A Dynamical Lifting Problem For Additive Polynomials

✨

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

The Big Picture: The "Time Travel" Problem

Imagine you have a machine (a mathematical formula) that takes a number, does something to it, and spits out a new number. You can run this machine over and over again. This is called a dynamical system.

Now, imagine you have two versions of this machine:

The "Dirty" Machine: It lives in a world with a weird rule called "characteristic $p$ " (think of it as a world where numbers wrap around like a clock, but with a specific, strange number of hours).
The "Clean" Machine: It lives in our normal world of "characteristic zero" (where numbers go on forever without wrapping around).

The Question: Can we take the "Dirty" machine, understand exactly how it works, and build a "Clean" machine that behaves exactly the same way?

In mathematics, this is called a lifting problem. Usually, if the machine is simple enough, you can build the Clean version. But this paper asks: What if the machine is chaotic and wild?

The Cast of Characters

The Additive Polynomials: These are special formulas (like $z^p - cz$ ) that act like a "magic trick" in the Dirty world. They have a very neat, predictable structure.
The Post-Critical Orbit: Imagine a ball rolling down a hill. The "critical points" are the spots where the ground is slippery or bumpy. The "post-critical orbit" is the path the ball takes after it hits those bumps.
- In the Dirty world, these special formulas have a very simple path: The ball hits a bump and just stays there forever. It's a tiny, finite loop.
The Monodromy Group: This is a fancy name for the "symmetry group" of the machine. It describes all the different ways you can shuffle the inputs and outputs while keeping the machine's rules intact. Think of it as the machine's "DNA."

The Discovery: The "Uncopyable" DNA

The author, Daniel Tedeschi, investigates these special "Additive Polynomials" in the Dirty world. He finds something surprising:

1. The DNA is Too Simple in the Dirty World
In the Dirty world, the symmetry group (the DNA) of these machines is very small and rigid. It's like a lock with only a few specific keys that fit. The machine is "Post-Critically Finite" (PCF), meaning the ball hits a bump and stops moving.

2. The DNA is Too Complex in the Clean World
When Tedeschi tries to build a "Clean" version of these machines (lifting them to characteristic zero), he hits a wall.

In the Clean world, if you try to make a machine that looks like the Dirty one, the symmetry group explodes. It becomes huge and complex.
The Analogy: Imagine you have a simple, flat paper cutout of a star (the Dirty machine). You try to fold it into a 3D origami star (the Clean machine). But the rules of 3D geometry force the paper to crumple and expand in ways the flat paper never did. The "shape" of the symmetry changes fundamentally.

The Verdict: You cannot lift these specific Dirty machines to the Clean world while keeping their "DNA" (symmetry group) the same. The paper proves that for these specific formulas, the answer is NO.

The "Shape" of the Problem

The paper also looks at the "Post-Critical Orbit" (the path of the ball).

In the Dirty world, many different machines can have the exact same path (the same "mapping scheme").
In the Clean world (over the complex numbers), if two machines have the same path, they are usually identical or very close to it. This is called Thurston Rigidity.
The Twist: In the Dirty world, this rigidity breaks! You can have a whole family of different machines that all have the exact same path. Tedeschi calculates exactly how big this family is (it turns out to be a specific dimension, $m$ ). This shows that the Dirty world is much more "flexible" and chaotic than the Clean world.

The "Green and Matignon" Lift (The Failed Attempt)

The paper mentions a famous construction by Green and Matignon. They found a way to lift a single step of these machines from Dirty to Clean.

The Result: They successfully built a Clean machine that looks like the Dirty one at first glance.
The Catch: When you run the machine over and over (iterate it), the Clean version goes crazy. The ball never stops; it wanders off to infinity. The Dirty version stayed put.
The Lesson: Just because you can lift the first step doesn't mean you can lift the whole story. The long-term behavior (the dynamics) changes completely.

