Continued fractions of Laurent series
DefinitionThe variable \(x\) of \(K((x^{-1}))\)#✓fully proved
Let \(K\) be a field. Formal Laurent series in \(x^{-1}\) over \(K\) are realized as Laurent series \(\sum_n c_n X^n\) in \(X = x^{-1}\), and the variable \(x\) is the series \(X^{-1}\), with coefficient \(1\) at \(X^{-1}\) and \(0\) elsewhere.
A series \(\alpha = \sum_n c_n x^{-n} \ne 0\) has absolute value \(|\alpha| = |x|^{-v(\alpha)}\), where its order \(v(\alpha)\) is the least \(n\) with \(c_n \ne 0\).
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Used by 8
TheoremThe variable \(x\) times \(x^{-1}\) is one#✓fully proved
Let \(K\) be a field. In \(K((x^{-1}))\), the product of the variable \(x\) and the series \(x^{-1}\) (coefficient \(1\) at \(x^{-1}\) and \(0\) elsewhere) is \(1\).
theorem laurentInvVar_mul_single_one :
laurentInvVar K * HahnSeries.single (1 : ℤ) (1 : K) = 1
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 1
DefinitionPolynomials as Laurent series in the inverse variable#✓fully proved
Let \(K\) be a field. The embedding \(K[x] \to K((x^{-1}))\) is the \(K\)-algebra homomorphism that sends \(x\) to the element \(x\) of \(K((x^{-1}))\), that is, it evaluates a polynomial at \(x\).
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 1 · Used by 13
TheoremThe polynomial variable embeds as the Laurent variable#✓fully proved
Let \(K\) be a field. The embedding \(K[x] \to K((x^{-1}))\) sends the polynomial \(x\) to the element \(x\) of \(K((x^{-1}))\).
theorem polyToLaurentInv_X : polyToLaurentInv (Polynomial.X : Polynomial K) = laurentInvVar K
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 2
DefinitionPolynomial part of a Laurent series#✓fully proved
Let \(K\) be a field and write a Laurent series \(\alpha \in K((x^{-1}))\) as \(\alpha = \sum_n c_n x^{-n}\). Its polynomial part is \([\alpha] = \sum_{n \le 0} c_n x^{-n}\), the sum of the terms in nonnegative powers of \(x\).
def laurentPolyPart (α : LaurentSeries K) : LaurentSeries K where
coeff n := if n ≤ 0 then α.coeff n else 0
isPWO_support' := α.isPWO_support.mono fun n hn => by
simp only [Function.mem_support, ne_eq] at hn
split_ifs at hn with h
· exact hn
· exact (hn rfl).elim
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1, (1.3).
TheoremCoefficients of the polynomial part of a Laurent series#✓fully proved
Let \(K\) be a field and write \(\alpha \in K((x^{-1}))\) as \(\alpha = \sum_n c_n x^{-n}\). For every integer \(n\), the coefficient of \(x^{-n}\) in the polynomial part \([\alpha]\) is \(c_n\) if \(n \le 0\) and \(0\) if \(n > 0\).
theorem laurentPolyPart_coeff (α : LaurentSeries K) (n : ℤ) :
(laurentPolyPart α).coeff n = if n ≤ 0 then α.coeff n else 0
def laurentPolyPart (α : LaurentSeries K) : LaurentSeries K where
coeff n := if n ≤ 0 then α.coeff n else 0
isPWO_support' := α.isPWO_support.mono fun n hn => by
simp only [Function.mem_support, ne_eq] at hn
split_ifs at hn with h
· exact hn
· exact (hn rfl).elim
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:61
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1, (1.3).
Uses 1
TheoremPolynomial parts of the variable and its inverse#✓fully proved
Let \(K\) be a field. In \(K((x^{-1}))\) the polynomial part of \(x\) is \(x\), and the polynomial part of \(x^{-1}\) is \(0\).
theorem laurentPolyPart_laurentInvVar :
laurentPolyPart (laurentInvVar K) = laurentInvVar K ∧
laurentPolyPart (HahnSeries.single (1 : ℤ) (1 : K)) = 0
def laurentPolyPart (α : LaurentSeries K) : LaurentSeries K where
coeff n := if n ≤ 0 then α.coeff n else 0
isPWO_support' := α.isPWO_support.mono fun n hn => by
simp only [Function.mem_support, ne_eq] at hn
split_ifs at hn with h
· exact hn
· exact (hn rfl).elim
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:61
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1, (1.3).
Uses 2
DefinitionComplete quotients of a Laurent series#✓fully proved
Let \(K\) be a field and \(\alpha \in K((x^{-1}))\). The complete quotients of the continued fraction expansion of \(\alpha\) are \(\alpha_0 = \alpha\) and \(\alpha_{n+1} = (\alpha_n - [\alpha_n])^{-1}\) for \(n \ge 0\), where \([\beta]\) is the polynomial part of \(\beta\) and \(0^{-1} = 0\). So the complete quotients become \(0\) after a finite expansion ends.
def laurentCompleteQuotient (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => α
| n + 1 => (laurentCompleteQuotient α n - laurentPolyPart (laurentCompleteQuotient α n))⁻¹
def laurentPolyPart (α : LaurentSeries K) : LaurentSeries K where
coeff n := if n ≤ 0 then α.coeff n else 0
isPWO_support' := α.isPWO_support.mono fun n hn => by
simp only [Function.mem_support, ne_eq] at hn
split_ifs at hn with h
· exact hn
· exact (hn rfl).elim
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:61
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1, (1.3).
