Transcendence of $p$-adic continued fractions and a quantitative $p$-adic Roth theorem

This paper advances the theory of $p$-adic continued fractions by proving that palindromic and quasi-periodic expansions converge to either transcendental numbers or quadratic irrationals without prior restrictions on partial quotients, while also establishing a quantitative $p$-adic version of Ridout's theorem and analyzing the growth of denominators for algebraic numbers.

The Gist

Imagine you are trying to guess a secret number. You have a special machine that spits out a sequence of clues (numbers) to help you get closer and closer to the truth. In the world of mathematics, this machine is called a Continued Fraction.

Usually, we use these machines with "real" numbers (like 3.14159...). But this paper explores what happens when we use a different kind of number system called $p$-adic numbers. Think of $p$-adic numbers as a universe where the rules of "closeness" are flipped. In our normal world, 100 is far from 0. In the $p$-adic world, if you keep multiplying by a specific prime number (like 2, 3, or 5), numbers get "closer" to zero. It's like a spiral staircase where the higher you go, the closer you get to the ground.

The authors, Anne Kalitzin and Nadir Murru, are trying to solve a mystery: When does this machine spit out a number that is "transcendental"?

What is a Transcendental Number?

Think of numbers as having two types of DNA:

1. Algebraic Numbers: These are numbers that can be described by a simple equation (like $x^2 - 2 = 0$, which gives you $\sqrt{2}$). They are the "rule-followers."
2. Transcendental Numbers: These are the "rebels." They cannot be described by any simple equation. Famous examples are $\pi$ and $e$. They are wild, unpredictable, and infinitely complex.

The goal of the paper is to build a machine (a continued fraction) that is guaranteed to produce a "rebel" number (transcendental) rather than a "rule-follower" (algebraic).

The Two Main Tricks the Authors Used

1. The "Palindrome" Trick (The Mirror Game)

Imagine you are writing a list of clues.

• Normal list: 1, 5, 2, 9, 4... (Random)
• Palindrome list: 1, 5, 2, 2, 5, 1, 3, 7, 7, 3... (It reads the same forwards and backwards in chunks).

In the past, mathematicians knew that if your list had these "mirror" patterns (palindromes) that kept getting longer and longer, the resulting number was likely a rebel (transcendental). However, previous rules were very strict: they said, "The clues must be huge numbers" or "The clues must be small in this specific way."

The Authors' Breakthrough:
They removed all those strict rules about the size of the clues. They proved that as long as the mirror patterns keep getting longer, the number is either a rebel (transcendental) or a very specific type of rule-follower (a quadratic irrational). It doesn't matter if the clues are big or small; the pattern of the mirror is what matters.

2. The "Repeating Loop" Trick (The Quasi-Periodic)

Imagine a song that repeats a chorus, but every time it comes back, the chorus is slightly longer or the gap between choruses changes.

• Chorus 1: A-B-C
• Chorus 2: A-B-C-A-B-C (Longer)
• Chorus 3: A-B-C-A-B-C-A-B-C (Even longer)

The authors looked at these "quasi-periodic" patterns. They found that if the loops grow fast enough (specifically, if the length of the repeating part grows faster than the gap between them), the number is almost certainly a rebel.

The "Speed Limit" for Approximation

To prove these things, the authors had to invent a new tool. They looked at a famous rule called Roth's Theorem (which is like a speed limit for how well you can guess a number).

• The Old Rule: "You can't get too close to an algebraic number without using a denominator that is impossibly huge."
• The Problem: In the $p$-adic world, no one had written down exactly how many times you could break this speed limit before it became impossible. It was like knowing there's a speed limit, but not knowing the exact number of tickets you can get before you're banned.

The Authors' New Tool:
They created a Quantitative Version of this rule. They calculated a precise "ticket limit." They showed that for any algebraic number, there is a strict, calculable maximum number of times you can get a "good guess" (a convergent) before the math breaks down.

