TheoremBase

A spanning orthonormal tuple would make the sum of the squared coordinates of each member of the orthonormal sequence equal to one, while Bessel's inequality forces each of those finitely many coordinates to tend to zero.

Proof

Each result cited is universally quantified over the data in its own statement. For s∈Rs\in\mathbb{R} we write s2=sss^{2}=ss, and ∣s∣|s| is the absolute value of ss.

Suppose, for a contradiction, that HH is finite-dimensional.

Step 1 (the space is not the zero space). Since (fj)j∈N(f_{j})_{j\in\mathbb{N}} is orthonormal, ⟨f1,f1⟩=1\langle f_{1},f_{1}\rangle=1 by Orthogonality, Orthogonal Complement and Orthonormal Families in a Real Inner Product Space (iii), that is, ∣f1∣2=1|f_{1}|^{2}=1 by Real Inner Product Spaces (ii). Were f1=0Hf_{1}=0_{H}, we should have ∣f1∣=0|f_{1}|=0 by Elementary Identities in a Real Inner Product Space (iii) and hence ∣f1∣2=0|f_{1}|^{2}=0, whereas 0β‰ 10\ne1 by claim 6 of Elementary Order Arithmetic in an Ordered Field. So f1β‰ 0Hf_{1}\ne0_{H}, and therefore Hβ‰ {0H}H\ne\{0_{H}\}.

Step 2 (a spanning orthonormal tuple). The set HH is a linear subspace of HH: it contains 0H0_{H} and is closed under the addition and scalar multiplication of HH. It is finite-dimensional by assumption and differs from {0H}\{0_{H}\} by step 1, so claim 2 of Gram-Schmidt Orthonormalisation in a Real Inner Product Space, applied with the inner product space HH in the role of EE and with L=HL=H, provides a natural number mm and an orthonormal mm-tuple e∈Hme\in H^{m}, with components e1,…,eme_{1},\dots,e_{m}, such that span⁑(e)=H\operatorname{span}(e)=H.

Step 3 (each member of the sequence has unit coordinate sum). Let P:Hβ†’HP:H\to H be the map of Projection onto the Span of an Orthonormal Tuple, and Coordinates on a Finite-Dimensional Subspace formed with this ee, namely Px=βˆ‘i=1m⟨x,ei⟩eiPx=\sum_{i=1}^{m}\langle x,e_{i}\rangle e_{i}, and let M=span⁑(e)M=\operatorname{span}(e), so that M=HM=H by step 2. By Projection onto the Span of an Orthonormal Tuple, and Coordinates on a Finite-Dimensional Subspace (i) we have Px=xPx=x for every x∈Mx\in M, hence for every x∈Hx\in H; and by Projection onto the Span of an Orthonormal Tuple, and Coordinates on a Finite-Dimensional Subspace (ii), ∣Px∣2=βˆ‘i=1m⟨x,ei⟩2|Px|^{2}=\sum_{i=1}^{m}\langle x,e_{i}\rangle^{2}. Taking x=fjx=f_{j} and using ∣fj∣2=⟨fj,fj⟩=1|f_{j}|^{2}=\langle f_{j},f_{j}\rangle=1, which holds by Orthogonality, Orthogonal Complement and Orthonormal Families in a Real Inner Product Space (iii) and Real Inner Product Spaces (ii),

βˆ‘i=1m⟨fj,ei⟩2=1forΒ everyΒ j∈N.\sum_{i=1}^{m}\langle f_{j},e_{i}\rangle^{2}=1\qquad\text{for every }j\in\mathbb{N}.

Step 4 (each coordinate tends to zero). Let ii be a natural number with 1≀i≀m1\le i\le m. By Orthonormal Expansions in a Real Hilbert Space (i), applied to the orthonormal sequence (fj)j∈N(f_{j})_{j\in\mathbb{N}} and to the vector eie_{i}, the series βˆ‘j=1∞⟨ei,fj⟩2\sum_{j=1}^{\infty}\langle e_{i},f_{j}\rangle^{2} converges; hence the sequence (⟨ei,fj⟩2)j∈N(\langle e_{i},f_{j}\rangle^{2})_{j\in\mathbb{N}} converges to 00 by claim 2 of Elementary Properties of Series of Real Numbers. Since ⟨fj,ei⟩=⟨ei,fj⟩\langle f_{j},e_{i}\rangle=\langle e_{i},f_{j}\rangle by Real Inner Product Spaces (i), the sequence (⟨fj,ei⟩2)j∈N(\langle f_{j},e_{i}\rangle^{2})_{j\in\mathbb{N}} converges to 00 as well.

Step 5 (contradiction). Apply Finite Sums of Real Numbers: Nonnegativity, Domination by the Sum, and Limits (iii) with N=mN=m to the mm sequences of step 4, indexed by ii: the sequence whose jj-th term is βˆ‘i=1m⟨fj,ei⟩2\sum_{i=1}^{m}\langle f_{j},e_{i}\rangle^{2} converges to βˆ‘i=1m0\sum_{i=1}^{m}0, which is 00 by claim 3 of Properties of Finite Sums applied with Ξ»=0\lambda=0. By step 3 that sequence is constant with value 11. Taking Ξ΅=1\varepsilon=1 in Limit of a Sequence of Real Numbers there is accordingly some index at which ∣1βˆ’0∣<1|1-0|<1; but 0≀10\le1 by claim 1 of Elementary Arithmetic in an Ordered Field, so ∣1βˆ’0∣=∣1∣=1|1-0|=|1|=1 by claim 1 of Properties of the Absolute Value in an Ordered Field, and 1<11<1 contradicts the irreflexivity of the strict order of Ordered Field.

Hence the assumption was false, and HH is not finite-dimensional.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…