Summary in a Nutshell

The Goal: Can we translate a specific type of chaotic math machine from a "weird number world" to our "normal number world" without changing its core identity?
The Answer: No.
Why? The "weird world" allows for a very tight, simple symmetry that simply cannot exist in the "normal world." When you try to force it into the normal world, the symmetry breaks, the machine becomes infinitely complex, and the ball never stops rolling.
The Takeaway: Some mathematical structures are so deeply tied to their "weird" environment that they cannot be translated to our world without losing their soul.

This paper is a "negative solution"—it tells us what cannot be done, which is just as important as telling us what can be done. It highlights the unique, wild nature of mathematics in "characteristic $p$ ."

1. Problem Statement and Context

The paper addresses a dynamical analogue of the classical lifting problem for Galois covers of algebraic curves.

Classical Context: In arithmetic geometry, the "lifting problem" asks whether a branched cover of curves defined over a field $k$ of characteristic $p > 0$ can be lifted to a cover over a characteristic zero discrete valuation ring (DVR) $R$ (e.g., the Witt vectors $W(k)$ ) such that the Galois group action is preserved. The Oort Conjecture (now a theorem by Obus, Wewers, and Pop) states that a cover admits a lift if and only if its inertia groups are cyclic.
Dynamical Context: The author extends this to dynamical systems (rational maps $f: \mathbb{P}^1 \to \mathbb{P}^1$ ). Instead of a single cover, one considers the infinite tower of iterates $\{f^n\}$ . The relevant invariant is the Iterated Monodromy Group (IMG), defined as the Galois group of the tower of splitting fields of $f^n(z) - t$ .
The Core Question: Given a dynamical system $f$ over $\mathbb{F}_p$ , does there exist a lift $\tilde{f}$ over a characteristic zero DVR such that the action of the geometric IMG of $\tilde{f}$ on the inverse limit of pre-images is equivalent to the action of the IMG of $f$ ?

The paper specifically investigates this problem for additive, separable polynomials over $\mathbb{F}_p$ (polynomials of the form $f(z) = \sum a_i z^{p^i}$ with $a_0 \neq 0$ ).

2. Methodology

The author employs a combination of arithmetic dynamics, Galois theory, and combinatorial group theory:

Iterated Monodromy Groups (IMG): The paper formalizes the IMG for dynamical systems as the inverse limit of the monodromy groups of the iterates $f^n$ . It distinguishes between the geometric IMG (over the algebraic closure) and the arithmetic IMG.
Additivity Properties: The study leverages the specific algebraic structure of additive polynomials. Key properties used include:
- The set of roots of $f^n(z) - t$ forms an $\mathbb{F}_p$ -vector space.
- Additive polynomials commute with scalar multiplication by elements of $\mathbb{F}_p$ .
- The derivative of an additive polynomial is a constant ( $f'(z) = a_0$ ), implying separability if $a_0 \neq 0$ .
Riemann-Hurwitz Formula: To analyze lifts to characteristic zero, the author uses the Riemann-Hurwitz formula to count critical points and ramification indices. This is used to prove that certain group actions (specifically free actions) cannot exist in characteristic zero for maps of sufficiently high degree.
Explicit Construction of Lifts: In Section 5, the author constructs explicit lifts of specific additive polynomials ( $z^p - cz$ ) to $\mathbb{Q}_p$ using roots of unity and the Green-Matignon construction, analyzing how the dynamical behavior changes upon lifting.

3. Key Contributions and Results

A. The Dynamical Oort Conjecture (Conjecture 1.2)

The author proposes a dynamical analogue of the Oort Conjecture:

Let $f$ be a PCF (post-critically finite) map over a perfect field of characteristic $p$ . If the inertia groups of the iterates are pro-cyclic, then $f$ admits a geometric IMG-lift to characteristic zero.

The paper notes that while this holds for tamely ramified maps of degree 2 (via Pink), it is unknown for higher degrees.

B. Dimension of Additive Classes (Theorem 1.3)

The author computes the dimension of the moduli space of dynamical systems linearly conjugate to additive, separable polynomials.

