XCIII.1 (2000)
Algebraic independence of the values of Mahler functions satisfying implicit functional equations
by
Bernd Greuel (K¨oln)
1. Introduction and results. In a sequence of three papers Mahler ([4]–[6]) discussed the transcendence and algebraic independence of values of functions in several variables satisfying a certain type of functional equation.
In his survey article [7], 37 years later, he stated three new problems. The third problem (for the first and second problem cf. Loxton and van der Poorten [3]) dealt with implicit functional equations of the type
(1) P (z, f (z), f (T z)) = 0
with T z = z
d, d ∈ Z, d ≥ 2 and a polynomial P (z, y, u) with coefficients in Q, the algebraic closure of Q. Nishioka [8] (cf. Chapter 1.5 in [11]) solved this problem for polynomial transformations T . In [9] she extended her method to functions in several variables and suitable generalizations of the transfor- mation T z = z
d.
Becker [1] generalized the result of Nishioka to algebraic transforma- tions T. T¨opfer gave in [15] a quantitative version of Becker’s result. In that article T¨opfer asked for a proof of the algebraic independence of the val- ues of several functions satisfying implicit functional equations at algebraic points.
In this paper we follow the proof of T¨opfer [15] and derive a lower bound for the transcendence degree of the values of functions f
1, . . . , f
msatisfy- ing a special system of implicit functional equations for the transformation T z = z
dwith an integer d ≥ 2. It should be easy to generalize the follow- ing result to polynomial or even rational or algebraic transformations T (cf.
Becker [1] and T¨opfer [14, 15]).
For the development of Mahler’s method in the last 15 years see the monograph of Nishioka [11] and the overview article of Waldschmidt [16] for further references.
2000 Mathematics Subject Classification: 11J82, 11J91.
[1]
Throughout the paper let K denote an algebraic number field and O
Kthe ring of integers in K. As usual we denote by α the house of an algebraic number α, which is the maximum of the absolute values of the conjugates of α. A denominator of an algebraic number α is a positive integer D such that Dα ∈ O
K. If P (z, y
1, . . . , y
m) =: P (z, y ) is a polynomial with complex coefficients, deg
zP =: d
zP denotes the partial degree of P with respect to z, deg
yP =: d
yP denotes the total degree in y := (y
1, . . . , y
m) and analogous notations in other cases. If the coefficients of P are algebraic, the height H(P ) of P is defined as the maximum of the houses of the coefficients of P , and the length L(P ) is the sum of the houses of the coefficients. In what follows let c, c
0, c
1, . . . and γ
0, γ
1, . . . denote positive constants which are independent of the parameters M, N, k, k
0, k
1used. For a vector µ ∈ C
mwe define |µ| := |µ
1| + . . . + |µ
m| and by N and N
0we denote the positive and nonnegative integers.
Theorem 1. Let f
1, . . . , f
mbe analytic in a neighborhood U of the ori- gin, algebraically independent over C(z) and suppose that the coefficients of their power series
f
i(z) = X
∞ j=0f
i,jz
j(i = 1, . . . , m) belong to a fixed algebraic number field K and satisfy
f
i,j≤ exp(c
0(1 + j
L)) and D
[c0(1+jL)]f
i,j∈ O
Kfor j ∈ N
0and i = 1, . . . , m with suitable constants D ∈ N and L ≥ 1. Let n ∈ N
mand β := n
1· . . . · n
m. Suppose that the functions f
1, . . . , f
msatisfy the functional equations
(2) a(z)f
j(z
d)
nj=
n
X
j−1 ν=0P
ν,j(z, f (z))f
j(z
d)
νwith polynomials a ∈ Q[z]\{0} and P
0,1, . . . , P
nm−1,m∈ Q[z, y ] and an integer d satisfying d > max{β
L, d
y(P )}, where d
y(P ) is defined by
d
y(P ) := max{deg
y(P
0,1), . . . , deg
y(P
nm−1,m)}.
