We present Topographical Orthogenetic Architecture (TOA) — a deterministic growth framework for structurally stable construction derived from the dm³ contact-geometric operator chain G = U ∘ F ∘ K ∘ C. Seismically and vibrationally stable architectural forms are characterized as fixed points of G within the Gronwall ball B(ε₀ = 1/3). We show that the growth constant previously stated as g = 33^(1/6) ≈ 1.814 in an earlier formulation is a geometric-mean approximation of the first six n-bonacci constants (φ₁ = 1, φ₂ ≈ 1.618, φ₃ ≈ 1.839, φ₄ ≈ 1.927, φ₅ ≈ 1.966, φ₆ ≈ 1.984), whose product is ≈ 33.5, recovering the original conjecture from first principles. We then present the superior graded-growth law, in which each structural stage uses the corresponding n-bonacci constant, producing multiple phononic band gaps simultaneously — the same Sierpiński carpet mechanism observed in Roman amphitheaters, Borobudur, Hindu stepwells, and the NASA O'Neill space colony designs. Empirical verification from structures surviving 100–11,000 years in active seismic zones is presented. We extend the framework to Martian and Lunar construction by replacing Earth's seismic spectrum with InSight-derived Martian modal data. TOA unifies ancient empirical knowledge and modern contact geometry into a single deterministic architecture algorithm deployable by autonomous robotic systems.
Keywords: contact geometry, seismic metamaterial, phononic crystal, n-bonacci constants, orthogenesis, multiplanetary architecture, dm³ framework, Whitney A₁ fold
An earlier formulation of this work [Grossi, Medium 2026] proposed a growth law for a deterministic architectural system: a hexagonal structure growing in six radial stages, each stage multiplying the previous dimension by g = 33^(1/6) ≈ 1.814. The insight — that building geometry should be internally determined by a mathematical growth law rather than externally imposed by stylistic or functional preference — is correct. However, the specific growth constant was asserted without derivation, and the connection to existing bodies of knowledge in structural mechanics, phononic crystal physics, and the dm³ contact-geometric framework was not made explicit.
This paper provides that grounding. We show that the proposed growth constant g ≈ 1.814 is a geometric-mean approximation of the first six n-bonacci constants that appear naturally in the dm³ operator chain. We then replace this uniform approximation with a graded growth law in which each structural stage is governed by its corresponding n-bonacci constant — a modification that generalizes the original single-ratio scheme into a true Sierpiński hierarchical structure with multiple phononic band gaps.
The motivation for this extension comes from two directions. First, from 1,200 years of empirical evidence: the ancient structures that survive in seismically active zones — Roman amphitheaters, Buddhist stupas, Hindu stepwells, Mayan pyramids — all implement multi-scale fractal geometries with Farey-neighbor spacing between structural levels, not uniform-ratio geometries. Second, from the 1975 NASA Space Settlement Design Study [O'Neill 1977], in which engineers derived from first principles that the optimal space habitat geometry under rotational stress is a rotationally symmetric shell (sphere or cylinder) — the identical geometric family that seismic physics and biological evolution converged on independently.
The paper is organized as follows. Section 2 reviews the dm³ fixed-point framework. Section 3 derives the n-bonacci growth law and shows its relationship to the original g^6 = 33 conjecture. Section 4 presents the TOA growth algorithm in both uniform and graded forms. Section 5 reviews the ancient empirical record as natural experiment. Section 6 presents the orthogenesis argument — why NASA 1975, bee hives, and Roman amphitheaters converge on the same geometry. Section 7 extends to Martian and Lunar construction. Section 8 concludes with the formal algorithmic specification.
The dm³ framework [Grossi 2026, doi:10.5281/zenodo.20682934] defines a contact-geometric operator chain on a contact 3-manifold M:
A structure S is G-stable if it is a fixed point of G applied to the local wave energy E:
The Gronwall radius ε₀ = 1/3 is the maximum perturbation from the fixed-point geometry before resonant energy accumulation begins and structural failure becomes likely. Any structure whose geometry deviates from the fixed-point form by more than 1/3 in normalized coordinates accumulates stress at the deviation point; the accumulation rate is governed by μ_max = −2, meaning the deviation decays at twice the natural rate when inside B(ε₀) and grows when outside it.
