Two-Sine Small Divisor Inequality

• Two-sine small divisor inequality is a condition in Minkowski–Funk transforms that quantifies when the spectral multipliers approach zero.
• It employs asymptotic expansions and Diophantine approximation techniques to connect the behavior of sine function sums with Sobolev space mapping properties.
• The results show that for almost every irrational parameter, small divisors cause the inverse transform to fail in achieving optimal Sobolev regularity.

The two-sine small divisor inequality arises in the spectral analysis of generalized Minkowski–Funk transforms on spheres and lies at the intersection of harmonic analysis, spectral theory, and metric Diophantine approximation. For irrational radii parameters $t = \cos(\beta\pi)$, this framework leads to the study of when certain “small denominators” can occur in the spectral multipliers governing the mapping properties of the inverse transform on Sobolev spaces. The inequality quantifies the proximity of weighted sums of two sine functions (with parameters depending on geometric and transform parameters) to zero, and its metric theory determines, for almost every $\beta$, the global behavior of the Sobolev mapping properties of the inverse operator.

1. Formal Statement and Origins

Let $n \ge 2$, $\alpha \in \mathbb{C}$, and define

$$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$

The generalized Minkowski–Funk transform $M_t^\alpha$ acting on functions on $\mathbb{S}^n$ leads to the study of its spectral multipliers $\mathfrak{m}^\alpha_t(j)$, whose asymptotics as $j\to\infty$ are given by

$$j^{\rho + 1}\, \mathfrak{m}^\alpha_t(j) = c_\beta (1 + O(j^{-1})) \Big[ j\sin(\pi(j\beta - r_0)) + \rho(\rho-1)\sin(\pi(j\beta - r_1)) + O(j^{-1}) \Big],$$

with shifts

$\beta$0

and $\beta$1. The core analytical object is the two-sine sum

$\beta$2

The two-sine small divisor inequality is then

$\beta$3

for some fixed $\beta$4 and infinitely many $\beta$5.

2. Metric Diophantine Properties and Full Measure Results

For irrational $\beta$6, the main metric theorem establishes that for Lebesgue-almost every $\beta$7 (i.e., on a full-measure subset),

$\beta$8

has infinitely many solutions $\beta$9 for any $n \ge 2$0 (Han et al., 14 Jan 2026). No Diophantine restrictions beyond irrationality are imposed, confirming that these small divisors occur generically in the parameter space.

The corresponding operator-theoretic corollary asserts that for almost every irrational $n \ge 2$1, the inverse $n \ge 2$2 fails to be bounded from $n \ge 2$3 to $n \ge 2$4. Thus, for generic irrational radii, the loss of regularity is dictated sharply by the two-sine inequality.

3. Proof Outline and Diophantine Reformulation

The proof strategy begins with the asymptotic expansion for the spectral multipliers, reducing the boundedness analysis to the behavior of $n \ge 2$5. Specifically, the question is whether $n \ge 2$6 can be uniformly bounded away from zero; the main result shows this is impossible for typical $n \ge 2$7.

A crucial analytic reduction rewrites the two-sine small divisor condition as a moving-target inhomogeneous Diophantine approximation problem: $n \ge 2$8 for infinitely many $n \ge 2$9 with $\alpha \in \mathbb{C}$0, $\alpha \in \mathbb{C}$1, and $\alpha \in \mathbb{C}$2 depending on parameters [(Han et al., 14 Jan 2026), Lemma 3.1]. Through this reduction, the problem is situated in a variant of Khintchine’s theorem, with restricted denominators and moving inhomogeneous shifts. The final metric result relies on verifying the technical smallness and periodicity hypotheses for the shifts, after which a two-parameter Khintchine-type theorem yields the desired full-measure conclusion.

4. Critical Cases and Endpoint Regularity Conjectures

For the critical cases $\alpha \in \mathbb{C}$3, the asymptotic formula for the spectral multipliers collapses to expressions involving a single term, and the small divisor inequality reduces to classical inhomogeneous Diophantine inequalities of Khintchine type. In these endpoints, the inequalities to be solved are of the form

$\alpha \in \mathbb{C}$4

and similarly for shifts without the $\alpha \in \mathbb{C}$5, all with exponent $\alpha \in \mathbb{C}$6. Rubin’s Conjectures 4.4 and 4.7 posited that the inverse transform cannot be bounded when $\alpha \in \mathbb{C}$7. Han and Rahimi confirm that, for almost every $\alpha \in \mathbb{C}$8, each such inequality admits infinitely many solutions, thus establishing the critical failure of endpoint Sobolev regularity for $\alpha \in \mathbb{C}$9 [(Han et al., 14 Jan 2026), Theorem 2.8].

5. Key Formulas and Technical Estimates

A summary of core analytic expressions central to the theory:

Expression	Formula	Relevance
Spectral asymptotics	$$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$0 (truncated)	Multipliers’ decay
Two-sine function	$$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$1	Small divisor analysis
Moving-target reformulation	$$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$2	Diophantine reduction

The analysis relies on metric theory for the divergence of $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$3 with $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$4, ensuring the limsup set for approximations has full measure and, by extension, small divisors occur abundantly.

6. Consequences for Sobolev Mapping and Resolution of Rubin’s Problems

The occurrence of small values of $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$5 translates directly into the failure of uniform lower bounds for the high-frequency spectral multipliers of $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$6, thereby obstructing the expected gain in smoothness for the inverse operator on the natural Sobolev scale. In general, the inverse fails to map $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$7 to $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$8 for almost every irrational $$\rho = \alpha + \tfrac{n-1}{2} \neq 0,1, \quad \theta = \beta\pi, \quad t = \cos\theta, \quad 0 < \beta < 1, \;\; \beta \neq \tfrac12.$$9, excluding the possibility of sharp Sobolev regularity improvement. In the endpoint cases $M_t^\alpha$0, the analysis confirms that even the minimal gain is unattainable for a set of $M_t^\alpha$1 of full measure, completing the metric analysis posed by Rubin’s small denominator conjectures (Problems 3.8, Conjectures 4.4 and 4.7) for generalized Minkowski–Funk transforms (Han et al., 14 Jan 2026).

7. Broader Significance and Theoretical Implications

The two-sine small divisor inequality exemplifies the intricate interplay between harmonic analysis and Diophantine approximation in the study of inversion problems for integral transforms with oscillatory spectral data. Its resolution for Minkowski–Funk transforms provides a definitive regularity threshold in terms of Sobolev exponents, precisely characterizing when bounded inversion is unattainable. This framework may be viewed as a model for similar small denominator phenomena in other contexts where operator spectra depend on irrational rotation parameters, suggesting broad applicability in spectral theory and related fields. The metric “full-measure” result underlines the inherent unavoidability of small denominators in generic settings, reinforcing the necessity of Diophantine techniques in harmonic analysis on homogeneous spaces (Han et al., 14 Jan 2026).