---
title: 'Weak Fan: Diverse Perspectives in Theory & Applications'
url: https://www.emergentmind.com/topics/weak-fan
type: topic
---

# Weak Fan: Diverse Perspectives in Theory & Applications

Searching arXiv for recent and relevant papers on "Weak Fan" across the senses present in the provided data.
{"query":"\"weak fan\" arXiv", "max_results": 10}
Using the arXiv search tool to retrieve papers mentioning "weak fan".
arxiv_search({"query":"\"weak fan\"","max_results":10})
In current research usage, **Weak Fan** is not a single universally fixed concept. The expression appears in at least four distinct technical settings: graph drawing, where it denotes the permissive version of fan-planarity; constructive and reverse mathematics, where it denotes WWKL-style weakenings of Brouwerian fan principles; Hodge-theoretic compactification, where it denotes a weak fan of nilpotent cones; and superconducting neuromorphic hardware, where it refers to historically poor fan-in or fan-out capability in SFQ logic [2308.08966] [1510.02141] [2507.14954] [2008.06409]. Across these settings, the common lexical element “fan” does not determine a common formal structure; the meaning is entirely domain-dependent.

## 1. Weak fan-planarity in graph drawing

In graph drawing, the relevant notion is **weak fan-planarity**. A drawing is weakly fan-planar if it avoids two forbidden crossing patterns but may contain a third. The underlying constraint is that if an edge \(e\) is crossed, then the crossing edges must form a **fan**: they all share a common endpoint, the **anchor** of \(e\), and cross \(e\) from the same side. The paper distinguishes three patterns. **Pattern (I)** forbids two edges crossing the same edge \(e\) from being non-adjacent. **Pattern (II)** requires two adjacent edges crossing \(e\) to cross from the same side. **Pattern (III)** is the additional configuration introduced in the later journal definition, and it is exactly the point at which weak and strong fan-planarity diverge [2308.08966].

The distinction is formalized as follows. **Weak fan-planarity** forbids Patterns (I) and (II), but allows Pattern (III). **Strong fan-planarity** forbids Patterns (I), (II), and (III). In the cell-based characterization, the crossing edges and the crossed edge \(e\) form two cells. In Pattern (II), one endpoint of \(e\) lies in the bounded cell. In Pattern (III), both endpoints of \(e\) lie in the bounded cell. Strong fan-planarity adds the requirement that both endpoints of \(e\) lie in the unbounded cell [2308.08966].

The inclusion is proper:
\[
\text{strongly fan-planar graphs} \subsetneq \text{weakly fan-planar graphs}.
\]
The separation is established by constructing a graph \(G\) with a weakly fan-planar drawing such that every weakly fan-planar drawing of \(G\) must contain at least one Pattern (III). The construction uses a **3-connected planar quadrangulation** \(G_0\), a gadget \(H\) inserted into every face of \(G_0\), and replacement of every red edge by a \(K_7\). This yields **Theorem 1**: there exists a weakly fan-planar graph that does not admit a strongly fan-planar drawing [2308.08966].

For density arguments, the paper introduces a **heart**, a triple \(e,e_\ell,e_r\) such that \(e_\ell\) and \(e_r\) share an endpoint \(u\), both cross \(e\), they realize Pattern (III), and the subcurve of \(e\) between the two crossings is uncrossed. **Lemma 2** states that if a weakly fan-planar drawing is not strongly fan-planar, then it contains a heart. The proof strategy then uses “valves” and a rerouting operation called **flipping a valve** to eliminate Pattern (III) configurations while controlling new crossings [2308.08966].