This new tool allowed them to prove that if your continued fraction grows too fast (like the repeating loops mentioned above), it would need more good guesses than the universe allows for an algebraic number. Therefore, it must be a transcendental number.

The "Denominator Growth" Discovery

Finally, they looked at the "denominators" of the guesses (the bottom numbers of the fractions).

• If the number is a "rebel" (transcendental), the denominators can grow wildly and unpredictably.
• If the number is a "rule-follower" (algebraic), the denominators have to grow at a specific, controlled pace.

The authors proved that for algebraic numbers in this $p$-adic world, the denominators cannot grow too fast. They established a "speed limit" for how quickly the bottom numbers of the fractions can expand. If they expand faster than this limit, the number is definitely transcendental.

Summary

In simple terms, this paper is like upgrading the rules of a guessing game:

1. Old Rules: You could only prove a number was a "rebel" if the clues were very specific sizes.
2. New Rules: You can prove a number is a "rebel" just by looking at the patterns (mirrors and loops), regardless of the size of the clues.
3. New Tool: They built a better "speedometer" to measure how fast you can guess a number, proving that if you guess too fast, you must be guessing a transcendental number.

They successfully removed the "restrictions" of previous math, making it easier to identify these mysterious, infinite numbers in the strange world of $p$-adic numbers.

Technical Summary

Here is a detailed technical summary of the paper "Transcendence of p-adic Continued Fractions and a Quantitative p-adic Roth Theorem" by Anne Kalitzin and Nadir Murru.

1. Problem Statement

The paper addresses the problem of determining the transcendence of numbers represented by p-adic continued fractions in the field of p-adic numbers $\mathbb{Q}_p$. While transcendence criteria for real continued fractions are well-established (relying on the growth of partial quotients, palindromic structures, or quasi-periodicity), the p-adic setting presents unique challenges due to the lack of a canonical floor function and different convergence behaviors.

Previous works (e.g., Ooto, Murru, etc.) had established transcendence results for p-adic continued fractions under restrictive assumptions, specifically requiring lower bounds on the p-adic norm of the partial quotients (e.g., $|b_i|_p \geq p^2$ or $p^4$) or specific boundedness conditions on the Archimedean norms. The authors aim to:

1. Remove these restrictive norm conditions on partial quotients.
2. Prove that palindromic and quasi-periodic p-adic continued fractions converge to either transcendental numbers or quadratic irrationals.
3. Establish a quantitative version of Ridout's Theorem (the p-adic analogue of Roth's Theorem) to facilitate these proofs and study the growth of denominators of convergents for algebraic numbers.

2. Methodology

The authors employ a combination of Diophantine approximation theory, linear algebra, and the Subspace Theorem in the p-adic context.

• Definitions: The paper considers two standard definitions of p-adic continued fractions: Ruban (digits in $\{0, \dots, p-1\}$) and Browkin (digits in $\{-\frac{p-1}{2}, \dots, \frac{p-1}{2}\}$).
• Palindromic Structure: For palindromic sequences, the authors utilize the symmetry of the matrix product of partial quotients. If a sequence is palindromic, the convergents satisfy specific relations (e.g., $A_n = B_{n-1}$), allowing the construction of linear forms that vanish if the number is algebraic of degree $\geq 3$.
• Quantitative Diophantine Approximation: The core methodological innovation is the derivation of a quantitative version of Ridout's Theorem. Unlike the classical qualitative statement (which says there are finitely many solutions), the authors provide an explicit upper bound on the number of solutions to the inequality $|\alpha - A/B|_p < |B|_\infty^{-(2+\epsilon)}$.
• Subspace Theorem: They apply the p-adic Subspace Theorem (Schlickewei) to show that if a number satisfies certain approximation properties derived from its continued fraction structure, the convergents must lie in a proper linear subspace, leading to a contradiction for algebraic numbers of degree $\geq 3$.
• Growth Analysis: For quasi-periodic fractions, they analyze the growth rate of the denominators of convergents ($B_n$) using the derived quantitative bounds to show that the "repetition" in the continued fraction forces the number to be algebraic of low degree or transcendental.

