Papex

Open problem · Mathematics · Open

Irrationality of Euler’s constant

Is Euler’s constant γ ≈ 0.5772 an irrational number?

Summary

Euler’s constant measures how far the harmonic series drifts from the natural logarithm and appears throughout analysis and number theory. It is not known to be irrational, let alone transcendental; if it were a fraction, its denominator would have to be enormous.

Why it is open

Jeffrey Lagarias's 2013 survey of Euler's constant states its irrationality as the first conjecture and calls it "a long-standing open problem" [1]. A stronger conjecture is that γ is not a period in the sense of Maxim Kontsevich and Don Zagier, that is, not an integral of a rational function with rational coefficients over a domain defined by polynomial inequalities with rational coefficients [1, 2]. Every algebraic number is a period, so this conjecture would make γ transcendental [1]. The irrationality of e^γ ≈ 1.781 is also unproved [1, 2].

The contrast with e and π shows part of the difficulty. Euler deduced in 1737 that e is irrational from the regular pattern of its continued fraction, and e and π were proved transcendental by Charles Hermite in 1873 and Ferdinand Lindemann in 1882 [1, 2]. The exponential function is an E-function, a class of power series introduced by Siegel in 1929, and π is a value of a G-function; transcendence results are usually associated with E-functions and irrationality results with G-functions [1, 8]. Euler's constant is known to be an "exponential period", given by integrals that involve exponentials, but its known integral representations "do not resolve the question whether γ is a period", and whether it has rational approximations of the G-function type is unknown [1]. Lagarias concludes that "Euler's constant appears sui generis" [1].

A resolution would be a proof that γ is irrational, or a fraction equal to it; no prize or body sets a further bar. This page's status follows Wikipedia's list of unsolved problems in mathematics, which on 15 October 2026 asked "Are Euler's constant γ and Catalan's constant G irrational? Are they transcendental?" [3].

What is known

Computation (1980–2023). In 1980 Richard Brent and Edwin McMillan computed γ and e^γ to 30,000 places and the start of their continued fractions, and proved that if either is a fraction its denominator exceeds 10^15000; the partial quotients they found agreed with the distribution expected for a random real number [1]. In 1998 Bruno Haible and Thomas Papanikolaou reported 475,006 partial quotients of the continued fraction of γ, which raised the bound for γ to 10^244663 [4]. On 29 April 2021 Syed Fahad computed 16,695,279,010 terms, as recorded in the On-Line Encyclopedia of Integer Sequences, whose entry gives no denominator bound from them [5]. As of 15 October 2026 the y-cruncher record list gave 1,337,000,000,000 decimal digits of γ (Andrew Sun, September 2023), against 314,000,000,000,000 for π [6].

Transcendence results (1968–2012). Mahler proved in 1968 that (π/2)·Y₀(2)/J₀(2) − γ is transcendental, where J₀ and Y₀ are Bessel functions [1]. Alexander Aptekarev was apparently the first to state explicitly, in a 2009 preprint, that at least one of γ and the Euler–Gompertz constant δ = ∫₀^∞ e^(−t)/(1 + t) dt ≈ 0.5963 is irrational; his preprint notes that this also follows from Andrei Shidlovskii's theorem on the algebraic independence of values of E-functions [7, 8]. In 2012 Tanguy Rivoal proved that the field generated by e, γ and δ has transcendence degree at least two, so that at least one of γ and δ is transcendental, and that the pair (γ, δ) cannot be approximated very well by fractions with a common denominator [1, 8]. Khodabakhsh and Tatiana Hessami Pilehrood obtained the transcendence statement at about the same time [1].

Families of constants (2010–2013). For the Euler–Lehmer constants γ(h, k), the analogues of γ for the integers congruent to h modulo k, M. Ram Murty and N. Saradha proved in 2010 that at most one is algebraic, and that if γ is algebraic the exception is γ(2, 4) = γ/4 [1]. In 2013 Murty and Anastasia Zaytseva proved that the generalised Euler constants of Diamond and Ford are all transcendental with at most one exception, which would be γ itself if γ were algebraic [1].

Approximations and criteria (2003–2013). In 2003 Jonathan Sondow, modifying F. Beukers's proof that ζ(3) is irrational, gave a criterion equivalent to the irrationality of γ, with computations for n < 2500 and n = 10,000 showing that a rational γ could not have a denominator dividing certain explicit integers [9]. In 2007 Aptekarev, A. I. Bogolyubskii, D. V. Khristoforov, V. G. Lysov and D. N. Tulyakov found integer sequences pₙ and qₙ, solutions of a third-order recurrence, with γ − pₙ/qₙ about −2π·e^(−2√(2n)) while qₙ grows roughly like (2n)! [1]. In 2013 the Hessami Pilehroods found approximations from a second-order recurrence and, from them, a continued fraction for γ (not a simple one) that they describe as the first non-trivial one converging sub-exponentially [10]. Lagarias writes that the 2013 approximations "are of insufficient quality to imply the irrationality" of γ [1].