Uses 1 · Used by 2
DefinitionPartial quotients of a Laurent series#✓fully proved
Let \(K\) be a field, \(\alpha \in K((x^{-1}))\) and \(n \ge 0\). The \(n\)-th partial quotient of \(\alpha\) is \(b_n = [\alpha_n]\), the polynomial part of the \(n\)-th complete quotient \(\alpha_n\) of \(\alpha\).
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
def laurentPolyPart (α : LaurentSeries K) : LaurentSeries K where
coeff n := if n ≤ 0 then α.coeff n else 0
isPWO_support' := α.isPWO_support.mono fun n hn => by
simp only [Function.mem_support, ne_eq] at hn
split_ifs at hn with h
· exact hn
· exact (hn rfl).elim
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:61
def laurentCompleteQuotient (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => α
| n + 1 => (laurentCompleteQuotient α n - laurentPolyPart (laurentCompleteQuotient α n))⁻¹
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:100
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1, (1.3).
Uses 2 · Used by 13
TheoremFirst step of the continued fraction of a Laurent series#✓fully proved
Let \(K\) be a field and \(\alpha \in K((x^{-1}))\). The partial quotient \(b_0\) of \(\alpha\) is its polynomial part \([\alpha]\), and the complete quotient \(\alpha_1\) is \((\alpha - [\alpha])^{-1}\), where \(0^{-1} = 0\).
theorem laurentPartialQuotient_zero (α : LaurentSeries K) :
laurentPartialQuotient α 0 = laurentPolyPart α ∧
laurentCompleteQuotient α 1 = (α - laurentPolyPart α)⁻¹
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
def laurentPolyPart (α : LaurentSeries K) : LaurentSeries K where
coeff n := if n ≤ 0 then α.coeff n else 0
isPWO_support' := α.isPWO_support.mono fun n hn => by
simp only [Function.mem_support, ne_eq] at hn
split_ifs at hn with h
· exact hn
· exact (hn rfl).elim
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:61
def laurentCompleteQuotient (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => α
| n + 1 => (laurentCompleteQuotient α n - laurentPolyPart (laurentCompleteQuotient α n))⁻¹
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:100
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1, (1.3).
DefinitionLaurent series with bounded partial quotients#✓fully proved
Let \(K\) be a field. A Laurent series \(\alpha \in K((x^{-1}))\) with partial quotients \(b_0, b_1, \dots\) has bounded partial quotients if there is an integer \(B\) such that for every \(n \ge 0\) and every integer \(k > B\) the coefficient of \(x^k\) in \(b_n\) is zero, that is, every nonzero \(b_n\) is a polynomial in \(x\) of degree at most \(B\).
def LaurentHasBoundedPartialQuotients (α : LaurentSeries K) : Prop :=
∃ B : ℤ, ∀ n : ℕ, ∀ m : ℤ, m < -B → (laurentPartialQuotient α n).coeff m = 0
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 1 · Used by 5
DefinitionNumerators of the convergents of a Laurent series#✓fully proved
Let \(K\) be a field and let \(\alpha \in K((x^{-1}))\) have partial quotients \(b_0, b_1, \dots\). The numerators of the convergents of \(\alpha\) are \(P_{-1} = 1\), \(P_0 = b_0\) and \(P_{n+1} = b_{n+1} P_n + P_{n-1}\) for \(n \ge 0\). The sequence is indexed with a shift of one: its term at \(n \ge 0\) is \(P_{n-1}\).
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 1 · Used by 8
DefinitionDenominators of the convergents#✓fully proved
Let \(K\) be a field, let \(\alpha \in K((x^{-1}))\) and let \(b_0, b_1, \dots\) be its partial quotients. The denominators of the convergents of \(\alpha\) are given by \(Q_{-1} = 0\), \(Q_0 = 1\) and \(Q_{n+1} = b_{n+1} Q_n + Q_{n-1}\) for \(n \ge 0\). They are indexed from \(0\) here: the \(k\)-th term of the sequence is \(Q_{k-1}\).
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 1 · Used by 8
TheoremThe convergent \(P_1 / Q_1\) of a Laurent series#✓fully proved
Let \(K\) be a field and let \(\alpha \in K((x^{-1}))\) have partial quotients \(b_0, b_1, \dots\). Let \(P_n\) and \(Q_n\) be the numerators and denominators of its convergents, given by \(P_{-1} = 1\), \(Q_{-1} = 0\), \(P_0 = b_0\), \(Q_0 = 1\), \(P_{n+1} = b_{n+1} P_n + P_{n-1}\) and \(Q_{n+1} = b_{n+1} Q_n + Q_{n-1}\). Then \(P_1 = b_1 b_0 + 1\) and \(Q_1 = b_1\).
theorem laurentConv_two (α : LaurentSeries K) :
laurentConvNum α 2 = laurentPartialQuotient α 1 * laurentPartialQuotient α 0 + 1 ∧
laurentConvDen α 2 = laurentPartialQuotient α 1
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
DefinitionIrrational Laurent series#✓fully proved
Let \(K\) be a field. A Laurent series \(\alpha \in K((x^{-1}))\) is irrational if \(Q \alpha \ne P\) for all polynomials \(P, Q \in K[x]\) with \(Q \ne 0\), where polynomials are embedded in \(K((x^{-1}))\) by sending \(x\) to the variable \(x\).
