Weyl's 1922 problem of space: a constructive proof of the uniqueness of the Pythagorean measure determination, and an explanation of the world's four dimensions and metric signature (3,1)

Hermann Weyl's 1922 paper 'Die Einzigartigkeit der Pythagoreischen Massbestimmung' proved that the postulates of the relativistic problem of space - a measure determination fixed in the infinitesimal neighborhood, and closure properties of the group of infinitesimal rotations - admit exactly the 'Pythagorean' form of the metric: the infinitesimal rotation group must be a special orthogonal group $SO(p,q)$ of a non-degenerate quadratic form of arbitrary signature, leaving only the dimension $n=p+q$ and the inertia index $q$ undetermined. Weyl himself stated that this proof fails the standards of the constructive continuum analysis he then shared with Brouwer, because it rests on case distinctions (notably from elementary divisor theory) that are undecidable over the continuum, and he deferred a constructive treatment. He also stated that the question of why the actual world's metric has dimension 4 and inertia index 1 admits, at present, only provisional answers. Both demands remain open. The later literature reworked the uniqueness theorem exclusively classically (Weyl 1923; Scheibe 1957; Laugwitz 1958; Klingenberg 1958; Freudenthal 1964; Kobayashi-Nagano 1965), and van Atten (2019) records the 1922 proof as classical in its dependence on the excluded middle. The physics literature addresses the dimension and signature of spacetime only through anthropic selection arguments (Tegmark 1997), which explicitly disclaim rigor. No constructive proof, constructive impossibility result, or fundamental derivation of dimension 4 and inertia index 1 is known.

Background

Since the nineteenth century, the 'problem of space' (Raumproblem) has asked how the metric structure of physical space can be determined from structural postulates rather than posited outright. Helmholtz argued that free mobility of rigid bodies forces a geometry of constant curvature; Lie recast the question in terms of transformation groups. In this tradition the task is to characterize which metric structures are compatible with a group of 'motions', and thereby to explain why measurement in space takes the 'Pythagorean' form - lengths governed by a quadratic differential expression $ds^2=\sum g_{ij}dx_i dx_j$ - rather than by some other law.

Relativity forced the question to be re-posed. The metric of special relativity is not positive definite but has signature $(3,1)$ - three positive and one negative direction, that is, three space directions and one time direction; in general relativity the metric is a field varying from point to point; and Weyl's 1918 'purely infinitesimal' geometry (expounded in his textbook 'Raum, Zeit, Materie') transferred even the comparison of lengths to an infinitesimal connection. Weyl therefore rebuilt the space problem infinitesimally in his 1922 paper, the source of the present problem. His postulates demand that the measure determination (Massbestimmung: the assignment of squared length to infinitesimal line elements, i.e., a first-order metric) be fixed in the infinitesimal neighborhood of a point, and that the admissible infinitesimal rotations - the linear transformations of the tangent space preserving the measure - form a group with the closure properties of a rotation group. His main theorem states that under these postulates the infinitesimal rotation group is, up to conjugation, a special orthogonal group $SO(p,q)$: the group of determinant-one linear maps preserving a non-degenerate quadratic form with $p$ positive and $q$ negative squares. Hence the measure determination is necessarily 'Pythagorean' - a quadratic form, undetermined only in dimension $n=p+q$ and inertia index $q$ (the number of negative squares).

The paper's closing section concedes a foundational defect. Weyl had in 1921 publicly sided with Brouwer's intuitionism, according to which the law of excluded middle (every mathematical statement is either true or false) is not valid for the continuum: one cannot in general decide whether a given real number equals zero, so mathematics on the continuum must be built constructively, with every case distinction justified by decidable data. Weyl's 1922 proof, however, proceeds through classical case distinctions - in his own words, the case distinctions of elementary divisor theory (the classification used to put matrix pencils into normal form) are, from the constructive standpoint, 'a particularly questionable starting point' - and he promised to return to the question once a constructive 'continuum analysis' had been further developed. In classical logic the proof stands; in the constructive mathematics Weyl demanded, it is not a proof at all.