Despite the strict inclusion, the exact edge bounds coincide. **Theorem 2** states that a weakly fan-planar graph with \(n\) vertices has at most
\[
5n-10
\]
edges. **Theorem 3** states that an \(n\)-vertex bipartite weakly fan-planar graph has at most
\[
4n-12
\]
edges. The same bounds were already known for strong fan-planarity. The same paper also situates weak fan-planarity relative to **adjacency-crossing graphs**, in which only Pattern (I) is forbidden, yielding the proper inclusion chain
\[
\text{strongly fan-planar} \subsetneq \text{weakly fan-planar} \subsetneq \text{adjacency-crossing}.
\]
A common misconception in this literature is that weak and strong fan-planarity are merely terminological variants; the separation theorem shows that they define genuinely different graph classes, even though their upper density bounds coincide [2308.08966].

## 2. Weak Fan in constructive and reverse mathematics

In constructive and reverse mathematics, **Weak Fan** is used in a different sense. In “Separating the Fan Theorem and Its Weakenings,” the expression is introduced in the final “Questions” section as the **weak analogue** of each of the fan principles already under discussion, by explicit analogy with **Weak Weak König’s Lemma (WWKL)**. The motivating contrast is: whereas \(WKL\) states that any bar contains an entire level of \(2^*\), \(WWKL\) states that any bar contains **half of a level**. The paper therefore proposes the scheme
\[
\mathrm{Weak\ FAN}_A,\quad \mathrm{Weak\ FAN}_c,\quad \mathrm{Weak\ FAN}_{\Pi^0_1},\quad \mathrm{Weak\ FAN}_{\mathrm{full}},
\]
rather than a single theorem with one canonical formalization [1510.02141].

This formulation weakens the **conclusion**, not the complexity class of admissible bars. By contrast, the ordinary fan-theorem hierarchy is organized by bar complexity:
\[
\mathrm{FAN}_A,\quad \mathrm{FAN}_c,\quad \mathrm{FAN}_{\Pi^0_1},\quad \mathrm{FAN}_{\mathrm{full}}.
\]
The strongest principle is
\[
\mathrm{FAN}_{\mathrm{full}}:\quad \text{Every bar is uniform,}
\]
where uniformity is expressed by
\[
\exists n\in\mathbb N\, \forall \alpha\in 2^{\mathbb N}\,\exists m<n\;(\alpha{\upharpoonright}m\in B).
\]
The weak versions retain the same bar classes but replace uniformity by the weaker WWKL-style largeness conclusion, described only as “contains half of a level” [1510.02141].

The implication structure given in the paper is straightforward in two directions. Any full fan principle implies its weak correlate, and weak principles inherit the usual downward hierarchy:
\[
\mathrm{Weak\ FAN}_{\mathrm{full}} \Rightarrow \mathrm{Weak\ FAN}_{\Pi^0_1} \Rightarrow \mathrm{Weak\ FAN}_c \Rightarrow \mathrm{Weak\ FAN}_A.
\]
Likewise,
\[
\mathrm{FAN}_{\Pi^0_1}\Rightarrow \mathrm{Weak\ FAN}_{\Pi^0_1},
\]
and similarly for the other variants. What remains open in that paper is whether there are **diagonal implications**, such as between \(\mathrm{FAN}_c\) and \(\mathrm{Weak\ FAN}_{\Pi^0_1}\). The authors do not provide Kripke-model separation theorems specifically for the weak hierarchy [1510.02141].

The background constructive landscape is supplied by Brouwer’s fan theorem itself. In Berger’s overview, Brouwer’s fan theorem is the statement that **every detachable bar is a uniform bar** on the binary fan. This places the weak versions within a larger network of compactness and continuity principles, including
\[
WKL \Rightarrow FAN,\qquad UC \Rightarrow FAN,\qquad MUC \Longleftrightarrow FAN.
\]
In that setting, \(FAN\) is a compactness principle on Cantor space, \(UC\) is the uniform continuity theorem for pointwise continuous functions \(F:0,1\to\mathbb N\), and \(WKL\) is constructively equivalent to \(LLPO\) [2001.00064]. This suggests that “Weak Fan” in reverse mathematics is best understood as a measure-theoretic weakening of a compactness principle rather than as a weakened tree structure.