Assume α ∈ Q
∗∩ U and a(α
dk) 6= 0 for all k ∈ N
0. Let m
0be the smallest integer satisfying
m
0≥ m log d − L(m + 1) log β 1 +
log βlog dlog β + log d + L(m + 1) 1 +
log βlog d+ m
(2 log β + log d
y(P )) . Then
trdeg
QQ(f
1(α), . . . , f
m(α)) ≥ m
0.
As an application of this theorem we obtain easily the following
Corollary 2. Under the assumptions of Theorem 1, if α, f and the parameters d, β and d
y(P ) satisfy for m > 1 the inequality
log d
y(P )
log d < 1 −
log βlog d2m
2− m − 1 + L(m + 1) 1 +
log βlog d(2m − 1) (m − 1) L(m + 1) 1 +
log βlog d+ m ,
then f
1(α), . . . , f
m(α) are algebraically independent.
Remarks. (i) Nishioka [8] proved the transcendence of f (α) under the condition d
2> n
2max{d, deg
y(P )}, where f satisfies the functional equation (1) and n = deg
u(P ).
Under the hypotheses of Theorem 1 we get the transcendence of f (α) only under the stronger condition d > max{n
√3+1, deg
y(P )}. The reason for this is that we have to construct a sequence of polynomials (Q
k)
k0≤k≤k1, where the difference k
1− k
0has to be relatively large (cf. Lemma 8). In the simpler case m = 1 it suffices to find just one integer k to obtain a contradic- tion. By an improvement of the method of proof we get the transcendence of f (α) under the condition d > max{n
2, deg
y(P )}, which coincides with the condition of Nishioka in the case d > deg
y(P ). Note that we have to assume d > d
y(P ) only for technical reasons (cf. formula (24)).
(ii) T¨opfer proved in [15] a transcendence measure for f (α) under the condition d > n max{n, deg
y(P )}.
(iii) For m ≥ 1 and β = 1 we get the result of Nishioka [10]. In [10]
one can also find a lot of applications. For other examples in this case, but d
y(P ) = 1, see Chirski˘ı [2] and T¨opfer [14].
Our next example deals with infinite products of the form f
n(z) :=
Y
∞ j=0(1 − z
dj)
nj, where d and n are positive integers with d ≥ 2.
Let 1 ≤ n
1< . . . < n
m(m ≥ 2). Then the functions f
niare analytic for
|z| < 1 and satisfy the functional equations
f
ni(z) = (1 − z)f
ni(z
d)
ni(i = 1, . . . , m).
Hence we have the following:
Corollary 3. Let 1 ≤ n
1< . . . < n
mbe integers and β := n
1· . . . · n
m. If α is algebraic with 0 < |α| < 1 and d is an integer with
log d > (2m
2− 1 + p
4m
4− 2m
2+ m) log β,
then the values
Y
∞ j=0(1 − α
dj)
nj1, . . . , Y
∞ j=0(1 − α
dj)
njmare algebraically independent over Q. Under the corresponding conditions on α, d and n we get the algebraic independence of
Y
∞ j=0(1 − α
dj), Y
∞ j=0(1 − α
dj)
2j, . . . , Y
∞ j=0(1 − α
dj)
nj.
Remark. Nishioka proved (Theorem 3.4.13 in [11]) the algebraic inde- pendence of
Y
∞ j=0(1 − α
dj) (d = 2, 3, . . .) for any algebraic number α with 0 < |α| < 1.
P r o o f (of Corollary 3). The algebraic independence of the functions f
n1, . . . , f
nmover C(z) will be shown in the last section.
By the remark after Lemma 4, f
n1, . . . , f
nmsatisfy the conditions for the houses and denominators of the coefficients in Theorem 1 for any L > 1.
Then the assumption of Corollary 3 follows immediately from Theorem 1 and Corollary 2.
