🔢 mathematics

Unbounded Gaps Between Ordinary and Equivariant Dehn Surgery Numbers

This paper proves that the difference between the equivariant and ordinary Dehn surgery numbers for 3-manifolds with involutions is unbounded, thereby resolving specific open problems in the K3 problem list and demonstrating that certain lens spaces require strictly more surgery components when symmetry is preserved.

Original authors: Qilong Guo, Chunxing Yan

Published 2026-08-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Qilong Guo, Chunxing Yan

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are an architect trying to build a complex 3D shape out of clay. In the world of mathematics, specifically a field called topology, these shapes are called "3-manifolds." They are like the surface of a sphere but extended into three dimensions, and they can be twisted, knotted, and twisted again in ways that are hard to visualize. To understand these shapes, mathematicians use a technique called Dehn surgery. Think of this as a magical construction kit: you start with a simple, perfect sphere (our universe, $S^3$ ), and you cut out some tubes (knots or links) and glue them back together with a specific twist. The number of tubes you need to cut and glue to build a specific shape is called the surgery number. The fewer tubes you need, the "simpler" the shape is considered to be.

Now, imagine your clay shape has a secret symmetry, like a snowflake that looks the same if you flip it over or spin it. This is called an involution (a symmetry that, when done twice, brings you back to the start). The big question mathematicians have been asking is: If you have a shape with a specific symmetry, does that symmetry force you to use more tubes in your construction kit than if you didn't care about the symmetry? In other words, is it harder to build a "symmetric" version of a shape than a "regular" one? For a long time, no one knew if the gap between the "regular" difficulty and the "symmetric" difficulty could grow infinitely large, or if there was a hard limit.

This paper, written by Qilong Guo and Chunxing Yan, answers that question with a resounding "yes." The authors prove that the difference between the number of tubes needed for a regular construction and the number needed for a symmetric one can be as huge as you want. They didn't just guess; they built a mathematical machine that creates an infinite family of shapes where the symmetric version requires exactly twice as many tubes as the regular version.

The Magic of the "Symmetry Tax"

To understand the authors' discovery, let's look at their construction. They start with a basic building block: a shape made from two interlinked rings, like a simple chain link. In the regular world, you can build this shape using just one tube (one knot). It's a quick, easy job. However, this shape has a special property: if you flip it, the two rings swap places.

The authors discovered that if you try to build this shape while respecting the flip symmetry (meaning your construction process must look the same after the flip), you cannot get away with just one tube. You are forced to use two tubes. It's like trying to tie a knot with your hands while wearing a blindfold that forces you to use both hands in perfect sync; the symmetry constraint makes the job twice as hard.

But the real magic happens when they stack these blocks together. The authors show that if you take $k$ copies of this basic block and glue them together, you get a new, larger shape.

The Regular Way: To build this giant shape without worrying about symmetry, you only need $k$ tubes (one for each block).
The Symmetric Way: To build the exact same shape while keeping the flip symmetry intact, you are forced to use $2k$ tubes.

The gap between the two numbers is $k$ . Since $k$ can be any number you choose (1, 10, 1,000, or a million), the gap is unbounded. There is no limit to how much harder the symmetric version is compared to the regular one.

Why This Matters

Before this paper, there was a famous list of unsolved problems in mathematics (known as the K3 problem list). Two of these problems asked:

Can the difference between the symmetric and regular surgery numbers get arbitrarily large?
Can the symmetric number be strictly larger than the regular number, even for simple "flipping" symmetries?

The authors proved that the answer to both is yes. They didn't just find one weird example; they found a whole infinite family of them.

They also tackled a trickier question: Do these weird examples only happen in "messy" shapes made of glued-together parts (reducible manifolds), or do they happen in "pure" shapes that can't be broken down (irreducible manifolds)? They found that even in the purest, most indivisible shapes known as lens spaces, the gap exists. They constructed an infinite list of these pure shapes where the symmetric version always requires more tubes than the regular version.

The Bottom Line

The paper provides a definitive proof that symmetry can be a heavy burden in the world of 3D shapes. It shows that the "cost" of maintaining a perfect symmetry in a construction can be arbitrarily high. Whether you need 2 tubes or 2,000,000 tubes, the authors have shown that the symmetric version will always demand more resources than the non-symmetric one, and the gap between them can grow as wide as you like. This settles a long-standing debate and opens the door to understanding just how complex symmetric shapes can truly be.

Technical Summary: Unbounded Gaps Between Ordinary and Equivariant Dehn Surgery Numbers

Problem Statement
The paper addresses two specific questions regarding the relationship between the integral Dehn surgery number, $DS(Y)$, and the equivariant integral Dehn surgery number, $EDS(Y, \phi)$ , for a closed oriented 3-manifold $Y$ and an orientation-preserving diffeomorphism $\phi$ .

$DS(Y)$ is the minimum number of components in an integral surgery description of $Y$ on a link in $S^3$ .
$EDS(Y, \phi)$ is the minimum number of components in a periodic surgery description that induces the specific diffeomorphism $\phi$ .

While it is known that $DS(Y) \leq EDS(Y, \phi)$ , the authors investigate the extent of this inequality. Specifically, they address Problems 1.15(b) and 1.15(c) from the K3 problem list (Baykur, Kirby, and Ruberman):

Can the difference $EDS(Y, \phi) - DS(Y)$ be arbitrarily large?
Can strict inequality $DS(Y) < EDS(Y, \phi)$ occur when $\phi$ is an involution?

