3Blue1Brown
November 8, 2024
TL;DR
Five geometry puzzles demonstrate how stepping into higher dimensions provides elegant solutions to seemingly lower-dimensional problems, from rhombus tilings to circle coverings to proving collinearity.
“With a little thinking, you can see pretty quickly that the answer has to be no, because for example, that very first tiling I showed you, which is like a squished rectangular grid, has no hexagons at all.”
“This move of rotating hexagons when you find them is really the same thing as adding or removing cubes, and that makes this way easier to think about.”
“The high-level idea is that along lines of latitude, that little patch of area gets stretched out, but along lines of longitude, it actually gets squished down because of the angle that it sits at, and those two effects happen to cancel out precisely.”
“What I'm jealous of, and a little bit sad about, is the thought that there might be problems in three dimensions where somehow squinting your eyes as a four-dimensional creature or even higher would offer some guiding light that, to you and me, is inaccessible.”
1. Rhombus Tiling Hexagons
A rhombus with 60° and 120° angles can tile the plane in multiple ways. The puzzle asks whether you can transform any tiling of a finite hexagon into any other using moves that rotate three-rhombus hexagons. The key insight is recognizing these tilings as projections of 3D cube stacks, making the answer yes with a maximum of n³ steps needed.
2. Tarski-Planck Problem: Covering a Circle with Strips
Given a circle of radius 1, cover it completely with strips (regions between parallel chords) and find the minimum sum of strip widths. While parallel strips sum to 2 (the diameter), the puzzle appears impossible to solve directly. The solution uses a 3D hemisphere projection where strip area equals π times width, proving the minimum is exactly 2.
3. External Tangents and Collinearity
For three circles, draw external tangents to each pair and mark their intersection points. These three points always lie on a single line. The proof uses cones (or pyramids) in 3D where each circle is the base, and the centers of similarity concept ensures collinearity regardless of circle positions.
4. Generalization and the Cone Method
The cone-based proof is more general than the sphere method, allowing circles to be nested. The concept of centers of similarity—points where scaling one shape produces another—unifies the proof and applies to any similar shapes, not just circles.
5. Tetrahedron Volume Formula
The challenge is deriving an explicit formula for tetrahedron volume from coordinates of its four vertices. The solution involves a 4×4 matrix determinant, with deeper insights available by considering the analogous 2D problem (triangle area) across dimensions.
6. 4D Hypercube Projection and Rhombic Dodecahedron
A 4D hypercube projected perpendicular to its diagonal (1,1,1,1) forms a 3D rhombic dodecahedron. This shape can tessellate 3D space, and the projection contains four 'squished cubes' (one per basis vector subset) that tile it.
7. The Rhombic Dodecahedron Tiling Puzzle
The absurd cocktail party puzzle: tile a large rhombic dodecahedron with unit rhombic dodecahedra, making moves that slide pieces through the center. Understanding this as a 4D hypercube projection reveals the maximum moves needed is n⁴.
8. Why Higher Dimensions Matter
Higher-dimensional constructs solve lower-dimensional problems: quaternions elegantly handle 3D rotations, 24D sphere packing relates to Voyager error correction, and high-dimensional statistics explain neural network scaling. However, intuition breaks down beyond 3D.
9. The Sadness of Inaccessible Intuition
While 2D and 3D problems allow visual intuition for higher-dimensional solutions, problems in 3D or higher that require stepping into 4+ dimensions lack intuitive guidance. Mathematicians must rely on pure analysis rather than the geometric intuition that guides lower-dimensional thinking.
10. Broader Applications of Dimensional Thinking
Higher-dimensional geometry appears unexpectedly in neural networks, compression algorithms, and error correction. The same dimensional-thinking pattern—finding hidden structure in higher spaces—solves diverse problems across mathematics and engineering.