Symmetry, duality, topology — the three invariants that survive when a 4-polytope transforms, each readable as a pattern in parallel coordinates.
A 4-polytope has many numbers attached: vertex counts, edge counts, faces, cells, angles, distances. Most change under rotation, duality, or deformation. The invariants are what survive.
Three invariants do most of the work. Symmetry: which rotations map the polytope to itself. Duality: what happens when vertices swap with cells and edges with faces. Topology: what's left when distances are forgotten — Euler characteristic and the crossing pattern between axes.
Each shows up as a readable pattern in parallel coordinates.
A symmetry leaves an object unchanged. Rotate a square 90° or reflect it across a diagonal — same square. The collection of such transformations forms a group.
A 4-polytope symmetry permutes vertices while preserving edge structure, so it permutes polylines while preserving the crossing pattern. The polyline bundle maps to itself.
Consider a polytope rotating continuously in a 4D rotation plane. At most orientations the polyline bundle looks different from its starting configuration. But at certain special angles the rotated polytope is indistinguishable from the original: every vertex has moved to a position previously occupied by some other vertex, and every edge maps to an edge. The polyline pattern is invariant.
For a rotation in the plane, the tesseract hits a symmetry every (the cyclic group ). The 24-cell, with its denser vertex arrangement, hits a symmetry every (the cyclic group ). These are subgroups of the full symmetry group, which has order for the tesseract and for the 24-cell.
Figure 1. Symmetry explorer. The wireframe projection (left) and parallel coordinates (right) rotate together. When the rotation maps the polytope to itself, the status bar flashes and the PC pattern is momentarily highlighted. The 24-cell has 6 symmetric orientations per full turn in the zw plane; the tesseract has 4.
The vertex figure of a polytope is the shape formed by the neighbors of a vertex. It encodes the local combinatorial structure: how many edges, faces, and cells meet at that point. For the tesseract, each vertex connects to 4 others and those 4 neighbors form a tetrahedron. For the 24-cell, each vertex connects to 8 others and those 8 neighbors form a cube. For the 16-cell, each vertex connects to 6 others and those neighbors form an octahedron.
In parallel coordinates, selecting a vertex highlights one polyline. Its neighbors are the polylines connected to it by edges. The vertex figure is visible as the local crossing pattern of these neighbor polylines.
Figure 2. Vertex figure visualization. Click any vertex in the wireframe (left) to highlight it and its edge-connected neighbors. The same vertices are highlighted in the PC view (right). Neighbors form the vertex figure: tetrahedron for the tesseract, cube for the 24-cell, octahedron for the 16-cell.
Every convex polytope has a dual, obtained by swapping vertices with cells and edges with faces. If a polytope has Schläfli symbol , its dual has symbol . The duality operation reverses the entire incidence structure: a vertex of the original becomes a cell of the dual, an edge becomes a face, a face becomes an edge, and a cell becomes a vertex.
For three-dimensional polytopes this is familiar: the cube (6 faces, 8 vertices, 12 edges) is dual to the octahedron (8 faces, 6 vertices, 12 edges). In four dimensions the duality is richer because we have four levels of structure to invert.
The tesseract has 16 vertices, 32 edges, 24 square faces, and 8 cubic cells. Its dual, the 16-cell , has 8 vertices, 24 edges, 32 triangular faces, and 16 tetrahedral cells. The numbers swap in complementary pairs: and .
When we visualize a polytope and its dual simultaneously, a distinctive pattern emerges. The dual polytope is constructed by placing a vertex at the centroid of each cell of the original. In the PC representation, the dual's vertex polylines thread through the gaps left by the original's polylines. The two families of polylines interleave, creating a woven pattern that encodes the incidence structure.
Figure 3. A polytope (blue) and its dual (red) shown in wireframe projection (left) and parallel coordinates (right). In the PC view, the dual's polylines interleave with the original's. For self-dual polytopes, the two patterns are identical up to rotation. Drag the wireframe to rotate.