The theorem also leaves the two parameters of the actual world unexplained: why is the metric of the physical world four-dimensional, and why does it have inertia index 1 (one 'time' direction, as in general relativity)? Weyl offered provisional answers only. For dimension: only in four dimensions does there exist the simplest integral invariant of his gauge geometry with the character of an action, built from the length curvature (the field strength of Weyl's length-transfer connection, which he identified with the electromagnetic field). For the inertia index: only when it equals 1 does the infinitesimal cone split into two mantles (two connected components), making a distinction between past and future possible; on this question Weyl wrote that at present no other answer can be given.

The mathematical line of the space problem became part of the classical foundations of differential geometry - the characterization of Riemannian structure, and in modern language the theory of $G$-structures (first-order geometric structures on manifolds). The constructive question connects it to constructive mathematics, the tradition of Brouwer and Bishop: mathematics without excluded middle, where existence claims must be witnessed explicitly. The dimension question lives on in physics and philosophy of physics as 'why does spacetime have 3+1 dimensions', discussed mainly through anthropic selection arguments - arguments that only worlds like ours can contain observers to observe them. The present problem concerns the pair of demands Weyl left open, which together state one objective: to make the determination of the world's metric structure compelling in the rigor of its proof and in its explanatory grip on the actual parameters $(n,q)=(4,1)$.

Problem Statement

Weyl's 1922 paper proves its uniqueness theorem classically and closes with two demands left open, which together state one objective: to make the determination of the world's metric structure compelling - rigorous by the strictest standards Weyl accepted, and explanatory of the actual world's parameters. The problem is to satisfy both demands.

(1) Constructive rigor of the uniqueness theorem. Weyl's theorem - under the postulates of the infinitesimal, relativistic problem of space, the group of infinitesimal rotations of a measure determination is, up to conjugation, the special orthogonal group $SO(p,q)$ of a non-degenerate quadratic form, so that the measure determination is necessarily 'Pythagorean', with only the dimension $n=p+q$ and inertia index $q$ left open - was obtained using case distinctions that are not valid in the constructive continuum analysis Weyl demanded (in his own diagnosis, chiefly the case distinctions of elementary divisor theory and the implicit decisions whether real quantities vanish or not). Provide this theorem with a proof acceptable in a pinned constructive foundation (for example, Bishop-style constructive mathematics or intuitionistic logic), making explicit and justifying every reformulation the constructive setting requires; or else establish a principled impossibility result for an explicitly stated constructive formulation of the statement, fixed in advance rather than tailored to the negative result.

(2) Explanation of the actual dimension and inertia index. Explain why the actual world's measure determination has dimension 4 and inertia index 1, going beyond Weyl's provisional answers (the simplest integral invariant with the character of action, built from the length curvature of his gauge geometry, exists only in four dimensions; only inertia index 1 splits the infinitesimal cone into two mantles and thereby grounds the past/future distinction) and beyond the currently available anthropic selection arguments, which explicitly disclaim rigor. The explanation must determine both parameters, $n=4$ and $q=1$, from principles that are stated independently of the target values and motivated on grounds other than the outcome, rather than presupposing either as an observed input or leaving it undetermined. The principles must admit other dimensions and signatures as non-vacuous alternatives, so that neither target value is presupposed or encoded in the choice of principle; the argument must be able to exhibit at least one other parameter setting $(n,q)$ that the same principles fail to select. Arguments that invoke only the existence of the length-curvature action invariant, or only the two-mantle splitting of the infinitesimal cone, presuppose one of the target values and so do not determine $(n,q)$ independently of them.

The two deliverables are coupled by the source: both are required to place the Pythagorean measure determination of the actual world on the footing Weyl considered compelling. Establishing the constructive-rigor demand (1), by a constructive proof or by the principled impossibility result, does not by itself establish the determination of $(n,q)$.

