Chapter 1 established the act of distinction as the primitive operation. Distance is the quantification of distinction: how distinct are two points? How many nested distinctions separate them? This chapter makes that quantification precise through the theory of metric spaces.
2.1 Distance as Quantified Distinction
A metric formalizes the intuitive notion of “how different” two points are. The stronger the distinction between $x$ and $y$, the larger their distance. The identity of indiscernibles — $d(x,y) = 0 \iff x = y$ — is the formal statement that zero distance means no distinction. If two points cannot be distinguished by the metric, they are the same point.
2.2 What Is a Metric?
Definition 2.1 (Metric, Metric space)
A metric on a set $X$ is a function $d: X \times X \to \mathbb{R}_{\geq 0}$ satisfying, for all $x,y,z \in X$:
Identity of indiscernibles: $d(x,y) = 0 \iff x = y$. No distance = no distinction.
Symmetry: $d(x,y) = d(y,x)$. Distinction is mutual.
Triangle inequality: $d(x,z) \leq d(x,y) + d(y,z)$. The distinction between $x$ and $z$ cannot exceed the sum of the distinctions via $y$.
$(X,d)$ is a metric space.
Example 2.2 (Euclidean metric)
On $\mathbb{R}$: $d(x,y)=|x-y|$. On $\mathbb{R}^n$: $d(\mathbf{x},\mathbf{y})=\sqrt{\sum (x_i-y_i)^2}$. This is the Archimedean metric — distinctions add linearly.
Example 2.3 (Discrete metric)
$d(x,y) = 0$ if $x=y$, $1$ otherwise. Every distinct point is equally distinct from every other. No gradations — just "same" or "different." This is the crudest possible metric, and it satisfies a property (the ultrametric inequality) that will become central in Chapter 3.
2.3 Open Balls and Topology
Definition 2.4 (Open ball)
$B(x,r) = \{y \in X \mid d(x,y) < r\}$. The set of points whose distinction from $x$ is less than threshold $r$.
In $\mathbb{R}$, $B(0,1)=(-1,1)$. In $\mathbb{R}^2$ with Manhattan metric, $B(\mathbf{0},1)$ is a diamond. The shape of balls depends on the metric — on how distinctions are structured.
Definition 2.5 (Open set, Topology)
$U \subseteq X$ is open if $\forall x \in U, \exists r > 0: B(x,r) \subseteq U$. The collection of all open sets is the topology — the structure of which distinctions are "close enough" to blur together.
2.4 Convergence and Completeness
Definition 2.6 (Convergent sequence)
$x_n \to x$ if $\forall \varepsilon > 0, \exists N, \forall n \geq N: d(x_n,x) < \varepsilon$. A sequence converges if its distinctions from the limit vanish.
Definition 2.7 (Cauchy sequence)
$(x_n)$ is Cauchy if $\forall \varepsilon > 0, \exists N, \forall m,n \geq N: d(x_m,x_n) < \varepsilon$. Terms eventually cannot be distinguished from each other at scale $\varepsilon$.
Definition 2.8 (Complete metric space)
A space is complete if every Cauchy sequence converges. Every sequence whose internal distinctions vanish actually approaches a point in the space. Gaps are filled.
Example 2.9
$\mathbb{Q}$ is not complete with the Archimedean metric. The sequence $1, 1.4, 1.41, 1.414, \ldots$ (converging to $\sqrt{2}$) is Cauchy in $\mathbb{Q}$ but has no rational limit. The distinction between successive terms vanishes, but the limit lies outside $\mathbb{Q}$ — a gap in the rational distinction-space.
Theorem 2.10 (Metric completion)
Every metric space has a unique completion. The completion of $\mathbb{Q}$ w.r.t. $d_\infty(x,y)=|x-y|_\infty$ is $\mathbb{R}$, the real numbers. But — and this is Ostrowski's revelation (Chapter 4) — there are infinitely many OTHER completions of $\mathbb{Q}$, each corresponding to a different way of measuring distinction: the $p$-adic numbers.
2.5 The Metric from Absolute Values
Any absolute value gives a metric: $d(x,y)=
x-y
$. The Archimedean absolute value $
\cdot
_\infty$ gives the familiar Euclidean metric. But as we previewed in Chapter 1, the $p$-adic absolute values $
\cdot
_p$ give radically different metrics — ultrametrics — where distinctions nest rather than add.
2.6 The Nature of the Metric Enterprise
Every metric embodies a theory of distinction. The Euclidean metric says distinctions add linearly — a step plus a step equals two steps. The discrete metric says distinctions are binary — same or different, no middle ground. The $p$-adic ultrametric says distinctions nest hierarchically — a distinction at a coarse scale contains finer distinctions within it.
Metric type
Distinction logic
Geometry
Euclidean
Additive distinctions
Continuous, connected
Discrete
Binary distinctions
Totally disconnected
$p$-adic
Nested distinctions
Hierarchical, tree-like
The rest of this work explores the third column — and argues it is the column that describes nature.