Large Galois Orbits Conjecture

• The Large Galois Orbits Conjecture is a principle asserting that the arithmetic complexity of special points and subvarieties forces their Galois orbits to grow polynomially with discriminant or other complexity measures.
• It encompasses various formulations across pure and mixed Shimura varieties, modular forms, and unlikely intersections, employing techniques like point counting, height bounds, and Galois cohomology.
• The conjecture plays a key role in proofs of André–Oort and Zilber–Pink type results by linking local arithmetic invariants to global Galois symmetries.

Searching arXiv for recent and foundational papers on the Large Galois Orbits Conjecture and related André–Oort/Zilber–Pink literature. arxiv_search(query="Large Galois Orbits Conjecture Shimura varieties newforms André-Oort Zilber-Pink", max_results=10) The Large Galois Orbits Conjecture denotes a family of lower-bound principles asserting that Galois orbits attached to arithmetically distinguished points or subvarieties are forced to be large once an appropriate complexity parameter becomes large. In its foundational Shimura-theoretic form, it predicts polynomial growth of the degree of the Galois orbit of a special subvariety in terms of the discriminant of the splitting field of its connected centre; later work reformulated the principle for special points, mixed Shimura varieties, generalised Hecke orbits, atypical intersections in $A_g$ and $Y(1)^n$, and, in a different but related sense, for Galois orbits of modular newforms classified by local types (Ullmo et al., 2012, Binyamini et al., 2021, Gao, 2013, Dieulefait et al., 2018).

1. Foundational Shimura-variety formulation

For a connected Shimura variety $S$ attached to a Shimura datum $(G,X)$ and level $K$, a special subvariety $Z\subset S$ arises from a Shimura subdatum $(H,X_H)$. If $T=Z(H)^\circ$ is the connected centre of $H$ and $L_T$ its splitting field, the Large Galois Orbits Conjecture in the sense of Ullmo–Yafaev states that there exist constants $Y(1)^n$0 and $Y(1)^n$1 such that

$Y(1)^n$2

Here the degree is taken in the Baily–Borel compactification with respect to the canonical ample line bundle $Y(1)^n$3. The conjecture is thus formulated for degrees of Galois orbits of special subvarieties, not only for orbit cardinalities of points (Ullmo et al., 2012).

The same paper gives a conditional theorem under the Generalized Riemann Hypothesis for CM fields. In that theorem, the lower bound involves two types of data: a global factor of the form $Y(1)^n$4, coming from reciprocity and class-field-theoretic control of the toric part, and a local factor built from $Y(1)^n$5-adic indices measuring the failure of the toric level to be maximal. The proof separates accordingly into a global CM part and a local level-raising part, then recombines them through degree-splitting and projection-formula arguments (Ullmo et al., 2012).

This formulation was designed for André–Oort. Ullmo–Yafaev present a dichotomy in which sequences of special subvarieties either have large Galois degrees or fall into an equidistribution regime governed by strongly special subvarieties. In that strategy, large Galois orbits provide the Galois-theoretic half of the argument, complementary to the ergodic and equidistribution methods of Clozel–Ullmo (Ullmo et al., 2012).

2. Complexity parameters and variant formulations

Across the literature, the same general heuristic is expressed through different complexity invariants and different ambient orbit problems. The principal formulations represented in the papers cited here are summarised below.

Setting	Complexity parameter	Lower-bound form
Special subvariety $Y(1)^n$6	$Y(1)^n$7	$Y(1)^n$8
Special point $Y(1)^n$9 on a Shimura variety	$S$0	$S$1
Special point $S$2 on a mixed Shimura variety	$S$3 and orbit of $S$4	$S$5
Point in a generalised Hecke orbit	$S$6	$S$7
Isolated PEL-type intersection point in $S$8	$S$9	$(G,X)$0
Unlikely point $(G,X)$1	$(G,X)$2	$(G,X)$3

These are not identical conjectures. Some concern special subvarieties, some special points, some Hecke orbits, and some atypical intersections. What they share is a lower-bound mechanism: the arithmetic complexity attached to the relevant datum must force polynomial growth of the Galois orbit. A distinct but related usage occurs for modular forms, where the problem is not a lower bound for one orbit size but a lower bound, and conjecturally an exact formula, for the number of global Galois orbits compatible with prescribed local data (Binyamini et al., 2021, Gao, 2013, Richard et al., 2021, Daw et al., 2023, Papas, 2024, Dieulefait et al., 2018).

