The Arithmetics of Forms

Replacing Binary Constructive Solid Geometry with Ternary Integer Arithmetic

A non-anthropomorphic mathematical framework rooted in the 1997 SGDL paradigm by Jean-François Rotgé. Physical structure is decomposed into elementary building blocks through pure integer arithmetic — no floating-point, no division, no approximation. Where Tarski proved geometry complete because it cannot define the integers, Rotgé reverses the relationship: using integers to reconstruct a total, decidable geometry — Exact Intelligence.

Rooting Spatial Logic in Pure Numbers

Standard systems patch geometry and logic together through syntax. The SGDL approach unifies geometry and topology upstream through the arithmetic of Primitive Recursive Functions. Space is quantified before a computer procedure is ever written. Three classical approaches have been proposed:

Mechanistic

Turing machines — step-by-step procedural instructions. The path taken by most CAD systems.

Used by traditional CAD

Linguistic

Formal grammars — Chomsky, Post. Shape described through rewriting rules and production systems.

Grammar-based systems

sgdl

Functional-Arithmetic

Recursive functions, lambda calculus — Skolem, Gödel, Kleene, Church. Spatial form as the output of pure recursive functions on natural integers.

The SGDL path

In the arithmetic approach, geometric and combinatorial models are not separate — they are unified through arithmetic, a pillar of metamathematics. All constructive and logical processes of spatial arrangement are formalized as algorithms that are exclusively and purely arithmetic.

Modern statistical AI — the dominant paradigm — is an extreme instrumentalism: its knowledge is the weighted correlations most predictive of training data. It will fail whenever the structure of a new problem differs from the training distribution. sgdl takes the opposite posture: understand why the solution space has the shape it has, identify the invariants that constrain it, and derive the answer from first principles. The first posture scales with compute. The second scales with understanding.

“Arithmetic is transmuted into a metalanguage of spatial forms.”

— J.F. Rotgé, doctoral thesis, 1997

Two Worlds That Need to Connect

Symbolic AI

  • Reasons with logic and formal rules
  • Based on programmed rules and knowledge
  • Performant in structured, well-defined domains
  • Explainable and traceable
  • Limited when dealing with raw, unstructured data

Statistical AI

  • Learns patterns directly from data
  • Powered by LLMs and neural networks
  • Performant for vision, translation, generation
  • Scales with compute and data
  • Approximative and black-box by nature

Neither approach alone can interact with the physical world. Their convergence gives rise to a new discipline: Physical AI.

The Emergence of Physical AI

Physical AI perceives the environment and acts in the real world. It powers robots, autonomous vehicles, surgical tools, and industrial automation — but also generative AI, synthetic cities, and low-altitude economy. It requires extreme reliability — there is no room for approximation when a robotic arm operates near a human body or a drone navigates urban airspace.

The Arithmetics of Forms provides exactly this: the deterministic, precise link between AI agents and deep-tech platforms that operate in physical space.

Symbolic AIReasons with rulesStatistical AILearns from dataPhysical AIPerceives, decides, actsSymbolic AI + Statistical AI = Physical AI

The Transmutation Principle

Geometry becomes arithmetic. Every spatial question becomes a calculation on integers.

The Computable Canvas

Every primitive geometry partitions space into three mathematical states. By assigning an integer “density” to these states, we transform physical space into an arithmetic canvas: density 0 means Void (outside), density 1 means Boundary, density 2 means Mass (inside).

The topological question “is this point inside?” becomes the arithmetic question “what is this number?” When multiple forms overlap, densities accumulate — the intersection of two solids has density 4, their shared boundary density 2.

The 1D Density Ray020110 = Void1 = Boundary2 = Mass

The Design Logic Operators

Geometric operations map exactly to arithmetic operations. Each spatial combination of forms reduces to a simple calculation on their density values — turning CAD into algebra.

DLaddUnionAB242a + bDLmulIntersectionAB040a × bDLdifDifferenceAB200a − bDLsabSym. Diff.AB202|a − b|Given: density(A) = 2, density(B) = 2

Geometry via Integer Addition

Shapes are literally added together like numbers. The resulting mathematical densities natively describe perfect topological regions without any complex geometric calculations. Multiplicity is preserved.

Addition Proof Diagram

F0F10 + 0 = 01 + 0 = 12 + 0 = 22 + 1 = 32 + 2 = 41 + 2 = 30 + 2 = 20 + 1 = 10 + 0 = 0

The Arithmetic Toolset

Purely mathematical functions natively generate distinct topological regions simply by filtering for specific integer results. Every geometric operation reduces to a simple calculation on density values.

