel: zoom · Shift/Right‑drag: pan  —  camera is independent of the 4D shape
Controls
Geometry  f(x,y,z,t)
3
−1
Slice structure view
4D time (shape morph)
0.19π
Auto‑orbit camera (view only)
Appearance
0.14
1.00
0.46
0.30
0.80
0.40
Disk classes
I · Slice‑disk membrane
II · Whitney disks
III · Casson tower
Overlays
§2 Double points
§4 Gauge field FA
§5 Framing ribbon + K′
§6 Exotic ℝ⁴ shells
4D image — shape for all t
Projection walls

4D Knot Traces — Theory & Equations

Every figure below is a live 3‑D render (not SVG), and each idea is also a toggle on the Simulator.

Smooth sphere ↔ cornered cube: one continuous bijection, two different smooth structures.

1. Two notions of "the same space"

A homeomorphism is a bijection \(f:X\to Y\) with \(f,f^{-1}\) continuous (\(C^0\)). A diffeomorphism also needs \(f,f^{-1}\) smooth (\(C^\infty\)). Every diffeomorphism is a homeomorphism, but a cube and a sphere are homeomorphic yet not diffeomorphic — the corners admit no tangent plane.

Two 2‑sheets inside a 4‑space meet in isolated points (red), since 2+2−4 = 0.

2. Why dimension four is special

Generic submanifolds of dimension \(p,q\) inside an \(n\)-manifold meet in dimension

\[ \dim(\Sigma_1\cap\Sigma_2)=p+q-n. \]

For two surfaces \((p=q=2)\): \(n=3\) gives a curve, \(n=4\) gives isolated points, \(n\ge5\) gives \(<0\) (generically disjoint). The Whitney trick cancels a pair of opposite double points by sliding a sheet over a Whitney disk — there is room for \(n\ge5\), but it stalls at \(n=4\). Overlay §2 marks those 0‑dimensional double points in red.

Each kinky handle (self‑intersection loop) spawns a smaller one — the infinite Casson tower.

3. Casson handles

Removing the leftover points needs stage‑1 Whitney disks; these self‑intersect, needing stage‑2 disks, and so on:

\[ \text{Stage }0 \to \text{Stage }1 \to \text{Stage }2 \to \cdots \]

The limit is a Casson handle — topologically the standard open 2‑handle \(D^2\times\mathbb{R}^2\) (Freedman) but not always smoothly standard. Disk classes I/II/III in the Simulator draw stages 0, 1 and 2 of exactly this tower.

Self‑dual curvature \(F_A=\star F_A\): field loops threaded around an instanton core.

4. Gauge theory: the PDE filter

Donaldson tested smoothness with Yang–Mills. For a connection \(A\) with curvature \(F_A=dA+A\wedge A\), (anti‑)self‑dual instantons satisfy

\[ F_A=\pm\star F_A, \]

where \(\star\) maps 2‑forms to 2‑forms precisely because \(2=4-2\). Seiberg–Witten replaces \(SU(2)\) by a \(U(1)\) connection and spinor \(\psi\):

\[ \mathcal{D}_A\psi=0,\qquad F_A^{+}=\sigma(\psi,\psi). \]

The compact moduli space yields invariants that jump when a Casson handle fails to smooth. Overlay §4 shows the self‑dual field loops around the disk.

A knot with its framed pushoff K′ — the attaching data of the trace \(X_n(K)\).

5. Knot traces and the Conway knot

Attach a 2‑handle \(D^2\times D^2\) to \(B^4\) along \(K\subset S^3=\partial B^4\) with integer framing \(n\):

\[ X_n(K)=B^4\cup_K\big(D^2\times D^2\big). \]

A knot is topologically slice if it bounds a continuous disk in \(B^4\), smoothly slice if a smooth one. Freedman showed the Conway knot is topologically slice. In 2020 Piccirillo built a sibling knot \(K'\) with the same trace \(X_0\), computed Rasmussen's \(s(K')\neq0\), and concluded the Conway knot is not smoothly slice. Overlay §5 draws the framed pushoff linking \(K\) exactly \(n\) times.

Nested charts: standard near the core, smoothly "jittered" toward infinity.

6. The anomaly: exotic \(\mathbb{R}^4\)

For every \(n\neq4\), \(\mathbb{R}^n\) carries a unique smooth structure. Only \(\mathbb{R}^4\) admits uncountably many — spaces all homeomorphic to standard \(\mathbb{R}^4\) yet pairwise non‑diffeomorphic. Overlay §6 nests shells that are clean near the origin and jittered far out.

Reading the simulator

  • Knot complexity sets the torus‑knot boundary \(T(2,2c{+}1)\) on the Clifford torus of \(S^3\subset\mathbb{R}^4\).
  • Framing n twists the cyan ribbon \(n\) times so the pushoff \(K'\) links \(K\) exactly \(n\) times.
  • 4D time t rotates the construction in the \(xw\) and \(yw\) planes (frequencies 1 and 2) and projects \(\mathbb{R}^4\!\to\!\mathbb{R}^3\); integer frequencies make it periodic over the slider \([0,2\pi]\). Camera pan/zoom/orbit never touch this shape.
  • 4D image overlays the faint union of the knot over all \(t\) so you watch the live shape sweep through its own orbit.
  • Disk classes I/II/III draw the slice‑disk membrane, the Whitney disks pairing double points, and the recursive Casson tower.

Bibliography — contributors to these ideas

  • Hassler Whitney (1944) — the Whitney trick; cancelling double points by an embedded disk in dimensions \(\ge 5\).
  • John Milnor (1956) — exotic 7‑spheres: first homeomorphic‑but‑not‑diffeomorphic manifolds.
  • Andrew Casson (1973) — Casson handles, the infinite towers of kinky handles standing in for missing Whitney disks.
  • Robion Kirby & Laurent Siebenmann (1977) — smoothing theory and the Kirby–Siebenmann invariant in high dimensions.
  • Michael Freedman (1982) — topological classification of simply‑connected 4‑manifolds; Casson handles are topologically standard. Fields Medal 1986.
  • Simon Donaldson (1983) — Yang–Mills gauge theory obstructing smooth structures; Donaldson's diagonalization theorem. Fields Medal 1986.
  • Clifford Taubes (1987) — uncountably many exotic smooth structures on \(\mathbb{R}^4\).
  • Robert Gompf (1980s–90s) — explicit exotic \(\mathbb{R}^4\) constructions and Kirby calculus.
  • Nathan Seiberg & Edward Witten (1994) — the Seiberg–Witten equations and their 4‑manifold invariants.
  • Jacob Rasmussen (2010) — the \(s\)‑invariant from Khovanov homology bounding the smooth slice genus.
  • Lisa Piccirillo (2020) — "The Conway knot is not slice," Annals of Mathematics, via a trace‑sibling knot.
  • John H. Conway — the Conway knot \(11n34\), topologically but not smoothly slice.