1.5.4 Scalar Tensor Representation

Scalar Tensor Representation encodes tensors as scalars using algebraic structures, simplifying complex relationships in physics and geometry.

Scalar Tensor Representation is the description of a tensor of type (0, 0), an object represented by a single component that carries no free indices and remains numerically unchanged under any admissible change of basis or coordinate system. It sits at the base of the tensor type hierarchy, beneath vectors of type (1, 0) and covectors of type (0, 1), and it supplies the model against which the transformation behavior of every higher-rank tensor is measured: a scalar is precisely the case in which the general tensor transformation law degenerates into the identity map.

Definition and Basic Properties

The Single-Component Object

A scalar tensor is specified in any given basis by exactly one number, drawn from the underlying field F over which the vector space V is defined. Unlike a vector, which requires n components relative to a basis of an n-dimensional space, or a covector, which likewise requires n components relative to the dual basis, a scalar requires no basis at all to be written down. Its value is the same number regardless of which basis of V is chosen.

Rank and Type

In the (p, q) classification of tensors, a scalar has p = 0 upper indices and q = 0 lower indices, giving it rank zero. It is the unique tensor type with no indices, and it occupies the one-dimensional space T^0_0(V), which is naturally identified with the field F itself.

T_{0}^{0} (V) ≅ F

The Transformation Law for Scalars

Trivial Invariance Under Basis Change

Every tensor of type (p, q) transforms according to a law involving p copies of the basis change matrix and q copies of its inverse, applied to each of the tensor's indices. Since a scalar has no indices, none of these factors appear, and the transformation law reduces to the statement that the component is left unchanged.

s' = s

Here s denotes the value of the scalar in the original basis and s′ denotes its value in the new basis; the two are equal by definition of what it means for an object to be a scalar. This equality is what is meant when a scalar is called invariant: the object itself does not depend on the coordinate description used to express it.

Contrast with Vector and Covector Transformation

A vector component transforms with a single factor of the basis change matrix, and a covector component transforms with a single factor of its inverse. Comparing these laws to the scalar case shows that the scalar transformation law is the p = q = 0 instance of the same general pattern, obtained by dropping every index-dependent factor.

\begin{matrix} v^{'} i & = & A_{j}^{i} v^{j} \\ ω_{'} i & = & B_{i}^{j} ω_{j} \\ s' & = & s \end{matrix}

Sources of Scalar Tensors

Full Contraction of Balanced Tensors

A scalar can be produced by fully contracting a tensor of type (p, p), pairing every upper index with a lower index until no free indices remain. The most familiar instance is the contraction of a vector with a covector, which yields a single number independent of basis.

s = ω_{i} v^{i}

Inner Products and Norms

When the vector space carries a metric tensor, contracting two vectors through the metric produces a scalar inner product, and contracting a vector with itself through the metric produces a scalar that serves as a squared norm.

s = g_{i j} u^{i} v^{j}

Determinants and Traces

Certain scalar-valued functions of higher-rank tensors, such as the trace of a type (1, 1) tensor or the determinant built from a set of vectors, are themselves scalars because their value does not depend on the basis used to compute them, even though the intermediate arrays of numbers used in the computation do depend on the basis.

Scalars as Fields

Scalar Fields Over a Manifold

When a scalar is assigned to every point of a manifold rather than to a single vector space, the resulting assignment is called a scalar field. Physical quantities such as temperature, pressure, density, and electric potential are commonly modeled as scalar fields, each point of the space being associated with one invariant number.

Distinction Between Pointwise Scalars and Constant Scalars

A scalar field need not take the same value at every point; what makes each of its values a scalar is that the value at a given point does not change when the coordinate system used to label points is changed, not that the field is constant across the manifold.

Role Within the Tensor Algebra

The Base Case of the Hierarchy

Within the graded tensor algebra T(V), formed as the direct sum of all spaces T^p_q(V), the scalar space T^0_0(V) occupies the grade-zero piece. Every other graded piece is built by tensor products of V and V*, while the grade-zero piece is generated by neither, standing outside the tensor product construction as its multiplicative identity.

s \otimes T = T \otimes s = s T

Scalars as Coefficients

Because multiplying any tensor by a scalar simply rescales its components without altering its type, scalars act as the coefficients of the tensor algebra, allowing tensors of the same type to be combined into linear combinations while remaining within that type.

Diagrammatic Summary

The diagram summarizes the defining feature of a scalar tensor representation: a single component that remains equal to itself, s′ = s, across every change of basis, and that commonly arises as the result of fully contracting tensors of balanced upper and lower rank.