Vector Invariance and Structural Closure of Julia-Type Iterations in Clifford Algebra
Orgest ZAKA · 2026
Read Vector Invariance and Structural Closure of Julia-Type Iterations in Clifford Algebra by Orgest ZAKA (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-20252.txt.
Introduction
Orgest Zaka introduces a Clifford setting for Julia dynamics via the geometric product. The iteration raises a vector to a power along a unit direction and adds a constant. Though the geometric product creates higher-grade parts, a grade-reduction mechanism collapses to the vector subspace. The Clifford Julia operator is closed on the space, defining a well-posed system in any dimension. The work extends Julia dynamics past the complex plane and into higher-dimensional geometric algebra, opening a new path for fractal-type dynamics in Clifford algebras.
Catalog record
| CREATOR | Orgest ZAKA |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39146 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.20252
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Vector Invariance and Structural Closure of Julia-Type Iterations in Clifford Algebra Orgest ZAKA Abstract . In this paper, we introduce a Clifford algebra framework for Julia-type dynamics driven by the geometric product. The nonlinear iteration f ( x → ) = ( x → ⋄ n → ) p ⋄ n → + c → , p ≥ 2 , f(\vec{x})=(\vec{x}\diamond\vec{n})^{p}\diamond\vec{n}+\vec{c},\qquad p\geq 2, is studied in a real n n -dimensional inner-product space V V , where x → , n → , c → ∈ V \vec{x},\vec{n},\vec{c}\in V and n → \vec{n} is a unit vector. The main result reveals a previously unreported invariance phenomenon: although the geometric product generates higher-grade multivector components at intermediate stages, a built-in grade-reduction mechanism ensures complete collapse back to the vector subspace. Consequently, the Clifford Julia operator is shown to be closed on V V , and the iteration defines a well-posed nonlinear dynamical system in arbitrary dimensions. This invariance is established through a structural decomposition of the Clifford product and an inductive closure argument, supported by explicit verification in low-dimensional cases and a general proof in ℝ n \mathbb{R}^{n} . The results demonstrate that classical Julia dynamics can be consistently extended beyond the complex plane into higher-dimensional geometric algebra without loss of geometric interpretability. The framework opens a new direction for fractal-type dynamics in Clifford algebras, providing a unified algebraic setting for higher-dimensional invariant-preserving iterative systems. 2000 Mathematics Subject Classification 15A66, 37F50, 30C35, 11S80 1. Introduction The generation of Julia-type fractal sets traditionally relies on the iteration of complex functions z\mapsto z^{p}+c . This paper proposes a generalization of this process to n -dimensional vector spaces V through Geometric Algebra (GA), exploiting the isomorphism between complex numbers and planes defined by the geometric product. In this context, we consider the geometric product operator \diamond , which unifies the inner product ( \cdot ) and the outer product ( \wedge ): \vec{a}\diamond\vec{b}=\vec{a}\cdot\vec{b}+\vec{a}\wedge\vec{b}. For a unit vector n\in