3. Special points, discriminants, and point-counting

Binyamini, Schmidt, and Yafaev formulate the large-orbit problem for special points on a connected Shimura variety $(G,X)$4 of adjoint type over its reflex field $(G,X)$5. For a special point $(G,X)$6, with Mumford–Tate torus $(G,X)$7, compact open $(G,X)$8, maximal compact $(G,X)$9, and splitting field $K$0 of $K$1, they define the discriminant invariant

$K$2

where $K$3. They then conjecture discriminant-negligible heights: $K$4 for every $K$5, where $K$6 is a Weil height on an algebraic compactification of $K$7 (Binyamini et al., 2021).

Assuming that height conjecture, they prove the large-orbit bound

$K$8

for constants $K$9 and $Z\subset S$0 depending only on $Z\subset S$1 and $Z\subset S$2. A key intermediate statement is a purely combinatorial inequality

$Z\subset S$3

so that any sufficiently subpolynomial height bound yields a polynomial Galois lower bound (Binyamini et al., 2021).

The methodological novelty is the replacement of Masser–Wüstholz isogeny estimates by point-counting on leaves of a foliation. The construction uses the principal $Z\subset S$4-bundle $Z\subset S$5 over $Z\subset S$6, the graph

$Z\subset S$7

inside a horizontal leaf, the absence of positive-dimensional algebraic subvarieties in $Z\subset S$8, and Binyamini’s leaf-counting theorem to bound algebraic points of bounded degree and height on $Z\subset S$9. This converts a height bound for special points into a Galois-orbit bound by comparison with the size of the zero-dimensional special subvariety $(H,X_H)$0 (Binyamini et al., 2021).

In the abelian-type case, the required height bound is supplied by the averaged Colmez formula. For $(H,X_H)$1, special points correspond to simple CM abelian varieties $(H,X_H)$2 with endomorphism field $(H,X_H)$3, and the Faltings height bound

$(H,X_H)$4

implies the corresponding Weil-height estimate. The resulting theorem recovers Tsimerman’s Galois lower bound and, via the Pila–Zannier strategy, yields André–Oort for any mixed Shimura variety whose pure part is $(H,X_H)$5, conditional only on the stated height conjecture in the general case (Binyamini et al., 2021).

4. Mixed Shimura varieties and the lift from pure to mixed

For a mixed Shimura datum $(H,X_H)$6 of abelian type, reflex field $(H,X_H)$7, neat level $(H,X_H)$8, and associated mixed Shimura variety

$(H,X_H)$9

Gao studies special points relative to the projection

$T=Z(H)^\circ$0

to the pure quotient $T=Z(H)^\circ$1. Any special point $T=Z(H)^\circ$2 can be written as $T=Z(H)^\circ$3 with $T=Z(H)^\circ$4 and $T=Z(H)^\circ$5. The mixed order $T=Z(H)^\circ$6 is the least integer $T=Z(H)^\circ$7 such that $T=Z(H)^\circ$8. Proposition 13.3 then states that for every $T=Z(H)^\circ$9 there exists $H$0, depending only on $H$1 and $H$2, such that

$H$3

This is the characteristic mixed large-orbit estimate (Gao, 2013).

The lower bound is unconditional once one has the corresponding pure bound for $H$4. In the paper’s discussion, the pure bound is known under GRH for all $H$5 of abelian type, and unconditionally for $H$6 with $H$7 by Tsimerman’s Brauer–Siegel methods. Consequently, if

$H$8

then one obtains

$H$9

Here $L_T$0 is the centre of the endomorphism ring of the CM torus attached to $L_T$1 (Gao, 2013).