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 1 · Used by 9
TheoremPolynomials are not irrational#✓fully proved
Let \(K\) be a field. For every polynomial \(P \in K[x]\), the image of \(P\) in \(K((x^{-1}))\) is not irrational.
theorem not_laurentIsIrrational_polyToLaurentInv (P : Polynomial K) :
¬ LaurentIsIrrational (polyToLaurentInv P)
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Section 1.
Uses 3
TheoremBaum–Sweet: cubics with bounded partial quotients#●ready to prove
Let \(p \in \mathbb{F}_2[x]\) have degree at least \(1\). If \(f \in \mathbb{F}_2((x^{-1}))\) satisfies \(f^3 + p^{-1} f + 1 = 0\), then \(f\) has bounded partial quotients.
public theorem_wanted baumSweet_cubic_boundedPartialQuotients (p : Polynomial (ZMod 2))
(hp : 1 ≤ p.natDegree) (f : LaurentSeries (ZMod 2))
(hf : f ^ 3 + (polyToLaurentInv p)⁻¹ * f + 1 = 0) :
LaurentHasBoundedPartialQuotients f
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def LaurentHasBoundedPartialQuotients (α : LaurentSeries K) : Prop :=
∃ B : ℤ, ∀ n : ℕ, ∀ m : ℤ, m < -B → (laurentPartialQuotient α n).coeff m = 0
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:128
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Theorem 1 (quoting Baum and Sweet, Corollary 3).
TheoremBaum–Sweet: unbounded partial quotients for higher degrees#●ready to prove
Let \(n > 1\) and let \(g \in \mathbb{F}_2[x]\) have degree at least \(1\). If \(f \in \mathbb{F}_2((x^{-1}))\) satisfies \(f^{2^n+1} + g^{-1} f + 1 = 0\), then \(f\) does not have bounded partial quotients.
public theorem_wanted baumSweet_unboundedPartialQuotients (n : ℕ) (hn : 1 < n)
(g : Polynomial (ZMod 2)) (hg : 1 ≤ g.natDegree) (f : LaurentSeries (ZMod 2))
(hf : f ^ (2 ^ n + 1) + (polyToLaurentInv g)⁻¹ * f + 1 = 0) :
¬ LaurentHasBoundedPartialQuotients f
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def LaurentHasBoundedPartialQuotients (α : LaurentSeries K) : Prop :=
∃ B : ℤ, ∀ n : ℕ, ∀ m : ℤ, m < -B → (laurentPartialQuotient α n).coeff m = 0
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:128
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Theorem 2 (quoting Baum and Sweet, Theorem 5).
TheoremBaum–Sweet: explicit continued fractions of roots#●ready to prove
Let \(n \ge 0\) and let \(P, Q \in \mathbb{F}_2[x]\) with \(P + Q^{2^n} \ne 0\) and \(P + Q^{2^n} \ne 1\). Then some \(f \in \mathbb{F}_2((x^{-1}))\) satisfies \(f^{2^n+1} + Q f^{2^n} + P f + QP + 1 = 0\) and has partial quotients \(b_0 = Q\) and \(b_k = P^{2^{(k-1)n}} + Q^{2^{kn}}\) for all \(k \ge 1\), that is, \(f = [Q; P + Q^{2^n}, P^{2^n} + Q^{2^{2n}}, P^{2^{2n}} + Q^{2^{3n}}, \dots]\).
The paper also claims that this root is unique, which fails: for \(n = 0\), \(P = x\) and \(Q = 0\) the equation \(f(f + x) = 1\) has the two roots \([0; x, x, \dots]\) and \([x; x, x, \dots]\).
public theorem_wanted baumSweet_exists_root_continuedFraction (n : ℕ)
(P Q : Polynomial (ZMod 2)) (h0 : P + Q ^ (2 ^ n) ≠ 0) (h1 : P + Q ^ (2 ^ n) ≠ 1) :
∃ f : LaurentSeries (ZMod 2),
f ^ (2 ^ n + 1) + polyToLaurentInv Q * f ^ (2 ^ n) + polyToLaurentInv P * f +
polyToLaurentInv Q * polyToLaurentInv P + 1 = 0 ∧
laurentPartialQuotient f 0 = polyToLaurentInv Q ∧
∀ k, 1 ≤ k → laurentPartialQuotient f k =
polyToLaurentInv (P ^ (2 ^ ((k - 1) * n)) + Q ^ (2 ^ (k * n)))
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Theorem 3 (quoting Baum and Sweet).
Uses 2
TheoremBasic properties of the convergents of a Laurent series#●ready to prove
Let \(K\) be a field and let \(\alpha \in K((x^{-1}))\) be irrational, with partial quotients \(b_0, b_1, \dots\) and convergents \(P_n / Q_n\), where \(P_{-1} = 1\), \(Q_{-1} = 0\), \(P_0 = b_0\), \(Q_0 = 1\), \(P_{n+1} = b_{n+1} P_n + P_{n-1}\) and \(Q_{n+1} = b_{n+1} Q_n + Q_{n-1}\). For a Laurent series \(y = \sum_m c_m x^{-m}\) let \(v(y)\) be its order, the least \(m\) with \(c_m \ne 0\), and let \(v(0) = 0\); thus \(|y| = |x|^{-v(y)}\) for \(y \ne 0\). For \(\beta \in K((x^{-1}))\) write \([b_k, \dots, b_n, \beta] = b_k + 1/(b_{k+1} + 1/(\cdots + 1/(b_n + 1/\beta)))\). Then: (i) for every \(n \ge 0\) and every \(\beta \ne 0\) such that \([b_k, \dots, b_n, \beta] \ne 0\) for all \(1 \le k \le n\), \((\beta P_n + P_{n-1}) / (\beta Q_n + Q_{n-1}) = [b_0, \dots, b_n, \beta]\); (ii) \(P_n Q_{n-1} - P_{n-1} Q_n = (-1)^{n+1}\) for every \(n \ge 0\); (iii) \(Q_{n-1} \ne 0\) and \(v(Q_n) < v(Q_{n-1})\) for every \(n \ge 1\); (iv) \(v(Q_n) = v(b_1) + \dots + v(b_n)\) for every \(n \ge 1\); (v) \(v(\alpha - P_n / Q_n) = -v(b_{n+1}) - 2 v(Q_n)\) for every \(n \ge 1\); (vi) \(v(Q_n \alpha - P_n) < v(Q_{n+1} \alpha - P_{n+1})\) for every \(n \ge 1\).
