---
title: 'Zariski Pair: Local vs Global Topology'
url: https://www.emergentmind.com/topics/zariski-pair
type: topic
---

# Zariski Pair: Local vs Global Topology

A Zariski pair is a pair of reduced plane curves in \(\mathbb{P}^2\) with identical local-combinatorial data but different global embedded topology. In the standard formulation, one requires the existence of a homeomorphism between tubular neighborhoods \((T(C_1),C_1)\cong (T(C_2),C_2)\), while there is no homeomorphism of pairs \((\mathbb{P}^2,C_1)\cong(\mathbb{P}^2,C_2)\); equivalently, the curves are locally indistinguishable along the curve but globally inequivalent as embeddings in the projective plane [1506.05345]. Since Zariski’s original sextic examples, the subject has become a central interface between singularity theory, arrangement theory, braid monodromy, and the topology of complements.

## 1. Definition, combinatorics, and embedded topology

For reduced plane curves \(C_1,C_2\subset\mathbb{P}^2\), the decisive distinction is between **combinatorics** and **embedded topology**. The embedded topology is the homeomorphism class of the pair \((\mathbb{P}^2,C)\), whereas the combinatorial type records only discrete data: the number of irreducible components, their degrees, the singularities of each component and of the total curve, and the incidence and intersection data among components [2307.01736]. In line or conic–line arrangements, this combinatorics is often encoded by an intersection lattice or incidence graph.

A common formalization is that \((C_1,C_2)\) is a Zariski pair when \(C_1\) and \(C_2\) have the same combinatorics, but there is no homeomorphism
\[
(\mathbb{P}^2,C_1)\longrightarrow(\mathbb{P}^2,C_2).
\]
The tubular-neighborhood formulation makes precise the idea that local singularity data and incidence relations agree, while the global embedding does not [1506.05345]. This separation between local and global data is the foundational phenomenon of the theory.

For reducible curves, especially arrangements, the combinatorics must also remember which local branches belong to which global components. In that sense, Zariski-pair theory is not merely about singularity types in isolation; it is about the failure of combinatorial equivalence to control the ambient topology of the embedding.

## 2. Historical origin and classical examples

The classical example goes back to Zariski’s irreducible sextics with six cusps: one sextic has its six cusps on a conic, while the other does not [1111.5924]. These curves have the same degree and the same local singularity types, but their complements have different topology. In one explicit modern presentation of this pair, the complement groups are
\[
\pi_1(\mathbb{P}^2\setminus B_1,*) \cong \mathbb{Z}/2\mathbb{Z} * \mathbb{Z}/3\mathbb{Z},\qquad
\pi_1(\mathbb{P}^2\setminus B_2,*) \cong \mathbb{Z}/6\mathbb{Z},
\]
so the Zariski-pair property is certified by non-isomorphic \(\pi_1\) [2509.08403]. Degree \(6\) is the smallest degree where this phenomenon occurs for irreducible plane curves [2410.04969].

Subsequent work moved beyond irreducible sextics to arrangements of lines and conics. Tokunaga constructed degree-\(7\) Zariski pairs with irreducible components consisting of lines and conics, and singularities restricted to nodes, tacnodes, and ordinary triple points [1111.5924]. Later work on degree-\(7\) conic–line arrangements exhibited a realization space with exactly two connected components and produced a \(\pi_1\)-equivalent Zariski pair, showing that even the full complement fundamental group may fail to determine the embedded topology [2307.01736].

This historical progression established two enduring principles. First, the phenomenon is not confined to highly singular irreducible curves: it already appears in low-degree mixed arrangements with smooth rational components. Second, the relevant topological distinctions can be subtler than non-isomorphic complement groups.

## 3. Invariants and detection mechanisms

The classical invariant is the complement fundamental group
\[
\pi_1(\mathbb{P}^2\setminus C),
\]
often computed via the Zariski–van Kampen theorem after choosing a suitable projection and extracting braid monodromy relations [2307.01736]. Alexander polynomials, higher homotopy data, and local-system cohomology also occur as standard distinguishing tools. More algorithmic work on conic–line arrangements now explicitly combines combinatorial generation, structural lemmas, projective-equivalence tests, and van Kampen computations to detect candidate pairs [2601.00463].