2. Preliminaries and auxiliary results. For µ ∈ N
0, µ ∈ N
m0and f
i(z) := P
∞j=0
f
i,jz
j(i = 1, . . . , m) we define f
i(z)
µ:=
X
∞ j=0f
i,j(µ)z
j, f
i,j(µ):= X
ν1,...,νµ∈N0
ν1+...+νµ=j
f
i,ν1· . . . · f
i,νµ, (3)
f (z)
µ:= f
1(z)
µ1· . . . · f
m(z)
µm= X
∞ j=0f
j(µ)z
j,
f
j(µ):= X
ν1,...,νm∈N0 ν1+...+νm=j
f
1,ν(µ11)· . . . · f
m,ν(µmm). (4)
Lemma 4. If f
i,j≤ exp(c
0(1 + j
L)) and D
[c0(1+jL)]f
i,j∈ O
Kfor i = 1, . . . , m and all j ∈ N
0with L ≥ 1 and D ∈ N, then for all µ ∈ N
0and µ ∈ N
m0the following assertions hold:
(i) f
i,j(µ)≤ exp(c
1(µ + j
L)), D
[c1(µ+jL)]f
i,j(µ)∈ O
K,
(ii) f
j(µ)≤ exp(c
2(|µ| + j
L)), D
[c2(|µ|+jL)]f
j(µ)∈ O
K.
P r o o f. Assertions (i) and (ii) are consequences of the identities (3) and (4) using the fact that the number of ν ∈ N
µ0with ν
1+ . . . + ν
µ= j is bounded by
j+µ−1µ−1≤ 2
j+µ.
Remark. If the functions f
1, . . . , f
msatisfy functional equations of type P
i(z, f
i(z), f
i(z
d)) = 0 (i = 1, . . . , m)
with polynomials P
i∈ Q[z, y, u]\{0} and deg
u(P
i) ≥ 1, we see that there exist an algebraic number field K, an explicit computable constant c > 0 and a positive integer D ∈ N such that for j ∈ N
0and all ε > 0:
(i) f
i,j∈ K,
(ii) f
i,j≤ exp(c(1 + j
1+ε)), (iii) D
1+jf
i,j∈ O
Khold, i.e. the conditions of Lemma 4 are fulfilled for all L > 1. For a proof of this remark see Lemma 1.5.3 of Nishioka [11] and Proposition 1 of Becker [1] for a more general result.
Lemma 5. For N ∈ N there exists a polynomial R ∈ O
K[z, y ]\{0} with the following properties:
(i) deg
zR ≤ N , deg
yR ≤ N , (ii) log H(R) ≤ c
3N
(m+1)L, (iii) ν := ord
0R(z, f (z)) ≥ c
4N
m+1for suitable constants c
3, c
4∈ R
+.
P r o o f. Put
R(z, y ) :=
X
N λ=0X
|µ|≤N
r
λ,µz
λy
µwith (N + 1)
N +mmunknown coefficients r
λ,µ. Then R(z, f (z)) :=
X
N λ=0X
|µ|≤N
r
λ,µz
λf (z)
µ= X
∞ h=0β
hz
h(say) with (cf. the identity (4))
(5) β
h=
min{h,N }
X
λ=0
X
|µ|≤N
r
λ,µf
h−λ(µ).
Assertion (iii) is equivalent to the condition β
h= 0 for 0 ≤ h < c
4N
m+1, and this yields at most [c
4N
m+1] + 1 equations in the
(N + 1)
N + m m
≥ 1
m! N
1+m> 2c
4N
m+1+ 1
unknowns r
λ,µfor a suitable constant c
4. After multiplication with a suit- able denominator D
[c2N(1+m)L]according to Lemma 4 the coefficients f
h−λ(µ)are algebraic integers and their houses are bounded by exp(c
5(N
(1+m)L)).
Siegel’s lemma (cf. Hilfssatz 31 in Schneider [12]) yields the assertion.
