---
title: Plany Kakeya Sets in Finite Fields
url: https://www.emergentmind.com/topics/plany-kakeya-sets
type: topic
---

# Plany Kakeya Sets in Finite Fields

A plany Kakeya set is a Kakeya configuration endowed with a local planar incidence constraint. In the finite-field model, a set $K\subset \mathbb F_q^n$ is Kakeya if it contains one full $q$-point line in every direction, and it is called plany when, for every point $x\in K$, there exists an affine $2$-plane $\Pi_x$ containing $x$ such that every Kakeya line through $x$ lies entirely in $\Pi_x$. This hypothesis is much stronger than the ordinary Kakeya condition, and in $\mathbb F_q^4$ it yields the lower bound $|K|\ge C q^{10/3}$ by the planebrush method [2507.09605]. The subject belongs to the broader Kakeya program, whose planar Euclidean side includes boundary-optimal sets with $\operatorname{Vol}_2(K(\epsilon))=\Theta(1/|\log\epsilon|)$ [1207.6389].

## 1. Classical Kakeya sets and the finite-field model

In the Euclidean plane, a Kakeya set is a planar set containing a unit-length line segment in every direction. A central refinement asks how small the area of the $\epsilon$-neighborhood
$$
K(\epsilon):=\{x\in \mathbb R^2:\operatorname{dist}(x,K)<\epsilon\}
$$
can be as $\epsilon\to 0$. The lower bound $\operatorname{Vol}_2(K(\epsilon))\ge c/|\log\epsilon|$ and constructions with $\operatorname{Vol}_2(K(\epsilon))=O(1/|\log\epsilon|)$ single out the boundary-optimal regime $\Theta(1/|\log\epsilon|)$ [1207.6389].

The finite-field analogue replaces segments by affine lines. A line in $\mathbb F_q^n$ is an affine translate of a $1$-dimensional subspace,
$$
\ell=\{a+\lambda v:\lambda\in\mathbb F_q\},
$$
where $a\in\mathbb F_q^n$ and $v\in \mathbb F_q^n\setminus\{0\}$. Directions are the $1$-dimensional subspaces, i.e. the projective space $P\mathbb F_q^n$. A finite-field Kakeya set $K\subset \mathbb F_q^n$ contains at least one line in every direction [2507.09605].

Within this framework, the adjective “plany” isolates a structured subcase in which the local bush of lines through each point is confined to a single affine $2$-plane. The resulting incidence geometry is the central input behind the four-dimensional planebrush bound.

## 2. The plany hypothesis in $\mathbb F_q^4$

Let $L$ be the family of Kakeya lines whose union is $K$. The family $L$, and hence $K$, is called plany if for every point $x\in K$ there exists a $2$-plane $\Pi_x\subset \mathbb F_q^n$ containing $x$ such that every line of $L$ passing through $x$ lies entirely in $\Pi_x$ [2507.09605].

This condition can be read as a pointwise restriction on incidence patterns. In a general Kakeya set, multiple lines through a point may spread across many transverse directions. In a plany set, the full bush through $x$ is trapped inside one affine plane. The paper presenting the finite-field planebrush method emphasizes that this is much stronger than the unstructured Kakeya condition and is precisely what allows a gain beyond the three-dimensional hairbrush exponent [2507.09605].

In $\mathbb F_q^4$, a Kakeya set carries one line in each of the $q^3+q^2+q+1$ directions. The main theorem states that if such a set is plany, then there is an absolute constant $C>0$, independent of $q$, such that
$$
|K|\ge C q^{10/3}.
$$
Equivalently, one sometimes writes $\dim(K)\ge 10/3$, since $q^{10/3}$ is of order $q^{3.333\ldots}$ [2507.09605].

The theorem is quantitative rather than merely structural: it does not classify plany Kakeya sets, but it forces any such set in four dimensions to occupy substantially more than $q^3$ points.

