Overview

  • Which polynomial records an exact critical period or tail?
  • Which irreducibility cases are known?
  • How do these parameter polynomials relate to dynatomic curves?
  • What changes in higher degree?

Critical points and parameters

For fc(z)=z2+cf_c(z)=z^2+c, the finite critical point is 00. As cc varies, its orbit is

0⟼c⟼c2+c⟼(c2+c)2+c⟼⋯ .0\longmapsto c\longmapsto c^2+c\longmapsto(c^2+c)^2+c\longmapsto\cdots.

It is periodic if it returns to 00, and strictly preperiodic if it enters a cycle without returning to 00. The tail length mm is the first time it enters the cycle; the period nn is the cycle length. For a strictly preperiodic critical point in this family, m≥2m\ge2.

These conditions are polynomial equations in cc. To obtain exact period and tail, one removes the factors belonging to shorter orbits. The resulting arithmetic questions concern the parameters cc, rather than roots of an iterate for a fixed parameter.

References 1

Gleason polynomials

Set a0(c)=0a_0(c)=0 and aj+1(c)=aj(c)2+ca_{j+1}(c)=a_j(c)^2+c. The Gleason polynomial is

Gn(c)=∏d∣nad(c)μ(n/d)∈Z[c].G_n(c)=\prod_{d\mid n}a_d(c)^{\mu(n/d)}\in\mathbb Z[c].

Its roots are exactly the parameters for which 00 has exact period nn. Möbius inversion gives an=∏d∣nGda_n=\prod_{d\mid n}G_d, and hence

deg⁡Gn=∑d∣nμ(n/d) 2d−1.\deg G_n=\sum_{d\mid n}\mu(n/d)\,2^{d-1}.

The conjecture that every GnG_n is irreducible over Q\mathbb Q says that the exact-period parameters form one Galois orbit. Ramadas records verification for n≤19n\le19. The 2026 necklace paper still states the general conjecture.

A stronger conjecture asks whether Gal⁡(Gn/Q)=Sdeg⁡Gn\operatorname{Gal}(G_n/\mathbb Q)=S_{\deg G_n} for every n≥1n\ge1. Schleicher states the symmetric-group conjecture for centers of fixed period.

Normalization in Schleicher’s statement

The displayed definition before Conjecture 7.2 uses fc∘n(c)−c=an(c)2f_c^{\circ n}(c)-c=a_n(c)^2. Its recursively defined exact-period factors are therefore Gn2G_n^2, which are not irreducible. We interpret the conjecture as concerning the reduced polynomial GnG_n, in accordance with the accompanying description of the centers. This is an explicit interpretation of the source, not its literal polynomial formula.

References 3

Example

The first three polynomials are

G1=c,G2=c+1,G3=c3+2c2+c+1.G_1=c,\qquad G_2=c+1,\qquad G_3=c^3+2c^2+c+1.

The reduction G3 mod 2=c3+c+1G_3\bmod2=c^3+c+1 has no root in F2\mathbb F_2, so it is irreducible. Its discriminant is −23-23; therefore its Galois group over Q\mathbb Q is S3S_3.

At period four,

G4=c6+3c5+3c4+3c3+2c2+1,G_4=c^6+3c^5+3c^4+3c^3+2c^2+1,G4 mod 2=(c2+c+1)(c4+c+1).G_4\bmod2=(c^2+c+1)(c^4+c+1).

The two factors are irreducible over F2\mathbb F_2. Thus this prime gives the cycle type (2)(4)(2)(4), but does not prove irreducibility over Q\mathbb Q. Buff–Floyd–Koch–Parry prove that the number of irreducible factors modulo 22 equals the number of real roots. Baker–Chen–Li–Qian give a later description using binary necklaces and construct a bijection. The group at period four can nevertheless be determined by combining other primes, as follows.

References 2

The Galois group at period four

Example

The polynomial G4=c6+3c5+3c4+3c3+2c2+1G_4=c^6+3c^5+3c^4+3c^3+2c^2+1 has Galois group S6S_6 over Q\mathbb Q.

Modulo 33, write g=c6−c2+1g=c^6-c^2+1. Direct reduction gives

c9≡c5−c3,c27≡−c,c729≡c(modg).c^9\equiv c^5-c^3,\qquad c^{27}\equiv-c,\qquad c^{729}\equiv c\pmod g.

Also gcd⁡(g,c27−c)=gcd⁡(g,c)=1\gcd(g,c^{27}-c)=\gcd(g,c)=1, and the identity

(−c4−c2+1)g+(c5−c3+c)(c5−c3−c)=1(-c^4-c^2+1)g+(c^5-c^3+c)(c^5-c^3-c)=1

holds in F3[c]\mathbb F_3[c]. Thus gcd⁡(g,c9−c)=1\gcd(g,c^9-c)=1. The finite-field irreducibility criterion proves that gg, and hence G4G_4, is irreducible.