Lemma 6. Let ν be as in Lemma 5 and β
hdenote the Taylor coefficients of R(z, f (z)) as in the proof. Then
(i) |β
h| ≤ exp(c
6(h + N
(1+m)L)) ≤ exp(c
7(h + ν
L)).
(ii) |β
ν| ≥ exp(−c
8ν
L).
(iii) Suppose that k ∈ N satisfies d
k≥ c
9ν
Lwith ν, N, L as above and a suitable constant c
9∈ R
+depending only on f and α. Then there exist constants c
10, c
11∈ R
+depending only on f and α such that
−c
10νd
k≤ log |R(T
k(α), f (T
k(α)))| ≤ −c
11νd
k, where T
k(α) denotes the kth iterate of T at the point α.
P r o o f. From (5) we get β
h=
min{h,N }
X
λ=0
X
|µ|≤N
r
λ,µf
h−λ(µ).
This representation together with Lemma 5 and the inequality |f
i,j| ≤ exp(γ
0(j + 1)) (notice that the functions f
1, . . . , f
mare analytic in a neigh- borhood of 0), hence |f
h(µ)| ≤ exp(γ
1(|µ| + h)) with γ
0, γ
1∈ R
+, implies the first estimate of Lemma 6.
For D, L, c
4as above and ν as in Lemma 5 we get (recall ν ≥ c
4N
1+m) D
[γ2(N +νL)]β
ν∈ O
Kand
β
ν≤ exp(γ
3(N
(1+m)L+ ν
L+ N )) ≤ exp(γ
4ν
L).
By a Liouville estimate we obtain the second part.
We now come to the last part of Lemma 6. By Lemma 5 we write R(T
k(α), f (T
k(α))) = β
ν(T
k(α))
ν1 +
X
∞ h=1β
h+νβ
ν(T
k(α))
hand by the assumption on k and the first two parts of Lemma 6 we get
X
∞ h=1β
h+νβ
ν(T
k(α))
h≤
X
∞ h=1exp(c
7(ν
L+ h) + c
8ν
L− γ
5hd
k)
≤ X
∞ h=1exp(γ
6ν
L− γ
7hd
k) < 1
2 .
Now the assertion follows from |T
k(α)|
ν= exp(−γ
8νd
k) and exp(−c
8ν
L) ≤
|β
ν| ≤ exp(2c
7ν
L).
Lemma 7. Let S, U
1, . . . , U
d∈ C satisfy S
d+ U
1S
d−1+ . . . + U
d= 0 and
−X
1≤ log |S| ≤ −X
2, log |U
i| ≤ Y (1 ≤ i ≤ d) for X
1, X
2, Y ∈ R
+. Then there exists j ∈ {1, . . . , d} such that
−dX
1− Y − log d ≤ log |U
j| ≤ −X
2+ Y + log d.
P r o o f. This is Lemma 4.2.3 of Wass [17].
Remark. The examples S
d+ U
d= 0 and S
d+ U
1S
d−1= 0 show that the bounds for |U
j| cannot be improved.
The proof of Theorem 1 depends on the following result from elimination theory, which can be found in T¨opfer [13, Theorem 1] with slight modifica- tions.
Lemma 8. Suppose ω ∈ C
mand K is an algebraic number field. Then there exists a constant c
12= c
12(ω, K) ∈ R
+with the following prop- erty: If there exist increasing functions ψ
1, ψ
2, Λ : N → R
+, real numbers Φ
2≥ Φ
1≥ c
12, positive integers k
0< k
1, m
0∈ {0, . . . , m} and polynomials (Q
k)
k0≤k≤k1∈ O
K[ y ] such that the following assumptions are satisfied:
(i) 1 ≤ ψ
1(k + 1)/ψ
2(k) ≤ Λ(k) and ψ
2(k) ≥ c
12(log H(Q
k) + deg
yQ
k) for k ∈ {k
0, . . . , k
1},
(ii) the polynomials (Q
k)
k0≤k≤k1satisfy, for k ∈ {k
0, . . . , k
1}, (a) deg
yQ
k≤ Φ
1,
(b) log H(Q
k) ≤ Φ
2,
(c) −ψ
1(k) ≤ log |Q
k(ω)| ≤ −ψ
2(k),
(iii) ψ
2(k
1) ≥ c
12Λ(k
1)
m0−1Φ
m10−1max{ψ
1(k
0), Φ
2}, then trdeg
QQ(ω) ≥ m
0.