Several finer invariants have been introduced for cases where \(\pi_1\) is insufficient. Tokunaga’s degree-\(7\) line–conic arrangements are distinguished through the existence or non-existence of elliptic \(D_{2p}\)-covers, equivalently by specific dihedral quotients of complement groups [1111.5924]. In degree \(7\) conic–line arrangements, a double cover branched along a quartic \(Q=C_1+C_2\) leads to the **splitting type** of a pair \((D_1,D_2;Q)\), encoded by intersection numbers such as \((0,2)\) or \((1,1)\); differing splitting types obstruct homeomorphisms of the corresponding embedded pairs [2307.01736].

For line arrangements, the boundary-manifold invariant
\[
\mathcal{I}(\mathcal{A},\xi,\gamma)
\]
uses a character \(\xi\) on \(H_1\) of the complement and a cycle \(\gamma\) in the incidence graph. It is computable from braided wiring diagrams and distinguishes arrangements with the same combinatorics but different oriented ordered topological type [1411.2300]. For conic–line arrangements with a unique conic, the **connected number** of a curve \(C\) with respect to a cyclic cover \(\Phi\) branched along \(B\),
\[
c_\Phi(C)=\#\pi_0\big(\Phi^{-1}(C\setminus B)\big),
\]
detects degree-\(9\) Zariski pairs when the relevant double covers produce different numbers of connected lift components [2410.04969].

At an even finer level, twisted Alexander polynomials associated with finite \(SU(2)\)-representations can distinguish complements whose abelian invariants and ordinary Alexander-type data coincide. This occurs for a Zariski pair of affine nodal curves of fiber type, where the position of nodes changes the complement group although all abelian invariants agree [2306.07359]. This suggests that Zariski-pair detection has evolved from purely group-theoretic obstructions to a layered arsenal of covering-theoretic, monodromic, and representation-theoretic invariants.

## 4. Conic–line arrangements as a testing ground

Conic–line arrangements have become one of the most productive laboratories for explicit Zariski-pair construction. They combine low degree, rigid incidence constraints, and computable topology, yet still exhibit moduli splitting invisible to coarse combinatorics.

| Paper | Configuration | Distinguishing mechanism |
|---|---|---|
| [1111.5924] | Degree \(7\), lines and conics | Elliptic \(D_{2p}\)-covers |
| [2307.01736] | Degree \(7\), three conics and one line | Splitting type; \(\pi_1\)-equivalent pair |
| [2410.04969] | Degree \(9\), unique conic and seven lines | Connected numbers |
| [2509.08403] | Degrees \(7\) and \(8\), conic–line arrangements | Milnor algebra / strong Ziegler pairs |
| [2601.00463] | Conic–line arrangements | Inductive computational classification |

Tokunaga’s degree-\(7\) examples established that even arrangements built from lines and conics with only nodes, tacnodes, and ordinary triple points can form Zariski pairs [1111.5924]. A later degree-\(7\) construction fixed a specific combinatorial type consisting of three smooth conics and a line, proved that its realization space has exactly two connected components, and showed that pairs crossing the two components are \(\pi_1\)-equivalent Zariski pairs [2307.01736]. In that example, the complements have isomorphic fundamental groups
\[
\pi_1(\mathbb{P}^2\setminus C_1)\cong \pi_1(\mathbb{P}^2\setminus C_3)\cong \mathbb{Z}^3,
\]
yet the embedded topologies differ.

Degree-\(9\) examples with a unique conic further show that one can obtain minimal Zariski pairs of \((7,1)\)-arrangements, distinguished by connected numbers of double covers rather than by a direct \(\pi_1\)-comparison [2410.04969]. More recently, examples of degree \(7\) and \(8\) conic–line arrangements were shown to be not only Zariski pairs but also **strong Ziegler pairs**, meaning that their Milnor algebras are non-isomorphic even though the combinatorics agree [2509.08403].

The newest algorithmic direction formulates a combinatorial condition that reformulates the tubular-neighborhood homeomorphism criterion, generates combinatorial equivalence classes by an inductive algorithm, and then filters possible Zariski pairs using structural lemmas, projective equivalence, and van Kampen computations [2601.00463]. This suggests a shift from isolated constructions toward systematic computational enumeration.