The fixed-point geometries are not unique — they form a family parameterized by the local wave spectrum. For isotropic wave spectra (equal energy in all directions), the fixed-point family consists of rotationally symmetric surfaces: spheres, cylinders, ellipsoids, and their fractal hierarchies. This is the mathematical reason that seismic stability, structural efficiency, and biological optimality all converge on the same geometric family.
The dm³ framework introduces a sequence of constants — the n-bonacci constants — that appear as the eigenvalue ratios of G at each level of structural hierarchy. The n-th n-bonacci constant φₙ is the unique positive real root greater than 1 of:
These constants are the optimal growth ratios between successive structural levels in a fractal hierarchy. A structure whose n-th level has linear dimension proportional to φₙ relative to the (n−1)-th level creates a phononic band gap at the n-th octave of the local wave spectrum — preventing resonant energy accumulation at that frequency band.
The original formulation proposed a uniform growth constant g = 33^(1/6) ≈ 1.814 for a 6-stage hexagonal structure. We now show that this value is recoverable from the dm³ framework as a geometric-mean approximation, and that it is superseded by the graded growth law.
A uniform-ratio structure with N stages and ratio g has final dimension g^N. In the graded framework, the final dimension after N stages is the product of all stage ratios:
The product of φ₁ through φ₆ is ≈ 22.4 when the first (trivial) stage uses ratio 1. If we begin with a non-trivial seed and index the six stages as φ₂ through φ₇ (where φ₇ ≈ 1.9919, the Heptabonacci constant), the product is:
The original g = 33^(1/6) ≈ 1.814 lies between the φ₂–φ₆ geometric mean (1.867) and the golden ratio φ₂ (1.618), and is close to the Tribonacci constant η ≈ 1.839. The most natural single-constant approximation to the n-bonacci growth ladder for a 6-stage structure is η, giving final dimension η^6 ≈ 34.1 — within 3% of the original conjecture of 33.
The graded law uses the n-bonacci sequence directly, with each stage governed by its corresponding constant. This is not merely an approximation improvement: it changes the qualitative behavior of the structure. A uniform-ratio structure has a single phononic band gap; the graded structure has N distinct band gaps, one per stage, covering N octaves of the wave spectrum simultaneously.
The full TOA algorithm generates a building from four parameters: seed radius r₀, number of stages N, local seismic spectrum S(f), and symmetry group Sym. The output is a complete geometric specification — no stylistic decisions required.
The wall thickness t_k = r_k / 3 = r_k · ε₀ at each stage is not arbitrary: it is the Gronwall radius condition applied to material geometry. A wall thinner than r/3 at stage k allows wave penetration at the k-th band-gap frequency; a wall thicker than r/3 increases material cost without improving stability. The ε₀ = 1/3 constant from the dm³ framework determines the optimal structural aspect ratio.
Connections between stages use golden-ratio elliptical arches (eccentricity e = 1/φ ≈ 0.618). This is the ellipse whose major/minor axis ratio equals φ — the same elliptical form that appears in Roman amphitheaters as the wave-routing perimeter. The arch crown at q* = 1 implements the Whitney A₁ fold condition: seismic energy arriving at the arch is redirected into compression rather than tension, where masonry and concrete perform best.
The TOA framework predicts that structures implementing the graded n-bonacci hierarchy will survive seismic events that destroy uniform-geometry structures. The ancient record provides 1,200 years of data from at least six independent building traditions across three tectonic regimes.
