Chapter 5: The Bruhat-Tits Tree

The $p$-adic numbers, as constructed in Chapter 4, are an algebraic object — a field complete with respect to an ultrametric absolute value. But algebra alone does not reveal the full geometric structure. The Bruhat-Tits tree $T_p$ is the geometric realization of $p$-adic space: an infinite, regular tree that makes the hierarchical, ultrametric organization of $\mathbb{Q}_p$ visible and computable.

This is the chapter where the Spencer-Brown distinction tree of Chapter 1 becomes a rigorous mathematical object. The Bruhat-Tits tree is not merely analogous to a hierarchy of distinctions — it IS a hierarchy of distinctions, encoded with the full structure of a non-Archimedean local field.

5.1 The Tree as the Geometry of Nested Distinctions

Why represent a field as a tree? The answer lies in the ultrametric property established in Chapter 3. In an ultrametric space, distinctions nest — two balls (distinction-bounded regions) are either disjoint, or one contains the other entirely. This hierarchical nesting is naturally encoded as a rooted tree, where:

Each vertex represents a distinction — a ball in $\mathbb{Q}_p$
Each edge represents a containment relation — one distinction nested inside another
Moving down the tree = moving to finer distinctions (smaller balls, higher $p$-adic precision)
Moving up the tree = coarse-graining (larger balls, coarser distinctions)

**Key Insight: The Tree as Distinction Manifold.** In Euclidean space, a ball of radius $$r$$ has a unique center and its boundary is a fuzzy sphere. In ultrametric space, every point in a ball is a center, and the "boundary" between nested balls has no interior — there is no continuous transition from one distinction-cluster to another. The only way to represent this hierarchical, non-overlapping structure faithfully is as a tree. The tree is the **distinction manifold** — the geometric form of Spencer-Brown's nested marks.

The Bruhat-Tits tree for $\mathbb{Q}_p$ is denoted $T_p$. It is an infinite tree where every vertex is connected to exactly $p + 1$ other vertices — it is $(p+1)$-regular. For $p = 2$, this is a trivalent (3-regular) tree; for $p = 3$, a 4-regular tree; and so on.

5.2 Lattice Construction of $T_p$

The standard construction of the Bruhat-Tits tree uses the language of lattices — an algebraic encoding of nested containment relations.

**Definition 5.1 (Lattices in $$\mathbb{Q}_p^2$$).** A **lattice** $$\Lambda \subset \mathbb{Q}_p^2$$ is a rank-2 $$\mathbb{Z}_p$$-submodule — equivalently, $$\Lambda = \mathbb{Z}_p e_1 \oplus \mathbb{Z}_p e_2$$ for some basis $$\{e_1, e_2\}$$ of $$\mathbb{Q}_p^2$$. Two lattices $$\Lambda_1, \Lambda_2$$ are **equivalent**, written $$\Lambda_1 \sim \Lambda_2$$, if one is a scalar multiple of the other: $$\Lambda_1 = \lambda \Lambda_2$$ for some $$\lambda \in \mathbb{Q}_p^\times$$.

The vertices of the Bruhat-Tits tree are precisely the equivalence classes of lattices under this relation. Each vertex is a “scale class” — a distinction at a particular level of resolution.

To define edges, we need adjacency — the finest possible containment relation. Two lattice classes $[\Lambda_1]$ and $[\Lambda_2]$ are adjacent if there exist representatives $\Lambda_1, \Lambda_2$ such that:

[ p \Lambda_1 \subsetneq \Lambda_2 \subsetneq \Lambda_1. ]

This says: $\Lambda_2$ is nested strictly between $\Lambda_1$ and $p\Lambda_1$ — a single-step refinement of the distinction.

**Theorem 5.2 (Structure of the Bruhat-Tits Tree).** The graph $$T_p$$ is an infinite, $$(p+1)$$-regular tree. Every vertex has exactly $$p + 1$$ neighbors, and there are no cycles — $$T_p$$ is a tree in the strict graph-theoretic sense.

**Proof (Sketch).** Fix a representative lattice $$\Lambda$$. The sublattices $$\Lambda'$$ satisfying $$p\Lambda \subset \Lambda' \subset \Lambda$$ correspond to proper non-trivial subspaces of the $$p$$-dimensional vector space $$\Lambda / p\Lambda \cong \mathbb{F}_p^2$$ over the finite field $$\mathbb{F}_p$$. The number of 1-dimensional subspaces of $$\mathbb{F}_p^2$$ is $$\frac{p^2 - 1}{p - 1} = p + 1$$, each corresponding to a neighbor — a possible refinement of the distinction $$\Lambda$$. That the resulting graph has no cycles follows from the ultrametric structure of the containment relations — any cycle would imply a violation of the ultrametric inequality, which is impossible for nested distinctions. ∎

5.3 Combinatorial Construction

For practical purposes, the lattice construction can be translated into a purely combinatorial description that makes the distinction structure transparent.

**Definition 5.3 (Combinatorial Bruhat-Tits Tree).** The tree $$T_p$$ can be constructed as follows: 1. Start with a **root vertex** $$v_0$$ — the coarsest distinction. 2. Each vertex has $$p + 1$$ children (neighbors, one of which is the parent — the containing distinction). 3. Recursively, for each leaf created, generate $$p$$ new children — finer distinctions nested within. 4. Continue ad infinitum. The resulting infinite graph is $$T_p$$.