3. Construction of an auxiliary function. Since the case β = 1 (i.e.
n
1= . . . = n
m= 1) was treated by Nishioka [10] we can assume β > 1.
The proof is rather long, so we give a short sketch of the main steps.
In the first step we show how the powers of f (α) can be reduced by using
the functional equations. In the second step we consider R(T
k(α), f (T
k(α)))
for a polynomial R and construct by induction a polynomial R
k, with de-
grees and height depending only on the degrees and height of R and on
d, β, d
y(P ) and k, such that |R
k(α, f (α))| has almost the same analytic
bounds as |R(T
k(α), f (T
k(α)))|. In the last step we use this polynomial R
kto construct a suitable sequence of polynomials Q
k∈ O
K[ y ] satisfying the
assumptions of Lemma 8 and prove Theorem 1 by Lemma 8.
For a real number a we define a
+:= max{a, 0} =
12(a + |a|).
Let K be an algebraic number field containing α, the coefficients of f
1, . . . , f
m(cf. the assumption of Theorem 1 and Lemma 4) and the co- efficients of the polynomials a, P
0,1, . . . , P
nm−1,m. Without loss of generality we can assume a ∈ O
K[z] and P
0,1, . . . , P
nm−1,m∈ O
K[z, y ].
In what follows let k ∈ N be fixed. Under the conditions of Theorem 1 on α, d and f we put for abbreviation
τ
κ:= α
dκ, ϕ
i,κ:= f
i(α
dκ) and ϕ
κ:= (f
1(α
dκ), . . . , f
m(α
dκ)).
For j = 1, . . . , m let P
nj,j:= a and we define the following notations:
d
z(P ) := max{deg
z(P
0,1), . . . , deg
z(P
nm,m)}, d
y(P ) := max{deg
y(P
0,1), . . . , deg
y(P
nm,m)},
L(P ) := max{L(P
0,1), . . . , L(P
nm,m)}.
Lemma 9. Suppose that k ∈ N and λ ∈ N
0. Then for all j = 1, . . . , m we have
(a(τ
k−1)f
j(τ
k))
λ=
n
X
j−1 i=0P
i,λ,j(k)(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
iwith polynomials P
i,λ,j(k)∈ O
K[z, y ] satisfying
d
z(P
i,λ,j(k)) ≤ (λ − i)
+d
z(P ), d
y(P
i,λ,j(k)) ≤ (λ − i)
+d
y(P ),
L(P
i,λ,j(k)) ≤ 2
(λ−nj)+L(P )
(λ−i)+.
P r o o f. For λ ∈ {0, . . . , n
j− 1} we choose P
i,λ,j(k)= δ
i,λ, where δ
i,kis the Kronecker symbol, and the assertions are obvious.
Let now λ = n
j+ l for l ∈ N
0. We show the assertion by induction on l.
This is obvious for l = 0 because of (2) and (a(τ
k−1)f
j(τ
k))
nj=
n
X
j−1 i=0P
i,j(τ
k−1, ϕ
k−1)a(τ
k−1)
nj−1−i(a(τ
k−1)f
j(τ
k))
i, with P
i,n(k)j,j
(z, y ) := P
i,j(z, y )a(z)
nj−1−i.