## 5. Variants and extensions

The basic notion has generated a substantial family of variants. An **arithmetic Zariski pair** consists of plane curves with Galois-conjugate defining equations over a number field, identical combinatorics, and different embedded topology [1411.2300]. Arithmetic examples are especially delicate because Galois conjugation preserves profinite and many algebro-geometric invariants. In fact, complement-equivalent arithmetic Zariski pairs are known: their complements are homeomorphic, but their embeddings in \(\mathbb{P}^2\) are not [1506.05345]. There are also arithmetic line-arrangement pairs whose complements have non-isomorphic discrete fundamental groups despite isomorphic profinite completions [1507.00190].

A different refinement is the **\(\pi_1\)-equivalent Zariski pair**, where the complements have isomorphic fundamental groups even though the embedded pairs are not homeomorphic [2307.01736]. On the algebraic side, a **strong Ziegler pair** is a pair of plane curves with equivalent combinatorics but distinct Jacobian syzygy modules \(\mathrm{AR}(B)\), equivalently non-isomorphic Milnor algebras in the examples considered [2509.08403]. This variant isolates the failure of combinatorics to determine the Jacobian-algebra structure, rather than the ambient topology.

The notion has also been lifted to isolated hypersurface singularities in \(\mathbb{C}^3\). Starting from a classical Zariski pair of degree-\(d\) plane curves \(f_0=0\), \(f_1=0\), one may add a common monomial \(z_i^{d+m}\) to obtain surface singularities with isolated singular points, the same monodromy zeta-function, and the same Milnor number; under suitable hypotheses these pairs become \(\mu^*\)-Zariski or \(\mu\)-Zariski pairs of surfaces, meaning that they lie in different connected components of the corresponding \(\mu^*\)-constant or \(\mu\)-constant strata [2204.14119]. Related constructions produce Zariski pairs of links and examples where tangent cones form a classical plane-curve Zariski pair even though the resulting \(3\)-manifold links are diffeomorphic [2105.03549; 2203.10684]. More recently, a non–Lê–Yomdin \(\mu\)-Zariski pair of surface singularities was constructed, extending the subject beyond the previously dominant Lê–Yomdin regime [2604.03018].

An opposite notion has also been proposed: an **anti-Zariski pair** consists of equisingular plane curves belonging to different connected components of the same equisingular family, but with homeomorphic embedded pairs \((\mathbb{P}^2,C)\) [2606.20268]. This reversal makes explicit that the topology of the embedding and the connectedness of equisingular strata are distinct layers of structure.

## 6. Realization spaces, equisingular strata, and current perspective

A modern viewpoint studies Zariski pairs via realization spaces. For a fixed combinatorial type of degree \(d\), the realization space is the quasi-projective subset of \(\mathbb{P}H^0(\mathbb{P}^2,\mathcal{O}(d))\) parameterizing curves with that combinatorics [2307.01736]. A Zariski pair often reflects the existence of at least two connected components of this realization space, but the relation is subtle: anti-Zariski pairs show that distinct components need not be topologically distinguishable [2606.20268].

Arithmetic examples reinforce the same point from another direction. Galois-conjugate curves can share étale and profinite information, and even some characteristic-variety data, while still failing to be homeomorphic as embedded curves [1506.05345]. Conversely, complement-equivalent arithmetic pairs show that the complement topology itself may be too coarse. This suggests that the embedded topology of plane curves sits between local combinatorics and algebro-geometric deformation data, but is not reducible to either.

Current work therefore moves along two complementary axes. One axis seeks stronger invariants—splitting types, connected numbers, twisted Alexander polynomials, boundary-manifold invariants, Milnor algebras, and covering-theoretic obstructions. The other seeks exhaustive generation and classification of candidate combinatorics, especially for conic–line arrangements, via inductive algorithms and symbolic computation [2601.00463]. Taken together, these developments indicate that the theory of Zariski pairs is no longer only a source of isolated counterexamples: it has become a structured program for understanding the non-combinatorial topology of algebraic curves and arrangements.

Source: https://www.emergentmind.com/topics/zariski-pair