## 3. The combinatorial estimates behind the $10/3$ exponent

The planebrush argument rests on two standard combinatorial inputs and one new four-dimensional estimate. The first input is Córdoba’s union-of-lines estimate: if $\{A_i\}_{i=1}^N$ are finite sets with $|A_i\cap A_j|\le 1$ for $i\ne j$, then
$$
\left|\bigcup_{i=1}^N A_i\right|
\ge
\sum_{i=1}^N |A_i|-\frac{N(N-1)}2.
$$
In particular, if $N\le 2q$ and every $A_i$ has size at least $q$, then
$$
\left|\bigcup_i A_i\right|\gtrsim Nq.
$$
This estimate is the finite-field “union-of-lines” lemma [2507.09605].

The second input is Wolff’s hairbrush bound in $\mathbb F_q^3$. If $L$ is a collection of distinct lines in $\mathbb F_q^3$, with no more than $2q$ of them in any single $2$-plane and with $|L|\le 3q^2$, then
$$
\left|\bigcup_{\ell\in L}\ell\right|\gtrsim |L|\,q^{1/2}.
$$
Its geometric mechanism is the classical hairbrush picture: select a stem line meeting many others, decompose by the $2$-planes through that stem, and apply Córdoba’s estimate in each slice [2507.09605].

The new ingredient is the planebrush lemma in $\mathbb F_q^4$. If $L$ is a plany family of lines in $\mathbb F_q^4$ such that at most $2q$ lie in any common $2$-plane, at most $3q^2$ lie in any common $3$-plane, and $|L|\le 4q^3$, then
$$
\left|\bigcup_{\ell\in L}\ell\right|\gtrsim |L|\,q^{1/3}.
$$
For the full Kakeya line set of a plany Kakeya set, these axioms are automatic. Combining them with the number of directions gives the lower bound of order $q^{10/3}$ [2507.09605].

The shift from the hairbrush gain $q^{1/2}$ in $\mathbb F_q^3$ to the planebrush gain $q^{1/3}$ in $\mathbb F_q^4$ is not a deterioration but a dimensional tradeoff: the ambient family contains on the order of $q^3$ directions, so $|L|q^{1/3}$ is exactly the scale needed to reach the exponent $10/3$.

## 4. Proof architecture of the planebrush method

The proof begins with multiplicity trimming. Let
$$
X=\bigcup_{\ell\in L}\ell,
\qquad
\mu(p)=\#\{\ell\in L:p\in \ell\}.
$$
The average multiplicity is $q|L|/|X|$. One discards points with multiplicity much smaller than this average, losing at most $1\%$ of each line’s points, and calls the remaining set $X'$ [2507.09605].

A base point $x_1\in X'$ is then selected by pigeonholing so that at least half the lines through $x_1$ still carry at least $q/2$ points of $X'$. By planiness, all lines through $x_1$ lie in a single $2$-plane $\Pi_{x_1}$. One next defines the planebrush $L_1$ as the subfamily of all lines of $L$ that meet or are parallel to $\Pi_{x_1}$. A second pigeonhole argument yields
$$
|L_1|\gtrsim q^3 |L|^2/|X|^2.
$$
This is the key enlargement step: the local planar bush through $x_1$ controls a large global subfamily [2507.09605].

The argument then splits into two cases.

In the first case, at least $50\%$ of the incidence pairs $(p,\ell)$ with $\ell\in L_1$ are unique inside $L_1$. A direct count gives
$$
|X|\ge |X'|\ge \tfrac12 q|L_1|
\gtrsim q^4 |L|^2/|X|^2,
$$
hence
$$
|X|\gtrsim |L|^{2/3}q^{4/3}\gtrsim |L|q^{1/3}.
$$
This is already the required bound [2507.09605].

In the second case, at least $50\%$ of the incidences come from points $p\in P_1'$ lying on at least two lines of $L_1$. The geometry of the planebrush implies that no line outside $L_1$ can pass through such a point, so $P_1'$ is disjoint from the union of $L\setminus L_1$. One then keeps only those lines of $L_1$ meeting $P_1'$ in at least $q/200$ points, obtaining a large subfamily $L_1'$. Foliate $\mathbb F_q^4$ by $3$-spaces containing $\Pi_{x_1}$, apply the three-dimensional hairbrush bound in each $3$-space to the lines of $L_1'$ inside it, and use Córdoba’s lemma for the lines lying wholly in $\Pi_{x_1}$. This yields
$$
|P_1'|\gtrsim |L_1'|q^{1/2}\gtrsim |L_1|q^{1/2}\gtrsim |L|q^{1/3}.
$$
The remaining lines outside $L_1$ are then handled by induction on $|L|$ [2507.09605].

