Regular Radial Limit Set in Complex & Geometric Analysis

• Regular radial limit set is a concept highlighting a subset where controlled radial behavior enforces unique boundary limits across various mathematical contexts.
• It encapsulates the structural rigidity seen in complex analysis, inner-function theory, holomorphic foliations, and CAT(0) geometry through specific radial approach mechanisms.
• This concept is practically significant for proving boundary uniqueness, algebraic integrability, and global rigidity in both analytic and geometric frameworks.

Searching arXiv for papers relevant to “regular radial limit set” and closely related usages of radial limit sets across several mathematical contexts. I’ll gather a small set of primary sources spanning the main meanings of radial limit sets: holomorphic foliations with radial Kupka sets, radial limits in complex analysis, inner functions, and CAT(0)/hyperbolic limit sets. “Regular radial limit set” is not a standardized term across the arXiv literature. Instead, it names a recurring pattern: a subset singled out by controlled radial behavior toward a boundary, a singular core, or a limit direction. In one-variable complex analysis it refers informally to subsets of the unit circle where radial limits exist, take prescribed values, or enforce boundary uniqueness; in inner-function theory it becomes a measure-theoretic set of boundary points where radial limits, and even minimal fine limits, exist; in holomorphic foliation theory the closest analogue is a compact connected Kupka set of radial transversal type; and in CAT(0) geometry the natural replacement is the horospherical limit set $\Sigma(M;A)$, which in geometrically finite hyperbolic cases coincides with the conical limit set (Danielyan, 2016, Ivrii et al., 2022, Calvo-Andrade, 2013, Bieri et al., 2013).

1. Terminological status and unifying scheme

Across these literatures, “radial” always refers to approach along distinguished rays, geodesics, or normal directions, while “regular” marks a subset on which the limiting behavior is constrained rather than arbitrary. The term is explicit in none of the cited papers as a universal definition; several sources state that it is “not a formal term” or “not a phrase explicitly used,” and then reconstruct its meaning from the ambient theory (Danielyan, 2016, Wu, 2013).

A common structural pattern emerges. One first specifies an ambient object: a holomorphic function on $\mathbb D$, an inner function of finite entropy, a codimension-one foliation on $\mathbb P^n$, or a group action on a proper CAT(0) space. One then defines a radial approach mechanism: $r\zeta$ in $\mathbb D$, thick approach regions at $\zeta\in\partial\mathbb D$, linear normal slices to a Kupka component, or horoballs/geodesic rays toward $\xi\in\partial_\infty M$. Finally, one isolates a subset on which radial behavior is rigid: existence of limits, uniqueness from boundary values, algebraic integrability, or module-theoretic control.

This suggests a family resemblance rather than a single invariant. In all cases, the regular radial set is the locus where local radial data determines global structure more strongly than generic boundary behavior would suggest.

2. Boundary sets in classical complex analysis

For analytic functions on the unit disc, the basic radial limit at $e^{i\theta}\in T$ is

$f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$

provided the limit exists. The associated radial limit set is

$$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$

with exceptional set $\mathbb D$0 (Danielyan, 2016).

In this setting, the regular radial limit set is controlled by classical theorems. Fatou’s theorem says that if $\mathbb D$1 is analytic and bounded on $\mathbb D$2, then the radial limit exists for almost all $\mathbb D$3. F. and M. Riesz show that if such an $\mathbb D$4 has radial boundary values $\mathbb D$5 on a set of positive Lebesgue measure, then $\mathbb D$6. For univalent functions, Beurling replaces measure by logarithmic capacity: radial limits exist outside a set of zero logarithmic capacity, and boundary values cannot vanish on a set of positive capacity for a nontrivial univalent function (Danielyan, 2016).

Danielyan’s central observation is that boundary uniqueness follows from large-set existence of radial limits. If a univalent or zero-free bounded analytic function has radial boundary values $\mathbb D$7 on a set $\mathbb D$8, then one can construct another function in the same class with no radial limit on $\mathbb D$9. In the zero-free bounded analytic case an explicit formula is

$\mathbb P^n$0

or, in the elementary variant, $\mathbb P^n$1 when $\mathbb P^n$2 and $\mathbb P^n$3 (Danielyan, 2016).

