Differential Geometry on Pointwise Affine Spaces

Dan Jonsson · 2025

Read Differential Geometry on Pointwise Affine Spaces by Dan Jonsson (2025) 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-2509-06748.txt.

Introduction

Curved spaces rebuilt without manifolds are presented by Dan Jonsson in a solo 2025 geometry paper. Jonsson defines pointwise affine spaces as a replacement for the manifold concept at the heart of differential geometry. Familiar notions then need fresh definitions inside the new setting.

Connections, torsion, the Riemann curvature tensor, vector fields, and ways to differentiate vector fields are all reintroduced in the pointwise affine framework. Examples show the resulting machinery at work. Jonsson reports that the notions and results generally take simpler and more transparent forms than in conventional treatments. Frameworks of this kind are judged over time: if the simpler forms help prove new results or clarify old ones, the reformulation earns its place beside the standard manifold language of geometry.

Catalog record

CREATORDan Jonsson
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE153 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-06748

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Differential Geometry on Pointwise Affine Spaces Dan Jonsson Abstract . We introduce an alternative formalization of curved spaces in which the concept of a pointwise affine space, as defined here, replaces that of a manifold. New or modified definitions of familiar notions from differential geometry such as connections, torsion, Riemann curvature, vector fields and differentiation of vector fields are presented, and examples of applications of the resulting framework are given. In general, the notions and results considered receive simpler and more transparent forms than in conventional approaches. 1. Introduction It may be said that there are two mathematical languages for differential geometry: one coordinate-oriented and one coordinate-free. The coordinate-oriented language is the oldest one, and bears evidence of its origins. During the early development of differential geometry, modern notions of abstract spaces had not yet crystallized. The first axiomatic definition of affine spaces was given by Weyl in 1918 [ 1 ] , and in spite of early contributions by Grassmann [ 2 ] and Peano [ 3 ] abstract vector space theory ”was actually not developed before 1920” [ 4 , p. 228] . By then, differential calculus on curved spaces, following Ricci and Levi-Civita [ 5 ] , had already been used in the theory of General Relativity. Differential calculus on \mathbb{R}^{n} regarded as a concrete flat space had morphed into a more modern differential calculus when \mathbb{R}^{n} came to be regarded as a possibly curved space. To clarify this point, we give a brief and simplified account of generalized differential calculus on \mathbb{R}^{n} . \mathbb{R}^{n} can be regarded as a space where points, vectors and their coordinates live. We define a point x\in\mathbb{R}^{n} as a tuple \mathopen{}\mathclose{{\left(x_{1},\ldots,x_{n}}}\right) , a vector v\in\mathbb{R}^{n} as a tuple \mathopen{}\mathclose{{\left(v_{1},\ldots v_{n}}}\right), and denote their their coordinates by \mathopen{}\mathclose{{\left(x^{1},\ldots,x^{n}}}\right) and \mathopen{}\mathclose{{\left(v^{1},\ldots v^{n}}}\right) , respectively. A coordinate system is a bijection \displaystyle\mathsf{C}:\mathbb{R}^{n}\times\mathbb{R}^{n}\rightarrow\mathbb{R}^{n}\times\mathbb{R}^{n}, \displaystyle\mathopen{}\mathclose{{\left(\mathopen{}\mathclose{{\left(x_{1},\ldots,x_{n}}}\right),\mathopen{}\mathclose{{\left(v_{1},\ldots,v_{n}}}\right)}}\right)\mapsto\mathopen{}\mathclose{{\left(\mathopen{}\mathclose{{\left(x^{1},\ldots,x^{n}}}\right),\mathopen{}\mathclose{{\left(v^{1},\ldots,v^{n}}}\right)}}\right) such that x^{i}=f^{i}\mathopen{}\mathclose{{\left(x}}\right) and v^{i}=f^{i}\mathopen{}\mathclose{{\left(v}}\right) for i=1,\ldots,n . A Cartesian coordinate system is a coordinate system such that x^{i}=f^{i}\mathopen{}\mathclose{{\left(x_{1},\ldots,x_{n}}}\right)=x_{i} and v^{i}=f^{i}\mathopen{}\mathclose{{\left(v_{1},\ldots,v_{n}}}\right)=v_{i} for i=1,\ldots,n . In particular, we let v\thicksim_{\mathsf{C}}v^{k} describe a situation where the vector v has coordinates \mathopen{}\mathclose{{\left(v^{1},\ldots,v^{n}}}\right) relative to \mathsf{C} . A vector field \boldsymbol{v} on \mathbb{R}^{n} is a function \boldsymbol{v}:\mathbb{R}^{n}\rightarrow\mathbb{R}^{n},\qquad x\mapsto v. Using the Einstein summation convention, we can define the directional derivative \nabla_{u}\boldsymbol{v}\mathopen{}\mathclose{{\left(x}}\right) of the differentiable vector field \boldsymbol{v} with respect to the vector u at the point x by requiring that (1.1) \nabla{}_{u}\boldsymbol{v}\mathopen{}\mathclose{{\left(x}}\right)\sim_{C}u^{i}\frac{\partial\boldsymbol{v}^{k}}{\partial x^{i}}\mathopen{}\mathclose{{\left(x}}\right), where C is a Cartesian coordinate system. However, if we substitute coordinates relative to a curvilinear (that is, not necessarily Cartesian) coordinate system \overline{C} in ( 1.1 ) then the definition of \nabla{}_{u}\boldsymbol{v}\mathopen{}\mathclose{{\left(x}}\right) may need to be given another form since, in general, \nabla_{u}\boldsymbol{v}\mathopen{}\mathclose{{\left(x}}\right)\nsim_{\overline{C}}\,\overline{u}^{i}\frac{\partial\overline{\boldsymbol{v}}^{k}}{\partial\overline{x}^{i}}\mathopen{}\mathclose{{\left(x}}\right). We