Part (iii) fails at \(n = 0\) because \(Q_{-1} = 0\).
MetaMathlibExt.laurentConvergent_properties · 18 lines
public theorem_wanted laurentConvergent_properties (α : LaurentSeries K)
(hα : LaurentIsIrrational α) :
(∀ n : ℕ, ∀ β : LaurentSeries K, β ≠ 0 →
(∀ k, 1 ≤ k → k ≤ n → ((List.range' k (n + 1 - k)).map (laurentPartialQuotient α)).foldr
(fun b t => b + t⁻¹) β ≠ 0) →
(β * laurentConvNum α (n + 1) + laurentConvNum α n) /
(β * laurentConvDen α (n + 1) + laurentConvDen α n) =
((List.range (n + 1)).map (laurentPartialQuotient α)).foldr (fun b t => b + t⁻¹) β) ∧
(∀ n : ℕ, laurentConvNum α (n + 1) * laurentConvDen α n -
laurentConvNum α n * laurentConvDen α (n + 1) = (-1) ^ (n + 1)) ∧
(∀ n : ℕ, 1 ≤ n → laurentConvDen α n ≠ 0 ∧
(laurentConvDen α (n + 1)).order < (laurentConvDen α n).order) ∧
(∀ n : ℕ, 1 ≤ n → (laurentConvDen α (n + 1)).order =
∑ i ∈ Finset.Icc 1 n, (laurentPartialQuotient α i).order) ∧
(∀ n : ℕ, 1 ≤ n → (α - laurentConvNum α (n + 1) / laurentConvDen α (n + 1)).order =
-(laurentPartialQuotient α (n + 1)).order - 2 * (laurentConvDen α (n + 1)).order) ∧
(∀ n : ℕ, 1 ≤ n → (laurentConvDen α (n + 1) * α - laurentConvNum α (n + 1)).order <
(laurentConvDen α (n + 2) * α - laurentConvNum α (n + 2)).order)
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 1.
TheoremFrobenius acts on continued fractions of Laurent series#●ready to prove
Let \(K\) be a finite field with \(q\) elements and let \(\alpha \in K((x^{-1}))\), with partial quotients \(b_0, b_1, \dots\) and convergent numerators and denominators \(P_n, Q_n\), where \(P_{-1} = 1\) and \(Q_{-1} = 0\). Then for every \(n \ge 0\) the \(n\)-th partial quotient of \(\alpha^q\) is \(b_n^q\), and for every \(n \ge -1\) the convergent numerator and denominator of \(\alpha^q\) of index \(n\) are \(P_n^q\) and \(Q_n^q\).
public theorem_wanted laurentPartialQuotient_pow_card [Fintype K] (α : LaurentSeries K) (n : ℕ) :
laurentPartialQuotient (α ^ Fintype.card K) n =
laurentPartialQuotient α n ^ Fintype.card K ∧
laurentConvNum (α ^ Fintype.card K) n = laurentConvNum α n ^ Fintype.card K ∧
laurentConvDen (α ^ Fintype.card K) n = laurentConvDen α n ^ Fintype.card K
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 2.
TheoremClose rational approximations and convergents#●ready to prove
Let \(K\) be a field and let \(\alpha \in K((x^{-1}))\) be irrational, with partial quotients \(b_0, b_1, \dots\) and convergents \(P_n / Q_n\), where \(P_{-1} = 1\), \(Q_{-1} = 0\), \(P_0 = b_0\), \(Q_0 = 1\), \(P_{n+1} = b_{n+1} P_n + P_{n-1}\) and \(Q_{n+1} = b_{n+1} Q_n + Q_{n-1}\). For a Laurent series \(y = \sum_m c_m x^{-m}\) let \(v(y)\) be its order, the least \(m\) with \(c_m \ne 0\), and let \(v(0) = 0\). Let \(P, Q \in K[x]\) be coprime with \(Q \ne 0\). (i) If \(v(Q \alpha - P) = s + \deg Q\) for some integer \(s \ge 1\), then there are \(n \ge 0\) and \(c \in K \setminus \{0\}\) with \(P = c P_n\), \(Q = c Q_n\) and \(v(b_{n+1}) = -s\). (ii) If \(v(Q \alpha - P) = \deg Q\), then there are \(n \ge 0\) and \(c, d \in K \setminus \{0\}\) with \(P = c P_n + d P_{n-1}\) and \(Q = c Q_n + d Q_{n-1}\).
