The Deep Structure — Two Ways of Measuring

p-adic Numbers

The tree distance $d_T(v,w) = b^{-k}$ is not an arbitrary construction. It is the natural distance on the $p$-adic numbers — one of the most important structures in number theory.

Definition (p-adic order). For a prime $p$, factor any rational $x \neq 0$ as $x = p^k \cdot \frac{a}{b}$, where $a, b$ are not divisible by $p$. Then $\text{ord}_p(x) = k$. Examples:

$\text{ord}_2(8) = 3$ (since $8 = 2^3$)
$\text{ord}_3(12) = 1$ (since $12 = 2^2 \cdot 3$)
$\text{ord}_2(1/2) = -1$
$\text{ord}_p(0) = \infty$ (convention)

Definition (p-adic absolute value). $|x|_p = p^{-\text{ord}_p(x)}$, with $|0|_p = 0$.

This is identical in form to the tree distance. Replace $b$ with $p$, and the common prefix length $k$ with the $p$-adic order of the difference: $d_p(x, y) = |x - y|_p$. Two numbers are $p$-adically close if their difference is divisible by a large power of $p$ — their base-$p$ expansions agree for many digits.

Key Properties of $|\cdot|_p$

$|x|_p \geq 0$, with equality only for $x = 0$
$|xy|_p = |x|_p \cdot |y|_p$ (multiplicativity)
$|x + y|_p \leq \max(|x|_p, |y|_p)$ — the strong triangle inequality!

The $p$-adic numbers $\mathbb{Q}_p$ are the completion of $\mathbb{Q}$ under $|\cdot|_p$, just as $\mathbb{R}$ is the completion under the ordinary absolute value. They form an ultrametric field — the algebraic counterpart of the tree $T_p$.

Ostrowski's Theorem

Ostrowski's Theorem (1916)

Every non-trivial absolute value on $\mathbb{Q}$ is equivalent either to:

The ordinary (Archimedean) absolute value $|\cdot|_\infty$, or
A $p$-adic absolute value $|\cdot|_p$ for some prime $p$.

There is no third way.

Any coherent notion of distance on the rational numbers — one that respects addition and multiplication — must be either Archimedean (continuous, additive) or ultrametric (hierarchical, maximum-based). The two ways of measuring are forced by the structure of number itself.

The p-adic Interpretation of the Threshold Principle

In $T_p$, the ball of radius $p^{-D}$ is the encoding cluster. Perturbations of $p$-adic size $\delta < p^{-D}$ cannot move the system out of this cluster. Perturbations with $\delta \geq p^{-D}$ can change the logical state — but must cross the corresponding energy barrier.

The physical energy barrier and the $p$-adic cluster radius are related:

$$\Delta E_D = E_0 \cdot p^{-\alpha D} = E_0 \cdot (p^{-D})^{\alpha}$$

In $p$-adic language: noise with $p$-adic size less than $p^{-D}$ is irrelevant to logical information encoded at depth $D$.

The Convergence of Independent Lines

Four independent lines of evidence, from unrelated fields, converge on a single statement:

Line 1: Metric Space Theory

The strong triangle inequality produces no-accumulation, isosceles triangles, and nested disjoint clusters. Ultrametric spaces form a distinct category with radically different geometry.

Line 2: Number Theory (Ostrowski)

The only absolute values on $\mathbb{Q}$ are Archimedean and $p$-adic. Ultrametric geometry is forced by algebraic consistency — not an arbitrary invention.

Line 3: Statistical Mechanics

Barrier crossing obeys the Arrhenius law: rate $\propto e^{-\Delta E / k_B T}$. Hierarchical barriers → exponential error suppression. The landscape must be ultrametric.

Line 4: Information Theory

Error correction cost: polynomial in Archimedean spaces, logarithmic in ultrametric spaces. Geometry directly determines information-theoretic efficiency.

To build a memory that endures without active correction, engineer the state space to be ultrametric.