Proof of Evaluation of Noncommutative Polynomials at Bounded Operators: a Unital Homomorphism Compatible with Adjoints, Substitution, Representations and the GNS Multiplication Operators
lemmalem:nc-polynomial-evaluation-operators-2026aProducts along words are multiplicative by induction on word length, and each remaining clause follows by checking that two linear maps agree on monomials (again by induction on word length) and invoking the uniqueness of linear extensions from monomials, with the transport clause using the finite-sum formula and the bound following from submultiplicativity of the operator norm.
Each result cited below is universally quantified over the data in its own statement. We work in the setting of Noncommutative Laws, Couplings and the Wasserstein Distance: Standing Notation and Complex Hilbert Spaces and Bounded Linear Maps: Standing Notation.
Throughout, is the -th component of and is the product along the letter ; the same notation is used for the other tuples below. Since also names a tuple in clause 4, we write for the successor of , as permitted by claim 1 of Natural Numbers. By Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §operations, is a complex vector space, and composition in it, being composition of maps, is associative, distributes over sums, commutes with scalar multiples and satisfies . Clauses 1 and 2 are proved for an arbitrary complex Hilbert space and an arbitrary -tuple of bounded operators on it; the later clauses apply them to other tuples and spaces.
Uniqueness principle. If are linear maps with for every , then , since both are linear maps with the prescribed values and such a map is unique by Linear Maps from Noncommutative Polynomials into Bounded Operators: Existence and Uniqueness from Values on Monomials §extension. In particular, by Evaluation of Noncommutative Polynomials at a Tuple of Bounded Operators §evaluation, a linear map with for every is the evaluation .
Clause 1. By The Algebra of Noncommutative Polynomials in Finitely Many Self-Adjoint Variables §monomials, Evaluation of Noncommutative Polynomials at a Tuple of Bounded Operators §evaluation and Evaluation of Noncommutative Polynomials at a Tuple of Bounded Operators §word-products, . For the letter is the word of length with value at , so by Evaluation of Noncommutative Polynomials at a Tuple of Bounded Operators §word-products, and .
We first show
If , then by Words over a Finite Alphabet: the Empty Word, Concatenation and Reversal §concatenation, and . Let have length . By Words over a Finite Alphabet: the Empty Word, Concatenation and Reversal §concatenation, has length , its restriction to is , and its value at is . Let be the map of Evaluation of Noncommutative Polynomials at a Tuple of Bounded Operators §word-products for the word . As by claim 3 of Basic Properties of Initial Segments of the Natural Numbers, the restriction of to satisfies the two conditions that define the map for , so it is that map by the uniqueness in Existence and Uniqueness of Iterates of a Binary Operation. Hence and .
Now let . If or , then is or by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §monoid, and because . For nonempty we argue by induction on the length of , using Principle of Induction for the Natural Numbers for the inductive set of those such that for every and every word of length . A word of length is a letter by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter, so by (R). Let and let have length . By Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter, with of length , and by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §monoid. By (R), , associativity of composition and (R) again,
so . Hence ; as every nonempty word has a length, for all .
Clause 2. Fix and let and for . Both maps are linear: is linear by Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §algebra, evaluation is linear, and is linear because composition distributes over sums and commutes with scalar multiples. For , Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §monomials and clause 1 give . By the uniqueness principle,
Now fix and let and for . Both are linear, by Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §algebra, the linearity of evaluation and the linearity of ; and for every . The uniqueness principle gives , that is, .
Clause 3. Assume that every is self-adjoint. As is a Hilbert space, every has an adjoint by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint, and adjoints are unique by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-unique; so each rule of Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus used below is an equality between adjoints. By that clause each is an adjoint of itself, so .
We first show for every . For this reads , which holds by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus, since by Words over a Finite Alphabet: the Empty Word, Concatenation and Reversal §reversal. For nonempty words we argue by induction on the length, using Principle of Induction for the Natural Numbers for the inductive set of those for which the identity holds for every word of length . For a letter, by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §reversal and , so . Let and let have length ; write with of length , by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter. By clause 1, the rule of Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus, , , clause 1 again and Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §reversal,
so . Hence , and the identity holds for every word.
