Orthogonal Set is Linearly Independent Set
Theorem
Let $\struct {V, \innerprod \cdot \cdot}$ be an inner product space over a real or complex subfield $\mathbb F$.
Let $S \subseteq V$ be an orthogonal set.
Let $\mathbf 0 \notin S$, where $\mathbf 0$ denotes the zero vector of $V$.
Then $S$ is a linearly independent set.
Proof
Let $n \in \N$.
Let $\lambda_1, \ldots, \lambda_n \in \mathbb F$, and let $u_1, \ldots, u_n \in S$ such that:
- $\ds \sum_{i \mathop= 1}^n \lambda_i u_1 = \mathbf 0$
To prove that $S$ is a linearly independent set, we must show that $\lambda_k = 0$ for all $k \in \set { 1, \ldots, n}$.
We calculate:
\(\ds \mathbf 0\) | \(=\) | \(\ds \ds \sum_{i \mathop= 1}^n \lambda_i u_1\) | ||||||||||||
\(\ds \leadsto \ \ \) | \(\ds \innerprod {\mathbf 0}{u_k}\) | \(=\) | \(\ds \ds \innerprod {\sum_{i \mathop= 1}^n \lambda_i u_i }{u_k}\) | take the inner product with $u_k$ on both sides | ||||||||||
\(\ds \leadsto \ \ \) | \(\ds 0\) | \(=\) | \(\ds \ds \innerprod {\sum_{i \mathop= 1}^n \lambda_i u_i }{u_k}\) | Inner Product with Zero Vector | ||||||||||
\(\ds \) | \(=\) | \(\ds \ds \sum_{i \mathop= 1}^n \lambda_i \innerprod {u_i}{u_k}\) | Inner Product Axiom $\text N 2$: Linearity in first argument | |||||||||||
\(\ds \) | \(=\) | \(\ds \lambda_k \innerprod {u_k}{u_k}\) | as $\innerprod {u_i}{u_k} = 0$ for $i \ne k$ |
As $u_k \ne \mathbf 0$ by assumption, it follows by Inner Product Axiom $\text N 3$: Non-Negative Definiteness that $\innerprod {u_k}{u_k} \ne 0$.
From Field has no Proper Zero Divisors, it follows that $0 = \lambda_k$ for all $k \in \set {1, \ldots, n}$.
$\blacksquare$
Sources
- 1994: Robert Messer: Linear Algebra: Gateway to Mathematics: $\S 4.4$