At two further primes the factorizations into distinct irreducibles are

G4≡(c−1)(c5+4c4−6c3−3c2−c−1)(mod13),G_4\equiv(c-1)(c^5+4c^4-6c^3-3c^2-c-1)\pmod{13},G4≡(c+7)(c2−13c+4)(c3+9c2−11c+10)(mod31).G_4\equiv(c+7)(c^2-13c+4)(c^3+9c^2-11c+10)\pmod{31}.

Frobenius supplies a 5-cycle and an element of cycle type (1)(2)(3)(1)(2)(3). In a transitive degree-six action, the 5-cycle makes the stabilizer of its fixed point transitive on the other points. Thus the group is doubly transitive, hence primitive. Cubing the (1)(2)(3)(1)(2)(3) element gives a transposition. A primitive group containing a transposition is S6S_6.

Finite-field checks

The non-linear factors at 1313 and 3131 all have prime degree dd. For each factor hh, the identities below give r=cp mod hr=c^p\bmod h and polynomials u,vu,v satisfying uh+v(r−c)=1uh+v(r-c)=1 in Fp[c]\mathbb F_p[c]. Repeated modular powering gives cpd≡c(modh)c^{p^d}\equiv c\pmod h. These two conditions prove irreducibility.

Modulo 13

h=c5+4c4−6c3−3c2−c−1.h=c^{5} + 4 c^{4} - 6 c^{3} - 3 c^{2} - c - 1.r=−3c4+4c3−c2+c+1.r=- 3 c^{4} + 4 c^{3} - c^{2} + c + 1.u=−5c3−3c2+5c−5.u=- 5 c^{3} - 3 c^{2} + 5 c - 5.v=−6c4+6c3−4c2−4.v=- 6 c^{4} + 6 c^{3} - 4 c^{2} - 4.

Cubic factor modulo 31

h=c3+9c2−11c+10.h=c^{3} + 9 c^{2} - 11 c + 10.r=14c2−13c−6.r=14 c^{2} - 13 c - 6.u=12c+9.u=12 c + 9.v=8c2−7c−11.v=8 c^{2} - 7 c - 11.

Quadratic factor modulo 31

h=c2−13c+4.h=c^{2} - 13 c + 4.r=13−c.r=13 - c.u=2.u=2.v=c+9.v=c + 9.

The displayed products equal G4G_4 modulo their primes. Distinct irreducible factors of different degrees are coprime, so both reductions are squarefree.

References 3

Misiurewicz polynomials

For fd,c(z)=zd+cf_{d,c}(z)=z^d+c, let aj=fd,c∘j(0)a_j=f_{d,c}^{\circ j}(0). Fix d,m≥2d,m\ge2, n≥1n\ge1, and a nontrivial dd-th root of unity ζ\zeta. The normalized Misiurewicz polynomial is

Gd,m,nζ(c)=∏k∣n(am+k−1−ζam−1)μ(n/k)(∏k∣nakμ(n/k))ϵ,ϵ={1n∣m−1,0n∤m−1.G_{d,m,n}^{\zeta}(c)=\frac{\prod_{k\mid n}(a_{m+k-1}-\zeta a_{m-1})^{\mu(n/k)}}{\left(\prod_{k\mid n}a_k^{\mu(n/k)}\right)^{\epsilon}},\qquad \epsilon=\begin{cases}1&n\mid m-1,\\0&n\nmid m-1.\end{cases}

It is monic in Z[ζ][c]\mathbb Z[\zeta][c]. Its roots have exact critical tail mm, period nn, and collision relation am+n−1=ζam−1a_{m+n-1}=\zeta a_{m-1}. The factor in the denominator removes periodic critical parameters.

Conjecture 1.1 of Benedetto–Goksel predicts irreducibility over Q(ζ)\mathbb Q(\zeta). The known cases they record include d=2,n≤3d=2,n\le3; d=3,n=2d=3,n=2; and prime-power dd with n=1n=1. For d=2d=2, write Gm,n=G2,m,nζ∣ζ=−1G_{m,n}=G_{2,m,n}^{\zeta}\big|_{\zeta=-1}. For example, G2,1=c+2G_{2,1}=c+2, whose root gives the orbit 0↦−2↦2↦20\mapsto-2\mapsto2\mapsto2.

References 1

Polynomial values and algebraic units

Let c0c_0 be a root of Gm,nG_{m,n} for the quadratic family z2+cz^2+c, with m≥2m\ge2, and suppose 1≤ℓ<n1\le\ell<n divides nn. The non-unit conjecture predicts that Gm,ℓ(c0)G_{m,\ell}(c_0) is not an algebraic unit. Benedetto–Goksel prove this when ℓ\ell is a prime at most 10241024, assuming both Gm,ℓG_{m,\ell} and Gm,nG_{m,n} are irreducible over Q\mathbb Q.