3. Key Contributions and Results

A. Removal of Norm Restrictions (Theorem A)

The authors prove that for a p-adic continued fraction $\alpha = [0, b_1, b_2, \dots]$ beginning with arbitrarily long palindromes:

• Condition: The Archimedean norms of the numerators and denominators of the convergents must grow slowly: $\max(|A_n|_\infty^{1/n}, |B_n|_\infty^{1/n}) < p^{1/4}$ for large $n$.
• Result: No restriction is placed on the p-adic norm $|b_i|_p$.
• Conclusion: $\alpha$ is either transcendental or a quadratic irrational. This generalizes previous results by removing the requirement that $|b_i|_p$ be large.

B. Quantitative p-adic Roth Theorem (Theorem 8 & Corollary 9)

The paper establishes a new quantitative bound for the number of good rational approximations to an algebraic number $\alpha \in \mathbb{Q}_p$ of degree $n \geq 2$.

• The Inequality: For $|\alpha - A/B|_p < |B|_\infty^{-(2+\epsilon)}$ with $|B|_\infty \geq |A|_\infty$.
• The Bound: The number of such solutions is bounded by:
  $$\text{Count} < 4\epsilon^{-1} \log \hat{C} + \exp(C_1 n^2 \epsilon^{-2})$$
  where $\hat{C}$ depends on the coefficients of the minimal polynomial of $\alpha$.
• Significance: This is the first explicit quantitative version of Ridout's theorem suitable for analyzing the growth of denominators in continued fractions, filling a gap in the literature.

C. Growth of Denominators for Algebraic Numbers (Theorem 12)

Using the quantitative Roth theorem, the authors derive a bound on the growth of the p-adic norm of the denominators of convergents for algebraic numbers.

• Result: If $\alpha$ is algebraic, the double logarithm of the p-adic norm of the denominator $B_k$ grows sub-linearly:
  $$\log \log |B_k|_p < \frac{c(\alpha) k}{(\log k)^{1/2}}$$
• Implication: This provides a necessary condition for a p-adic continued fraction to represent an algebraic number. If the denominators grow faster than this bound, the number must be transcendental.

D. Quasi-Periodic Continued Fractions (Theorem B & Theorem 15)

The authors extend the results to quasi-periodic continued fractions (sequences containing arbitrarily long repetitions).

• Conditions:
  1. $|A_i|_\infty \leq |B_i|_\infty$ for large $i$.
  2. The length of the repeating block $k_i$ is bounded relative to the start of the block $n_i$ ($k_i < C n_i$).
  3. The repetition length $\lambda_i$ grows sufficiently fast: $\lim_{i \to \infty} \frac{\log \lambda_i (\log n_i)^{1/2}}{n_i} = \infty$.
• Conclusion: Under these conditions, $\alpha$ is either transcendental or a quadratic irrational. This result is closer to the classical Baker/Maillet results for real numbers than previous p-adic analogues.

4. Significance

• Generalization: The paper significantly broadens the scope of known transcendence criteria for p-adic numbers by eliminating the need for restrictive lower bounds on the p-adic norm of partial quotients.
• Methodological Advancement: The derivation of a quantitative p-adic Roth theorem is a major theoretical contribution. It provides a powerful tool for future research in p-adic Diophantine approximation, allowing for precise counting of solutions rather than just existence.
• Unification: The results apply to both major definitions of p-adic continued fractions (Ruban and Browkin), unifying the theory across different constructions.
• Bridge to Classical Theory: By establishing a p-adic version of the Davenport-Roth result on the growth of denominators, the paper strengthens the parallel between real and p-adic Diophantine approximation, showing that similar structural constraints (palindromes, quasi-periodicity) lead to similar transcendence outcomes in both fields.

In summary, Kalitzin and Murru have successfully removed technical barriers in p-adic transcendence theory and provided the necessary quantitative tools (a refined Roth theorem) to analyze the arithmetic properties of p-adic continued fractions more deeply.