Methodology
The authors employ a combination of homological obstructions, equivariant surgery theory, and explicit geometric constructions.

Homological Lower Bounds:
The paper establishes a lower bound for $EDS(Y, \tau)$ based on the action of the involution $\tau$ on the first homology group $H_1(Y; \mathbb{Z})$ .
- Using Lemma 2.7 and Lemma 2.8, the authors relate the number of surgery components to the generators of the homology of the link exterior.
- Crucially, Lemma 2.10 and Corollary 2.11 show that if $\tau_* \neq \pm \text{id}$ on $H_1(Y; \mathbb{Z})$ , then $EDS(Y, \tau) \geq 2$ .
- The authors refine this by analyzing the action of $\tau$ over different prime fields $\mathbb{F}_\ell$ . They define $b^\delta_\ell(Y, \tau)$ as the dimension of the $\delta$ -eigenspace of $\tau_*$ on $H_1(Y; \mathbb{F}_\ell)$ for $\delta \in \{+1, -1\}$ .
- Proposition 4.2 proves the key inequality: $EDS(Y, \tau) \geq b^+(Y, \tau) + b^-(Y, \tau)$ , where $b^\delta(Y, \tau)$ is the maximum dimension over all odd primes. This relies on the fact that the action of the involution on the meridian basis of the link exterior is a signed permutation, whose eigenspace dimensions are field-independent (Lemma 4.1).
Geometric Construction (The Basic Block):
The authors construct a specific manifold $Y$ using a $\rho$ -periodic Hopf link in $S^3$ with framings $(+4, +4)$ .
- The involution $\rho$ is a half-turn exchanging the two components of the Hopf link.
- The resulting manifold $Y$ is identified as the lens space $L(15, 4)$ .
- Homological calculation shows $H_1(Y; \mathbb{Z}) \cong \mathbb{Z}/15$ , and the induced involution $\tau_*$ acts as multiplication by $-4$.
- Since $-4 \not\equiv \pm 1 \pmod{15}$ , $DS(Y)=1$ (via a single knot surgery description), but $EDS(Y, \tau) \geq 2$ . Specifically, modulo 5, $\tau_*$ acts as $+1$ (contributing to $b^+$ ), and modulo 3, it acts as $-1$ (contributing to $b^-$ ).
Arbitrarily Large Gaps (Theorem 1.1):
To achieve an arbitrary gap $k$ , the authors construct $Y_k$ as the connected sum of $k$ copies of the basic block $Y$ .
- The involution $\tau_k$ acts as $\tau$ on each summand.
- The homology becomes $H_1(Y_k; \mathbb{Z}) \cong (\mathbb{Z}/15)^k$ .
- The equivariant surgery number is bounded below by $b^+ + b^- \geq k + k = 2k$ .
- The ordinary surgery number is bounded by $k$ (via the split union of $k$ single-knot surgeries).
- This yields $DS(Y_k) = k$ and $EDS(Y_k, \tau_k) = 2k$ , creating a gap of $k$ .
Irreducible Examples (Theorem 1.2):
To address the case of irreducible manifolds, the authors utilize lens spaces $Z_a = L(p_a, q_a)$ where $p_a = (a-1)(a^2+1)$ and $q_a = a^2$ .
- They show $DS(Z_a) = 1$ using Moser's theorem on torus knot surgeries.
- They construct an involution $\sigma_a$ via coordinate exchange $(z_1, z_2) \to (z_2, z_1)$ , which descends to the quotient because $q_a^2 \equiv 1 \pmod{p_a}$ .
- The action of $\sigma_a$ on $H_1$ is multiplication by $q_a$ . For $a \geq 3$ , $1 < q_a < p_a - 1$ , ensuring $\sigma_{a*} \neq \pm \text{id}$ .
- Consequently, $EDS(Z_a, \sigma_a) \geq 2 > 1 = DS(Z_a)$ . Since $p_a$ is strictly increasing, this provides an infinite family of pairwise non-homeomorphic irreducible lens spaces.

Key Contributions and Results

Theorem 1.1: For every integer $k \geq 1$ , there exists a pair $(Y_k, \tau_k)$ such that $DS(Y_k) = k$ and $EDS(Y_k, \tau_k) = 2k$ . This confirms that the difference $EDS - DS$ is unbounded, even for involutions.
Theorem 1.2: There exist infinitely many pairwise non-homeomorphic irreducible lens spaces $Z$ admitting involutions $\sigma$ such that $DS(Z) < EDS(Z, \sigma)$ .
Homological Inequality: The paper establishes a robust lower bound for equivariant surgery numbers based on the dimensions of eigenspaces of the involution acting on homology over various prime fields (Proposition 4.2).

Significance
The paper provides affirmative answers to Problems 1.15(b) and 1.15(c) in the K3 problem list. It demonstrates that the requirement for a surgery description to be equivariant with respect to a specific involution can impose a significant topological cost, increasing the number of required surgery components by an arbitrarily large factor compared to the non-equivariant case. Furthermore, it shows that this phenomenon is not restricted to reducible manifolds (connected sums) but occurs within the class of irreducible lens spaces. The work bridges equivariant surgery theory with homological invariants, providing a concrete method to detect obstructions to low-component periodic surgery descriptions.

The Magic of the "Symmetry Tax"

Why This Matters

The Bottom Line

Technical Summary: Unbounded Gaps Between Ordinary and Equivariant Dehn Surgery Numbers

More like this