The value is a product of differences between c0c_0 and roots of Gm,ℓG_{m,\ell}. A non-unit product shows that at least one difference is a non-unit. To conclude this for every difference, one needs further Galois information, such as irreducibility over Q(c0)\mathbb Q(c_0).

References 1

A higher-degree exception

The quadratic conjecture does not extend unchanged to every degree. Buff uses the family fa(z)=azD+1f_a(z)=az^D+1, D≥2D\ge2. Its exact-period-three parameter polynomial is reducible over Q\mathbb Q if and only if D≡1(mod6)D\equiv1\pmod6. In that case it has exactly two irreducible factors, one equal to a2+a+1a^2+a+1.

This is a statement in the parameter aa for the displayed normalization. For D=2D=2 and a≠0a\ne0, the change w=azw=az conjugates az2+1az^2+1 to w2+aw^2+a, linking Buff’s normalization to the quadratic family above.

References 1

Dynatomic curves and parameter polynomials

For fc(z)=z2+cf_c(z)=z^2+c, define the dynatomic polynomial

Φn(c,z)=∏d∣n(fc∘d(z)−z)μ(n/d).\Phi_n(c,z)=\prod_{d\mid n}(f_c^{\circ d}(z)-z)^{\mu(n/d)}.

The affine curve Φn(c,z)=0\Phi_n(c,z)=0 is smooth and irreducible over C\mathbb C. Smoothness is due to Douady–Hubbard; irreducibility was proved by Bousch and independently by Lau–Schleicher. Buff–Tan give accessible new proofs using quadratic differentials and kneading sequences. Its intersection with z=0z=0 is the equation Gn(c)=0G_n(c)=0. Irreducibility of the curve does not imply irreducibility of this intersection over Q\mathbb Q.

A different curve, Per⁡n(0)\operatorname{Per}_n(0), parametrizes quadratic rational maps with an nn-periodic critical point. Ramadas proves, for n≥4n\ge4, that irreducibility of GnG_n over Q\mathbb Q implies irreducibility of Per⁡n(0)\operatorname{Per}_n(0) over C\mathbb C. Her argument uses a smooth rational point on a suitable projective completion.

References 6

Preperiodic curves and rational points

For zd+cz^d+c, the irreducible factors of fc∘n(z)−fc∘m(z)f_c^{\circ n}(z)-f_c^{\circ m}(z), n>mn>m, define curves parametrizing preperiodic points. Doyle–Poonen prove that, over a field kk containing all dd-th roots of unity and with characteristic not dividing dd, each such irreducible component is geometrically irreducible.

They also prove that the gonalities of these distinct curves tend to infinity: for each bound, only finitely many have gonality below it. Here gonality is the least degree of a nonconstant map over kk from the smooth projective curve to P1\mathbb P^1. Over a function field K/kK/k, this gives a uniform bound for preperiodic points in extensions of bounded degree when cc is not algebraic over kk; for finite kk, that restriction on cc can be removed. Over number fields of bounded degree, a uniform bound holds for preperiodic points of bounded eventual period.

References 1

Questions

OpenComposition and iteration

Are all quadratic Gleason polynomials irreducible?

Scope. G_n(c) over Q, for every integer n≥1

The conjecture asks whether the exact-period parameters form a single Galois orbit for every nn. Verification through n=19n=19 is recorded by Ramadas; the 2026 necklace paper still states the general conjecture.

References 2
Discuss this question →
OpenComposition and iteration

Do all quadratic Gleason polynomials have full symmetric Galois group?

Scope. Every integer n≥1; the reduced exact-period polynomial G_n over Q

Define a0(c)=0a_0(c)=0, aj+1(c)=aj(c)2+ca_{j+1}(c)=a_j(c)^2+c, and Gn(c)=∏d∣nad(c)μ(n/d)G_n(c)=\prod_{d\mid n}a_d(c)^{\mu(n/d)}. The conjecture asserts that GnG_n is irreducible and

Gal⁡(Gn/Q)=Sdeg⁡Gn.\operatorname{Gal}(G_n/\mathbb Q)=S_{\deg G_n}.

Schleicher’s Conjecture 7.2 is interpreted here for these reduced polynomials. Its preceding displayed definition introduces squares; the article gives the normalization calculation. Periods three and four have groups S3S_3 and S6S_6, respectively, by the displayed examples.

References 1
  • Schleicher2017Galois · Conjecture 7.2 and preceding definition, p. 28Reduced-polynomial interpretation, not a literal repetition of the source formula.Source
Discuss this question →
OpenComposition and iteration

Are the normalized Misiurewicz polynomials irreducible?

