Roth's Theorem (Diophantine Approximation)¶
For each fixed real algebraic irrational and positive exponent margin, only finitely many rationals approximate it more closely than denominator power 2 plus that margin.
Core Idea¶
Roth's theorem says that for each fixed real algebraic irrational α and each fixed ε > 0, only finitely many distinct reduced rationals p/q with q > 0 satisfy \(0<|α-p/q|<q^{-(2+ε)}\). The theorem allows finitely many exceptionally close fractions; it gives no general method to list them or compute their largest denominator.[ref-f3b30dfce4a9][ref-2c827137270f]
The exponent 2 is the sharp boundary. The theorem needs a positive margin chosen before the rationals vary. Changing ε with q or setting ε = 0 does not give the same finite-exception conclusion.[^ref-f3b30dfce4a9]
Scope of Application¶
Use the theorem when a problem supplies a real algebraic irrational target, reduced rational approximants, and an error estimate at one fixed exponent above 2. It can show that certain Diophantine equations have finitely many integer solutions: solutions of \(x^3-2y^3=1\) with \(y\ne0\) yield rational approximants \(x/y\) to the algebraic irrational cube root of 2 with error below \(|y|^{-5/2}\). Roth rules out infinitely many such solutions; it does not supply their full list.[^ref-f3b30dfce4a9]
It can also support a transcendence proof by contradiction. The partial sums of \(ξ=\sum_{n\ge0}2^{-3^n}\) are distinct reduced rationals with denominator \(q_N=2^{3^N}\) and error below \(2q_N^{-3}\). The series is irrational because its binary digits have unbounded gaps between 1s. If it were algebraic, these fractions would violate Roth for the fixed margin ε = ½; hence it is transcendental.[^ref-09826598cc1a]
Clarity¶
One unusually good fraction is consistent with the theorem. An infinite family of distinct fractions at one fixed positive exponent gap is the relevant contradiction. The theorem is a class-wide statement about algebraic irrationals, whereas an irrationality measure describes the approximability of an individual number. It does not say that every number with exponent 2 is algebraic.[ref-f3b30dfce4a9][ref-09826598cc1a]
Manages Complexity¶
The theorem compresses infinitely many possible denominator tests into one qualitative guarantee: fix the target and margin, compare each rational error with the denominator power, and conclude that only finitely many pass. This makes finiteness arguments possible without enumerating approximants. The bare theorem is ineffective for a general explicit largest denominator or number of exceptions; special estimates need additional work.[ref-f3b30dfce4a9][ref-2c827137270f]
Abstract Reasoning¶
For a forward use, start with a known algebraic irrational and show that infinitely many proposed objects would produce infinitely many too-close reduced fractions. For a contradiction use, first prove the target irrational, construct infinitely many distinct fractions meeting one fixed stronger-than-exponent-2 test, then reject the assumption that the target is algebraic. The cubic equation and sparse series fill those two different roles. Neither finite example collection proves Roth's universal theorem.[ref-f3b30dfce4a9][ref-09826598cc1a]
Knowledge Transfer¶
The same reasoning transfers among number-theoretic settings when the real algebraic target, rational denominator, ordinary absolute error, fixed positive margin and finite-versus-infinite conclusion remain literal. A polynomial factorization can supply close fractions in an equation; truncation tails can supply them in a series. Describing another process as “hard to approximate” is only an analogy unless these arithmetic roles are present. This named theorem is a proved mathematical statement; its proposed parent is Formal theorem, whose broader identity does not depend on Roth's specific exponent boundary.
Relationships to Other Abstractions¶
Current abstraction Roth's Theorem (Diophantine Approximation) Domain-specific
Parents (1) — more general patterns this builds on
-
Roth's Theorem (Diophantine Approximation) is a kind of Formal theorem Domain-specific
Roth's Theorem is a proved mathematical statement with a specific algebraic rational-approximation finiteness claim.
Hierarchy path (1) — routes to 1 parentless root
- Roth's Theorem (Diophantine Approximation) → Formal theorem
Neighborhood in Abstraction Space¶
Roth's Theorem (Diophantine Approximation) sits in a sparse region of the domain-specific corpus (77th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Number-Theoretic Properties & Tests (22 abstractions)
Nearest neighbors
- Arakelian's Theorem — 0.83
- Trial Division — 0.83
- Average Order of an Arithmetic Function — 0.83
- Least Common Multiple — 0.82
- Arithmetic–Geometric Mean — 0.82
Computed from structural-signature embeddings · 2026-10-17
References¶
[^ref-f3b30dfce4a9]: K. F. Roth, “Rational Approximations to Algebraic Numbers,” Mathematika 2, no. 1 (1955): 1–20, DOI 10.1112/S0025579300000644, original published article PDF, especially printed pp. 1–2 for inequality, theorem, sharpness and homogeneous-form corollary. The full article was retained, but its general proof was not independently reproduced.
[^ref-09826598cc1a]: Michel Waldschmidt, “Words and Transcendence,” in Analytic Number Theory (2009), 449–470, author-hosted typeset chapter, §4.1, printed pp. 463–464, Theorem 4.2 and the explicit Σg(−3n) example. This is an author exposition of the theorem and example, not a second independent proof of Roth.
[^ref-2c827137270f]: Michael A. Bennett, “Effective Measures of Irrationality for Certain Algebraic Numbers,” Journal of the Australian Mathematical Society (Series A) 62 (1997): 329–344, DOI 10.1017/S144678870000104X, original article PDF, especially introduction pp. 329–330 on the noncomputable Roth constant in that account and special effective bounds. No numerical bound from damaged OCR is used here.