References

1. Hermann Weyl, 'Die Einzigartigkeit der Pythagoreischen Maßbestimmung', Mathematische Zeitschrift 12 (1922), 114-146. DOI 10.1007/BF01482071. URL: https://doi.org/10.1007/BF01482071
2. Hermann Weyl, 'Über die neue Grundlagenkrise der Mathematik', Mathematische Zeitschrift 10 (1921), 39-79. DOI 10.1007/BF02102305. URL: https://doi.org/10.1007/BF02102305
3. Hermann Weyl, 'Raum, Zeit, Materie: Vorlesungen über allgemeine Relativitätstheorie' (4th edition), Springer, 1921. DOI 10.1007/978-3-662-02044-9. URL: https://doi.org/10.1007/978-3-662-02044-9
4. Hermann Weyl, 'Mathematische Analyse des Raumproblems: Vorlesungen, gehalten in Barcelona und Madrid', Springer, 1923. DOI 10.1007/978-3-642-92431-6. URL: https://doi.org/10.1007/978-3-642-92431-6
5. Hermann Weyl, 'Zur Charakterisierung der Drehungsgruppe', Mathematische Zeitschrift 17 (1923), 293-320. DOI 10.1007/bf01504349. URL: https://doi.org/10.1007/bf01504349
6. Erhard Scheibe, 'Über das Weylsche Raumproblem', Journal für die reine und angewandte Mathematik 197 (1957), 162-207. DOI 10.1515/crll.1957.197.162. URL: https://doi.org/10.1515/crll.1957.197.162
7. Detlef Laugwitz, 'Über eine Vermutung von Hermann Weyl zum Raumproblem', Archiv der Mathematik 9 (1958), 128-133. DOI 10.1007/BF02287072. URL: https://doi.org/10.1007/BF02287072
8. Wilhelm Klingenberg, 'Eine Kennzeichnung der Riemannschen sowie der Hermiteschen Mannigfaltigkeiten', Mathematische Zeitschrift 70 (1958), 300-309. DOI 10.1007/BF01558595. URL: https://doi.org/10.1007/BF01558595
9. Hans Freudenthal, 'Lie groups in the foundations of geometry', Advances in Mathematics 1 (1964), 145-190. DOI 10.1016/0001-8708(65)90038-1. URL: https://doi.org/10.1016/0001-8708(65)90038-1
10. Shoshichi Kobayashi and Tadashi Nagano, 'On a fundamental theorem of Weyl-Cartan on G-structures', Journal of the Mathematical Society of Japan 17 (1965), 84-101. DOI 10.2969/jmsj/01710084. URL: https://doi.org/10.2969/jmsj/01710084
11. Erhard Scholz, 'The problem of space in the light of relativity: the views of H. Weyl and E. Cartan' (2014). arXiv:1310.7334. URL: https://arxiv.org/abs/1310.7334
12. Erhard Scholz, 'The changing faces of the Problem of Space in the work of Hermann Weyl' (2016). arXiv:1608.04003. URL: https://arxiv.org/abs/1608.04003
13. Mark van Atten, 'Weyl and Intuitionistic Infinitesimals', in: Julien Bernard and Carlos Lobo (eds.), 'Weyl and the Problem of Space', Springer, 2019, 125-159. DOI 10.1007/978-3-030-11527-2_5; arXiv:1803.05014. URL: https://doi.org/10.1007/978-3-030-11527-2_5
14. Silvia De Bianchi, 'From the Problem of Space to the Epistemology of Science: Hermann Weyl's Reflection on the Dimensionality of the World', in: Julien Bernard and Carlos Lobo (eds.), 'Weyl and the Problem of Space', Springer, 2019, 189-209. DOI 10.1007/978-3-030-11527-2_7. URL: https://doi.org/10.1007/978-3-030-11527-2_7
15. Max Tegmark, 'On the dimensionality of spacetime', Classical and Quantum Gravity 14 (1997), L69-L75. DOI 10.1088/0264-9381/14/4/002; arXiv:gr-qc/9702052. URL: https://doi.org/10.1088/0264-9381/14/4/002
16. Taro Asuke, 'Formal frames and deformations of affine connections', Tohoku Mathematical Journal (2) 77 (2025), 269-297. DOI 10.2748/tmj.20230615; arXiv:2206.00336. URL: https://doi.org/10.2748/tmj.20230615