## 3. Approximate fans, almost-fans, and terminological instability

A related but not identical strand of the literature uses **weak fan principle** language for variants of Brouwerian compactness based on approximate or almost-fan structures. In “The Principle of Open Induction on Cantor space and the Approximate-Fan Theorem,” the **Approximate-Fan Theorem (AppFT)** is presented as a weak fan principle. An **approximate-fan-law** is a spread-law \(B\) such that, for each \(n\), the set of nodes of length \(n\) compatible with \(B\) is **bounded-in-number**. In the explicit version, there exists a function \(\gamma\) such that
\[
\forall n\, \bigl(\{t\in \mathbb{N}^n \mid B(t)=0\}\text{ has at most }\gamma(n)\text{ members}\bigr).
\]
The theorem is formulated as: in an explicit approximate fan, every thin bar is **almost-finite** [1408.2493].

The same paper places AppFT in a broad intuitionistic reverse-mathematical network. It states that
\[
\mathrm{BIM}\vdash \mathrm{AppFT}\rightarrow \mathrm{OI}([0,1]),
\]
that \(\mathrm{OI}(2^{\mathbb N})\) implies the Fan Theorem, and that the converse fails. It also gives equivalent or closely related formulations involving strong bars, the contrapositive of Bolzano–Weierstrass in \(\mathbb N^{\mathbb N}\), Ramsey-theoretic principles \(\mathrm{IRT}(k)\), and \(\mathrm{Asc}\), the contrapositive of Ascoli’s Lemma [1408.2493].

However, the terminology is not stable across authors. In “The Fan Theorem, its strong negation, and the determinacy of games,” the paper explicitly notes that **some authors have called FT the Weak Fan Theorem WFT**, but declines to adopt that usage. That work distinguishes \(FT\), \(FT^+\), \(AppFT\), \(ALMFAN\), and \(AlmFT\), and emphasizes that there is no single uncontested target for the phrase “weak fan” in this area [1311.6988].

This terminological instability is substantive rather than merely stylistic. One line of work treats weak fan principles as WWKL-style weakenings of the **conclusion** of a fan theorem [1510.02141]. Another treats approximate or almost-fan principles as weakening the **ambient branching hypothesis** while preserving a compactness-type conclusion [1408.2493]. A plausible implication is that the phrase “Weak Fan” in intuitionistic analysis should always be qualified by the exact scheme or structural assumption being used.

## 4. Weak fan structures in Hodge-theoretic compactification

In Hodge theory, a **weak fan** is a collection of nilpotent cones organizing boundary data for degenerations of period maps. In the explicit two-parameter K3 example, the relevant vector space is
\[
V=H^2(X,\mathbb Q),
\]
with cup-product pairing \(Q\) of signature \((3,19)\), and the period domain is the type IV domain
\[
\Omega = \{ [\omega] \in \mathbb{P}(V_{\mathbb{C}}) \mid Q(\omega,\omega)=0,\; Q(\omega,\overline{\omega})>0 \}.
\]
A two-parameter family over \((\Delta^*)^2\) produces commuting monodromy logarithms \(N_1,N_2\) with
\[
N_i^2=0,\qquad [N_1,N_2]=0,
\]
and associated cone
\[
\sigma=\mathbb{R}_{\ge 0}N_1+\mathbb{R}_{\ge 0}N_2.
\]
The weak fan is then
\[
\Sigma=\{\sigma,\rho_1,\rho_2,\{0\}\},
\]
where \(\rho_1=\mathbb{R}_{\ge 0}N_1\) and \(\rho_2=\mathbb{R}_{\ge 0}N_2\) [2507.14954].

The axioms emphasized in this example are **face closure**, **compatibility**, and a **weaker intersection condition** than in ordinary toric-fan theory. The faces of \(\sigma\) are exactly
\[
\{0\},\quad \rho_1,\quad \rho_2,\quad \sigma,
\]
and the weak fan is closed under taking faces. Each cone corresponds to a class of limiting mixed Hodge structures and to a boundary stratum: \(\rho_1\) and \(\rho_2\) correspond to one-parameter degenerations, while \(\sigma\) corresponds to the simultaneous degeneration \((t_1,t_2)\to(0,0)\) [2507.14954].

