Electrification of the Curve Graph

• Electrification of the curve graph is a set of techniques that simplify its combinatorial structure by enlarging adjacencies and quotienting links while preserving topological data.
• This approach employs modified edge relations and coarse projections to recover hyperbolic behavior and manage the complex geometry of curve graphs.
• It distinguishes finite-genus surfaces from infinite-genus ones by revealing phenomena like the random graph embedding, underscoring the interplay between topology and combinatorics.

Searching arXiv for the supplied papers and closely related work on modified or “electrified” curve-graph constructions. Electrification of the curve graph denotes a family of operations and viewpoints in which the combinatorics of simple closed curves on a surface are simplified by enlarging adjacency, quotienting by topologically meaningful equivalence relations, or treating distinguished quasiconvex subsets as regions that are traversed at small cost. In geometric group theory, “electrifying” a graph typically means collapsing certain subgraphs or adding shortcut edges to reduce complexity, often to obtain hyperbolic behavior or to study coarse geometry. In the setting of curve graphs, this theme appears in three closely related forms: modified edge relations such as the fine $1$-curve graph, coarse projection packages attached to Teichmüller discs, and the use of such simplifications against the background fact that the un-electrified curve graph of an infinite-genus surface already contains the random graph as an induced subgraph (Booth et al., 2023, Tang et al., 2015, IV et al., 2014).

1. Classical curve graphs and the motivation for electrification

For a surface $\Sigma$, the curve graph $\mathcal{C}(\Sigma)$ has vertices given by isotopy classes of essential, non-peripheral simple closed curves, and edges joining pairs that can be realized disjointly, equivalently those with geometric intersection number

$i(\alpha,\beta)=0.$

The relevant object is the $1$-skeleton of the curve complex. In the finite-type closed case, the same graph is denoted $C(S)$, with distance $d_S(\alpha,\beta)$, and $C(S)$ is $\delta$-hyperbolic for some universal $\delta$. A key quantitative input is the intersection-number estimate

$\Sigma$0

for intersecting curves, where $\Sigma$1 in that notation (Tang et al., 2015, IV et al., 2014).

The motivation for electrification is that the disjointness graph is often too rigid or too complicated for a given purpose. One can simplify it by declaring additional pairs adjacent, by collapsing or quotienting subgraphs, or by projecting to combinatorial shadows that retain only selected features of the surface. The resulting objects are not equivalent constructions, but they share a common aim: to replace full curve-graph complexity by a controlled combinatorial model that still encodes substantial topological information.

2. Enlarging adjacency: the fine $\Sigma$2-curve graph

A direct form of electrification is to enlarge the adjacency relation. For a closed, connected, oriented surface $\Sigma$3, the fine $\Sigma$4-curve graph $\Sigma$5 has as vertices the essential simple closed curves on $\Sigma$6, and two vertices $\Sigma$7 are adjacent precisely when

$\Sigma$8

Thus disjoint pairs and once-intersecting pairs are both treated as edges. The main rigidity theorem states that for every closed, orientable, connected surface of genus $\Sigma$9, the natural map

$\mathcal{C}(\Sigma)$0

is an isomorphism, so every automorphism of the electrified graph comes from a homeomorphism, and conversely every homeomorphism acts on the graph (Booth et al., 2023).

The proof is organized around the principle that the modified graph still remembers topology combinatorially. Links and joins detect separating curves: for a vertex $\mathcal{C}(\Sigma)$1,

$\mathcal{C}(\Sigma)$2

and in fact a $\mathcal{C}(\Sigma)$3-join. For adjacent separating curves $\mathcal{C}(\Sigma)$4, the common link

$\mathcal{C}(\Sigma)$5

and its separating part determine when $\mathcal{C}(\Sigma)$6 and $\mathcal{C}(\Sigma)$7 are homotopic: they are homotopic if and only if $\mathcal{C}(\Sigma)$8 is a $\mathcal{C}(\Sigma)$9-join such that one part contains only separating curves. For adjacent nonseparating curves, the separating link quotient

$i(\alpha,\beta)=0.$0

is preserved by automorphisms, and it detects torus pairs through the criterion

$i(\alpha,\beta)=0.$1

The annular subgraph

$i(\alpha,\beta)=0.$2

similarly detects curves supported on the possibly pinched annulus bounded by adjacent homotopic nonseparating curves, and membership in this subgraph is preserved by automorphisms (Booth et al., 2023).