With \(|y| = |x|^{-v(y)}\), the hypothesis of (i) reads \(|Q \alpha - P| = |x|^{-s} |Q|^{-1}\) and its conclusion \(|b_{n+1}| = |x|^s\); the hypothesis of (ii) reads \(|Q \alpha - P| = |Q|^{-1}\).
public theorem_wanted laurent_eq_convergent_of_order (α : LaurentSeries K)
(hα : LaurentIsIrrational α) (P Q : Polynomial K) (hPQ : IsCoprime P Q) (hQ : Q ≠ 0) :
(∀ s : ℕ, 1 ≤ s →
(polyToLaurentInv Q * α - polyToLaurentInv P).order = s + Q.natDegree →
∃ (n : ℕ) (c : K), c ≠ 0 ∧
polyToLaurentInv P = HahnSeries.C c * laurentConvNum α (n + 1) ∧
polyToLaurentInv Q = HahnSeries.C c * laurentConvDen α (n + 1) ∧
(laurentPartialQuotient α (n + 1)).order = -(s : ℤ)) ∧
((polyToLaurentInv Q * α - polyToLaurentInv P).order = Q.natDegree →
∃ (n : ℕ) (c d : K), c ≠ 0 ∧ d ≠ 0 ∧
polyToLaurentInv P =
HahnSeries.C c * laurentConvNum α (n + 1) + HahnSeries.C d * laurentConvNum α n ∧
polyToLaurentInv Q =
HahnSeries.C c * laurentConvDen α (n + 1) + HahnSeries.C d * laurentConvDen α n)
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 3 (generalizing Baum and Sweet, Lemma 1).
TheoremDistance from an irrational Laurent series to its convergents#●ready to prove
Let \(K\) be a field, let \(\alpha \in K((x^{-1}))\) be irrational with convergents \(P_n/Q_n\) (\(n \ge 0\)), and for nonzero \(y \in K((x^{-1}))\) let \(|y| = |x|^{\deg y}\), where \(\deg y\) is the largest exponent of \(x\) occurring in \(y\) and \(|x| > 1\) is fixed. (i) If \(|\alpha| \ge 1\), then \(|\alpha - P_n/Q_n| < |\alpha|\) for all \(n \ge 0\). (ii) If \(|\alpha| < 1\), then \(|\alpha - P_0/Q_0| = |\alpha|\) and \(|\alpha - P_n/Q_n| < |\alpha|\) for all \(n \ge 1\).
public theorem_wanted laurent_sub_convergent_order (α : LaurentSeries K)
(hα : LaurentIsIrrational α) :
(α.order ≤ 0 → ∀ n : ℕ,
α.order < (α - laurentConvNum α (n + 1) / laurentConvDen α (n + 1)).order) ∧
(0 < α.order →
(α - laurentConvNum α 1 / laurentConvDen α 1).order = α.order ∧
∀ n : ℕ, 1 ≤ n →
α.order < (α - laurentConvNum α (n + 1) / laurentConvDen α (n + 1)).order)
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 4.
Uses 3
TheoremLasjaunias's bound on the next partial quotient#●ready to prove
Let \(\mathbb{F}_q\) be a finite field, let \(A, B, C, D \in \mathbb{F}_q[x]\) with \(\Delta = AD - BC \ne 0\) and \(C \ne 0\), and let \(\alpha \in \mathbb{F}_q((x^{-1}))\) be irrational with \(\alpha = (A\alpha^q + B)/(C\alpha^q + D)\). Write \(|y| = |x|^{\deg y}\) for the absolute value on \(\mathbb{F}_q((x^{-1}))\), where \(\deg y\) is the largest exponent of \(x\) occurring in \(y \ne 0\), and write \(P_n/Q_n\) for the convergents and \(b_n\) for the partial quotients of \(\alpha\). Let \(n \ge 0\), let \(P, Q \in \mathbb{F}_q[x]\) be coprime with \(Q \ne 0\) and \(P/Q = P_n/Q_n\), and suppose \(|\alpha - P/Q| < |\alpha - A/C|\). Put \(R_1 = DP - BQ\), \(S_1 = -CP + AQ\) and \(\delta_1 = \gcd(R_1, S_1)\). If \(R_1/\delta_1\) or \(S_1/\delta_1\) is not a \(q\)-th power in \(\mathbb{F}_q[x]\), then \(|b_{n+1}| < |\Delta|\,|\delta_1|^{-2}\).
The hypothesis \(C \ne 0\), which makes \(A/C\) meaningful, is implicit in the paper.
MetaMathlibExt.lasjaunias_partialQuotient_bound · 15 lines
public theorem_wanted lasjaunias_partialQuotient_bound [Fintype K] [DecidableEq K] (A B C D : Polynomial K)
(hΔ : A * D - B * C ≠ 0) (hC : C ≠ 0) (α : LaurentSeries K) (hα : LaurentIsIrrational α)
(hfix : α = (polyToLaurentInv A * α ^ Fintype.card K + polyToLaurentInv B) /
(polyToLaurentInv C * α ^ Fintype.card K + polyToLaurentInv D))
(n : ℕ) (P Q : Polynomial K) (hPQ : IsCoprime P Q) (hQ : Q ≠ 0)
(hconv : polyToLaurentInv P * laurentConvDen α (n + 1) =
polyToLaurentInv Q * laurentConvNum α (n + 1))
(hclose : (α - polyToLaurentInv A / polyToLaurentInv C).order <
(α - polyToLaurentInv P / polyToLaurentInv Q).order)
(hpow : (¬ ∃ T : Polynomial K,
(D * P - B * Q) / gcd (D * P - B * Q) (-C * P + A * Q) = T ^ Fintype.card K) ∨
¬ ∃ T : Polynomial K,
(-C * P + A * Q) / gcd (D * P - B * Q) (-C * P + A * Q) = T ^ Fintype.card K) :
-(laurentPartialQuotient α (n + 1)).order <
((A * D - B * C).natDegree : ℤ) - 2 * (gcd (D * P - B * Q) (-C * P + A * Q)).natDegree
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 5 (quoting Lasjaunias, Lemma 1).
