Poincaré's 1884 program on linear equations with algebraic coefficients: effective determination of the accessory parameters of the uniformizing second-order Fuchsian equation for general types, and the unproven convergence criterion for his series construction (Poincaré, 1884)

Poincaré's memoir « Sur les groupes des équations linéaires » (Acta Mathematica 4, 1884) reduced the integration of every linear ordinary differential equation with algebraic coefficients to the study of a second-order Fuchsian equation on an algebraic curve, and closed with six acknowledged gaps. Of these gaps, the following are resolved or subsumed: the general existence question became Hilbert's 21st problem, with Poincaré's own parameter count giving the negative answer for scalar Fuchsian equations, Plemelj's 1908 theorem giving existence for regular singularities, and Bolibruch's counterexample plus the Kostov–Bolibruch irreducible theorem settling Fuchsian systems; the promised series development of the uniform functions was delivered by Poincaré's immediately following zétafuchsiennes memoir; the « whole-plane » disposition is subsumed by elliptic-function theory and the 1907 uniformization theorem; the case-by-case continuity method was superseded by rigorous uniformization proofs. The surviving open core is the effective determination of the accessory parameters of the uniformizing second-order equation for a general (non-symmetric) Fuchsian type — treated explicitly in the modern literature only in special families, such as four-punctured spheres, genus-0 variational formulas, finitely many arithmetic groups, and adjacent torus and Heun-class cases — together with Poincaré's unproven sufficient condition for his own general construction: finiteness of an integral over the fundamental region of the associated subgroup, which implies the convergence of his series and corresponds, in modern terms, to convergence of the exponent-1 Poincaré series over that subgroup, a convergence guaranteed when its critical exponent is strictly below 1. By Poincaré's own remark, proving this would both certify his conditional algorithm and replace the case-by-case existence proof with a unified direct argument.

Background

In 1884 Henri Poincaré published « Sur les groupes des équations linéaires » (Acta Mathematica 4, 201–312), the last-but-one of his memoirs on automorphic functions, with the goal of integrating every linear differential equation with algebraic coefficients. The terminology: a Fuchsian group is a discrete group of Möbius transformations preserving a circle and acting properly discontinuously inside it; the quotient of the disk by such a group is a Riemann surface. A Fuchsian function is an automorphic function invariant under such a group, and Kleinian functions are the analogous functions for groups acting properly discontinuously on a domain of the sphere that is not a disk. Poincaré's earlier memoir (1882) had constructed these functions from theta-fuchsian series $\sum_\gamma H(\gamma z)(\gamma'(z))^{m}$ , whose convergence he proved for weight $m \ge 2$ . Given a linear equation with coefficients rational in $x$ and $y$ on an algebraic curve $\phi(x,y)=0$ , each singular point carries a determinantal equation (indicial equation) whose roots are the local exponents. A type is the data of the curve, the singular points, and the determinantal equations, with exponent differences equal to zero or to an aliquot part of unity (a reciprocal of an integer) — the condition for a single-valued uniformizing ratio. A point where the coefficients are singular but all solutions remain single-valued is an apparent singularity. By a parameter count Poincaré shows that a prescribed group cannot in general be realized by an equation of order greater than two free of apparent singularities, so he restricts to the second-order equation $d^{2}v/dx^{2}=\theta(x)v$ with $\theta$ rational on the curve; the coefficients of $\theta$ not fixed by the local exponents are the accessory parameters, and their determination is the memoir's « third problem ». His main theorem: for every Fuchsian type the accessory parameters can be disposed so that the ratio $z$ of two integrals makes $x$ and $y$ single-valued (uniform) functions of $z$ ; consequently the integrals of any linear equation with algebraic coefficients are uniform functions of $z$ . The existence proof is by the continuity method — deforming types and showing the realized set is simultaneously open and closed — applied first to symmetric types (those invariant under an anti-conformal reflection) and extended case by case. For symmetric types Poincaré computes the accessory coefficients with unlimited approximation by successive reflections of a fundamental polygon (§XVII). For general types (§XVIII) he reduces the computation to a logarithmic potential series, formula (7), summed over an explicitly constructed subgroup $g$ of a Fuchsian group, whose fundamental region $Q_0$ has infinitely many sides; the convergence of this series is not established, and he reduces it to the finiteness of an integral, formula (11), over $Q_0$ . The memoir's conclusion (§XX) states the disposition theorem in three cases — the disk-Fuchsian disposition (one and only one way), the Kleinian disposition not extending to the whole plane (infinitely many ways), and the Fuchsian-or-Kleinian disposition existing in the whole plane (one and only one way, explicitly left undemonstrated) — and ends by deferring the series development of the new functions to a promised final memoir. The modern theory of Fuchsian groups equips such sums with a finer invariant: the critical exponent of a group $g$ is the infimum of the exponents $s$ for which the Poincaré series $\sum_{\gamma\in g}|\gamma'(z)|^{s}$ converges, so Poincaré's series (7) over $g$ is controlled by whether the critical exponent of $g$ is strictly below 1.

