8.13.4 Tensor Index Position Type Signal

Tensor Index Position Type Signal is the informational role played by the complete pattern of upper and lower indices on a tensor symbol in announcing, without any further declaration, the algebraic type $(p,q)$ of that tensor — where $p$ counts the number of upper (contravariant) indices and $q$ counts the number of lower (covariant) indices. Where the transformation signal carried by a single index position tells a reader how that one component behaves under a change of coordinates, the type signal is the aggregate reading of every index position on a symbol at once, giving the classification of the whole object rather than of any single slot.

From Individual Positions to an Aggregate Type

Counting Upper and Lower Slots

Given a tensor symbol, the type signal is read directly by tallying the superscripts and subscripts attached to it. A symbol $T^{ij}{}_{k}$ carries two upper indices and one lower index, so the position pattern alone signals that this object is of type $(2,1)$:

No accompanying statement is required to establish this; the notation itself is the complete declaration of type.

Type as a Basis-Independent Classifier

The pair $(p,q)$ signaled by index position is invariant under change of coordinates — a tensor that is type $(2,1)$ in one basis remains type $(2,1)$ in every other basis, because a coordinate change transforms each individual component according to its own position but never alters how many indices are upper versus lower. The type signal therefore classifies the tensor as an abstract object, independent of any particular coordinate representation.

What the Type Signal Determines

Rank as a Derived Quantity

The total rank of a tensor, $p + q$, is read directly from the same position pattern that yields the type. A type $(1,1)$ tensor and a type $(2,0)$ tensor both have rank 2, but the type signal distinguishes them further by specifying how that rank is distributed between contravariant and covariant slots — information the rank alone does not carry.

Legality of Operations

The type signaled by index position determines which operations are legally available on a tensor. Only a tensor with at least one upper and one lower index (that is, $p \geq 1$ and $q \geq 1$) can be contracted with itself without invoking a metric; a tensor of type $(2,0)$, having no lower index available, requires the metric to produce any scalar invariant from its own components. The type signal is therefore consulted before manipulation, not merely after, to determine what operations are even meaningful.

Compatibility for Products and Sums

Two tensors can only be added if the position pattern on each signals the identical type $(p,q)$; a type $(1,1)$ tensor cannot be added to a type $(2,0)$ tensor because their index position signals disagree, even if both happen to have the same total rank. Multiplication, by contrast, is unrestricted between any two types and produces a new tensor whose type signal is the sum of the two factors' types component-wise, so that a type $(1,0)$ tensor multiplied by a type $(0,1)$ tensor yields a type $(1,1)$ tensor.

Type Signal Changes Under Raising and Lowering

Shifting Type While Preserving Rank

Applying the metric to raise a lower index or lower an upper index changes the type signal while preserving the total rank. Raising one index of a type $(0,2)$ tensor $T_{ij}$ produces a type $(1,1)$ tensor $T^{i}{}_{j}$:

The rank remains 2 throughout, but the type signal shifts from $(0,2)$ to $(1,1)$, reflecting a real change in which position each slot occupies, even though the underlying geometric object — once a metric is fixed — is often regarded as essentially the same entity presented in a different type.

Full Range of Equivalent Presentations

A single geometric tensor of total rank 2 in a metric space can, through repeated raising and lowering, be presented in any of the four possible position patterns of that rank: $(2,0)$, $(1,1)$ with either ordering, and $(0,2)$. The type signal distinguishes these four presentations from one another notationally, even while the metric establishes that they correspond to a single invariant object.

Reading Type Signals in Composite Expressions

Type of a Contraction Result

When an implicit contraction removes one upper and one lower index from a tensor, the type signal of the result is read by subtracting one from both $p$ and $q$ of the original. Contracting a type $(2,1)$ tensor over one upper–lower pair produces a type $(1,0)$ result, and the position pattern of the contracted expression confirms this directly: the two indices that formed the contracted pair vanish from the notation entirely, leaving only the free indices whose positions signal the reduced type.

Type Consistency as a Well-Formedness Check

Because every term in a valid tensor equation must carry the same free-index type signal, checking that the upper and lower positions of the free indices match across every term of an equation is a direct and sufficient test of that equation's dimensional and algebraic consistency, independent of what the terms actually compute.

Role Within Index Position Notation

The type signal is the synthesis of every individual index position signal into a single classifying statement about a tensor as a whole. Where a lone index position tells a reader how one component transforms, the aggregate type signal tells a reader what kind of object the entire symbol represents, what operations are available to it, and what other tensors it may be validly combined with — making it the highest-level piece of information that index position notation is capable of conveying.