In the induction step the assertion follows from
(a(τ
k−1)f
j(τ
k))
nj+l+1= (a(τ
k−1)f
j(τ
k))
nj+l(a(τ
k−1)f
j(τ
k))
=
n
X
j−1 i=0P
i,n(k)j+l,j
(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
i+1=
n
X
j−2 i=0P
i,n(k)j+l,j(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
i+1+ P
n(k)j−1,nj+l,j
(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
nj=
n
X
j−2 i=0P
i,n(k)j+l,j(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
i+1+ P
n(k)j−1,nj+l,j(τ
k−1, ϕ
k−1)
×
n
X
j−1 i=0P
i,n(k)j,j(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
i=
n
X
j−1 i=0P
i,n(k)j+l+1,j(τ
k−1, ϕ
k−1)(a(τ
k−1)f
j(τ
k))
i. So we get
P
i,n(k)j+l+1,j(z, y ) := P
i−1,n(k) j+l,j(z, y ) + P
n(k)j−1,nj+l,j(z, y )P
i,n(k)j,j(z, y ), where P
−1,n(k)j+l,j
(z, y ) := 0.
By induction it follows that P
i,n(k)j+l+1,j
∈ O
K[z, y ] and d
z(P
i,n(k)j+l+1,j
) ≤ (n
j+ l + 1 − i)d
z(P ), d
y(P
i,n(k)j+l+1,j
) ≤ (n
j+ l + 1 − i)d
y(P ), L(P
i,n(k)j+l+1,j
) ≤ 2
l+1L(P )
nj+l+1−i.
In the reduction step we replace R(τ
k, ϕ
k) =: R
0(τ
k, ϕ
k) for an arbitrary polynomial R ∈ O
K[z, y ] inductively by R
l(τ
k−l, ϕ
k−l) and finally get a polynomial R
kwith almost the same bounds for |R
k(α, f (α))|, the degrees and the height of R
kas R
0.
Lemma 10. Suppose k ∈ N and R ∈ O
K[z, y ]. Then there exists a poly- nomial
R
∗(z, u, y ) := X
µ∈M
R
∗µ(z, u)y
µ∈ O
K[z, u, y ] with M := {0, 1, . . . , n
1− 1} × . . . × {0, 1, . . . , n
k− 1} and
d
yj(R
∗) ≤ n
j− 1 (j = 1, . . . , m), d
z(R
µ∗) ≤ dd
z(R) + d
z(P )d
y(R), d
u(R
µ∗) ≤ d
y(P )d
y(R),
L(R
µ∗) ≤ L(R)L(P )
dy(R)2
dy(R)such that
a(τ
k−1)
dy(R)R(τ
k, ϕ
k) = R
∗(τ
k−1, ϕ
k−1, a(τ
k−1)ϕ
k).
P r o o f. From the representation R(z, y ) :=
d
X
z(R) i=0X
|j|≤dy(R)
R
i,jz
iy
jwe get, by Lemma 9, a(τ
k−1)
dy(R)R(τ
k, ϕ
k) =
d
X
z(R) i=0X
|j|≤dy(R)
R
i,jτ
kia(τ
k−1)
dy(R)−|j|(a(τ
k−1)ϕ
k)
j= X
µ∈M
R
∗µ(τ
k−1, ϕ
k−1)(a(τ
k−1)ϕ
k)
µ, where
R
µ∗(z, u) :=
d
X
z(R) i=0X
|j|≤dy(R)
R
i,jz
dia(z)
dy(R)−|j|P
µ(k)1,j1,1
(z, u) · . . . · P
µ(k)m,jm,m
(z, u).
Now the bounds for the partial degrees d
yjare obvious. From Lemma 9 we get
d
z(R
∗µ) ≤ dd
z(R) + d
z(P )d
y(R) + d
z(P ) max
n X
mi=1
(j
i− µ
i)
+− j
i: |j| ≤ d
y(R) o
≤ dd
z(R) + d
z(P )d
y(R)
and similarly we derive the upper bound for d
u. The length can be bounded in an analogous way by
L(R
∗µ) ≤ L(R)2
max{Pmi=1(ji−ni)+:|j|≤dy(R)}L(P )
dy(R)≤ L(R)L(P )
dy(R)2
dy(R).