thus need to complement ( 1.1 ) by a correction term, which we express by means of a system of scalars \Gamma_{ij}^{k} for each point x , so that we obtain a coordinate-system independent definition (1.2) \nabla_{u}\boldsymbol{v}\mathopen{}\mathclose{{\left(x}}\right)\sim_{\overline{C}}\,\overline{u}^{i}\frac{\partial\overline{\boldsymbol{v}}^{k}}{\partial\overline{x}^{i}}\mathopen{}\mathclose{{\left(x}}\right)+\overline{u}^{i}\overline{\boldsymbol{v}}^{j}\Gamma_{ij}^{k}\mathopen{}\mathclose{{\left(x}}\right). By the construction of \Gamma_{ij}^{k} , we have \Gamma_{ij}^{k}\mathopen{}\mathclose{{\left(x}}\right)=0 for all i,j,k,x if \overline{C} is Cartesian, so \Gamma_{ij}^{k} measures the discrepancy between \overline{C} and a Cartesian coordinate system. We have thus generalized ( 1.1 ), a formula for calculating a directional derivative in \mathbb{R}^{n} , valid only for a Cartesian coordinate system, to a formula ( 1.2 ), valid for a curvilinear coordinate system. There is, however, another way to look at ( 1.2 ). In Einstein’s and Grossmann’s Entwurf article [ 6 , p. 23] , the latter makes a profound observation: Ricci und Levi-Cività haben, ausgehend von den Christoffel’schen Resultaten, ihre Methoden der absoluten, d.h. vom Koordinatensystem unabhängigen Differentialrechnung entwickelt, die gestatten, den Differentialgleichungen der mathematischen Physik eine invariante Form zu geben. Da aber die Vektoranalysis des auf beliebige krummlinige Koordinaten bezogenen euklidischen Raumes formal identisch ist mit der Vektoranalysis einer beliebigen [—] Mannigfaltigkeit , so bietet es keine Schwierigkeiten, die vektoranalytischen Begriffsbildungen [---] auszudehnen auf die vorstehende allgemeine Theorie von Einstein. [emphasis supplied] 1 1 1 Starting from the results of Christoffel, Ricci and Levi-Cività have developed their methods of absolute, i.e., coordinate-system independent, differential calculus that allows the differential equations of mathematical physics to be given an invariant form. However, as vector analysis in a Euclidean space with arbitrary curvilinear coordinate s is formally identical to vector analysis in an arbitrary manifold , it does not present any difficulties to extend the conceptual apparatus of vector analysis [so as to apply to] Einstein’s general theory presented above. In other words, the change of coordinates of u and v can depend on a change of the coordinate system (”passive transformation”), or it can depend on a change of the vectors u and v themselves (”active transformation”). These are interchangeable points of view. In the latter case, the vectors are transformed as a consequence of a deformation of \mathbb{R}^{n} to a curved space. Thus, in the first case, \Gamma_{ij}^{k} describe the ”curvature” of the coordinate system; in the second case \Gamma_{ij}^{k} describe the curvature of the space itself. Thus, early differential geometry was characterized by • a coordinate-oriented mathematical language, • a setting for differential calculus in the form of \mathbb{R}^{n} regarded as a possibly curved space, • a tool \Gamma_{ij}^{k} , nowadays known as a connection, for describing the curvature of \mathbb{R}^{n} . The coordinate-oriented approach, as amplified by the tensor formalism [ 5 ] , is quite powerful, but has two significant weaknesses. The first one is that \mathbb{R}^{n} has a globally trivial topology (shape): it cannot naturally represent a sphere, a torus etc. The second weakness derives precisely from the fact that a coordinate-oriented mathematical language is used. For example, in modern algebra a vector is conceived as simply an element of a vector space, not, in a more convoluted way, as an object the coordinates of which change in a particular way when the coordinate system changes in a particular way. A related reproach is that the ”débauches d’indices” (Cartan) used in the coordinate-oriented language tends to obscure the geometric meaning of objects and operations. Thus, the emphasis in the further development of differential geometry has fallen on elaborating an appropriate notion of a manifold that generalizes \mathbb{R}^{n} so as to accommodate spaces with more general shapes, and on proposing coordinate-free formulations whenever feasible. Several sophisticated frameworks have been developed, characterized by • an abstract, algebraically oriented mathematical language, • a setting for differential calculus in the form of a fiber bundle (such as a tangent bundle, a vector bundle, a frame bundle or a principal bundle) the base space of which is a differentiable manifold, a generalization of \mathbb{R}^{n} (sometimes \mathbb{C}^{n}) that is locally but not necessarily globally similar to \mathbb{R}^{n} ( \mathbb{C}^{n}) , • an abstract notion of a connection on the fiber bundle involved. The price in the form of increased complexity paid to generalize \mathbb{R}^{n} and reduce the use of coordinates is, alas, high. While this increased complexity serves a purpose for advanced applications, there are many simpler applications where it is redundant. Worse, a significant disadvantage of the coordinate-oriented approach persists: the geometric meaning of abstract algebraic expressions is often as obscure as that of coordinate expressions. Is it really not possible to find a simpler and simultaneously more transparent coordinate-free formulation of differential geometry? Following up on previous

Read the complete text