The Tesseract: The Shadow of a Four-Dimensional Cube

Building a tesseract by doubling

The cleanest way into four dimensions is not to imagine it but to build it, one dimension at a time, by a single repeated move: take the shape you have, slide a copy of it a unit step in a brand-new direction perpendicular to everything so far, and connect each old point to its moved twin. This is the prism (Cartesian-product) construction, and it manufactures the whole family of hypercubes.

• Start with a point (2⁰ = 1 vertex, dimension 0).
• Slide it and join: a segment (2 vertices, 1 edge).
• Slide the segment sideways: a square (4 vertices, 4 edges).
• Slide the square upward: a cube (8 vertices, 12 edges).
• Slide the cube along a fourth, mutually perpendicular axis: a tesseract (16 vertices, 32 edges).

The vertex count simply doubles at every step — 1, 2, 4, 8, 16 — because each move keeps the old copy and adds an identical new one, giving 2ⁿ vertices for the n-cube. Concretely, the 16 vertices of a unit tesseract are exactly the points whose coordinates are each 0 or 1: every string in {0,1}⁴, from (0,0,0,0) to (1,1,1,1). Two vertices are joined by an edge precisely when their coordinate strings differ in a single place, so each vertex has 4 neighbors and the edges are the 32 pairs at Hamming distance 1. That combinatorial skeleton — the hypercube graph Q₄ — is the tesseract, whether or not we can picture it filling space.

Counting the pieces: 16, 32, 24, 8

A tesseract is bounded not by flat faces but by solid cells, and the full census of its parts follows one formula. The number of m-dimensional faces of the n-cube is

• (number of m-faces) = C(n, m) · 2ⁿ⁻ᵐ,

where C(n, m) is the binomial coefficient. The reasoning: an m-face is fixed by choosing which m of the n coordinates are allowed to vary — that is C(n, m) ways — and then fixing each of the remaining n − m coordinates at 0 or 1, which is 2ⁿ⁻ᵐ ways. For n = 4 this yields, exactly:

• Vertices (m = 0): C(4,0)·2⁴ = 16.
• Edges (m = 1): C(4,1)·2³ = 32.
• Square faces (m = 2): C(4,2)·2² = 24.
• Cubic cells (m = 3): C(4,3)·2¹ = 8.

These four numbers satisfy the four-dimensional Euler–Poincaré relation, the higher analogue of the cube's V − E + F = 2. For a 4-polytope the alternating sum of the boundary counts vanishes: V − E + F − C = 16 − 32 + 24 − 8 = 0, reflecting that the tesseract's surface is topologically a 3-sphere, and odd-dimensional spheres have Euler characteristic 0.

The local picture is just as tidy. At each vertex sit C(4,1)=4 edges, C(4,2)=6 faces, and C(4,3)=4 cells; the four cells meeting there arrange as a tetrahedron — the tesseract's vertex figure. Around every edge, exactly 3 cubes meet. These facts are packed into its Schläfli symbol {4,3,3}: cells are cubes {4,3}, and three of them ring each edge. The tesseract is one of the six regular 4-polytopes, the four-dimensional counterparts of the five Platonic solids.

Why we only ever see the shadow

Our retinas and screens are three-dimensional at most, so a four-dimensional object cannot be shown to us whole. What we can show is a projection — a shadow that discards one dimension, exactly as a lamp casts a flat shadow of a solid cube onto a wall. The famous 'cube inside a cube' image of a tesseract is a Schlegel diagram, a perspective projection taken from a viewpoint just outside one of the eight cubic cells.

The mechanism is ordinary perspective, run one dimension higher. Place the 4D eye on the w-axis at distance d, and send each point (x, y, z, w) to the 3D point obtained by dividing the first three coordinates by its distance from the eye:

• (x, y, z, w) ↦ (x, y, z) ÷ (d − w).

Points with larger w are 'nearer' the eye and project bigger; points with smaller w project smaller. The cell lying at w = +1 therefore becomes a large outer cube, while the opposite cell at w = −1 becomes a small cube nested inside it. The remaining six cells — each straddling both w values — project to the six frustum-shaped solids that fill the gap between inner and outer cube, joining their corresponding corners. Nothing is missing: all 8 cells, 24 faces, 32 edges, and 16 vertices are present in the shadow; only the fourth dimension has been flattened away, which is why the inner and outer cubes look unequal even though in 4D they are congruent.

A parallel (orthographic) projection of the same tesseract instead yields two equal, overlapping cubes offset by the shadow of the w-axis — the other picture people recognize. Both are honest shadows; they differ only in where the light sits.

Rotation in four dimensions: planes, not axes

Here is the first place four dimensions genuinely breaks our intuition. In 3D we say an object spins around an axis — but that is a coincidence of three dimensions. A rotation actually fixes a plane and turns the two remaining directions into each other; in 3D the leftover 'axis' is simply the line perpendicular to that plane. In 4D there is no single perpendicular line — the complement of a plane is another plane — so rotations are best described by their plane of rotation.