Problem Statement

The surviving open core of Poincaré's 1884 program, after the resolutions summarized above, is the following two-part problem. Primary objective: given a general Fuchsian type — an algebraic curve $\phi(x,y)=0$ of unrestricted genus with prescribed singular points and determinantal equations whose exponent differences are zero or aliquot parts of unity, with no symmetry assumption — determine the accessory parameters of the second-order equation $d^{2}v/dx^{2}=\theta(x)v$ in that type in effective form: an unconditional procedure admitting unlimited approximation (or an explicit formula) that produces the unique disposition for which the ratio $z$ of two integrals makes $x$ and $y$ automorphic functions of $z$ (the disk-Fuchsian disposition), together with a proof of correctness for all types. Poincaré achieved this only for symmetric types (§XVII, by successive polygon reflections); his general-case reduction (§XVIII) expresses the coefficients through the logarithmic series (7) over the explicitly constructed subgroup $g$ but is conditional on the unproven convergence of that series, and the modern literature supplies effective methods only in special families (genus-0 curves with few punctures, finitely many arithmetic groups, and adjacent torus and Heun-class cases), so the general case is open. Secondary objective (independent analytic core): prove, or refute, Poincaré's sufficient condition — that the integral (11) over the fundamental region $Q_0$ of the subgroup $g$ attached to the type by the construction of §§XVIII–XIX of the memoir is finite, so that his series (7), an exponent-1 Poincaré series over $g$ , converges — for the subgroups arising in the construction; in modern terms such convergence is guaranteed when the critical exponent of $g$ is strictly below 1 and impossible once it strictly exceeds 1, so the question is where the subgroups of the construction lie. A proof certifies Poincaré's conditional algorithm unconditionally and, by his own remark, replaces the case-by-case continuity method with a unified direct existence proof; a refutation (a type for which the integral (11) is infinite, i.e. for which the series (7) diverges) would show that his series construction cannot be salvaged as such and would delimit where effective determination must proceed by other means. Explicitly out of scope, as resolved or subsumed: existence of realizing equations in the monodromy sense (Hilbert's 21st problem and its answers), the whole-plane disposition (uniformization), and the series expansion of the uniform functions (zétafuchsiennes memoir). The two parts may be attacked independently, and a complete answer to either resolves the task: an effective determination of the accessory parameters of a general type together with its correctness proof, or a decision, affirmative or negative, on the convergence criterion. A proof of the secondary also discharges the conditional step of the primary's general-case reduction and therefore answers both parts; a refutation of the secondary answers the secondary alone and leaves the primary's effective determination to other means.