DLadd

+

Addition

Merge volumes — densities accumulate

F0F102420

DLmul

×

Multiplication

Logical conjunction — only shared mass

F0F101410

DLdif

Positive Difference

Carve away — F0 minus F1's influence

F0F102100

DLsab

|−|

Absolute Subtraction

Symmetric exclusion — XOR of forms

F0F102020

Purely mathematical functions natively generate distinct topological regions simply by filtering for specific integer results.

Live Design Logic Operations the Arithmetics of Forms in action

Drag to rotate · Click for details

2
DLadd

Union = Addition

Merge volumes — densities are added together.

d=3∂=98m=22Σ=120
DLmul

Intersection = Multiplication

Keep the common part — densities are multiplied.

d=2∂=6m=2Σ=8
DLdif

Difference = Subtraction

Remove material — densities are subtracted.

d=2∂=50m=7Σ=57
DLsab

Symmetric Diff. = Abs. Subtraction

Exclusive zone — absolute difference of densities.

d=2∂=100m=14Σ=114

DLadd

Union = Addition

Merge volumes — densities are added together.

DLmul

Intersection = Multiplication

Keep the common part — densities are multiplied.

DLdif

Difference = Subtraction

Remove material — densities are subtracted.

DLsab

Symmetric Diff. = Abs. Subtraction

Exclusive zone — absolute difference of densities.

DLgcd

Greatest Common Volume = GCD

Largest shared structure — GCD of densities.

DLlcm

Smallest Enclosing Vol. = LCM

Minimal container — LCM of densities.

Arithmetizing Kleene's Ternary Logic: In the SGDL paradigm, we do not write complex geometric algorithms to find intersections or unions. We simply ask the primitive recursive function to evaluate the minimum or maximum integer between two density arrays. Logic is completely reduced to arithmetic. The dual operator DLdua inverts matter (swapping interior and exterior) while preserving boundaries — a self-inverse involution that validates the arithmetized De Morgan laws.

Union Max(x, y)
Intersection Min(x, y)
Implication Max(Pos(x), y)

The Architecture of Forms

Primitive Recursive Functions occupy the central position in the SGDL system. From this arithmetic nucleus, eight categories of Forms radiate outward — each representing a different facet of the same mathematical object.

Encoding

FORMES

ENCODÉES

Geometry

FORMES

PRÉCONTRAINTES

FORMES

POLYÉDRIQUES

FORMES

QUADRATIQUES

FONCTIONS

P. RÉCURSIVES

Arithmetic

FORMES

LOGIQUES

FORMES

SIMPLICIALES

Programming

FORMES

PROGRAMMÉES

FORMES

VIRTUELLES

Algorithmic Synthesis: The SGDL ArchitectureMathematical FoundationSimplicial & Quadratic Forms — Projective hyper-surfacesArithmetic EnginePrimitive Recursive Functions (PRF) — Σ, Π, successorLogic LayerProjective Volume Programming — SCHEME/LISP — DL operatorsVirtual FormsDigital mock-ups, synthetic images, encoded BigNumsA unified systemLogic, topology, and geometryin a single functional languageFrom projective hyper-surfaces through recursive functions to virtual forms — a single arithmetic pipeline

Left Arm — Geometry

Quadratic Forms — Projective quadric surfaces: the geometric primitives, constructed synthetically without algebraic considerations.

Polyhedral Forms — Logical (and therefore arithmetic) combinations of degenerate quadratic forms. Polyhedra emerge from the same formalism as curved surfaces.

Preconstrained Forms — Forms governed by projective configurations — spatial scaffolds that transmit geometric constraints to the objects they support.

Right Arm — Arithmetic

Logical Forms — The equivalence between arithmetic operators and logical operators acting on Forms. Design Logic (DLadd, DLmul, DLdif...) is defined here.

Simplicial Forms — The junction between arithmetized logic and topology. Simplicial complexes are built from logical combinations of Forms, completing the arithmetic model of volumetric objects.

Downward — Programming

Programmed Forms — Functional programming via lambda calculus, integrating recursive arithmetic functions. The concepts of the first two sections are formalized into executable code — including a database system driven by the Arithmetics of Forms.

Virtual Forms — The final output: synthetic images and physical models generated by the SGDL system, validating the theory through millions of computations per image.

Upward — Encoding

Encoded Forms — The culmination of the pipeline: every volumetric structure is encoded into a single BigNum via Gray metacurves. The integer is the form; the form is the integer. This is the transmutation made concrete.

Adapted from Fig. 1.4 of J.F. Rotgé's doctoral thesis, “L'Arithmétique des Formes”, Université de Montréal, 1997.

The Ariadne Exolanguage

The autonomous symbolic spatial reasoning system where all SGDL theories converge into a single operational platform.

