Unit II: Non-Euclidean Geometries — Summary

Analytical Digest of Part II, Chapter III (Les géométries non euclidiennes)

1. The Collapse of the Parallel Postulate Chimera

Distinguishing analytical a priori common notions ($A = B \wedge B = C \Rightarrow A = C$) from genuinely geometric postulates, Poincaré recounts how Lobachevsky (1829–1840) and János Bolyai (1832) ended twenty-one centuries of vain attempts to prove Euclid's Fifth Postulate by assuming multiple parallels through an exterior point and developing a contradiction-free hyperbolic geometry in which:

  • The angle sum of every triangle is strictly less than $\pi$, with angular defect $\Delta = \pi - (\alpha + \beta + \gamma)$ proportional to area;

  • Non-congruent similar figures cannot exist (absolute length scale);

  • Riemann (1854/1867) subsequently constructed spherical/elliptic geometry by abandoning both Postulate V and uniqueness of lines through antipodal points, yielding unbounded yet finite space with zero parallels and triangle angle sums exceeding $\pi$.

2. Beltrami’s Pseudosphere & Poincaré’s Translation Dictionary

How do we prove that infinite deduction in Lobachevsky’s system can never hit a hidden contradiction? Beltrami (1868) embedded two-dimensional constant-negative-curvature regions on pseudospherical surfaces, and Poincaré provides an elementary three-dimensional (and planar disk/half-plane) model-theoretic proof of relative consistency via his two-column Translation Dictionary:

  • Hyperbolic space / plane / straight line $\longleftrightarrow$ interior above/within a fundamental horizon $\Omega$, spheres/circles cutting $\Omega$ orthogonally;

  • Hyperbolic angle $\longleftrightarrow$ ordinary Euclidean angle (conformal invariance);

  • Hyperbolic distance $\longleftrightarrow$ natural logarithm of the projective cross-ratio $\chi(P, Q; U, V)$ with the ideal intersections $U, V \in \partial\Omega$.

Any contradiction derived from Lobachevsky’s axioms would translate word-for-word into a contradiction among theorems of ordinary Euclidean circle geometry!

3. Implicit Axioms, the Fourth Geometry, Lie Groups & Hilbert’s Continuum

  • Critique of Mill & Superposition: Answering John Stuart Mill, Poincaré exposes the circularity of defining geometric equality via rigid superposition and isolates the implicit axiom of rigid-body mobility (defining the straight line as the fixed axis of an invariable rotation).

  • The Fourth (Indefinite / Lorentzian) Geometry: Allowing local rotation while denying that a ray can rotate continuously into its opposite ray yields a coherent Fourth Geometry in which a real straight line can be perpendicular to itself (isotropic null line).

  • Sophus Lie & David Hilbert: By Sophus Lie’s group theorem, requiring free mobility of rigid figures restricts constant-curvature metric geometries to a finite family (excluding generic variable-curvature Riemannian metrics), while Veronese and Hilbert’s Grundlagen (1899) generate non-Archimedean geometries over third-order continua.

4. The Conventionalist Verdict: “Definitions in Disguise”

Geometric axioms cannot be Kantian synthetic a priori judgments (for we can conceive and build coherent non-Euclidean geometries, whereas we cannot build a coherent non-inductive arithmetic), nor can they be empirical laws (for ideal lines do not exist in matter, and exact postulates would be falsified daily by imperfect solids). Therefore, geometric axioms are conventions (*définitions déguisées*): guided by experience with natural solids, free subject to consistency, and judged not by truth or falsity—a meaningless question like asking whether Cartesian coordinates are “truer” than polar coordinates—but by simplicity and convenience (commodité).