The proof combines several ingredients. First, Galois cohomology and group-theoretic volume compare the full orbit size with the size of a Hecke-translation stabilizer of the unipotent factor $L_T$2 and with the pure orbit size. Second, local index bounds show that $L_T$3 grows like a power of $L_T$4. Third, the Ax–Lindemann theorem for mixed Shimura varieties, proved earlier in the paper by o-minimal point-counting and boundary-volume estimates, rules out unexpected semi-algebraic components unless they are weakly special. Finally, Pila–Wilkie counting turns a hypothetical failure of the lower bound into the production of excess weakly special subvarieties, contradicting Hodge-genericity (Gao, 2013).

The arithmetic significance is twofold. Proposition 13.3 shows that large Galois orbits in the pure quotient lift to large Galois orbits in the mixed variety, up to the mild factor $L_T$5. It also recovers and generalizes Silverberg’s lower bound for torsion points on CM abelian varieties. Combined with the pure-part lower bounds and the Pila–Zannier strategy, it yields André–Oort for any mixed Shimura variety whose pure part sits inside $L_T$6, unconditionally for $L_T$7 and on GRH for all $L_T$8 (Gao, 2013).

5. Zilber–Pink, multiplicative degeneration, and atypical intersections

A further development places large Galois orbits in the Zilber–Pink setting. Daw–Orr formulate a PEL-type Large Galois Orbits conjecture on $L_T$9. For a PEL-type special subvariety $Y(1)^n$00, its complexity is

$Y(1)^n$01

for a very general point $Y(1)^n$02. If $Y(1)^n$03 is irreducible, Hodge-generic, and of codimension exceeding a prescribed threshold, the conjecture predicts that any point $Y(1)^n$04 which is an isolated component of $Y(1)^n$05, with $Y(1)^n$06, satisfies

$Y(1)^n$07

When $Y(1)^n$08, this recovers the CM-point case (Daw et al., 2023).

The paper proves this conjecture for Hodge-generic curves in $Y(1)^n$09 possessing multiplicative degeneration. Here multiplicative degeneration means that the closure of the curve in the Baily–Borel compactification meets the zero-dimensional boundary stratum $Y(1)^n$10, equivalently that after finite base change the universal abelian scheme extends to a semi-abelian scheme whose special fibre at the cusp is $Y(1)^n$11. The proof uses André’s $Y(1)^n$12-functions method. Formal uniformisation

$Y(1)^n$13

is lifted to rigid and complex-analytic uniformisations, producing period $Y(1)^n$14-functions $Y(1)^n$15 and $Y(1)^n$16. Extra endomorphisms yield polynomial relations among their evaluations at a point $Y(1)^n$17; André’s Theorem E then gives a height bound $Y(1)^n$18, and Masser–Wüstholz isogeny estimates convert that into the required Galois lower bound. The corollaries include the full Zilber–Pink statement for curves in $Y(1)^n$19 with multiplicative degeneration and new cases in higher genus (Daw et al., 2023).

Papas establishes an analogous large-orbit statement for curves $Y(1)^n$20 in the setting of unlikely intersections. For points $Y(1)^n$21 satisfying two independent modular relations

$Y(1)^n$22

under the condition that one pair mixes a CM-coordinate with a singular coordinate of $Y(1)^n$23, the paper proves that there are effectively computable constants $Y(1)^n$24 and $Y(1)^n$25, depending only on $Y(1)^n$26, such that

$Y(1)^n$27

The argument again proceeds through $Y(1)^n$28-functions, archimedean period relations, a $Y(1)^n$29-adic avoidance lemma, and isogeny estimates, starting from the height bound

$Y(1)^n$30

These large-orbit estimates feed directly into new Zilber–Pink cases for curves in $Y(1)^n$31, including the cases where all but at most one boundary coordinate are singular and the remaining coordinate is a CM-point, and certain mixed-singular configurations in $Y(1)^n$32 (Papas, 2024).