TheoremDifference identities for a Möbius–Frobenius fixed point#●ready to prove
Let \(K\) be a finite field with \(q\) elements, let \(A, B, C, D \in K[x]\) with \(\Delta = AD - BC \ne 0\), and let \(\alpha \in K((x^{-1}))\) be irrational with \(\alpha = (A \alpha^q + B) / (C \alpha^q + D)\). Let \(P, Q \in K[x]\) with \(Q \ne 0\), put \(R_1 = DP - BQ\), \(S_1 = -CP + AQ\), \(R_2 = AP^q + BQ^q\) and \(S_2 = CP^q + DQ^q\), and assume \(S_1 \ne 0\) and \(S_2 \ne 0\). Then \(\alpha^q - R_1 / S_1 = \Delta (Q \alpha - P) / ((A - C \alpha) S_1)\) and \(\alpha - R_2 / S_2 = \Delta (Q \alpha - P)^q / ((C \alpha^q + D) S_2)\).
For \(C \ne 0\) the conditions \(S_1 \ne 0\) and \(S_2 \ne 0\) say \(P/Q \ne A/C\) and \((P/Q)^q \ne -D/C\). The source also assumes \(\gcd(A, B, C, D) = 1\) and that \(P/Q\) is a convergent of \(\alpha\) in lowest terms; the identities do not need these hypotheses.
MetaMathlibExt.laurentMobius_sub_identities · 15 lines
public theorem_wanted laurentMobius_sub_identities [Fintype K] (A B C D : Polynomial K)
(hΔ : A * D - B * C ≠ 0) (α : LaurentSeries K) (hα : LaurentIsIrrational α)
(hfix : α = (polyToLaurentInv A * α ^ Fintype.card K + polyToLaurentInv B) /
(polyToLaurentInv C * α ^ Fintype.card K + polyToLaurentInv D))
(P Q : Polynomial K) (hQ : Q ≠ 0) (hS₁ : -C * P + A * Q ≠ 0)
(hS₂ : C * P ^ Fintype.card K + D * Q ^ Fintype.card K ≠ 0) :
α ^ Fintype.card K - polyToLaurentInv (D * P - B * Q) / polyToLaurentInv (-C * P + A * Q) =
polyToLaurentInv (A * D - B * C) * (polyToLaurentInv Q * α - polyToLaurentInv P) /
((polyToLaurentInv A - polyToLaurentInv C * α) * polyToLaurentInv (-C * P + A * Q)) ∧
α - polyToLaurentInv (A * P ^ Fintype.card K + B * Q ^ Fintype.card K) /
polyToLaurentInv (C * P ^ Fintype.card K + D * Q ^ Fintype.card K) =
polyToLaurentInv (A * D - B * C) *
(polyToLaurentInv Q * α - polyToLaurentInv P) ^ Fintype.card K /
((polyToLaurentInv C * α ^ Fintype.card K + polyToLaurentInv D) *
polyToLaurentInv (C * P ^ Fintype.card K + D * Q ^ Fintype.card K))
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 6.
Uses 2
TheoremDegree identities at convergents of a Möbius–Frobenius fixed point#●ready to prove
Let \(K\) be a finite field with \(q\) elements, let \(A, B, C, D \in K[x]\) with \(\Delta = AD - BC \ne 0\) and \(C \ne 0\), and let \(\alpha \in K((x^{-1}))\) be irrational with \(\alpha = (A \alpha^q + B) / (C \alpha^q + D)\). Let \(b_0, b_1, \dots\) be the partial quotients of \(\alpha\) and \(P_n / Q_n\) its convergents, where \(P_{-1} = 1\), \(Q_{-1} = 0\), \(P_0 = b_0\), \(Q_0 = 1\), \(P_{n+1} = b_{n+1} P_n + P_{n-1}\) and \(Q_{n+1} = b_{n+1} Q_n + Q_{n-1}\). For a Laurent series \(y = \sum_m c_m x^{-m}\) let \(v(y)\) be its order, the least \(m\) with \(c_m \ne 0\), and let \(v(0) = 0\); thus \(|y| = |x|^{-v(y)}\) for \(y \ne 0\). Let \(n \ge 0\), let \(P, Q \in K[x]\) be coprime with \(Q \ne 0\) and \(P Q_n = Q P_n\), and put \(R_1 = DP - BQ\), \(S_1 = -CP + AQ\), \(R_2 = AP^q + BQ^q\) and \(S_2 = CP^q + DQ^q\), with the degree of the zero polynomial taken to be \(0\). Then: (i) if \(v(\alpha - A/C) < v(\alpha - P/Q)\), then \(v(S_1 \alpha^q - R_1) = \deg S_1 - \deg \Delta - v(b_{n+1})\); (ii) if \(v(\alpha^q + D/C) < q \, v(\alpha - P/Q)\), then \(v(S_2 \alpha - R_2) = \deg S_2 - \deg \Delta - q \, v(b_{n+1})\).