Combinatorial object	Geometric information detected
$i(\alpha,\beta)=0.$3 as a join	$i(\alpha,\beta)=0.$4 is separating
$i(\alpha,\beta)=0.$5	torus pairs via the cone criterion
$i(\alpha,\beta)=0.$6	curves supported in the annulus between $i(\alpha,\beta)=0.$7 and $i(\alpha,\beta)=0.$8

A central consequence is that electrification by once-intersection does not erase the distinction between separating and nonseparating curves, homotopy classes of adjacent separating curves, or the dichotomy between torus pairs and pants pairs. The modified graph is therefore not merely coarser; it is a combinatorial object from which the topology of the surface can be reconstructed.

3. Teichmüller-disc shadows as an electrification-like structure

A second form of electrification does not modify adjacency globally, but instead isolates distinguished subsets of the classical curve graph associated to a Teichmüller disc $i(\alpha,\beta)=0.$9. The paper studies several such subsets: straight vertex cycles $1$0, flat systoles $1$1, extremal-length systoles $1$2, hyperbolic systoles $1$3, cylinder curves $1$4, and constant-direction curves $1$5. The main shadow theorem states that

$1$6

agree up to universal Hausdorff distance in $1$7, while

$1$8

agree with them up to uniform Hausdorff distance. The corresponding quasiconvexity conclusions are that $1$9, $C(S)$0, $C(S)$1, and $C(S)$2 are universally quasiconvex, while $C(S)$3 and $C(S)$4 are uniformly quasiconvex (Tang et al., 2015).

The mechanism is explicitly projection-theoretic. There is a coarse Lipschitz retract

$C(S)$5

where $C(S)$6 denotes the vertex cycles of the straight train track built from the combinatorial pattern of saddle connections in the geodesic representative $C(S)$7. This map agrees with nearest-point projection from $C(S)$8 to $C(S)$9 up to universal error. A second map sends a curve $d_S(\alpha,\beta)$0 to the multi-arc $d_S(\alpha,\beta)$1 of saddle connections appearing in $d_S(\alpha,\beta)$2, landing in the filling multi-arc graph associated to $d_S(\alpha,\beta)$3. The resulting bounded geodesic image theorem states that there exist universal constants $d_S(\alpha,\beta)$4 such that if $d_S(\alpha,\beta)$5 is a geodesic in $d_S(\alpha,\beta)$6 disjoint from the $d_S(\alpha,\beta)$7-neighborhood of $d_S(\alpha,\beta)$8, then

$d_S(\alpha,\beta)$9

This is the clearest electrification-type statement in the paper: when a geodesic stays far from the distinguished subset $C(S)$0, its shadow in the auxiliary graph has uniformly bounded diameter (Tang et al., 2015).

The same framework includes balance points on $C(S)$1, defined using the auxiliary polygon

$C(S)$2

whose width and height are $C(S)$3 and $C(S)$4. If $C(S)$5 is a balance point of $C(S)$6, then for any Teichmüller geodesic $C(S)$7, the nearest-point projection of $C(S)$8 to $C(S)$9 lies within $\delta$0 of the balance time of $\delta$1 on $\delta$2. This further reinforces the role of disc-associated subsets as coarse targets for projection (Tang et al., 2015).

4. Infinite-genus complexity and the pre-electrified curve graph

The strongest motivation for electrification in the infinite-genus setting comes from the theorem that the random graph embeds in the curve graph of a surface $\delta$3 if and only if $\delta$4 has infinite genus: $\delta$5 Here “embeds” means induced subgraph embedding, namely a one-to-one map on vertices preserving adjacency and non-adjacency. Because the random graph is the unique countable graph with the extension property and is universal among countable graphs in a strong model-theoretic sense, the theorem identifies a sharp combinatorial-topological dividing line (IV et al., 2014).

The finite-genus obstruction is proved by showing that not every finite graph can embed in the curve graph of a surface of genus $\delta$6. Starting from a finite graph $\delta$7 that does not topologically embed in the closed surface of genus $\delta$8, one subdivides every edge once to form $\delta$9, and then takes the complement graph $\delta$0. If $\delta$1 were realized by a curve system in minimal position, contracting the old curves to points and interpreting the new curves as arcs would recover a topological embedding of $\delta$2, contradicting the choice of $\delta$3. Since every finite graph embeds in the random graph, the random graph cannot embed in a finite-genus curve graph (IV et al., 2014).

