This video provides an intuitive, first-principles explanation of what a tensor is, inspired by Richard Feynman's pedagogical approach. Starting with the familiar vector form of Ohm's Law, it explores the concept of conductivity in anisotropic crystals, where the resulting current density is not always aligned with the applied electric field. This misalignment reveals that conductivity cannot be described by a single number (a scalar) but requires a matrix of components—four in 2D and nine in 3D—leading to the rediscovery of a rank 2 tensor. The video generalizes this concept, defining scalars and vectors as rank 0 and rank 1 tensors, respectively. It then identifies other physical quantities like moment of inertia and stress as rank 2 tensors by applying the same principle of misaligned vector proportionality. The explanation extends to higher-rank tensors by examining the piezoelectric effect (rank 3) and elasticity (rank 4). Finally, the video defines a tensor as a multi-linear, coordinate-independent "machine" that relates vectors and other tensors, concluding with why this mathematical framework is essential for describing spacetime curvature in Einstein's field equations.
Rediscovering Tensors: An Intuitive Journey
00:00:01 Beyond Scalars and Vectors: The video sets out to intuitively build the concept of a tensor from scratch, moving beyond the high-school definitions of scalars (magnitude only) and vectors (magnitude and direction).
01:25 A Familiar Starting Point: The explanation begins with the vector form of Ohm's Law, where current density (J) is proportional to the electric field (E). The proportionality constant is conductivity.
02:48 The Anisotropic Crystal Problem: For materials like graphite that conduct differently in different directions (anisotropic), a single number for conductivity is insufficient. The direction of the electric field matters.
05:07 The Key Insight: Misaligned Vectors: The crucial discovery is that in anisotropic materials, the resulting current density vector (J) is not necessarily in the same direction as the applied electric field vector (E). This is the fundamental property that requires a tensor.
07:36 Why More Numbers Are Needed: In a general coordinate system, an electric field applied along one axis (e.g., the x-axis) can produce a current with components along multiple axes (both x and y). Therefore, to calculate the total current in any single direction, you need to account for contributions from the electric field in all directions.
10:40 The Birth of a Rank 2 Tensor: This relationship requires a collection of numbers that can be organized into a matrix. In 2D, you need 4 numbers (2x2 matrix), and in 3D, you need 9 numbers (3x3 matrix). This object, requiring n² components in n-dimensions, is a rank 2 tensor.
14:16 Identifying Other Tensors: The same logic applies to other physical phenomena where two vectors are proportional but not always aligned:
Moment of Inertia: Relates angular velocity to angular momentum, which are not always parallel for complex rotating objects.
Stress: Relates the orientation of a surface (area vector) to the force acting upon it, which are not always aligned.
18:01 Understanding Higher Rank Tensors: The concept is extended to relationships involving other tensors:
Piezoelectric Constant (Rank 3): Relates a rank 2 tensor (stress) to a rank 1 vector (electric field).
Elastic Constant (Rank 4): Relates a rank 2 tensor (stress) to another rank 2 tensor (strain).
20:33 What is a Tensor?: A tensor is formally defined as a multi-linear machine that maps vectors or tensors to other vectors or tensors, with the crucial property that the physical relationship it describes is independent of the coordinate system used.
21:57 Application in General Relativity: Tensors are essential for Einstein's field equations because describing spacetime curvature requires a tool that can relate how movement in one direction causes space to stretch or squeeze in all other directions simultaneously.