The Ariadne Exolanguage is an autonomous symbolic spatial reasoning system born from a mechanism discovered in the 1990s at the Université de Montréal. It unifies the five pillars of machine spatial knowledge — Logic, Topology, Morphogenesis, Indexing, and Encoding — through elementary arithmetic.

It is a triple system enabling simultaneously the representation, the genesis (morphogenesis), and the communication of spatial information. Designed from the ground up for machines (AI2AI), it is inherently non-anthropomorphic and self-evolving.

Belonging to the class of mathematically-oriented Domain-Specific Languages (DSL), it represents for everything concerning spatial reasoning the ultimate evolution of the coupling of Lambda Calculus and Primitive Recursive Functions.

Logic

Space Logic

Topology

Densities

Morpho.

Quadrics

Indexing

MCG

Encoding

MALVES

1D → nD: The Space-Filling Bridge

How does an integer become a spatial coordinate — and vice versa? Space-Filling Curves provide the bijective mapping between a one-dimensional integer thread and multi-dimensional space.

1D Thread & Integers0123456789Sequential data sourceFolding BeginsSpatial transformationSpace-Filling DensityUniversal Combinatorial KeySpace-Filling Curves (SFC) & Gray Meta-Curves (MCG)SFCs translate multi-dimensional spatial coordinates into a single, unique integer — and vice-versa

A Space-Filling Curve (SFC) visits every cell of an n-dimensional grid in a single, continuous path — establishing a total order over all spatial points. The SGDL system uses families of Gray metacurves (MCG) that preserve locality: points close in space stay close along the 1D thread. This locality-preserving property is essential for both efficient volumetric computation and cryptographic encoding.

Because the mapping is bijective, every n-dimensional volumetric form can be encoded as a single integer — and every integer can be decoded back into the original form. This is the mechanism behind Encoded Forms, the topmost arm of the SGDL Architecture.

“Every spatial structure can be reduced to a single integer. The integer IS the form.”

— The SGDL Encoding Principle

Beyond B-Rep and CSG

Traditional Constructive Solid Geometry (CSG) relies on binary set theory. A point is either inside a shape (1), or outside (0). This fails at the boundary — in binary logic, the boundary requires complex self-referential calculation (intersection of the closure and the complement closure). Furthermore, set union destroys multiplicity: if Shape A and Shape B overlap, the shared geometric density is completely lost.

The Arithmetics of Forms replaces this binary illusion with ternary integer arithmetic based on Kleene's three-valued logic. Every natural integer maps to a unique volumetric form, and every form maps back to its integer — with the full richness of arithmetic operations preserved as geometric operations.

The Set-Theory Impasse: Three Limits of CSGABBinary LimitSet theory has only 2 states (∈ / ∉).Impossible to encode the boundary.Multiplicity LossA ∪ B loses the overlap: the region containing two physical entities has no density..Post-RegularizationBoolean operations generate degenerate cases requiring heavy algorithmic corrections..Binary set theory cannot represent boundaries, multiplicity, or avoid degenerate edge cases

The Illusion

Binary Set Theory (CSG)

Shape AShape B11100A point is either inside (1) or outside (0).Overlap density is lost: 1 + 1 = 1

The Reality

Ternary Arithmetic (SGDL)

Shape AShape B0214120Three states: outside (0), boundary (1), inside (2).Multiplicity preserved: 2 + 2 = 4

The boundary problem: In binary CSG, the boundary requires complex self-referential calculation (intersection of the closure and complement closure). In SGDL, the boundary is simply the integer 1 — inherent to the representation, not computed after the fact.

CSG (Legacy) SGDL (Arithmetic)
Space States 2 (Inside / Outside) 3 (Outside / Boundary / Inside)
Logic Basis Binary Boolean Kleene Ternary Logic
Mathematical Root Set Theory Primitive Recursive Functions
Boundary Handling Requires complex topological calculation Inherent to the integer 1
Multiplicity Lost in union operations Preserved via integer addition

B-Rep

Boundary Representation

Lists surfaces — an extensional catalog of faces, edges, and vertices. No algebra on the geometry itself.

CSG

Constructive Solid Geometry

Boolean operations on primitives — powerful, but limited to a fixed vocabulary of shapes.

sgdl

Arithmetics of Forms

Integer ↔ Form bijection — full arithmetic on volumes. Shapes, operations, and encodings in a single computable system.

Key Properties

What makes the Arithmetics of Forms fundamentally different from statistical approaches.

Deterministic

Unlike probabilistic models, every computation produces exact, reproducible results with no randomness.

Precise

All computation uses exclusively natural integers — no floating-point numbers, no division, no rounding errors. Projective homogeneous coordinates ensure perfect precision through every calculation.

