✦ For everyone, free.

Practical knowledge for real and everyday life

Home

6.2.3 Tensor Valence Classification Area

Tensor Valence Classification Area explores how tensors are categorized by their valence, defining their structure and how they transform under coordinate changes.

Tensor Valence Classification Area is the branch of tensor classification that records the specific, ordered sequence of contravariant and covariant argument slots a tensor exposes, refining type classification's bare totals (p, q) into a full left-to-right sequence of variances such as "up, down, up," and treating two tensors with identical totals but different slot orderings as classified differently. This area supplies the finest-grained classification among order, type, and valence, since valence alone determines exactly which slot a given contraction, symmetrization, or substitution operation acts upon.


Defining Valence as an Ordered Sequence

The Sequence of Variances

For a tensor written with indices in a fixed order, for example T^{i}{}_{j}{}^{k}, the valence is the sequence recording the variance of each slot in that order:

valence (T) = ( , , )

with marking a contravariant (upper) slot and marking a covariant (lower) slot, read in the same left-to-right order the indices are written.

Why Order of Slots, Not Just Totals, Matters

A tensor of type (2,1) could in principle have any of three distinct valences depending on where its single lower slot falls relative to its two upper slots: (↑,↑,↓), (↑,↓,↑), or (↓,↑,↑); type classification alone cannot distinguish these three tensors, but valence classification can.

Counting Distinct Valences for a Given Type

For a type (p, q), the number of distinct valences consistent with that type equals the number of ways to interleave p up-arrows and q down-arrows in a sequence of length p + q:

number of valences = (p+q)! p!q!

so, for example, type (2,1) admits exactly 3!/(2!1!) = 3 distinct valences, matching the three interleavings listed above.


Notating and Visualizing Valence

Arrow-Sequence Notation

Within this area, valence is often drawn as a horizontal row of up-arrows and down-arrows, one per slot, giving an immediate visual comparison between tensors that share a type but differ in arrangement.

(↑,↑,↓): (↑,↓,↑): (↓,↑,↑):

Slot-Position Labels

Alongside arrow diagrams, this area uses explicit position labels — "first slot," "second slot," and so on — to refer unambiguously to a single argument of a tensor, a level of precision that neither order nor type classification can offer on its own.


Operational Consequences of Valence

Contraction Requires a Specific Pair of Slots

Contracting "the first upper slot against the second lower slot" is only a well-posed instruction once the full valence sequence is known; specifying only the type (p, q) leaves the choice of which upper slot pairs with which lower slot ambiguous whenever p or q exceeds one.

Valence Change Under Contraction

valence after contracting slots i and j = original valence with positions i , j removed

removes exactly those two labeled positions from the sequence and closes the resulting gap, preserving the relative order of every remaining slot.

Valence Under Tensor Product Is Concatenation

Forming A ⊗ B appends the valence sequence of B, in order, after the valence sequence of A, so that the valence of the product is fully determined by, and recoverable from, the valences of the two factors together with the order in which they were multiplied.


Relation to the Other Classification Areas

Valence Refines Type, Which Refines Order

Order gives only the length of the valence sequence; type gives only the count of up-arrows and down-arrows within it; valence gives the full sequence itself, making it strictly the most informative of the three areas while remaining fully consistent with, and derivable into, the coarser two.

Valence Says Nothing About Component Values or Symmetry

A fixed valence only fixes the arrangement of variances; it says nothing about the specific numbers stored in the tensor's components, nor about whether those components happen to satisfy a symmetry pattern under exchange of same-variance slots, both of which belong to separate classification areas layered on top of a valence already fixed.


Why Valence Classification Is Useful

Precision Required for Multi-Slot Operations

Any tensor with more than one upper slot or more than one lower slot needs valence, not just type, to specify unambiguously which particular slots participate in a contraction, a symmetrization, or an explicit substitution of one argument.

A Natural Stopping Point in the Classification Hierarchy

Because valence records the full arrangement of variances and nothing more, it forms a natural, complete stopping point for classification purposes based purely on slot structure, with any further distinction between tensors necessarily coming from their actual component values rather than from their combinatorial shape.