Two polytopes are self-dual in four dimensions. The 5-cell (the 4D simplex) is dual to itself: swapping vertices and cells gives back the same combinatorial type, since all 5 vertices connect to all others. But the simplex is self-dual in every dimension, so this is not surprising.
The 24-cell is self-dual: 24 vertices ↔ 24 octahedral cells, 96 edges ↔ 96 faces. The 24-cell has no analogue in any other dimension — the only regular non-simplex self-dual polytope that exists.
In 3D, only the tetrahedron is self-dual (a simplex). In 4D, the 5-cell is self-dual for the same simplex reason. The 24-cell stands alone — 1152 symmetries, octahedral cells, cube vertex figure, uniquely dense PC signature.
The combinatorial structure of a polytope is governed by a topological invariant. For any convex 4-polytope, the Euler characteristic satisfies:
This is the four-dimensional analogue of the familiar for 3-polytopes. The alternating sum always vanishes for convex 4-polytopes, regardless of how many vertices, edges, faces, or cells they have. It is a topological invariant: it does not change under continuous deformations that preserve the cell structure.
Each of our 4-polytopes achieves with a different set of numbers. The 5-cell does it with 5, 10, 10, 5. The tesseract uses 16, 32, 24, 8. The Euler relation provides a consistency check: if you count vertices, edges, faces, and cells and the alternating sum is not zero, you have miscounted.
| Polytope | |||||
|---|---|---|---|---|---|
| 5-cell | 5 | 10 | 10 | 5 | 0 |
| Tesseract | 16 | 32 | 24 | 8 | 0 |
| 16-cell | 8 | 24 | 32 | 16 | 0 |
| 24-cell | 24 | 96 | 96 | 24 | 0 |
Figure 4. The Euler characteristic for each regular 4-polytope. Despite having vastly different numbers of elements, all convex 4-polytopes share the same invariant. Notice how duality swaps V with C and E with F: the tesseract's row is the 16-cell's row reversed.
Duality and the Euler characteristic. If a polytope has , its dual has . The alternating sum is unchanged because reversing the order of an even number of terms preserves the alternating sign pattern. Duality is an Euler-preserving involution.
Duality is not just a static relationship. We can visualize the transition from a polytope to its dual by smoothly morphing the vertex coordinates. Start with the 16 vertices of the tesseract. Move them continuously toward the 8 vertices of the 16-cell (with the remaining 8 polylines merging into their partners). The polyline count drops from 16 to 8, the edge count from 32 to 24, while faces rise from 24 to 32 and cells from 8 to 16. Through all of this, .
Figure 5. Smooth morph from the tesseract to its dual, the 16-cell. As vertex positions interpolate, polylines move continuously. The counts change discretely at t = 1, but the Euler invariant holds throughout.
A topological deformation stretches or bends a shape without tearing it or gluing new parts. It changes the positions of vertices (and therefore the positions of polylines in parallel coordinates) but preserves the combinatorial structure: which vertices are connected by edges, which edges bound faces, and which faces enclose cells.
In the parallel coordinates view, a topological deformation moves polylines around, changes their slopes and intersection positions, but does not change which pairs of polylines intersect between which pairs of axes. The pattern of crossings — the incidence structure — is a topological invariant.
Drag the slider to deform the tesseract's vertices. Polylines move, but the crossing pattern is preserved.
Figure 6. Topological deformation of the tesseract. The slider applies a smooth nonlinear deformation to the vertex coordinates, moving polylines in parallel coordinates. Edge connectivity (which polylines cross between which axes) is preserved: topology is invariant, geometry is not.
Two 4-polytopes are topologically equivalent if they share the same incidence structure — in PC, the same crossing pattern between axis pairs. Crossing positions change under deformation; their existence does not.
Symmetry, duality, and topology aren't independent. A self-dual polytope has a symmetry that swaps vertices with cells. Duality preserves Euler characteristic. Topology is the floor (any continuous deformation); symmetry is the ceiling (every rotation).
In parallel coordinates, each shows up differently: symmetry as temporal invariance under rotation, duality as interleaved bundles, topology as persistent crossings.
The axes here are exactly 4. What happens as dimension climbs to 10, 100, 1000? Next.