MetaMathlibExt.laurentMobius_order_identities · 20 lines
public theorem_wanted laurentMobius_order_identities [Fintype K] (A B C D : Polynomial K)
(hΔ : A * D - B * C ≠ 0) (hC : C ≠ 0) (α : LaurentSeries K) (hα : LaurentIsIrrational α)
(hfix : α = (polyToLaurentInv A * α ^ Fintype.card K + polyToLaurentInv B) /
(polyToLaurentInv C * α ^ Fintype.card K + polyToLaurentInv D))
(n : ℕ) (P Q : Polynomial K) (hPQ : IsCoprime P Q) (hQ : Q ≠ 0)
(hconv : polyToLaurentInv P * laurentConvDen α (n + 1) =
polyToLaurentInv Q * laurentConvNum α (n + 1)) :
((α - polyToLaurentInv A / polyToLaurentInv C).order <
(α - polyToLaurentInv P / polyToLaurentInv Q).order →
(polyToLaurentInv (-C * P + A * Q) * α ^ Fintype.card K -
polyToLaurentInv (D * P - B * Q)).order =
((-C * P + A * Q).natDegree : ℤ) - (A * D - B * C).natDegree -
(laurentPartialQuotient α (n + 1)).order) ∧
((α ^ Fintype.card K + polyToLaurentInv D / polyToLaurentInv C).order <
Fintype.card K * (α - polyToLaurentInv P / polyToLaurentInv Q).order →
(polyToLaurentInv (C * P ^ Fintype.card K + D * Q ^ Fintype.card K) * α -
polyToLaurentInv (A * P ^ Fintype.card K + B * Q ^ Fintype.card K)).order =
((C * P ^ Fintype.card K + D * Q ^ Fintype.card K).natDegree : ℤ) -
(A * D - B * C).natDegree -
Fintype.card K * (laurentPartialQuotient α (n + 1)).order)
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 7, (3.7) and (3.8).
TheoremLasjaunias's criterion for unbounded partial quotients#●ready to prove
Let \(\mathbb{F}_q\) be a finite field, let \(A, B, C, D \in \mathbb{F}_q[x]\) with \(\Delta = AD - BC \ne 0\) and \(C \ne 0\), and let \(\alpha \in \mathbb{F}_q((x^{-1}))\) be irrational with \(\alpha = (A\alpha^q + B)/(C\alpha^q + D)\). Write \(|y| = |x|^{\deg y}\) for the absolute value on \(\mathbb{F}_q((x^{-1}))\), where \(\deg y\) is the largest exponent of \(x\) occurring in \(y \ne 0\), and write \(P_n/Q_n\) for the convergents and \(b_n\) for the partial quotients of \(\alpha\). Suppose that for some \(n \ge 0\) there are coprime \(P, Q \in \mathbb{F}_q[x]\) with \(Q \ne 0\) and \(P/Q = P_n/Q_n\) such that \(|\alpha - P/Q|^q < |\alpha^q + D/C|\), \(|\alpha - P/Q|^{q-1} < |\Delta|\,|C\alpha^q + D|^2\) and \(|b_{n+1}|^{q-1} > |\Delta|\). Then \(\alpha\) does not have bounded partial quotients.
In the paper these conditions read \(|\alpha - P/Q| < |\alpha^q + D/C|^{1/q}\), \(|\alpha - P/Q| < |\Delta|^{1/(q-1)} |C\alpha^q + D|^{2/(q-1)}\) and \(|b_{n+1}| > |\Delta|^{1/(q-1)}\).
MetaMathlibExt.lasjaunias_unbounded_criterion · 15 lines
public theorem_wanted lasjaunias_unbounded_criterion [Fintype K] (A B C D : Polynomial K)
(hΔ : A * D - B * C ≠ 0) (hC : C ≠ 0) (α : LaurentSeries K) (hα : LaurentIsIrrational α)
(hfix : α = (polyToLaurentInv A * α ^ Fintype.card K + polyToLaurentInv B) /
(polyToLaurentInv C * α ^ Fintype.card K + polyToLaurentInv D))
(n : ℕ) (P Q : Polynomial K) (hPQ : IsCoprime P Q) (hQ : Q ≠ 0)
(hconv : polyToLaurentInv P * laurentConvDen α (n + 1) =
polyToLaurentInv Q * laurentConvNum α (n + 1))
(hF : (α ^ Fintype.card K + polyToLaurentInv D / polyToLaurentInv C).order <
Fintype.card K * (α - polyToLaurentInv P / polyToLaurentInv Q).order)
(h₁ : -((Fintype.card K : ℤ) - 1) * (α - polyToLaurentInv P / polyToLaurentInv Q).order <
((A * D - B * C).natDegree : ℤ) -
2 * (polyToLaurentInv C * α ^ Fintype.card K + polyToLaurentInv D).order)
(h₂ : ((A * D - B * C).natDegree : ℤ) <
-((Fintype.card K : ℤ) - 1) * (laurentPartialQuotient α (n + 1)).order) :
¬ LaurentHasBoundedPartialQuotients α
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def laurentConvDen (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 0
| 1 => 1
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvDen α (n + 1) + laurentConvDen α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:148
def laurentConvNum (α : LaurentSeries K) : ℕ → LaurentSeries K
| 0 => 1
| 1 => laurentPartialQuotient α 0
| n + 2 => laurentPartialQuotient α (n + 1) * laurentConvNum α (n + 1) + laurentConvNum α n
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:137
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
def LaurentHasBoundedPartialQuotients (α : LaurentSeries K) : Prop :=
∃ B : ℤ, ∀ n : ℕ, ∀ m : ℤ, m < -B → (laurentPartialQuotient α n).coeff m = 0
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:128
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Lemma 8 (quoting Lasjaunias).