Frugal

CPU-native processing — no GPU clusters or data centers required. Orders of magnitude more efficient.

Economical

Dramatically lower infrastructure costs compared to GPU-dependent statistical approaches.

Unconditional

Projective geometry eliminates all singular configurations. The same algorithm works for every case — no IF/ELSE branching, no special-case handling. Parallel lines, degenerate surfaces, and edge cases are all treated uniformly.

Total

All operations are built on Primitive Recursive Functions (PRF), which are guaranteed to terminate. Every computation produces a result — no infinite loops, no undefined behavior, no crashes.

Where Approximation Is Not an Option

In continuous spatial domains, statistical approximation is not a necessary truth — it is a temporary instrumental limitation. The Arithmetics of Forms makes it obsolete in critical applications.

Autonomous Navigation

Obstacle recognition with 100% certainty in O(1) time — no training data, no rounding errors, no probabilistic confidence intervals.

Digital Twins & GIS

Regulatory boundaries as “high signs” — topologically exact, legally valid delimitations replacing the fuzzy zones of probabilistic segmentation.

Medical Imaging

Certification of organic structure connectivity via the topological invariant λ, eliminating false positives inherent to statistical segmentation models.

1 + 1 = 2

“We transform the notion of volume into integer arithmetic without division. The theoretical manipulation of whole numbers replaces set operations on volumes.”

— The SGDL Paradigm / L'Arithmétique des Formes

By abandoning the binary constraints of set theory and embracing the ternary logic of integer arithmetic, the boundary problem vanishes. Logical operations and physical operations become mathematically identical. Arithmetic IS Geometry.

A Fortress of 21 Interconnected Patents

Since 2018, sgdl has focused its strategy on building a comprehensive intellectual property portfolio. Today, 21 patents form an interlocking system of protection covering the complete pipeline — from spatial encoding and volumetric computation to cryptographic applications.

Spatial Encoding & Indexation

Patents covering the MALVES coding system and multidimensional data indexing methods that transform physical objects into mathematical representations, enabling precise geolocation and spatial reasoning at any scale.

Cryptographic Applications

Patents on MCG (MetaCurve Gray) cryptographic stencils, providing unique encryption capabilities derived from space-filling curve signatures.

Volumetric Computation

Patents protecting the core Arithmetics of Forms pipeline — from projective quadric definitions through Design Logic operators to BigNum encoding.

All 21 patents are interconnected and mature, forming a unified system ready for industrial deployment across robotics, autonomous vehicles, medical imaging, defense, and materials science.

From Space Stations to Synthetic Cities

The Arithmetics of Forms is not theoretical speculation. It has been deployed in industrial systems since the late 1990s, validated in peer-reviewed publications and mission-critical environments.

Space — CSA / ISS

ISS Canadarm2 & Dextre Simulator

Developed for the Canadian Space Agency to study astronaut skill degradation in orbit. SGDL provided pixel-exact rendering, cast shadow computation, and real-time collision detection for the complete ISS Mobile Servicing System — including Canadarm2's 7-DOF arm and Dextre's dual arms.

<2 MB

Entire ISS model compressed

P4 1.8 GHz

Real-time on a single laptop

50+ DOF

Elastic joints simulated

IAC-04-IAA.A.5.03 — Published at the International Astronautical Congress, 2004

Urban — IEEE 2007

Universal Solid 3D Format for Urban Simulation

A polynomial-based approach replacing conventional polygonal formats. Unifies CAD/CAM, GIS, AEC, and photogrammetry in a single volumetric representation — from individual barn details to entire states, with extremely compact data footprints.

~25,000 km²

Virtual Vermont — statewide

Real-time

On Bull Novascale parallel systems

Multi-layer

Topography, GIS, cadastral, transport

Virtual Vermont Project — statewide 3D simulation integrating topography, bathymetry, orthophotography, transport, and land use with UVM Transportation Center
Heritage — Roman, Gothic, Art Deco architectural reconstruction from libraries
Cnidaria Architecture — Medusa, Atlantis, Unda systems for real-time parallel processing

2007 Urban Remote Sensing Joint Event — IEEE, co-authored with J. Farret

SGDL-Scheme: the Manifesto (2000) — Presented at the Scheme and Functional Programming Workshop at Rice University, SGDL-Scheme is an absolutely pure functional language for volume programming: a geometric extension of Scheme where volume objects are manipulated as arithmetic expressions, compiled into optimized real-time pipelines, and evaluated at any point in space using only integer arithmetic.

The Platonic Solids

The Arithmetics of Forms operates on volumetric primitives. Explore the five Platonic solids — the foundational building blocks of geometric reasoning.