Next let for . By Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §adjoint, the linearity of evaluation and Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus, for and ,
so is linear. By Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §adjoint, the identity just proved and Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §reversal, for every . By the uniqueness principle, for every . By Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus, is the adjoint of , that is, . Finally, let . Then by The Algebra of Noncommutative Polynomials in Finitely Many Self-Adjoint Variables §self-adjoint, so ; thus is an adjoint of itself, and it is self-adjoint by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus.
Clause 4. Let and for . The map is linear as the composite of the linear map of Substitution of Noncommutative Polynomials into the Variables §substitution with evaluation at , and is evaluation at . By Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §values, , while . By the uniqueness principle it suffices to show for every . For , by Substitution of Noncommutative Polynomials into the Variables §word-products, and by clause 1 and Evaluation of Noncommutative Polynomials at a Tuple of Bounded Operators §word-products. For nonempty words we argue by induction on the length, using Principle of Induction for the Natural Numbers for the inductive set of those for which the identity holds for every word of length . For a letter, by Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §values, and by clause 1 for the tuple ; so . Let and let have length , with of length (Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter). By Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §values, clause 2 for , , and clause 1 for ,
so . Hence , and .
Clause 5. Write for the -tuple in ; clause 1 applies to it with in place of . We first show that and for every . For , and . For nonempty words we argue by induction on the length, using Principle of Induction for the Natural Numbers for the inductive set of those for which both statements hold for every word of length . For a letter, and by clause 1 for and for , so . Let and let have length , with of length (Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter). Then by clause 1 and closure under composition, and by the multiplicativity of , , and clause 1 for . So , and .
Next, sums in and are the finite sums of Finite Sum Notation in a Vector Space. We show by induction on , using Principle of Induction for the Natural Numbers for the inductive set of those such that for every map the sum lies in and , that . By claim 1 of Properties of Finite Sums of Vectors the two sums for are and , so . Let , let , and let be the restriction of to , which is contained in and contains by claims 1 and 3 of Basic Properties of Initial Segments of the Natural Numbers. By claim 1 of Properties of Finite Sums of Vectors, applied in and in ,
Since , closure of under sums and additivity of give . Hence .
Now let . If , then and by Linear Maps from Noncommutative Polynomials into Bounded Operators: Existence and Uniqueness from Values on Monomials §formula, , and . If , let and be as in Linear Maps from Noncommutative Polynomials into Bounded Operators: Existence and Uniqueness from Values on Monomials §formula. That clause, applied to evaluation at (the linear map with values ) and to evaluation at (with , on ), gives
The terms lie in by the first step and closure under multiplication by complex numbers, and by homogeneity of and the first step. Since , we get and .
Clause 6. Since , there is a real with , by Noncommutative Laws of Finitely Many Self-Adjoint Variables with a Norm Bound §law; so Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation applies with . The space is a complex Hilbert space by The Complex GNS Space of a Tracial State on Noncommutative Polynomials §gns, and for every by Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §multiplication. Write for the -tuple . The maps and from to are linear, the second by Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §algebra. By the uniqueness principle on it suffices to show for every . For , by Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §algebra. For nonempty words we argue by induction on the length, using Principle of Induction for the Natural Numbers for the inductive set of those for which the identity holds for every word of length . For a letter, by The Algebra of Noncommutative Polynomials in Finitely Many Self-Adjoint Variables §monomials and clause 1 for , so . Let and let have length , with of length (Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter). Then by Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §monomials, and by Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §algebra, , and clause 1 for ,
so . Hence , and for every .
Clause 7. Every operator norm is nonnegative: is a bound for by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §least-bound, and bounds are nonnegative by Bounded Linear Maps between Complex Normed Spaces and the Operator Norm §bounded. We argue by induction on , using Principle of Induction for the Natural Numbers for the inductive set of those with for every word of length . A word of length is a letter by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter, and by clause 1 and claim 1 of Properties of Natural Number Powers in a Field; so . Let and let have length ; write with of length , by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter. By clause 1 and Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §operations,
The second inequality is claim 5 of Elementary Arithmetic in an Ordered Field with the nonnegative factor ; the third is the same claim with the nonnegative factor , applied to (which holds as ), together with commutativity of multiplication; the final equality is claim 1 of Properties of Natural Number Powers in a Field. Hence , so , which is clause 7.
Loading…
Prerequisites
01001f81-2d16-436c-906a-fb27498c14af