The $p+1$ neighbors of any vertex consist of:

1 parent — the distinction that contains this one (coarser scale)
$p$ children — the distinctions nested inside this one (finer scale)

This is exactly the Spencer-Brown structure: a distinction (mark) can contain further distinctions. The tree is the complete form of all possible nested distinctions at prime $p$.

5.4 The Boundary: Where Distinctions Become Infinitely Fine

**Definition 5.4 (Boundary of $$T_p$$).** The **boundary** $$\partial T_p$$ is the set of equivalence classes of infinite geodesic rays starting at a fixed basepoint — infinite paths moving consistently toward finer and finer distinctions, with no termination.

The boundary $\partial T_p$ is homeomorphic to the Cantor set — a totally disconnected, perfect, compact metric space. Its Hausdorff dimension is $\dim_H(\partial T_p) = \log(p) / \log(p) = 1$ (for the standard metric), but with the natural visual metric, it is $\log(p+1)/\log(p) \approx 1$.

**Theorem 5.5 (Boundary = $$\mathbb{P}^1(\mathbb{Q}_p)$$).** $$\partial T_p \cong \mathbb{P}^1(\mathbb{Q}_p) \cong \mathbb{Q}_p \cup \{\infty\}$$. The boundary of the tree IS the p-adic projective line.

Why this matters. The bulk (the tree) is discrete and hierarchical — the realm of quantum distinctions. The boundary is the continuous limit where distinctions become infinitely fine — the realm of classical observers. Measurement (Chapter 6, the Monna map) is the projection from the discrete tree onto its continuous boundary.

5.5 Tree Geometry and Ultrametric Distance

On $T_p$, distance between vertices is measured in edge-counting distance. For vertices $v, w$, $d_T(v,w)$ is the number of edges on the unique geodesic path connecting them. Key property: the projection of tree distance to $\mathbb{Q}_p$ recovers the $p$-adic ultrametric:

For boundary points corresponding to $x, y \in \mathbb{Q}_p$: [ |x-y|p = p^{-d{\text{deep}}(x,y)} ] where $d_{\text{deep}}(x,y)$ is the depth from the root at which the geodesics to $x$ and $y$ diverge — the depth of their lowest common ancestor, i.e., the deepest distinction they share.

5.6 Ratio-Based Generalization: $T_{N,q}$

The Bruhat-Tits tree generalizes beyond prime-indexed trees. Define a tree $T_{N,q}$ with:

Branching number: $N$ (each vertex has $N+1$ neighbors; $N$ children + 1 parent)
Scaling ratio: $q > 1$ (distance between adjacent tree levels is $\log q$)
Depth: $d$ (number of levels from root to deepest vertex)

For $T_p$: $N = p$, $q = p$. For generalized trees: $N$ and $q$ are independent parameters. Possible ratios include:

$q = e$ (natural exponential — continuous dynamics)
$q = \pi$ (circular geometry)
$q = \varphi = (1+\sqrt{5})/2$ (golden ratio — biological scaling)
$q = \alpha^{-1} \approx 137.036$ (fine-structure constant — electromagnetic hierarchy)

Different ratios correspond to different physical domains, each with its own distinction tree.

5.7 Automorphisms: The Symmetries of Distinction

The automorphism group of $T_p$ is $\mathrm{PGL}(2,\mathbb{Q}_p)$ — the projective general linear group over the $p$-adic numbers. These are the symmetries that preserve the distinction structure of the tree.

$\mathrm{PGL}(2,\mathbb{Q}_p)$ acts transitively on vertices, edges, and directed edges of $T_p$, and its action extends to the boundary $\mathbb{P}^1(\mathbb{Q}_p)$ as fractional linear transformations:

[ z \mapsto \frac{az + b}{cz + d}, \quad ad - bc \neq 0. ]

In the continuum limit (large depth), these approximate Lorentz transformations — emergent Lorentz symmetry from tree automorphisms. Violations scale as $\delta c/c \sim q^{-d}$, suppressed by tree depth.

5.8 The Tree as the Arena of Physics

The Bruhat-Tits tree is not merely a mathematical curiosity — it is the fundamental geometric arena for non-Archimedean physics. Every physical concept we will develop has a tree-theoretic interpretation:

Physical concept	Tree interpretation
Spacetime	Vertices and edges of $T_{N,q}$ (Chapter 9)
Quantum state	Wavefunction $\psi: V(T) \to \mathbb{C}$ (Chapter 6)
Momentum	Boundary point $\xi \in \partial T$ (Chapter 6)
Dynamics	Tree automorphisms and path integrals (Chapters 6, 13)
Gauge fields	Edge group elements (Chapter 10)
Gravity	Tree geometry fluctuations (Chapter 12)
Measurement	Boundary projection — Monna map (Chapter 6)
Error correction	Hierarchical nesting (Chapter 13)

The tree is the distinction manifold — the geometric form taken by Spencer-Brown’s mark when iterated across all scales. Physics, in the ultrametric framework, is the study of how distinctions evolve on this tree.

Next: Chapter 6: Ultrametric Quantum Mechanics →

Chapter 5: The Bruhat-Tits Tree