TheoremBounded partial quotients for a cubic of Ayadi et al#●ready to prove
Let \(\ell \ge 2\) and let \(A \in \mathbb{F}_2[x]\) be a polynomial of degree \(2\ell\) whose constant and linear coefficients are both \(1\), that is, \(A = x^{2\ell} + \sum_{i=2}^{2\ell-1} e_i x^i + x + 1\) with \(e_i \in \mathbb{F}_2\). If \(\alpha \in \mathbb{F}_2((x^{-1}))\) has positive degree in \(x\) (the largest exponent of \(x\) occurring in \(\alpha\) is positive) and satisfies \(x^\ell \alpha^3 + A \alpha^2 + 1 = 0\), then \(\alpha\) has bounded partial quotients.
public theorem_wanted ayadi_cubic_boundedPartialQuotients (ℓ : ℕ) (hℓ : 2 ≤ ℓ)
(A : Polynomial (ZMod 2)) (hA : A.natDegree = 2 * ℓ) (hA₁ : A.coeff 1 = 1)
(hA₀ : A.coeff 0 = 1) (α : LaurentSeries (ZMod 2)) (hdeg : α.order < 0)
(hroot : laurentInvVar (ZMod 2) ^ ℓ * α ^ 3 + polyToLaurentInv A * α ^ 2 + 1 = 0) :
LaurentHasBoundedPartialQuotients α
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def LaurentHasBoundedPartialQuotients (α : LaurentSeries K) : Prop :=
∃ B : ℤ, ∀ n : ℕ, ∀ m : ℤ, m < -B → (laurentPartialQuotient α n).coeff m = 0
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:128
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Corollary 1.
TheoremBaum–Sweet cubic: partial quotients of degree at most 2#●ready to prove
If \(\alpha \in \mathbb{F}_2((x^{-1}))\) satisfies \(x\alpha^3 + \alpha + x = 0\), then every partial quotient \(b_n\) of \(\alpha\), \(n \ge 0\), has degree at most \(2\) in \(x\).
public theorem_wanted baumSweet_x_cubic_partialQuotient_degree_le_two
(α : LaurentSeries (ZMod 2))
(hroot : laurentInvVar (ZMod 2) * α ^ 3 + α + laurentInvVar (ZMod 2) = 0) :
∀ n : ℕ, ∀ m : ℤ, m < -2 → (laurentPartialQuotient α n).coeff m = 0
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
def laurentPartialQuotient (α : LaurentSeries K) (n : ℕ) : LaurentSeries K :=
laurentPolyPart (laurentCompleteQuotient α n)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:109
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Corollary 2.
Uses 2
TheoremSmall roots of a hyperquadratic equation have unbounded partial quotients#●ready to prove
Let \(\mathbb{F}_q\) be a finite field, let \(\ell \ge 1\) and let \(D \in \mathbb{F}_q[x]\) with \(D(0) \ne 0\). Write \(|y| = |x|^{\deg y}\) for the absolute value on \(\mathbb{F}_q((x^{-1}))\), where \(\deg y\) is the largest exponent of \(x\) occurring in \(y \ne 0\). If \(\alpha \in \mathbb{F}_q((x^{-1}))\) is irrational, satisfies \(x\alpha^{q+1} + D\alpha + x^\ell = 0\) and \(|\alpha| < |x|^{-((q-1)\ell + 1)}\), then \(\alpha\) does not have bounded partial quotients.
public theorem_wanted lasjaunias_unboundedPartialQuotients_of_small [Fintype K] (ℓ : ℕ)
(hℓ : 1 ≤ ℓ) (D : Polynomial K) (hD : D.coeff 0 ≠ 0) (α : LaurentSeries K)
(hα : LaurentIsIrrational α)
(hroot : laurentInvVar K * α ^ (Fintype.card K + 1) + polyToLaurentInv D * α +
laurentInvVar K ^ ℓ = 0)
(hsmall : ((Fintype.card K : ℤ) - 1) * ℓ + 1 < α.order) :
¬ LaurentHasBoundedPartialQuotients α
def LaurentIsIrrational (α : LaurentSeries K) : Prop :=
∀ P Q : Polynomial K, Q ≠ 0 → polyToLaurentInv Q * α ≠ polyToLaurentInv P
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:168
def laurentInvVar (K : Type*) [Field K] : LaurentSeries K :=
HahnSeries.single (-1 : ℤ) (1 : K)
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:30
def polyToLaurentInv (p : Polynomial K) : LaurentSeries K :=
Polynomial.aeval (laurentInvVar K) p
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:45
def LaurentHasBoundedPartialQuotients (α : LaurentSeries K) : Prop :=
∃ B : ℤ, ∀ n : ℕ, ∀ m : ℤ, m < -B → (laurentPartialQuotient α n).coeff m = 0
WantedExt/NumberTheory/LaurentSeriesBoundedPartialQuotientsWanted.lean:128
Sources
- Tuangrat Chaichana, Natthaphong Chitmung, and Vichian Laohakosol, Certain Algebraic Formal Laurent Series with Bounded Partial Quotients, INTEGERS 26 (2026), Article A57, Theorem 6.
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.