A common feature of these Zilber–Pink applications is that the lower bound is attached not to all points of the ambient variety, but to isolated or atypical intersection points satisfying additional geometric hypotheses such as multiplicative degeneration or CM/singular mixing. This shows that “large Galois orbits” is not a single uniform statement even within Shimura-theoretic geometry; it is a framework adapted to the particular unlikely-intersection problem under consideration (Daw et al., 2023, Papas, 2024).

6. Hecke orbits, modular forms, and the generalized Maeda picture

Richard–Yafaev introduce a different large-orbit framework for points in a generalised Hecke orbit. For a Shimura datum $Y(1)^n$33, a point $Y(1)^n$34, and $Y(1)^n$35, they consider the affine $Y(1)^n$36-variety

$Y(1)^n$37

where $Y(1)^n$38 is the inclusion. Choosing lattices in $Y(1)^n$39 and $Y(1)^n$40, they define a finite height

$Y(1)^n$41

Under the weakly adelic Mumford–Tate hypothesis, Theorem 6.4 and Proposition 3.6 imply that for every $Y(1)^n$42 in the generalised Hecke orbit,

$Y(1)^n$43

and hence

$Y(1)^n$44

with $Y(1)^n$45 after absorbing mild polynomial factors into the constant. This large-orbit theorem is then inserted into the Pila–Zannier strategy to prove the generalised André–Pink–Zannier conjecture under the same Mumford–Tate assumption (Richard et al., 2021).

A separate but influential use of the phrase “Large Galois Orbits Conjecture” occurs for newforms. In $Y(1)^n$46, a newform $Y(1)^n$47 has coefficient field $Y(1)^n$48, and its Galois orbit has size $Y(1)^n$49. At each prime $Y(1)^n$50, Dieulefait–Pacetti–Tsaknias attach two local invariants: the inertial Weil–Deligne type $Y(1)^n$51 and the minimal Atkin–Lehner sign $Y(1)^n$52. Writing

$Y(1)^n$53

for the number of Galois-conjugacy classes of admissible local pairs $Y(1)^n$54, they prove that when $Y(1)^n$55 is either a prime power or square-free, then for all sufficiently large weights $Y(1)^n$56,

$Y(1)^n$57

For $Y(1)^n$58,

$Y(1)^n$59

and for $Y(1)^n$60 an explicit periodic sequence is obtained. The lower bound is proved using existence theorems for newforms with prescribed local data, together with control of CM forms as $Y(1)^n$61 grows (Dieulefait et al., 2018).

The conjectural strengthening is that this lower bound is actually an equality for sufficiently large $Y(1)^n$62. Numerical evidence reported in the same paper indicates equality in virtually all tested prime-power and square-free cases, with one systematic discrepancy at $Y(1)^n$63, where the lower bound is $Y(1)^n$64 but $Y(1)^n$65 non-CM orbits are found for $Y(1)^n$66. This remains unexplained. The earlier paper of Dieulefait–Tsaknias formulates the generalized Maeda picture in broader terms: $Y(1)^n$67, the number of non-CM Galois orbits in $Y(1)^n$68, should eventually be constant in $Y(1)^n$69; the limiting function $Y(1)^n$70 should be multiplicative; and once local types and Atkin–Lehner signs are fixed, there should be exactly one global orbit for $Y(1)^n$71. In that language, the only expected obstruction to a full symmetric Galois group is a small “trivial” abelian quotient $Y(1)^n$72 contained in the coefficient field (Dieulefait et al., 2018, Dieulefait et al., 2016).

This modular-forms usage is structurally analogous to the Shimura-variety conjectures but not identical to them. Instead of bounding the size of a single orbit by a discriminant or a height, it predicts that local inertial data and involution signs account for all orbit splitting. The common theme is maximality of the global Galois action once the obvious local or toric constraints have been imposed (Dieulefait et al., 2016, Dieulefait et al., 2018).