Lemma 11. Suppose that R
∗∈ O
K[z, u, y ] is the polynomial in Lemma 10. Then there exist polynomials U
1, . . . , U
β∈ O
K[z, u] such that
R
∗β+ U
1R
∗β−1+ . . . + U
β= 0 at the point (z
0, u
0, y
0) := (τ
k−1, ϕ
k−1, a(τ
k−1)ϕ
k) and
d
z(U
l) ≤ βdd
z(R) + βd
z(P )(d
y(R) + |n|), d
u(U
l) ≤ βd
y(P )(d
y(R) + |n|),
L(U
l) ≤ exp(c
13(d
z(R) + d
y(R)))H(R)
β.
P r o o f. With R
∗(z, u, y ) := P
µ∈M
R
µ∗(z, u)y
µas in Lemma 10 we put for ν ∈ M ,
R
∗(τ
k−1, ϕ
k−1, a(τ
k−1)ϕ
k)(a(τ
k−1)ϕ
k)
ν= X
µ∈M
R
∗µ(τ
k−1, ϕ
k−1)(a(τ
k−1)ϕ
k)
µ+ν= X
λ∈M
R
λ,ν(τ
k−1, ϕ
k−1)(a(τ
k−1)ϕ
k)
λ, with (cf. Lemma 9)
R
λ,ν(z, u) := X
µ∈M
R
∗µ(z, u)P
λ(k)1,µ1+ν1,1
(z, u) · . . . · P
λ(k)m,µm+νm,m
(z, u).
The degrees and length of R
λ,νcan be bounded by Lemmas 9 and 10:
d
z(R
λ,ν) ≤ max
µ∈M
n
d
z(R
∗µ) + X
m j=1d
z(P
λ(k)j,µj+νj,j
) o
≤ dd
z(R) + d
z(P )d
y(R) + d
z(P ) max
µ∈M
n X
mj=1
(µ
j+ ν
j− λ
j)
+o
≤ dd
z(R) + d
z(P )(d
y(R) + |n| + |ν| − |λ|).
Similarly
d
u(R
λ,ν) ≤ d
y(P )(d
y(R) + |n| + |ν| − |λ|),
L(R
λ,ν) ≤ L(R)L(P )
dy(R)+|n|+|ν|−|λ|2
dy(R)+|ν|≤ γ
1L(R)γ
2dy(R), where the constants γ
1, γ
2∈ R
+depend only on P and n.
Thus the system of β linear equations with β unknowns, X
λ∈M
{R
λ,ν(τ
k−1, ϕ
k−1) − δ
λ,νR
∗(τ
k−1, ϕ
k−1, a(τ
k−1)ϕ
k)}ω
λ= 0, where
δ
λ,ν:=
n 1 if λ = ν, 0 else
is the generalized Kronecker symbol, has for ω := (ω
λ)
λ∈Ma nontrivial solution
ω
λ:= (a(τ
k−1)ϕ
k)
λ.
Hence the determinant of the matrix of coefficients must vanish at the point
(z
0, u
0, y
0) := (τ
k−1, ϕ
k−1, a(τ
k−1)ϕ
k), and the expansion of the determinant
with respect to the powers of R
∗(τ
k−1, ϕ
k−1, a(τ
k−1)ϕ
k) implies
0 = det(R
λ,ν− δ
λ,νR
∗)
λ,ν∈M= ±(R
∗β+ U
1R
∗β−1+ . . . + U
β) with polynomials U
l∈ O
K[z, u ].
Since the polynomials U
lare sums of products of the form R
λ1,σ(λ1)· . . . · R
λs,σ(λs),
where λ
1, . . . , λ
s∈ M are pairwise distinct and σ := (σ
1, . . . , σ
m) is a per- mutation of {0, . . . , n
1− 1} × . . . × {0, . . . , n
m− 1}, for l ∈ {1, . . . , β} we get
d
u(U
l) ≤ max
σ
n X
λ∈M
d
u(R
λ,σ(λ)) o
≤ βd
y(P )(d
y(R) + |n|)
because X
λ∈M
(|λ| − |σ(λ)|) = 0.