Result: Let $A_m$ be the subset of the moduli space $M_{p^m}(\mathbb{F}_p)$ containing classes with an additive, separable representative. Then $\dim_{\mathbb{F}_p}(A_m) = m$ .
Significance: This contrasts sharply with Thurston Rigidity in characteristic zero (over $\mathbb{C}$ ), where fixing the post-critical graph structure typically yields a 0 or 1-dimensional space. In characteristic $p$ , wild ramification allows the space of systems with a fixed mapping scheme to have arbitrarily large dimension.

C. Structure of the IMG for Additive Polynomials (Theorem 1.4)

For an additive, separable polynomial $f$ of degree $p^m$ :

The monodromy group of the $n$ -th iterate is isomorphic to $(\mathbb{Z}/p\mathbb{Z})^{mn}$ .
The action of this group on the fiber $f^{-n}(t)$ is free (the stabilizer of any point is trivial).
Consequently, the geometric IMG is the inverse limit: $\text{IMG}_{\mathbb{F}_p}(f) \cong \varprojlim (\mathbb{Z}/p\mathbb{Z})^{mn}$ .

D. Negative Solution to the Lifting Problem (Theorem 1.5)

This is the paper's primary negative result.

Result: An additive, separable polynomial $f \in \mathbb{F}_p[z]$ does not admit a lift to characteristic zero that preserves the geometric IMG action.
Reasoning:
1. In characteristic $p$ , the IMG action is free (Theorem 1.4).
2. In characteristic zero, for a polynomial of degree $d = p^k$ (with $k \ge 3$ ), the Riemann-Hurwitz formula implies that the monodromy action cannot be free (Lemma 4.5). Specifically, there must be unramified points in the fiber, forcing the inertia group to have fixed points.
3. Since the action in characteristic zero cannot be free, it cannot be equivalent to the free action in characteristic $p$ . Thus, no such lift exists.

E. Analysis of Specific Lifts (Section 5)

The author constructs explicit lifts $\tilde{f}_s$ of the family $f_c(z) = z^p - cz$ to $\mathbb{Q}_p$ .

Dynamical Discrepancy: While $f_c$ is post-critically finite (PCF), the lifted maps $\tilde{f}_s$ are post-critically infinite (PCI) for almost all $s$ (Proposition 1.1).
Group Growth: The monodromy groups of the lifts grow much faster ( $\log_p |\text{Mon}| \approx p^{n-1}$ ) compared to the original additive polynomials ( $\log_p |\text{Mon}| = n$ ). This confirms that the combinatorial invariants (mapping schemes) are not preserved under lifting for these wildly ramified systems.

4. Significance

Counter-Example to Heuristics: The paper demonstrates that heuristics suggesting "tamely ramified behavior extends to positive characteristic" fail for wildly ramified dynamical systems. The specific structure of additive polynomials creates a "dynamical obstruction" to lifting that is not present in the classical curve-covering context.
Failure of the Dynamical Oort Conjecture: The results imply that the hypotheses of the proposed Dynamical Oort Conjecture (pro-cyclic inertia) are not sufficient for the existence of an IMG-lift. The additive polynomials satisfy the inertia condition (inertia groups are cyclic/pro-cyclic) but fail to lift due to the global constraint of the free action.
Moduli Space Geometry: The dimension computation (Theorem 1.3) highlights a fundamental difference between the moduli spaces of dynamical systems in characteristic 0 and characteristic $p$ . In characteristic $p$ , wild ramification creates "large" families of systems with identical post-critical structures, whereas characteristic 0 is rigid.
New Obstruction Mechanism: The paper identifies the freeness of the monodromy action as a new, purely dynamical obstruction to lifting, distinct from the classical obstructions related to the structure of inertia groups alone.

In summary, Tedeschi establishes that for additive polynomials over $\mathbb{F}_p$ , the "wild" nature of the ramification creates a dynamical structure (a free monodromy action) that is incompatible with the geometric constraints of characteristic zero, thereby providing a definitive negative solution to the dynamical lifting problem for this class of polynomials.