The proof is therefore neither a purely local incidence argument nor a direct global counting lemma. Its central mechanism is the extraction of a large subfamily organized around one distinguished affine $2$-plane.

## 5. Geometric intuition and the Euclidean comparison

In the unstructured finite-field Kakeya problem, the only universal information is that no two Kakeya lines share a direction. The summary exposition emphasizes that this yields the trivial bound $|K|\gtrsim q^2$ by Córdoba and, in dimensions $n\ge 3$, the hairbrush-scale bound $\gtrsim q^{(n+2)/2}$ [2507.09605].

The plany hypothesis adds a local flatness principle. At every point $x$, the local fan of lines lies in a plane $\Pi_x$. Geometrically, this means that nearby incidences can be organized around common flats, so the proof can aggregate line families not merely around a stem line, as in a hairbrush, but around a distinguished $2$-plane, hence the term “planebrush” [2507.09605].

This same plany regime appears in the Euclidean four-dimensional theory. Katz and Zahl used a planebrush argument to prove that Kakeya sets in $\mathbb R^4$ have Hausdorff dimension at least $3.059$, and in the special plany case their argument gives the stronger lower bound $10/3$. The finite-field treatment is presented as a nontechnical model of that mechanism: it omits the real-analysis technicalities involving tubes, $\epsilon$-removal, and multilinear restriction, but recovers exactly the exponent $10/3$ [2507.09605].

A plausible implication is that the finite-field planebrush result should be read less as an isolated counting theorem than as a structural model for how planar concentration of incidences can force additional volume in four-dimensional Kakeya problems.

## 6. Relation to other Kakeya variants

A common source of confusion is that several distinct notions carry the name “Kakeya,” but they address different geometric and analytic questions.

The planar “Kakeya property” studied by Csörnyei, Héra, and Laczkovich concerns rigid motions rather than incidence in all directions: a set $A\subset \mathbb R^2$ has property $(K)$ if it can be continuously moved to a different position within a set of arbitrarily small area. For closed sets with this property, the union of the nontrivial connected components can be covered by a null union of parallel lines or a null union of concentric circles; in particular, a closed connected set with property $(K)$ lies on a single line or a single circle [1802.00286]. This is a classification theorem for sweepable planar sets, not a plany incidence theorem.

Curved Kakeya sets replace lines by curved families. Yang and Zhong construct a compact set in $\mathbb R^2$ of measure $0$ containing a piece of a parabola of every aperture between $1$ and $2$, and generalize the construction to suitable $C^2$ families satisfying a cinematic-curvature condition [2408.01917]. The geometry is governed by “cut-and-slide” tangency compression rather than pointwise planiness.

Directional Kakeya-type sets can also be formulated via lacunarity. Kroc and Pramanik define finite-order lacunarity for direction sets in $\mathbb R^{d+1}$ and show, in the planar case, that a direction set is sublacunary if and only if it admits Kakeya-type sets, equivalently if and only if the associated directional maximal operators are unbounded on every $L^p$, $1\le p<\infty$ [1404.6241]. This is a characterization of direction sets, not a four-dimensional planebrush phenomenon.

Finite affine-plane Kakeya sets form yet another branch. In an affine plane of order $q$, a Kakeya set is the union of $q+1$ pairwise non-parallel lines. De Boeck and Van de Voorde show that Kakeya sets of size asymptotically at least
$$
q^2-q\sqrt q+\tfrac32 q+o(q)
$$
contain a large knot, meaning a point lying on many of the defining lines [2003.08480]. Here the dominant issue is the upper end of the size spectrum and knot multiplicity, again distinct from the plany hypothesis in $\mathbb F_q^4$.

Taken together, these variants show that “Kakeya” is not a single problem but a family of tightly connected geometric regimes. Plany Kakeya sets occupy the regime where local planar concentration of line directions becomes strong enough to force the four-dimensional lower bound $q^{10/3}$, making the planebrush method the natural analogue of the hairbrush argument for structured incidence configurations [2507.09605].

Source: https://www.emergentmind.com/topics/plany-kakeya-sets