The dichotomy is sharp. On sets of positive measure, or positive capacity in the univalent case, radial behavior is rigid enough to force uniqueness. On measure-zero sets, the theory is maximally flexible: Privalov constructs zero-free bounded analytic functions tending to $\mathbb P^n$4 at every point of any null set, and Lusin constructs bounded analytic functions with no radial limit at any point of any null set (Danielyan, 2016).

3. Inner functions, finite entropy, and minimal fine regularity

For an inner function $\mathbb P^n$5 of finite entropy, the relevant boundary measure is not Lebesgue measure but the singular measure $\mathbb P^n$6 appearing in the singular inner factor of $\mathbb P^n$7. The radial regular set is

$\mathbb P^n$8

A stronger set is

$\mathbb P^n$9

with $r\zeta$0 (Ivrii et al., 2022).

Theorem 1.1 states that for $r\zeta$1-almost every $r\zeta$2, the radial limit $r\zeta$3 exists and lies in $r\zeta$4. Theorem 1.2 strengthens this: for $r\zeta$5-almost every $r\zeta$6, $r\zeta$7 has a thick limit, equivalently a minimal fine limit, at $r\zeta$8 (Ivrii et al., 2022). Thus the regular radial limit set is full with respect to the critical-boundary measure: $r\zeta$9

The proof is tied to the geometry of Beurling–Carleson sets and angular derivatives. The singular measure $\mathbb D$0 is supported on a countable union of Beurling–Carleson sets, and thickness at a boundary point is equivalent to a nonzero angular derivative for the corresponding Riemann map. The fundamental lemma

$\mathbb D$1

links critical structure to hyperbolic distortion and yields finite hyperbolic length of image radii at $\mathbb D$2-a.e. point (Ivrii et al., 2022).

This regular radial set is not merely descriptive. It is what makes the push-forward term $\mathbb D$3 in the singular value measure

$\mathbb D$4

well defined, and it is essential in proving continuity of $\mathbb D$5 under stable convergence in $\mathbb D$6 (Ivrii et al., 2022).

4. Radial cores in holomorphic foliations

In codimension-one holomorphic foliations on $\mathbb D$7, $\mathbb D$8, the closest analogue of a regular radial limit set is a compact connected Kupka set of radial transversal type. A foliation of Chern class $\mathbb D$9 is represented by

$\zeta\in\partial\mathbb D$0

with $\zeta\in\partial\mathbb D$1. Its Kupka set is

$\zeta\in\partial\mathbb D$2

and when $\zeta\in\partial\mathbb D$3 is compact, connected, and equals the codimension-two part of the singular set, it is locally a product with a planar singular foliation (Calvo-Andrade, 2013).

The radial transversal type is the linear form

$\zeta\in\partial\mathbb D$4

Locally near $\zeta\in\partial\mathbb D$5 there are coordinates $\zeta\in\partial\mathbb D$6 such that

$\zeta\in\partial\mathbb D$7

with $\zeta\in\partial\mathbb D$8 nowhere vanishing. Transversely, leaves are given by $\zeta\in\partial\mathbb D$9, so the Kupka set functions as a smooth invariant radial core (Calvo-Andrade, 2013).

The global rigidity is striking. Theorem 0.1 states that for $\xi\in\partial_\infty M$0 and $\xi\in\partial_\infty M$1, the set of foliations with compact connected Kupka set of radial transversal type is nonempty if and only if $\xi\in\partial_\infty M$2 is even. In that case there exist homogeneous polynomials

$\xi\in\partial_\infty M$3

such that

$\xi\in\partial_\infty M$4

Equivalently, the foliation has rational first integral

$\xi\in\partial_\infty M$5

and its leaves are fibers of a pencil of hypersurfaces (Calvo-Andrade, 2013).