Four coordinates x, y, z, w give C(4,2) = 6 coordinate planes: xy, xz, xw, yz, yw, zw. A rotation by angle θ in, say, the xw-plane mixes only those two coordinates and leaves y and z untouched:

• x′ = x cosθ − w sinθ, w′ = x sinθ + w cosθ (y, z fixed).

Because there are six independent planes, the rotation group SO(4) is 6-dimensional (in general dim SO(n) = n(n−1)/2). Four dimensions also permits something with no 3D counterpart: a double rotation, spinning in two completely orthogonal planes (for example xy and zw) at once, with two independent angles. When those two angles are equal the motion is isoclinic — every point turns through the same angle — and these special rotations are the geometry behind unit quaternions and the double cover Spin(4) = SU(2) × SU(2). For visualizing a tesseract, though, the crucial case is the simplest: a single rotation in a plane that involves the w-axis.

The inside-out illusion, explained

Now combine the two facts — the shadow is set by each vertex's w-coordinate, and a 4D rotation can change that w. Rotate the tesseract in the xw-plane. As θ grows, vertices trade x for w and w for x: a vertex that started deep in the fourth dimension at w = −1 (part of the small inner cube) swings toward larger w, while a vertex at w = +1 (the big outer cube) swings toward smaller w. Feed those changing w-values through the projection (x, y, z) ÷ (d − w) and the consequence is vivid:

• The inner cube swells as its vertices gain w and project larger.
• Its corners pass outward through the faces of the outer cube.
• At a quarter turn the two cubes have swapped roles entirely — the former inside is now the outside, and vice versa.

To a 3D viewer the figure appears to turn inside out, an eversion that no rigid solid object could perform: you cannot pass a smaller cube out through the walls of a larger one without tearing something. But nothing is torn. In four dimensions the tesseract is turning as rigidly as a wheel; every edge keeps its length and every angle its measure. The paradox lives entirely in the shadow, which must distort because it has one dimension too few to record the motion faithfully.

The right way to feel this is by analogy one step down. Watch the flat shadow of a rotating 3D cube: its outline stretches, edges appear to lengthen and shrink, near and far faces swap places, and a square-in-square silhouette can seem to invert — all while the actual cube stays perfectly rigid. The tesseract's inside-out shadow is precisely that phenomenon promoted from the 2D wall to our 3D world.

Shadows versus slices, and where hypercubes show up

A projection is not the only way to bring a tesseract into 3D; the other is a cross-section, and the two must not be confused. A shadow flattens the whole object at once; a slice keeps only the points that lie in a chosen 3D hyperplane, and the slice's shape changes as the plane sweeps through. Passing a hyperplane through the tesseract corner-first produces a growing-then-shrinking sequence point → tetrahedron → octahedron → tetrahedron → point, while a cell-first cut yields a cube that appears, holds, and vanishes. (This is the higher analogue of slicing a cube: corner-first it goes point → triangle → hexagon → triangle → point.)

Two more constructions round out the picture. Unfolding a tesseract into 3D — the analogue of flattening a cube's surface into a paper cross of six squares — gives a solid arrangement of eight cubes; there are exactly 261 distinct nets (compared with 11 for the ordinary cube), and the cross-shaped one is what Salvador Dalí painted in Crucifixion (Corpus Hypercubus), 1954. The tesseract's dual — swap vertices for cells — is the 16-cell, with 8 vertices, 24 edges, 32 faces, and 16 tetrahedral cells, its counts a mirror image of the tesseract's own.

History and use. Ludwig Schläfli classified the regular polytopes of every dimension in the 1850s (his manuscript was published in full only in 1901); Washington Irving Stringham drew the six four-dimensional cases in 1880; and the name 'tesseract' — from the Greek for 'four rays' — was coined by Charles Howard Hinton in 1888. Far from a curiosity, the hypercube graph Qₙ is a workhorse of computing: it is the wiring diagram of hypercube interconnection networks in parallel supercomputers, the vertices of an n-bit Gray code trace a Hamiltonian cycle around it, and n-dimensional cubes model the state spaces of Boolean logic, error-correcting codes, and high-dimensional data. Understanding how a tesseract casts and rotates its shadow is, in the end, a rehearsal for reasoning about dimensions we will never be able to see.

The measure polytopes: doubling from a point to a tesseract, and the counting laws each obeys

Object	Dimension n	Vertices (2ⁿ)	Edges (n·2ⁿ⁻¹)	Facets (2n)
Point	0	1	0	0
Segment	1	2	1	2 endpoints
Square	2	4	4	4 edges
Cube	3	8	12	6 square faces
Tesseract	4	16	32	8 cubic cells