| Structure | Date | Tectonic Zone | TOA Feature Implemented | Status (2026) |
|---|---|---|---|---|
| Colosseum, Rome | ~80 CE | Apennine (high) | Elliptical perimeter (wave cloak) + 3-order column ladder | Standing (partial) |
| Arena di Verona | ~30 CE | Alpine/Apennine | Elliptical perimeter + radial arch network | Standing (complete) |
| Borobudur, Java | ~800 CE | Java subduction (extreme) | 72-stupa Helmholtz array + 3-tier octave spacing | Standing (1,200 yr) |
| Bagan survivors, Myanmar | 9th–13th c. | Sagaing fault (high) | Bell-dome on octagonal Sierpiński base | ~2,000 of 10,000 standing |
| Rani ki Vav, Gujarat | ~1063 CE | Rann of Kutch (extreme) | Farey-sequence terrace proportions + fractal descent | Standing (UNESCO) |
| El Castillo, Chichén Itzá | ~900 CE | Caribbean plate (moderate) | Stepped pyramid = acoustic Bragg grating (frequency sweep) | Standing |
| Göbekli Tepe, Turkey | ~9500 BCE | East Anatolian fault (high) | Circular enclosures + T-pillars (partial implementation) | Standing (11,000 yr) |
In each case, the surviving structure implements multiple TOA features. The structures that collapsed in the same seismic events — documented in regional archaeological records — implemented single-scale geometries: uniform column spacing, flat roofs, square foundations. The earthquake ran a controlled experiment over centuries. The table above is the result.
The survival probability P(S, T) of a structure S over time T in a seismic zone is, to first approximation, determined by the distance d(S, Fix(G)) from S to the nearest G-stable fixed point:
The TOA geometry — rotationally symmetric shell, fractal interior hierarchy, n-bonacci stage ratios — appears in five independent systems that share no common ancestor, no cultural connection, and no shared design methodology. This is orthogenesis: convergent evolution toward a geometric attractor from different initial conditions.
| System | Selection Pressure | Method | Convergent Form | Time Scale |
|---|---|---|---|---|
| Roman amphitheaters | Mediterranean seismicity | Iterative construction + earthquake selection | Ellipse + radial arch + graded column order | ~300 years |
| Buddhist stupa tradition | Asian seismicity + cultural transmission | Iterative construction across 5 traditions | Bell dome on stepped octagonal base | ~1,500 years |
| Bee hive (Apis mellifera) | Gravitational + wind + thermal + predator | Natural selection | Hexagonal cells in spherical outer shell | ~50 million years |
| NASA 1975 Space Settlement Study | Rotational stress + radiation + meteorite | First-principles structural engineering | Sphere (Bernal) + Cylinder (O'Neill) | 10 weeks |
| dm³ framework (this paper) | Abstract contact-geometric stability | Mathematical derivation | Fixed points of G = U∘F∘K∘C in B(ε₀) | Instantaneous |
In the summer of 1975, Gerard O'Neill (Princeton) directed a 10-week study at NASA Ames Research Center with 28 engineers, physicists, and biologists. Their task: design a self-sufficient habitat for 10,000 people in space, from first principles, using only known physics and available materials [O'Neill 1977; NASA SP-413].
The constraints were: artificial gravity via rotation, structural integrity under centrifugal load, radiation shielding, meteorite protection, thermal control, and human habitat requirements. No architectural tradition was consulted. No stylistic preference was imposed.
Three designs emerged. All three were rotationally symmetric shells: the Bernal Sphere (500m diameter), the O'Neill Cylinder (8km length, 3.2km diameter), and a scaled-up cylinder (32km × 6.4km). The engineers had independently derived what seismic physics, biological evolution, and the dm³ framework all predict: the G-stable geometry for any isotropic loading spectrum is a surface of revolution — a sphere or cylinder.
The structural reason is uniform hoop stress. A rotating shell under centrifugal load develops internal tension. In a sphere or cylinder, this stress is identical at every point on the shell — no stress concentrations, no preferred failure axis. In any non-symmetric shape (cube, prism, polygon), stress concentrates at vertices and edges. The sphere and cylinder minimize the maximum stress for a given mean stress — they are the most efficient shells under any isotropic load. This is the same reason the amphitheater ellipse routes seismic energy without concentration, and the bee hive sphere minimizes material per unit volume: all three are fixed points of the relevant loading operator.
Within the O'Neill colony, the NASA engineers specified hexagonal windows and hexagonal internal structural subdivisions. The bee hive uses hexagonal cells. The dm³ hexagonal growth algorithm produces a hexagonal floor plan. The reason is the same in all three cases: the regular hexagon is the unique tiling of the plane that minimizes perimeter per unit area (the honeycomb conjecture, proved by Hales 1999) and has no preferred resonant axis (6-fold symmetry → equal stress distribution in all planar directions). For any planar structure subject to isotropic in-plane loading, the hexagon minimizes material and eliminates stress concentration simultaneously. It is the two-dimensional fixed point.