References

H. Poincaré (1884). Sur les groupes des équations linéaires. Acta Mathematica 4, 201–312. DOI 10.1007/BF02418420. https://doi.org/10.1007/BF02418420
H. Poincaré (1884). Mémoire sur les fonctions zétafuchsiennes. Acta Mathematica 5, 209–278. DOI 10.1007/BF02421560. https://doi.org/10.1007/BF02421560
H. Poincaré (1882). Mémoire sur les fonctions fuchsiennes. Acta Mathematica 1, 193–294. DOI 10.1007/BF02592135. https://doi.org/10.1007/BF02592135
H. Poincaré (1916). Œuvres de Henri Poincaré, tome II. Gauthier-Villars, Paris. (Source edition of both 1884 memoirs, pp. 300–462; full text at https://henripoincare.fr/files/original/03d1277ebc8ed5a59bb1650cef698fad15399710.pdf, scan also archived as Internet Archive item https://archive.org/details/uvresdehenripoin02poin)
J. Plemelj (1908). Riemannsche Funktionenscharen mit gegebener Monodromiegruppe. Monatshefte für Mathematik und Physik 19, 211–245. DOI 10.1007/BF01736697. https://doi.org/10.1007/bf01736697
J. Gray (1986). Linear Differential Equations and Group Theory from Riemann to Poincaré. Birkhäuser Boston. DOI 10.1007/978-1-4899-6672-8. https://doi.org/10.1007/978-1-4899-6672-8
D. V. Anosov, A. A. Bolibruch (1994). The Riemann-Hilbert Problem. Aspects of Mathematics, Vieweg+Teubner. DOI 10.1007/978-3-322-92909-9. https://doi.org/10.1007/978-3-322-92909-9
A. A. Bolibruch (1995). The Riemann-Hilbert Problem and Fuchsian Differential Equations on the Riemann Sphere. In: Proceedings of the International Congress of Mathematicians (Zürich, 1994), 1159–1168. DOI 10.1007/978-3-0348-9078-6_109. https://doi.org/10.1007/978-3-0348-9078-6_109
H.-P. de Saint-Gervais (2016). Uniformization of Riemann Surfaces: Revisiting a hundred-year-old theorem. Heritage of European Mathematics, European Mathematical Society. ISBN 9783037191453. DOI 10.4171/145. https://doi.org/10.4171/145
T. Bothner (2021). On the origins of Riemann–Hilbert problems in mathematics. Nonlinearity 34, R1–R73. DOI 10.1088/1361-6544/abb543. https://doi.org/10.1088/1361-6544/abb543 (preprint arXiv:2003.14374, https://arxiv.org/abs/2003.14374)
R. R. Gontsov, V. A. Poberezhnyi (2008). Various versions of the Riemann-Hilbert problem for linear differential equations. Russian Mathematical Surveys 63(4), 603–639. DOI 10.1070/rm2008v063n04abeh004547. https://doi.org/10.1070/rm2008v063n04abeh004547
P. G. Zograf, L. A. Takhtadzhyan (1988). On Liouville's equation, accessory parameters, and the geometry of Teichmüller space for Riemann surfaces of genus 0. Mathematics of the USSR-Sbornik 60(1), 143–161. DOI 10.1070/sm1988v060n01abeh003160. https://doi.org/10.1070/sm1988v060n01abeh003160
J. Hofmann (2012). Uniformizing differential equations of arithmetic (1;e)-groups. arXiv:1205.5680. https://arxiv.org/abs/1205.5680
G. Bogo (2022). Accessory Parameters for Four-Punctured Spheres. SIGMA 18, 024. DOI 10.3842/SIGMA.2022.024. https://doi.org/10.3842/SIGMA.2022.024 (preprint arXiv:2004.02971, https://arxiv.org/abs/2004.02971)
T. Moy (2025). Joyce structures from quadratic differentials on the sphere. arXiv:2509.05275. https://arxiv.org/abs/2509.05275
P. Menotti (2012). Accessory parameters for Liouville theory on the torus. Journal of High Energy Physics 2012(12), 001. DOI 10.1007/JHEP12(2012)001. https://doi.org/10.1007/jhep12(2012)001