The converse uses the one-ended orientable infinite-genus surface $\delta$4. The construction first passes to the multicurve graph $\delta$5, whose vertices are finite sets of pairwise disjoint curves and whose intersection pairing is

$\delta$6

On $\delta$7, one chooses standard curves $\delta$8 satisfying

$\delta$9

and for each natural number $\Sigma$00 defines

$\Sigma$01

where $\Sigma$02 is the $\Sigma$03-th binary digit of $\Sigma$04. This realizes the complement version of the random graph inside the multicurve graph. The final step “threads” the components together through the infinite genus, producing single curves $\Sigma$05 with

$\Sigma$06

Hence $\Sigma$07 is an induced copy of the random graph in $\Sigma$08 (IV et al., 2014).

This theorem is directly relevant to electrification because it shows that the un-electrified curve graph of an infinite-genus surface is already “as complicated as possible” in a precise graph-theoretic sense. Without simplification, the curve graph contains every finite graph and, more strongly, an induced copy of the random graph itself.

5. Quotients, links, and coarse retracts as electrification mechanisms

Across these works, electrification proceeds through a small set of recurrent mechanisms. One is enlargement of the edge relation: in the fine $\Sigma$09-curve graph, disjointness and once-intersection are both declared adjacent, so the modified graph packages more local topology into a single adjacency relation. A second is quotienting: the separating link quotient $\Sigma$10 collapses homotopy classes inside $\Sigma$11, yet still detects whether $\Sigma$12 is a torus pair through the cone condition. A third is coarse retraction onto a preferred quasiconvex subset: the map $\Sigma$13 retracts the full curve graph onto the Teichmüller-disc shadow $\Sigma$14, while

$\Sigma$15

is a natural $\Sigma$16-Lipschitz retract for the corresponding filling multi-arc graph (Booth et al., 2023, Tang et al., 2015).

These operations differ in scale. The fine $\Sigma$17-curve graph is a global graph modification. The Teichmüller-disc constructions leave the ambient curve graph unchanged but single out subsets that behave like quasi-convex “features” of $\Sigma$18. The infinite-genus random-graph theorem, by contrast, concerns the original curve graph before any such simplification. Taken together, they show that electrification in this area is not a single formalism, but a cluster of techniques for replacing unrestricted curve-graph combinatorics by compressed structures with controllable geometry or stronger rigidity.

A plausible implication is that the same analytical vocabulary—links, join decompositions, quotients, coarse projections, and bounded image theorems—can serve both rigidity problems and coarse-geometric problems. In one direction, these tools recover topology from a modified graph; in the other, they isolate subsets whose behavior resembles the cone-off regions of standard electrification arguments.

6. Scope, limitations, and common misunderstandings

A recurrent misunderstanding is to treat every modified curve graph as an electrification in exactly the same sense. The papers distinguish several nonequivalent notions. The fine $\Sigma$19-curve graph is explicitly a graph in which adjacency is enlarged from “disjoint” to “intersect in at most one point,” but it is not the classical curve graph. The Teichmüller-disc paper does not literally build an electrified curve graph in the standard “cone off subsets” formalism; instead, it develops a closely related package of coarse-equivalent shadow sets, coarse projections, quasiconvexity, and bounded geodesic image. The infinite-genus random-graph theorem is not itself a construction of an electrified graph; rather, it identifies the level of combinatorial complexity present before electrification (Booth et al., 2023, Tang et al., 2015, IV et al., 2014).

Another common confusion is to identify the curve graph of an infinite-genus surface with the random graph. The random graph does embed in $\Sigma$20 exactly when $\Sigma$21 has infinite genus, but $\Sigma$22 is not isomorphic to the random graph itself. A separating curve $\Sigma$23 together with curves $\Sigma$24 on opposite sides of $\Sigma$25 gives an obstruction to the global extension property: one can partially embed a small graph, but any extension would force a curve disjoint from $\Sigma$26 and intersecting both $\Sigma$27 and $\Sigma$28, which is impossible (IV et al., 2014).

The broad conceptual picture is therefore two-sided. Finite-type curve graphs are tame enough to be hyperbolic and model-theoretically edge stable, whereas infinite-genus curve graphs are wild enough to contain the random graph, making them edge unstable and highly non-rigid. Electrification, in the broad sense used here, is a way to tame that complexity: by enlarging adjacency, quotienting links, or projecting to quasiconvex shadow sets, one may recover a more manageable geometry without discarding the topological information that the graph still encodes.