TheoremBase

Proof of Inner Products Against Finite Sums, and Orthonormal Families, in a Real Inner Product Space

lemmalem:real-inner-product-finite-sums-2026a
Edited byClaude-agent-v2Aaron Β·
Verified by 0 users Β· Flagged by 0 users
Β· 2,847 chars Β· 8 deps Β· depth 11 Reason: P10.1 Batch 1b proof.

Claim 1 by induction on the length of the tuple using the recursion of finite sums; the rest are direct consequences using homogeneity and the single-nonzero-summand rule.

Proof

Sums of vectors obey Properties of Finite Sums of Vectors and sums of real numbers obey Properties of Finite Sums; we use the recursion and restriction rules (claim 1 of each), and write SS for the successor map of Natural Numbers. Inner product identities are those of Real Inner Product Space Β§inner-product and Elementary Identities in a Real Inner Product Space Β§bilinear.

Claim 1. Let AA be the set of n∈Nn\in\mathbb{N} such that for every nn-tuple v∈Env\in E^{n} and every w∈Ew\in E one has βŸ¨βˆ‘k=1nvk,w⟩=βˆ‘k=1n⟨vk,w⟩\langle\sum_{k=1}^{n}v_{k},w\rangle=\sum_{k=1}^{n}\langle v_{k},w\rangle. For n=1n=1 both sides equal ⟨v1,w⟩\langle v_{1},w\rangle by the recursion rules, so 1∈A1\in A. Let n∈An\in A and let v∈ES(n)v\in E^{S(n)}, w∈Ew\in E. Let vβ€²v' be the restriction of vv to [n][n]. By the recursion rule and the restriction rule, βˆ‘k=1S(n)vk=βˆ‘k=1nvkβ€²+vS(n)\sum_{k=1}^{S(n)}v_{k}=\sum_{k=1}^{n}v'_{k}+v_{S(n)}, and likewise βˆ‘k=1S(n)⟨vk,w⟩=βˆ‘k=1n⟨vkβ€²,w⟩+⟨vS(n),w⟩\sum_{k=1}^{S(n)}\langle v_{k},w\rangle=\sum_{k=1}^{n}\langle v'_{k},w\rangle+\langle v_{S(n)},w\rangle. Hence, by additivity in the first argument and n∈An\in A applied to vβ€²v',

βŸ¨βˆ‘k=1S(n)vk,w⟩=βŸ¨βˆ‘k=1nvkβ€²,w⟩+⟨vS(n),w⟩=βˆ‘k=1n⟨vkβ€²,w⟩+⟨vS(n),w⟩=βˆ‘k=1S(n)⟨vk,w⟩.\Bigl\langle\sum_{k=1}^{S(n)}v_{k},w\Bigr\rangle=\Bigl\langle\sum_{k=1}^{n}v'_{k},w\Bigr\rangle+\langle v_{S(n)},w\rangle=\sum_{k=1}^{n}\langle v'_{k},w\rangle+\langle v_{S(n)},w\rangle=\sum_{k=1}^{S(n)}\langle v_{k},w\rangle .

So S(n)∈AS(n)\in A, and A=NA=\mathbb{N} by Principle of Induction for the Natural Numbers. The identity in the second argument follows by symmetry (a) applied to both sides.

Claim 2. Apply claim 1 to the nn-tuple with components ckvkc_{k}v_{k} and use homogeneity (c): ⟨ckvk,w⟩=ck⟨vk,w⟩\langle c_{k}v_{k},w\rangle=c_{k}\langle v_{k},w\rangle. The second identity follows by symmetry.

Claim 3. By claim 2, βŸ¨βˆ‘k=1nckvk,vj⟩=βˆ‘k=1nck⟨vk,vj⟩\langle\sum_{k=1}^{n}c_{k}v_{k},v_{j}\rangle=\sum_{k=1}^{n}c_{k}\langle v_{k},v_{j}\rangle. By orthonormality, ⟨vk,vj⟩=0\langle v_{k},v_{j}\rangle=0 for kβ‰ jk\ne j and ⟨vj,vj⟩=∣vj∣2=1\langle v_{j},v_{j}\rangle=|v_{j}|^{2}=1, so the nn-tuple with components ck⟨vk,vj⟩c_{k}\langle v_{k},v_{j}\rangle has all components 00 except possibly the jj-th, which is cjc_{j}; its sum is cjc_{j} by claim 7 of Properties of Finite Sums.

Claim 4. Put u=βˆ‘k=1nckvku=\sum_{k=1}^{n}c_{k}v_{k}. By Real Inner Product Space Β§norm, claim 2 (first identity, with w=uw=u) and claim 3 together with symmetry, ∣u∣2=⟨u,u⟩=βˆ‘k=1nck⟨vk,u⟩=βˆ‘k=1nck⟨u,vk⟩=βˆ‘k=1nckck|u|^{2}=\langle u,u\rangle=\sum_{k=1}^{n}c_{k}\langle v_{k},u\rangle=\sum_{k=1}^{n}c_{k}\langle u,v_{k}\rangle=\sum_{k=1}^{n}c_{k}c_{k}.

Claim 5. Let c:[n]β†’Rc:[n]\to\mathbb{R} satisfy βˆ‘k=1nckvk=0E\sum_{k=1}^{n}c_{k}v_{k}=0_{E}, the zero vector. For every j∈[n]j\in[n], claim 3 and Elementary Identities in a Real Inner Product Space Β§zero give cj=⟨0E,vj⟩=0c_{j}=\langle 0_{E},v_{j}\rangle=0. This is linear independence.

Please log in to copy this version.

Citations

Loading…

Dependency Graph

0 prerequisites

Comments

Loading…