V ( \vec{n}^{2}=\vec{n}\diamond\vec{n}=1 ), the Julia iteration is rewritten in vector form as: (1.1) \vec{x}_{k+1}=(\vec{x_{k}}\diamond\vec{n})^{p}\diamond\vec{n}+\vec{c} where x,n,c\in V . Although the geometric product in the equation ( 1.1 ) generates multivectors of different degrees at intermediate steps, we propose that the specific structure of this operator preserves degree invariance. The main goal of this paper is to prove mathematically that this iteration is closed within the vector space V , providing a rigorous method for generating fractals in high dimensions without the need for artificial projections. The present work is situated at the intersection of geometric algebra, fractal geometry, and higher-dimensional dynamical systems. The foundational algebraic framework is based on Clifford Algebra and Geometric Calculus as developed in [ 3 , 4 ] , which provide the natural language for extending classical complex dynamics to multivector settings. Fractal theory and iterative complex dynamics are grounded in the classical results of Falconer [ 2 ] , while higher-dimensional and hypercomplex generalizations of fractal constructions are addressed in [ 8 , 9 , 7 ] . A significant part of the mathematical background of this work is also rooted in the author’s previous contributions to affine and Desarguesian plane geometry and their associated algebraic structures. In particular, earlier studies have developed algebraic models for point-line systems, collineations, and transformation groups in affine planes, together with their induced skew-field and group structures [ 15 , 16 , 17 , 18 , 19 , 20 ] . Further developments have focused on invariant-preserving transformations, cross-ratio geometry, and Dyck-type group structures on polygonal cycles, revealing deep connections between geometric configurations and algebraic operations [ 21 , 22 , 23 , 24 , 25 ] . These results were complemented by studies on endomorphism algebras, translation groups, and skew-field constructions in affine settings [ 26 , 27 , 28 , 29 ] . Collectively, these works establish a coherent geometric-algebraic framework that naturally motivates the present extension into Clifford algebra and higher-dimensional geometric dynamics, where classical affine invariants are generalized into multivector-valued transformations and iterative structures. Recent advances in the use of geometric algebra for fractal generation and computational geometry, particularly in [ 6 , 5 , 11 , 12 ] , motivate the algebraic formulation adopted in this paper. Applications in computer graphics, robotics, and geometric modeling further support the relevance of this framework, as discussed in [ 1 , 10 ] . The SIGGRAPH course notes [ 14 ] and related technical reports [ 13 ] provide additional computational and geometric perspectives. Together, these works establish a solid theoretical and applied background for extending Julia-type iterations into the Clifford algebra setting, while highlighting the gap addressed in this paper: the preservation of vector subspace invariance under nonlinear geometric product iterations in arbitrary dimensions. The main objective is to determine whether the iteration preserves the vector subspace 2. Explicit Low-Dimensional Cases Before