Main approaches

Apéry-type approximations. Find integer linear forms in 1 and γ that tend to zero, as Apéry did for ζ(3) in 1979 [1]. Strongest evidence: explicit integer recurrences for γ found in 2007 and 2013 [1, 10]. Main difficulty: Lagarias calls finding approximations fast enough to certify irrationality "a challenge"; Apéry's approximations to ζ(2) and ζ(3) are of the G-function type, while for γ "the current expectation is that this is not possible" [1].

Irrationality criteria. Restate irrationality as a property of explicit integers or integrals that can be tested [9]. Strongest evidence: Sondow's checks for n < 2500 and n = 10,000 [9]. Main difficulty: his criterion holds if and only if γ is rational, so it restates the problem exactly, and checks cover only finitely many n [9].

Transcendence methods for E-functions. Use Shidlovskii's theory and Hermite–Padé approximants to the exponential and related functions [1, 8]. Strongest evidence: Rivoal's theorem on e, γ and δ [8]. Main difficulty: the results are disjunctions; they show that γ or δ is irrational, or transcendental, without saying which [7, 8].

Linear forms in logarithms. Write combinations of Euler–Lehmer constants as logarithms of integers and apply Baker's bounds for linear forms in logarithms [1]. Strongest evidence: the "at most one exception" theorems of 2010 and 2013 [1]. Main difficulty: the one exception that the method allows is where γ would sit if it were algebraic [1].

Open sub-questions

  1. A certified 2021 bound. What denominator bound for a rational γ follows from the 16,695,279,010-term continued fraction of 2021, once its partial quotients are certified correct, as the 1980 and 1998 expansions were [1, 4, 5]?
  2. Partial quotients at scale. Do the partial quotients of that expansion follow the distribution expected for a random real number as closely as the terms computed in 1980 did [1, 5]?
  3. Beyond the 2007 recurrence. Is there a recurrence with polynomial coefficients whose integer solutions satisfy qₙγ − pₙ → 0, beating the error e^(−2√(2n)) and the growth (2n)! of the 2007 approximations [1]?
  4. Separating γ from δ. Can the Hermite–Padé construction behind Rivoal's simultaneous approximation bound for (γ, δ) be modified to give linear forms in 1 and γ alone [8]?
  5. Sondow's inequality. Does Sondow's sufficient condition, that the fractional part of log S(n) exceeds 2^(−n), hold for every n up to 10⁶, extending the checks for n < 2500 and n = 10,000 [9]?

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.

Key references

  1. Lagarias, J. C. (2013). Euler’s constant: Euler’s work and modern developments. Bulletin of the American Mathematical Society 50(4), 527–628. 10.1090/S0273-0979-2013-01423-X · Checked 15 Oct 2026
  2. Waldschmidt, M. (2023). On Euler’s constant. Lecture slides, Sorbonne Université, Institut de Mathématiques de Jussieu, 2 March 2023. Source · Checked 15 Oct 2026
  3. Wikipedia. List of unsolved problems in mathematics (revision of 10 October 2026). Source · Checked 15 Oct 2026
  4. Haible, B. & Papanikolaou, T. (1998). Fast multiprecision evaluation of series of rational numbers. In Algorithmic Number Theory (ANTS-III), Lecture Notes in Computer Science 1423, Springer, 338–350. 10.1007/BFb0054873 · Checked 15 Oct 2026
  5. OEIS Foundation. A002852: Continued fraction for Euler's constant (or Euler-Mascheroni constant) gamma. The On-Line Encyclopedia of Integer Sequences, revision 103 (4 August 2026). Source · Checked 15 Oct 2026
  6. Yee, A. J. Records set by y-cruncher. numberworld.org. Source · Checked 15 Oct 2026
  7. Aptekarev, A. I. (2009). On linear forms containing the Euler constant. Preprint, arXiv:0902.1768 (version 2, 28 February 2009), not peer reviewed. Source · Checked 15 Oct 2026
  8. Rivoal, T. (2012). On the arithmetic nature of the values of the gamma function, Euler’s constant, and Gompertz’s constant. Michigan Mathematical Journal 61(2), 239–254. 10.1307/mmj/1339011525 · Checked 15 Oct 2026
  9. Sondow, J. (2003). Criteria for irrationality of Euler’s constant. Proceedings of the American Mathematical Society 131(11), 3335–3344. 10.1090/S0002-9939-03-07081-3 · Checked 15 Oct 2026
  10. Hessami Pilehrood, K. & Hessami Pilehrood, T. (2013). On a continued fraction expansion for Euler’s constant. Journal of Number Theory 133(2), 769–786. 10.1016/j.jnt.2012.08.016 · Checked 15 Oct 2026

Page reviewed 15 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in mathematics. Checked 08 Oct 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Records on this problem

No public record addresses this problem yet.

More in Mathematics