The reason the structure is called **weak** is not combinatorial weakness in the sense of sparse data, but the relaxation of the rigid toric intersection condition. The admissibility criterion is tailored to nilpotent orbit theory and compatibility of LMHS rather than to purely polyhedral geometry. In this setting, “weak fan” is therefore a Hodge-theoretic generalization of fan-like boundary combinatorics rather than a weakening of Brouwer’s fan theorem or of graph-theoretic fan-planarity [2507.14954].

## 5. Weak fan in superconducting neuromorphic circuits

In superconducting neuromorphic hardware, “weak fan” has a hardware-architectural meaning. The paper uses **fan-out** for the number of output lines driven by a neuron’s spike, \(N_{\mathrm{FO}}\), and **fan-in** for the number of input signals summed at a neuron’s input, \(N_{\mathrm{FI}}\). It explicitly distinguishes the two: **fan-out is digital**, because the problem is to copy an SFQ pulse to many outputs, whereas **fan-in is analog**, because weighted synaptic inputs must be summed and compared with threshold [2008.06409].

The expression “weak fan” in this context refers to the historically poor fan-in or fan-out capability of standard SFQ logic, typically only **2 or 3**. The paper’s central claim is that superconducting neuromorphic circuits can substantially relax this weakness. For fan-out, splitter trees or current-based reamplification via JTLs allow very large replication. Simulations reached **1-to-128** in a nested 7-layer tree and **1-to-16,384** in a 14-layer tree, with **1-to-10,000**-class fan-out described as a realistic demonstrated or simulated level. The conclusion is that fan-out has **no fundamental physics limit** in the architectures considered; it is limited by junction count, chip area, power dissipation, delay, and layout constraints [2008.06409].

Fan-in is more limited because signal current decays roughly like \(1/N_{\mathrm{FI}}\), threshold margin must be preserved, and crosstalk from inactive branches becomes important. The paper gives the practical criterion that the signal should be at least about **20\% of \(I_{c2}\)**. Under WRSPICE simulations, current-based fan-in reaches about
\[
N_{\mathrm{FI}}\approx 20,
\]
whereas flux-based fan-in performs much better: with \(I_{c2}=20\,\mu\text{A}\), the maximum fan-in is **over 100**, and with \(I_{c2}=6\,\mu\text{A}\), it is **over 300**. The key scaling advantage is the factor \(I_{c1}/I_{c2}\), which the paper notes can be **50 or more** [2008.06409].

Here, then, “weak fan” does not name a theorem or a geometric object. It names a practical limitation of legacy SFQ circuitry that neuromorphic circuit design seeks to overcome.

## 6. Disambiguation and scope

Across these literatures, **Weak Fan** is best treated as a homonym. In graph theory it designates the permissive side of the weak/strong fan-planarity distinction and is characterized by allowing Pattern (III) while retaining the same edge-density bounds as strong fan-planarity [2308.08966]. In constructive mathematics it denotes either a WWKL-style weakening of a fan principle or, in adjacent usage, one of several approximate or almost-fan compactness principles [1510.02141] [1408.2493]. In Hodge theory it denotes a weak fan of nilpotent cones indexing boundary strata of degenerations [2507.14954]. In superconducting hardware it denotes historically poor fan-in or fan-out capability in SFQ logic [2008.06409].

A recurrent source of confusion is the transfer of terminology across these domains. In intuitionistic analysis, even within a single field, authors differ on whether FT itself should ever be called a “Weak Fan Theorem,” and some explicitly reject that nomenclature [1311.6988]. A plausible implication is that the term should not be used without a domain qualifier: **weakly fan-planar**, **Weak \(\mathrm{FAN}\)**, **weak fan structure**, and **weak fan-in/fan-out** are not variants of one concept but separate technical usages sharing only a word.

Source: https://www.emergentmind.com/topics/weak-fan