By analogy we obtain d
z(U
l) ≤ max
σ
n X
λ∈M
d
z(R
λ,σ(λ)) o
≤ βdd
z(R) + βd
z(P )(d
y(R) + |n|).
The length of U
lcan be bounded by
L(U
l) ≤ β! max{L(R
λ,ν) : λ, ν ∈ M }
β, with
L(R
λ,ν) ≤ exp(c
13(d
z(R) + d
y(R)))H(R).
Lemma 11 is proved.
Now the necessary tools for the reduction step from R
0to R
kare com- plete, and we prove for j = 0, . . . , k the existence of polynomials R
j∈ O
K[z, y ] such that for j = 0,
(6) d
z(R
0) := d
1,0, d
y(R
0) := d
2,0, log H(R
0) := H
0, exp(−ψ
1(0)) ≤ |R
0(τ
k, ϕ
k)| ≤ exp(−ψ
2(0)), and for j ≥ 1:
d
y(R
j) =: d
2,j≤ βd
y(P )(d
2,j−1+ |n|), (7)
d
z(R
j) =: d
1,j≤ βdd
1,j−1+ βd
z(P )(d
2,j−1+ |n|), (8)
log H(R
j) =: H
j≤ βH
j−1+ c
14(d
1,j−1+ d
2,j−1).
(9)
Here the constant c
14> 0 depends only on f and α and (10) exp(−ψ
1(j)) ≤ |R
j(τ
k−j, ϕ
k−j)| ≤ exp(−ψ
2(j)).
The functions ψ
1, ψ
2satisfy for j ≥ 1 the following recurrence equalities:
ψ
1(j) := βψ
1(j − 1) + βH
j−1+ c
15(d
1,j−1+ d
k−jd
2,j−1) + log β, (11)
ψ
2(j) := ψ
2(j − 1) − βH
j−1− c
16(d
1,j−1+ d
2,j−1) − log β
(12)
provided that
(13) ψ
2(0) ≥ c
17β
k(H
0+ d
k(d
1,0+ d
2,0)),
where c
15, c
16, c
17∈ R
+are suitable constants depending only on f and α.
The existence of the polynomials will be proved in the next section. First we will derive upper bounds for d
1,j, d
2,j, H
jand ψ
1(j) and a lower bound for ψ
2(j).
Obviously (7) implies
d
2,j≤ γ
0(βd
y(P ))
j(d
2,0+ |n|) ≤ c
18(βd
y(P ))
jd
2,0,
and for d
1,jwe get inductively (note that d > d
y(P ) by the condition of Theorem 1)
d
1,j≤ (βd)
jd
1,0+ βd
z(P )
j−1
X
i=0
(βd)
i(d
2,j−i−1+ |n|) ≤ c
19(βd)
j(d
1,0+ d
2,0).
For H
j, the logarithm of the height of R
j, we get in a similar way H
j≤ β
jH
0+ γ
1j−1
X
i=0
β
i(d
1,j−i−1+ d
2,j−i−1) ≤ β
jH
0+ c
20(d
1,0+ d
2,0)(βd)
j. Now we can easily deduce from (11) and the above estimates that
ψ
1(k) = β
kψ
1(0) (14)
+
k−1
X
i=0
β
i{βH
k−i−1+ c
15(d
1,k−i−1+ d
id
2,k−i−1) + log β}
≤ β
kψ
1(0) + kβ
kH
0+ c
21(βd)
k(d
1,0+ d
2,0).
In a similar way (cf. (13)) we can derive a lower bound for ψ
2(k):
ψ
2(k) = ψ
2(0) −
k−1
X
i=0