5.21.3 Tensor Product Parenthesis Suppression

Tensor Product Parenthesis Suppression is the standard notational convention of writing a multi-factor tensor product V₁ ⊗ V₂ ⊗ ... ⊗ Vₙ without any parentheses at all, a practice licensed by the associativity structure of the tensor product, which guarantees that every possible parenthesization of the same sequence of factors is connected by a canonical, coherent isomorphism. Parenthesis suppression is not a mere shorthand adopted for convenience alone; it reflects the mathematical fact that, up to canonical isomorphism, there is only one tensor product of a given ordered list of vector spaces, regardless of the order in which the pairwise tensor products used to build it are actually performed.

The Notational Convention

From Parenthesized to Unparenthesized Expressions

For three factors, associativity provides a canonical isomorphism (U ⊗ V) ⊗ W ≅ U ⊗ (V ⊗ W), and parenthesis suppression is the decision to write both sides simply as U ⊗ V ⊗ W, treating the canonical isomorphism as an identification rather than as a nontrivial map requiring explicit mention every time it is used.

Extension to Arbitrarily Many Factors

For n factors, there are many distinct ways to insert parentheses into V₁ ⊗ V₂ ⊗ ... ⊗ Vₙ (their count is given by the Catalan numbers), and parenthesis suppression asserts that all of them may be identified with the single unparenthesized expression, since any two parenthesizations are connected by a composite of grouping isomorphisms.

Why the Suppression Is Justified

The Coherence Theorem

Mac Lane's coherence theorem for tensor products (and more generally for symmetric monoidal categories) guarantees that any two sequences of associativity isomorphisms connecting one parenthesization of V₁ ⊗ ... ⊗ Vₙ to another parenthesization must be equal as maps. Without this coherence, parenthesis suppression would be unsafe: different regrouping paths might produce different identifications, and dropping parentheses could silently introduce an ambiguity.

Reduction to Pairwise Regroupings

Any full regrouping of n factors can be decomposed into a sequence of elementary triple-factor regroupings (each an instance of the associativity isomorphism α applied to three consecutive factors), so coherence for the whole multi-factor product follows from coherence at the level of three factors, verified directly via the pentagon condition for four factors and extended inductively.

Diagram of Suppressed Parentheses

Consequences of the Convention

Multi-Index Notation Without Ambiguity

Parenthesis suppression is the reason a general element of V₁ ⊗ V₂ ⊗ ... ⊗ Vₙ can be described in coordinates by a single multi-index array (c_{i₁ i₂ ... iₙ}) relative to chosen bases, without needing to specify which nested grouping the indices "belong to" — the coherence isomorphisms guarantee the coordinates match up identically regardless of grouping.

Simplification of Tensor Product Map Notation

The convention extends to maps as well: f₁ ⊗ f₂ ⊗ ... ⊗ fₙ : V₁ ⊗ ... ⊗ Vₙ → V₁′ ⊗ ... ⊗ Vₙ′ is written and manipulated without parentheses, since the induced maps between different groupings commute with the associativity isomorphisms by naturality, so no grouping-dependent correction terms are ever needed.

Limits of the Suppression

Order of Factors Still Matters

Parenthesis suppression concerns only the grouping of factors, not their order: U ⊗ V and V ⊗ U are related by the separate symmetry isomorphism, not by associativity, and this isomorphism, while canonical, is a genuinely different identification that is not absorbed into parenthesis suppression; reordering factors must still be tracked explicitly.

Suppression Is a Convention, Not a Literal Identity

Even though parentheses are omitted, the underlying spaces (U ⊗ V) ⊗ W and U ⊗ (V ⊗ W) remain, strictly speaking, distinct set-theoretic constructions; parenthesis suppression is the statement that they are canonically isomorphic and may be safely conflated, not that they are equal by definition, a distinction that matters in fully formal or foundational treatments of the tensor product.

Significance of Parenthesis Suppression

Practical Simplification Across Mathematics and Physics

Parenthesis suppression is what allows tensor product notation to remain readable and usable in practice, from the abstract algebra of multilinear forms to the explicit index notation of physics and engineering, where tensors of high order are manipulated constantly without any need to track a specific grouping of their tensor factors.

A Concrete Instance of Categorical Coherence

Parenthesis suppression is a paradigmatic and widely encountered example of a coherence phenomenon in category theory: a situation where many a priori different canonical isomorphisms are shown to agree, licensing a simplified notation that treats formally distinct but canonically identified constructions as one and the same.