The TOA framework is not Earth-specific. The fixed-point condition G(E) = E depends on the local wave energy spectrum S(f), which differs between planets. Algorithm 1, Step 3 requires only that the seed radius r₀ be calibrated so that the six stage-band-gaps align with the six dominant frequencies of the local loading spectrum. For Mars and the Moon, this calibration uses available seismic data.
NASA's InSight lander (2018–2022) provided the first direct measurements of Martian seismicity. The dominant marsquake frequencies cluster in two regimes: low-frequency events (0.1–1 Hz) from deep interior modes, and high-frequency scattering (1–10 Hz) from regolith surface waves [Lognonné et al. 2020]. Mars has no plate tectonics but experiences thermal cycling stress (diurnal temperature variation ~100°C) at frequencies of 10⁻⁵ to 10⁻⁴ Hz.
The Martian TOA building is approximately 4× larger than the Earth equivalent for the same structural performance. This is not a disadvantage for space construction: the lower Martian gravity (0.38g) means the increased size carries no weight penalty, and the absence of wind (thin CO₂ atmosphere) removes one loading term. Autonomous construction by swarm robots — a key feature of the original Medium article — is directly applicable.
The Moon has no seismic activity in the conventional sense but experiences meteorite microimpact loading (broadband, 0.01–100 Hz) and thermal cycling stress (lunar day ~14 Earth days, ΔT ~300°C). The dominant loading frequency from thermal cycling is ~8×10⁻⁷ Hz — extremely low frequency. TOA on the Moon should be calibrated to meteorite impact loading, giving r₀_Moon ≈ r₀_Earth × 0.1 (much smaller, higher-frequency band gaps needed for impact resistance).
The Lunar TOA building is therefore a compact hexagonal structure (~3.4m radius at 6 stages with r₀ = 0.1m) — essentially a modular unit that can be fabricated from sintered regolith using existing lunar ISRU (in-situ resource utilization) technology and tiled at any scale.
The graded TOA growth law is algorithmically simple: a robot requires only the current stage index k and the corresponding n-bonacci constant φ_{k+1} to determine all geometric parameters of the next stage. No central planning, no global state — each stage is locally determined from the previous one. This is the architectural equivalent of the bee's rule: build a hexagonal cell with the same ratio as the last one. The bee doesn't know it's implementing the dm³ framework; neither does the robot need to.
We have shown that Topographical Orthogenetic Architecture, originally conjectured in [Grossi, Medium 2026] as a hexagonal growth system with constant ratio g = 33^(1/6), is recoverable from and generalized by the dm³ contact-geometric framework. The key results are:
Result 1. The growth constant g ≈ 1.814 ≈ η (Tribonacci constant) is the geometric-mean approximation of the first six n-bonacci constants, giving final dimension η^6 ≈ 34.1 ≈ 33 — recovering the original conjecture from first principles.
Result 2. The superior graded growth law uses the n-bonacci constants φ₂, φ₃, …, φ_{N+1} at successive stages, creating N distinct phononic band gaps covering N octaves of the loading spectrum — the same mechanism responsible for the survival of ancient fractal structures in seismically active zones.
Result 3. The TOA geometry — rotationally symmetric outer shell, fractal interior hierarchy, hexagonal plan — is the attractor of at least five independent optimization processes: Mediterranean seismic iteration (Roman amphitheaters), Asian Buddhist construction tradition (stupas), biological evolution (bee hives), NASA first-principles structural engineering (1975 space colony study), and the dm³ mathematical framework. This convergence constitutes orthogenesis: independent arrival at the same fixed point from different initial conditions.
Result 4. The framework extends to Martian and Lunar construction by recalibrating the seed radius r₀ to align band gaps with the local loading spectrum. Mars requires ~4× larger structures; the Moon requires ~0.1× smaller compact modules, both robotically constructible from local materials.
The ancient builders knew this without the formalism. The mathematicians can now prove it. The robots can build it.