Scope. d,m≥2; n≥1; ζ≠1 a d-th root of unity; base field Q(ζ)

Conjecture 1.1 asserts irreducibility of Gd,m,nζG_{d,m,n}^{\zeta}. Even the full quadratic family remains unresolved in the cited 2026 paper. Removing lower-period factors and fixing ζ\zeta are part of the statement.

References 2
Discuss this question →
OpenComposition and iteration

Are the remaining Misiurewicz values always non-units?

Scope. m≥2; 1≤ℓ<n with ℓ|n; c₀ a root of quadratic G_{m,n}

Must Gm,ℓ(c0)G_{m,\ell}(c_0) fail to be an algebraic unit? The 2026 result covers prime ℓ≤1024\ell\le1024 under irreducibility assumptions, and leaves the general case open.

References 2
Discuss this question →
Research directionComposition and iteration

Determine the Galois groups of Misiurewicz polynomials.

Scope. Normalized G_{d,m,n}^ζ over Q(ζ), with d,m≥2, n≥1, ζ^d=1 and ζ≠1

For fd,c(z)=zd+cf_{d,c}(z)=z^d+c, let Gd,m,nζG_{d,m,n}^{\zeta} be the monic polynomial whose roots have exact critical tail mm, period nn, and collision relation am+n−1=ζam−1a_{m+n-1}=\zeta a_{m-1}, where aj=fd,c∘j(0)a_j=f_{d,c}^{\circ j}(0). Determine its permutation group over Q(ζ)\mathbb Q(\zeta). The article gives the normalization removing periodic critical parameters. The general irreducibility problem remains open; the symmetric-group conjecture for quadratic Gleason polynomials is a separate question.

References 1
Discuss this question →

Related topics

References

  1. The non-unit conjecture for Misiurewicz parameters

    Benedetto, Robert L. and Goksel, Vefa. Acta Arithmetica, 224(2), 117–139, 2026.

    Read the source
  2. Moduli Spaces of Quadratic Maps: Arithmetic and Geometry

    Ramadas, Rohini. International Mathematics Research Notices, 2024(15), 11364–11370, 2024.

    Read the source
  3. Necklaces, permutations, and periodic critical orbits for quadratic polynomials

    Baker, Matthew and Chen, Andrea and Li, Sophie and Qian, Matthew. European Journal of Combinatorics, 134, 104352, 2026.

    Read the source
  4. Internal Addresses of the Mandelbrot Set and Galois Groups of Polynomials

    Schleicher, Dierk. Arnold Mathematical Journal, 3, 1–35, 2017.

    Read the source
  5. Factoring Gleason polynomials modulo 2

    Buff, Xavier and Floyd, William and Koch, Sarah and Parry, Walter. Journal de théorie des nombres de Bordeaux, 34(3), 787–812, 2022.

    Read the source
  6. Galois groups over Q and factorizations mod p

    Conrad, Keith. Author-hosted expository note.

    Read the source
  7. Finite Fields

    Lidl, Rudolf and Niederreiter, Harald. Edition 2. Cambridge University Press, Encyclopedia of Mathematics and its Applications, vol. 20, 1997.

    Read the source
  8. Recognizing Galois groups S_n and A_n

    Conrad, Keith. Author-hosted expository note.

    Read the source
  9. Misiurewicz polynomials and dynamical units, part II

    Benedetto, Robert L. and Goksel, Vefa. Research in Number Theory, 10, 58, 2024.

    Read the source
  10. On postcritically finite unicritical polynomials

    Buff, Xavier. New York Journal of Mathematics, 24, 1111–1122, 2018.

    Read the source
  11. Étude dynamique des polynômes complexes. Première partie

    Douady, Adrien and Hubbard, John H. Publications mathématiques d’Orsay 84-02, Université de Paris-Sud, Orsay, Report 84-02, 1984.

    Read the source
  12. Étude dynamique des polynômes complexes. Deuxième partie

    Douady, Adrien and Hubbard, John H. Publications mathématiques d’Orsay 85-04, Université de Paris-Sud, Orsay, Report 85-04, 1985.

    Read the source
  13. Sur quelques problèmes de dynamique holomorphe

    Bousch, Thierry. Université de Paris-Sud, Orsay, 1992.

    Read the source
  14. Internal addresses in the Mandelbrot set and irreducibility of polynomials

    Lau, Eike and Schleicher, Dierk. Stony Brook Institute for Mathematical Sciences, Preprint 1994/19, 1994.

    Read the source
  15. The quadratic dynatomic curves are smooth and irreducible

    Buff, Xavier and Tan, Lei. Frontiers in Complex Dynamics: In Celebration of John Milnor's 80th Birthday, Princeton University Press, 49–72, 2014.

    Read the source
  16. Gonality of dynatomic curves and strong uniform boundedness of preperiodic points

    Doyle, John R. and Poonen, Bjorn. Compositio Mathematica, 156(4), 733–743, 2020.

    Read the source