Penrose graphical notation

Multilinear algebra

In the language of multilinear algebra, each shape represents a multilinear function. The lines attached to shapes represent the inputs or outputs of a function, and attaching shapes together in some way is essentially the composition of functions.

Tensors

In the language of tensor algebra, a particular tensor is associated with a particular shape with many lines projecting upwards and downwards, corresponding to abstract upper and lower indices of tensors respectively. Connecting lines between two shapes corresponds to contraction of indices. One advantage of this notation is that one does not have to invent new letters for new indices. This notation is also explicitly basis-independent.

Matrices

Each shape represents a matrix, and tensor multiplication is done horizontally, and matrix multiplication is done vertically.

Metric tensor

The metric tensor is represented by a U-shaped loop or an upside-down U-shaped loop, depending on the type of tensor that is used.

Levi-Civita tensor

The Levi-Civita antisymmetric tensor is represented by a thick horizontal bar with sticks pointing downwards or upwards, depending on the type of tensor that is used.

Structure constant

The structure constants ( γ a b c ) of a Lie algebra are represented by a small triangle with one line pointing upwards and two lines pointing downwards.

Contraction of indices

Contraction of indices is represented by joining the index lines together.

Symmetrization

Symmetrization of indices is represented by a thick zig-zag or wavy bar crossing the index lines horizontally.

Antisymmetrization

Antisymmetrization of indices is represented by a thick straight line crossing the index lines horizontally.

Determinant

The determinant is formed by applying antisymmetrization to the indices.

Covariant derivative

The covariant derivative ( ∇ ) is represented by a circle around the tensor(s) to be differentiated and a line joined from the circle pointing downwards to represent the lower index of the derivative.

Tensor manipulation

The diagrammatic notation is useful in manipulating tensor algebra. It usually involves a few simple "identities" of tensor manipulations.

For example, ε a . . . c ϵ a . . . c = n ! , where n is the number of dimensions, is a common "identity".

Riemann curvature tensor

The Ricci and Bianchi identities given in terms of the Riemann curvature tensor illustrate the power of the notation

Extensions

The notation has been extended with support for spinors and twistors.