establishing the general vector invariance result in arbitrary dimension, it is useful to examine explicitly the cases of \mathbb{R}^{2} and \mathbb{R}^{3} . These computations illustrate how the geometric product generates intermediate multivector terms while the final expression remains entirely in the vector subspace. The two-dimensional case reproduces the classical quadratic complex structure inside Clifford algebra, whereas the three-dimensional case shows how higher-grade terms cancel naturally through antisymmetry of the exterior product. Proposition 1 (Vector Invariance in \mathbb{R}^{2} ) . Let \vec{x},\vec{n},\vec{c}\in\mathbb{R}^{2}, where \vec{n} is a unit vector satisfying \vec{n}\diamond\vec{n}=1. For every integer p\geq 2 , define f(\vec{x})=(\vec{x}\diamond\vec{n})^{p}\diamond\vec{n}+\vec{c}. Then f:\mathbb{R}^{2}\to\mathbb{R}^{2} is well-defined. More precisely, (\vec{x}\diamond\vec{n})^{p}\diamond\vec{n} is always a vector in \mathbb{R}^{2} . Consequently, every iterate of the associated Clifford Julia iteration remains inside the vector space \mathbb{R}^{2} . Proof. We work in \mathbb{R}^{2} with orthonormal basis \{\vec{e}_{1},\vec{e}_{2}\} and choose \vec{n}=\vec{e}_{1} . Let \vec{x}=x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2}. Case p=1 . We compute (\vec{x}\diamond\vec{e}_{1})\diamond\vec{e}_{1}=\vec{x}\diamond(\vec{e}_{1}\diamond\vec{e}_{1}). Since \vec{e}_{1}\diamond\vec{e}_{1}=1 , it follows that (\vec{x}\diamond\vec{e}_{1})\diamond\vec{e}_{1}=\vec{x}, which is a vector in \mathbb{R}^{2} . Case p=2 . We now consider the first nontrivial case p=2 . To get a clear presentation of this, let’s start by writing this on the components for \vec{n}=\vec{e}_{1} and p=2, (also remember that \vec{x}=x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2} ) we obtain for the product term \displaystyle(\vec{x}\diamond\vec{e}_{1})^{2}\diamond\vec{e}_{1} \displaystyle=(\vec{x}\diamond\vec{e}_{1})\diamond(\vec{x}\diamond\vec{e}_{1})\diamond\vec{e}_{1} \displaystyle=(\vec{x}\diamond\vec{e}_{1}\diamond\vec{x})\diamond(\vec{e}_{1}\diamond\vec{e}_{1}) \displaystyle=(\vec{x}\diamond\vec{e}_{1}\diamond\vec{x})\diamond(\vec{e}_{1}\cdot\vec{e}_{1}+\vec{e}_{1}\wedge\vec{e}_{1}) \displaystyle=\vec{x}\diamond\vec{e}_{1}\diamond\vec{x}.\quad (because \overrightarrow{e_{1}}\cdot\vec{e}_{1}=1 and \vec{e}_{1}\wedge\vec{e}_{1}=0 ) \displaystyle=\Big((x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2})\diamond\vec{e}_{1}\diamond(x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2})\Big) \displaystyle=\Big[x_{1}(\vec{e}_{1}\diamond\vec{e}_{1})+x_{2}(\vec{e}_{2}\diamond\vec{e}_{1})\Big]\diamond(x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2}) \displaystyle=\Big[x_{1}(\vec{e}_{1}\cdot\vec{e}_{1}+\vec{e}_{1}\wedge\vec{e}_{1})+x_{2}(\vec{e}_{2}\cdot\vec{e}_{1}+\vec{e}_{2}\wedge\vec{e}_{1})\Big]\diamond(x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2}) \displaystyle=\Big[x_{1}+x_{2}(\vec{e}_{2}\wedge\vec{e}_{1})\Big]\diamond(x_{1}\vec{e}_{1}+x_{2}\vec{e}_{2}). (but \vec{e}_{2}\cdot\vec{e}_{1}=0 and \vec{e}_{1}\cdot\vec{e}_{1}=1. ) \displaystyle=x_{1}\Big[x_{1}+x_{2}(\vec{e}_{2}\wedge\vec{e}_{1})\Big]\diamond\vec{e}_{1}+x_{2}\Big[x_{1}+x_{2}(\vec{e}_{2}\wedge\vec{e}_{1})\Big]\diamond\vec{e}_{2} \displaystyle=\Big[x_{1}^{2}+x_{1}x_{2}(\vec{e}_{2}\wedge\vec{e}_{1})\Big]\diamond\vec{e}_{1}+\Big[x_{1}x_{2}+x_{2}^{2}(\vec{e}_{2}\wedge\vec{e}_{1})\Big]\diamond\vec{e}_{2} \displaystyle=x_{1}^{2}\vec{e}_{1}+x_{1}x_{2}(\vec{e}_{2}\wedge\vec{e}_{1})\diamond\vec{e}_{1}+x_{1}x_{2}\vec{e}_{2}-x_{2}^{2}(\vec{e}_{1}\wedge\vec{e}_{2})\diamond\vec{e}_{2}, usage: (\vec{e}_{2}\wedge\vec{e}_{1})\diamond\vec{e}_{1}=\vec{e}_{2},\quad\quad\vec{e}_{2}\wedge\vec{e}_{1}=-\,\vec{e}_{1}\wedge\vec{e}_{2},\quad\text{and}\quad(\vec{e}_{1}\wedge\vec{e}_{2})\diamond\vec{e}_{2}=\vec{e}_{1}, substituting, we get: \displaystyle=x_{1}^{2}\vec{e}_{1}+x_{1}x_{2}\vec{e}_{2}+x_{1}x_{2}\vec{e}_{2}-x_{2}^{2}\vec{e}_{1} \displaystyle=x_{1}^{2}\vec{e}_{1}+2x_{1}x_{2}\vec{e}_{2}-x_{2}^{2}\vec{e}_{1} \displaystyle=(x_{1}^{2}-x_{2}^{2})\vec{e}_{1}+2x_{1}x_{2}\vec{e}_{2}. Hence, (\vec{x}\diamond\vec{e}_{1})^{2}\diamond\vec{e}_{1}=(x_{1}^{2}-x_{2}^{2})\vec{e}_{1}+2x_{1}x_{2}\vec{e}_{2}, which is again a vector in \mathbb{R}^{2} . Inductive step. Assume that for some p\geq 2 the expression (\vec{x}\diamond\vec{e}_{1})^{p-1}\diamond\vec{e}_{1} is a vector in \mathbb{R}^{2} . Hence there exist scalars a,b\in\mathbb{R} such that (\vec{x}\diamond\vec{e}_{1})^{p-1}\diamond\vec{e}_{1}=a\vec{e}_{1}+b\vec{e}_{2}. Now consider \displaystyle(\vec{x}\diamond\vec{e}_{1})^{p}\diamond\vec{e}_{1} \displaystyle=(\vec{x}\diamond\vec{e}_{1})\diamond(a\vec{e}_{1}+b\vec{e}_{2}) \displaystyle=(x_{1}+x_{2}\vec{e}_{2}\diamond\vec{e}_{1})\diamond(a\vec{e}_{1}+b\vec{e}_{2}). Using the Clifford relations in \mathbb{R}^{2} : \vec{e}_{1}\diamond\vec{e}_{1}=1,\quad\vec{e}_{2}\diamond\vec{e}_{2}=1,\quad\vec{e}_{2}\diamond\vec{e}_{1}=-\vec{e}_{1}\diamond\vec{e}_{2}, we expand term by term: \displaystyle(\vec{x}\diamond\vec{e}_{1})^{p}\diamond\vec{e}_{1} \displaystyle=x_{1}(a\vec{e}_{1}+b\vec{e}_{2})+x_{2}(\vec{e}_{2}\diamond\vec{e}_{1})\diamond(a\vec{e}_{1}+b\vec{e}_{2}). Now compute the second part: \displaystyle(\vec{e}_{2}\diamond\vec{e}_{1})\diamond\vec{e}_{1} \displaystyle=\vec{e}_{2}, \displaystyle(\vec{e}_{2}\diamond\vec{e}_{1})\diamond\vec{e}_{2} \displaystyle=-\vec{e}_{1}. Therefore, \displaystyle(\vec{x}\diamond\vec{e}_{1})^{p}\diamond\vec{e}_{1} \displaystyle=x_{1}a\vec{e}_{1}+x_{1}b\vec{e}_{2}+x_{2}a\vec{e}_{2}-x_{2}b\vec{e}_{1} \displaystyle=(x_{1}a-x_{2}b)\vec{e}_{1}+(x_{1}b+x_{2}a)\vec{e}_{2}. Hence, (\vec{x}\diamond\vec{e}_{1})^{p}\diamond\vec{e}_{1}\in\mathbb{R}^{2}. This shows that the operator preserves the vector space structure in \mathbb{R}^{2} . ∎ Proposition 2 (Vector Invariance in \mathbb{R}^{3} ) . Let \vec{x},\vec{n},\vec{c}\in\mathbb{R}^{3}, where \vec{n} is a unit vector satisfying \vec{n}\diamond\vec{n}=1. For every integer p\geq 2 , define f(\vec{x})=(\vec{x}\diamond\vec{n})^{p}\diamond\vec{n}+\vec{c}. Then f:\mathbb{R}^{3}\to\mathbb{R}^{3} is well-defined. In particular, (\vec{x}\diamond\vec{n})^{p}\diamond\vec{n} belongs to the grade- 1 subspace of the Clifford algebra associated with \mathbb{R}^{3} . Therefore, despite the appearance of bivector and trivector terms in intermediate computations, the final expression reduces to a vector in \mathbb{R}^{3} . Proof. Without loss of generality, we choose an orthonormal basis \{\vec{e}_{1},\vec{e}_{2},\vec{e}_{3}\}