The radial condition also determines moduli. The closure of $\xi\in\partial_\infty M$6 is exactly the rational component $\xi\in\partial_\infty M$7, so radial Kupka foliations form an open dense subset of an irreducible component of $\xi\in\partial_\infty M$8 (Calvo-Andrade, 2013). Here the “regular radial limit set” is not a boundary subset but a compact singular submanifold whose radial normal model forces global algebraic integrability.

5. CAT(0) spaces, conical points, and horospherical regularity

For a discrete group $\xi\in\partial_\infty M$9 acting by isometries on a proper CAT(0) space $e^{i\theta}\in T$0, the classical orbit limit set is

$e^{i\theta}\in T$1

Bieri–Geoghegan define the horospherical limit set of a finitely generated $e^{i\theta}\in T$2-module $e^{i\theta}\in T$3 by

$e^{i\theta}\in T$4

equivalently, the set of boundary points along which every $e^{i\theta}\in T$5 has representatives supported arbitrarily deep in every horoball at $e^{i\theta}\in T$6 (Bieri et al., 2013).

In hyperbolic terminology, a point $e^{i\theta}\in T$7 is conical if some sequence $e^{i\theta}\in T$8 stays at uniformly bounded distance from a geodesic ray $e^{i\theta}\in T$9. These are the standard radial limit points. The paper’s central hyperbolic identification is Theorem 10.10: if $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$0 is geometrically finite, then

$f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$1

is the set of all conical limit points of $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$2, and

$f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$3

In this setting, the regular radial limit set is precisely the conical part of the classical limit set (Bieri et al., 2013).

The same framework extends beyond rank one. For arithmetic lattices on higher-rank symmetric spaces, $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$4 is the complement of the rational spherical building at infinity. For abelian groups acting on Euclidean space, the complement of $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$5 is the radial projection of a tropical variety associated to $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$6 (Bieri et al., 2013). Regular radial directions are therefore those avoiding the building or tropical obstruction.

6. Related regimes and comparative scope

Other radial theories sharpen the contrast between regularity and maximal irregularity. For elliptic operators $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$7 on strictly starlike domains in $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$8, $f(e^{i\theta})=\lim_{r\to1^-}f(re^{i\theta}),$9, Gauthier–Shirazi show that if

$$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$0

and each $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$1 has empty fine interior, then for every continuous $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$2 there exists an $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$3-harmonic $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$4 such that

$$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$5

for all $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$6, while for almost every $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$7 the radial cluster set is maximal: $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$8 Here a prescribed regular radial set coexists with almost-everywhere wild radial behavior on the complement (Gauthier et al., 2022).

In function-space universality, the same opposition appears. For compact $$E_{\mathrm{rad}}(f)=\Big\{e^{i\theta}\in T:\lim_{r\to1^-}f(re^{i\theta})\text{ exists}\Big\},$$9 with $\mathbb D$00, dense $\mathbb D$01 subsets of $\mathbb D$02 contain functions whose values along a single subsequence $\mathbb D$03 satisfy

$\mathbb D$04

uniformly on $\mathbb D$05 for arbitrary $\mathbb D$06 and compact $\mathbb D$07 (Maronikolakis, 2021). This does not produce a regular radial set; rather, it shows that on sufficiently small sets radial data can be universally nonunique.

A comparable arithmetic version occurs for unilateral $\mathbb D$08-series

$\mathbb D$09

with periodic $\mathbb D$10 and polynomial $\mathbb D$11 of positive degree. Radial limits at roots of unity reduce to twisted coefficients $\mathbb D$12; when $\mathbb D$13 has mean zero, the radial limit exists and is given by a finite linear combination of the values of $\mathbb D$14 over one period (Kurşungöz, 2015). This is another precise instance of a regular radial limit set determined by arithmetic cancellation.

Taken together, these theories show that “regular radial limit set” is best read as a context-dependent designation for the locus where radial approach is controlled strongly enough to force uniqueness, integrability, or structural rigidity. The ambient category changes—from Hardy theory to inner functions, foliations, CAT(0) boundaries, elliptic PDE, and $\mathbb D$15-series—but the organizing principle remains the same: a radial geometry selects a subset on which limiting behavior is not merely existent, but structurally decisive.