Metric Defines Norm iff it Preserves Linear Structure
From ProofWiki
Theorem
Let $V$ be a vector space over a valued field $k$.
Let $d : V \times V \to k$ be a metric on $V$.
Then the function $\| v \| := d(v,0)$ is a norm on $V$ if and only if for all $x,y,z \in V$, $\lambda \in k$:
- $(1): \quad d(x+z,y+z) = d(x,y)$ (homogeneity or translation invariance)
- $(2): \quad d(\lambda x, \lambda y) = | \lambda | d(x,y)$ (the enlargement property)
Proof
Suppose first that $d$ satisfies the hypotheses 1. and 2..
Since $d(u,v) \geq 0$ for all $u,v \in V$, $\| u \| = d(u,0) \geq 0$ for all $u \in V$.
Moreover if $\| u \| = 0$ then $d(u,0) = 0$, so $u = 0$.
Now let $\lambda \in K$, $u \in V$.
Then, using the enlargement property of $d$:
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \Vert \lambda v \Vert\) | \(=\) | \(\displaystyle d(\lambda v, 0 )\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the definition of $\vert \cdot \vert$ | ||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(=\) | \(\displaystyle d(\lambda v, \lambda \cdot 0)\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | |||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(=\) | \(\displaystyle \vert\lambda\vert d(v,0)\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the enlargement property | ||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(=\) | \(\displaystyle \vert\lambda\vert \Vert v \Vert\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the definition of $\Vert \cdot \Vert$ |
Finally if $u,v \in V$, then we have
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \Vert u + v \Vert\) | \(=\) | \(\displaystyle d(u + v, 0)\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the definition of $\Vert \cdot \Vert$ | ||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(=\) | \(\displaystyle d(u, -v)\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By translation invariance of $d$ | ||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\leq\) | \(\displaystyle d(u, 0) + d(0,-v)\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the triangle inequality for metrics. | ||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(=\) | \(\displaystyle d(u, 0) + d(v,0)\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the enlargement property | ||
| \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | \(=\) | \(\displaystyle \Vert u \Vert + \Vert v \Vert\) | \(\displaystyle \) | \(\displaystyle \) | \(\displaystyle \) | By the definition of $\Vert \cdot \Vert$ |
$\blacksquare$