Geometric Ratios as Generators of Hierarchical Structure: An Arithmetic Audit of π, φ, and e

In the paper's words

A recurring conjecture in the ultrametric-physics program holds that hierarchical—i.e., ultrametric—structure "emerges" when a scale ratio q is set equal to a fundamental geometric constant: π, the golden ratio φ = (1+√5)/2, or Euler's number e.

The abstract, verbatim, and its opening sentence is the claim. Every line below is quoted from the text.

Numbers it reports

1. Entropy: $H = -(\varphi ^{-1} \ln \varphi ^{-1} + \varphi ^{-2} \ln \varphi ^{-2}) = \ln \varphi \cdot (\varphi ^{-1} + 2\varphi ^{-2}) = 0.481212 \times (0.618034 + 0.763932) = 0.481212 \times 1.381966$ = 0.66502 nats , which is 0.66502/0.69315 = 95.9% of the binary maximum ln 2.

   from 4.4 Test D: self-similar ultrametric constructions

2. Self-similarity: $\varphi$ gapless ($1/\varphi + 1/\varphi ^{2} = 1$) with entropy 0.66502 nats (95.9% of binary max) and Cantor dimension $\ln 2/\ln \varphi ^{2} \approx 0.72024$; e gives dimension $\ln 2 \approx 0.69315$ at ratio 1/e but no exact tiling; $\pi$ gapped, dimension 0.30275 at $r = 1/\pi ^{2}$.

   from 5. Results

3. Using $\varphi ^{n} = F_{n} \varphi + F_{n-1}$: $\varphi ^{3} = 2\varphi + 1 = 4.236068, \varphi ^{-3} = 0.236068$, so [$3]_{\varphi } = 4.236068 - 0.236068$ = 4 exactly. $\varphi ^{4} = 3\varphi + 2 = 6.854102, \varphi ^{-4} = 0.145898, [4]_{\varphi } = 6.708204 = 3\sqrt{5}$ (3 × $2.236068 = 6.708204$ ✓).

   from 4.1 Test A at $q = \varphi$: exact algebraic structure

4. By Hurwitz's theorem the constant $1/\sqrt{5}$ is optimal and attained exactly by numbers equivalent to $\varphi$: $\varphi$ is the real number most resistant to rational approximation. $\pi 's$ profile is dominated by the single anomalously deep partial quotient 292 (its convergent 355/113 has error ≈ 2.7×10^(−7) at denominator only 113)—an unbalanced hierarchy with one dominant coarse split. e's profile is regular but unbounded (linear growth of every third partial quotient), an inexact hierarchy: patterned, but generated by no finite algebraic rule.

   from 4.3 Test B: continued-fraction hierarchies

5. The question audited here, inherited from the ULTRAMETRIC PHYSICS compilation [12], is deceptively simple: when a scale ratio q is one of the "geometric" constants $\pi \approx 3.14159, \varphi \approx 1.61803$, or $e \approx 2.71828$, does hierarchical structure emerge in a way that it does not for generic q?

   from 1. Introduction

6. The fine-structure-constant work [11], which reframes $\alpha$ as a cross-ratio of measurable electron length scales with explicit falsifiability conditions, is the methodological template.

   from 1. Introduction

How it is done

1. The embedding–extension–interpolation paper [3] proves ultrametric versions of the Arens–Eells isometric embedding theorem, the Hausdorff extension theorem, and the Niemytzki–Tychonoff compactness characterization, establishing that ultrametric spaces—zero-dimensional analogues of metric spaces—support the same structural infrastructure as ordinary metric theory.

   from 2. Background and Related Work

2. Throughout, $\varphi = (1+\sqrt{5})/2$ = 1.61803399…, e = 2.71828183…, $\pi$ = 3.14159265…; auxiliary constants $\ln 2 = 0.69314718, \ln \varphi = 0.48121183$.

   from 3. Methods

3. $e - e^{-1} = 2.718282 - 0.367879 = 2.350402$; [$2]_{e} = 3.086161$; [$3]_{e} = (20.085537 - 0.049787)/2.350402 = 20.035750/2.350402$ ≈ 8.524391 . $\pi - \pi ^{-1} = 3.141593 - 0.318310 = 2.823283$; [$2]_{\pi } = 3.459903$; [$3]_{\pi } = (31.006277 - 0.032252)/2.823283 = 30.974025/2.823283$ ≈ 10.970926 . e is transcendental (Hermite) and $\pi$ is transcendental (Lindemann); by Lindemann–Weierstrass, $e^{n}$ is transcendental for every nonzero algebraic n, and the same theorem applied to $e^{i\pi } = -1$ forces $\pi$ transcendental, hence $\pi ^{n}$ transcendental for each $n \geq 1$.

   from 4.2 Test A at $q = e$ and $q = \pi$: failure by transcendence

4. By Hurwitz's theorem the constant $1/\sqrt{5}$ is optimal and attained exactly by numbers equivalent to $\varphi$: $\varphi$ is the real number most resistant to rational approximation. $\pi 's$ profile is dominated by the single anomalously deep partial quotient 292 (its convergent 355/113 has error ≈ 2.7×10^(−7) at denominator only 113)—an unbalanced hierarchy with one dominant coarse split. e's profile is regular but unbounded (linear growth of every third partial quotient), an inexact hierarchy: patterned, but generated by no finite algebraic rule.

   from 4.3 Test B: continued-fraction hierarchies

What would falsify it

1. Explicit falsification conditions are stated.

   from Abstract

2. The fine-structure-constant work [11], which reframes $\alpha$ as a cross-ratio of measurable electron length scales with explicit falsifiability conditions, is the methodological template.

   from 1. Introduction

3. The fine-structure work [11] shows how a constant can be reframed as a geometric object with explicit falsifiability conditions; we adopt the same discipline for $\varphi$, whose defining quadratic relation is the exact algebraic self-similarity driving our strongest result.

   from 2. Background and Related Work

Where the claim stops

1. We therefore treat Test E as complementary: it refines, but does not replace, the algebraic verdict.

   from 6. Discussion

2. (4) That nature preferentially selects $\varphi$ as a design parameter is not established by any cited work; our result establishes an arithmetic distinction among the candidates, not a physical mechanism.

   from 6. Discussion

3. One headline claim was retracted in the course of the audit: the candidate Fibonacci ultrametric d(m,n) = 2^(−gcd(m,n)) fails the strong triangle inequality, so the divisibility tree is downgraded to a heuristic hierarchy, with the valuation-based construction $d(m,n) = 2^(-v_{p}(\gcd(m,n)))$ as the correct ultrametric alternative.

   from 7. Conclusion

Where to read

1. Abstract396 words · 2 min
2. 1. Introduction388 words · 2 min
3. 2. Background and Related Work643 words · 3 min
4. 3. Methods392 words · 2 min
5. 4. Analysis1,042 words · 5 min
6. 5. Results290 words · 1 min
7. 6. Discussion520 words · 2 min
8. 7. Conclusion273 words · 1 min